Ethiopia Big Book

30 YEARS OF WAR AND
FAMINE IN ETHIOPIA
A Division of Human Rights Watch

Ct^yrighted image
Copyrighted material
EVIL DAYS
Thirty Years of War and Famine in Ethiopia
An Africa Watch Report
Human Rights Watch

New York Washington
• • Los Angeles • London
This On©
3QQC-
Copyright © September 1991 by Human Rights Watch
All rights reserved.
Printed in the United States of America
Cover design by Dan NdH and Debonh Thomas.
Afrka Watch
Africa Watch was established in 1988 to monitor and promote observance of
internationally recognized human rights in Africa. The chair of Africa Watch is
William Carmichael and the vice chair is Alice Brown. Rakiya Omaar is the
executive director; Alex de Waal is the associate director; Janet Fleischman and
Karen Sorensen are research associates; Nicola Jefferson is a Sandler Fellow;
Urmi Shah and Ben Penglase are associates.
Libiaiy of Coafftas Catalog-in-Fnhlicatiflp Data
Evil days: thirty years of war and famine in Ethiopia,

p. cm.— (An Africa Watch report)
ISBN 1-56432-038-3 $15.00
:
I.Ethiopia— and government — 1974-1991.

^Politics Ethiopia— 2. Politics
and government — 1889-1974. 3.1nsurgency— — —20th
Êthiopia ^History century.
4. Eritrea (Ethiopia)--Hi8tory—Revolution, 1962-1991. 5.
Coiinterinsurgency—Êthiopia—History—20th centuy. Hnman 6.
Rights—Ethiopia— irtniy—20fli cealgy. 7.Flimmrii FWrinpia—IBatoiy^-20ai
century. 8. Food relief— Ethiopia—l]isloiy—20lfa oemmy. L Human R^gblB
Watch (Organization) U. Series.
DT387.95.E9 1991
963.07—dc20 91-76256 OP
CONTENTS
Preface page i
Glossary of AGronynis vii
Maps ix
Intioduction 1
X. Background to War and Famine in Ethiopia 19
2. Scorched Earth in Eritrea, 1961-77 39
3. Rebellion and Famine in the North under Haile Selassie 55
4. Insurection and Invasion in the Southeast, 1962-78 65
5. The Secret Wais to Crush the Southeast, 1978-84 81
6. The Red Terror 101
7. Total War in Eritrea, 1978-84 113
8. Counter-Insurgency and Famine in Tigray

and its Borderlands, 1980-84 133
9. "Economic War" on the Peasants and Famme 157
10. War and the Use of Relief as a Weapon

in Eritrea, 1984-88 177
11. Starving Tigray, 1984-88 195
12. Resettlement 211
13. Vfflagization, 1984-90 231
14. Eritrea under Siege, 1988-91 237
15. Aimed Decision: The North, 1988-91 255
16. The Politics of ReUef, 1989-91 277
17. Tlie Rage of Numbers: Meagistu's Soldieis 291
18. Wars within Wars: The Western and

Southwestern Lowlands 317
19. Divide and Misrule: Tte East, 1984-91 347
20. Western Policy towards Ethiopia 359
AMca Watch's Recooimendations 379
PREFACE
This report was initially planned at a time when it was not possible
for Africa Watch to undertake research in government-held areas of
Ethiopia, and access to rebel-held aieas bad not yet been obtained.
Primary leseaich consisted of interviews witfi refiigees and other
Ethiopians abroad. After the fall of the Mengistu government, access is
now possible, and extensive research in all parts of Eôpia would be
able to paint a much more detailed picture of the abuses associated with
the war. However, for reasons of time, that research remains to be done.
This should be the task of an investigative commission set up by the new
government.
Instead, the report relies heavily on secondary sources, including relief
workers, journalists, and others who have travelled to Ethiopia or who
have talked to Ethiopians. As a result, there are many blank areas: whole
campaigns, particularly m
the south, have scarcely been documented at
all in this report. Another result is tfiat m
some cases the incidents
reported cannot be fully cross-checked with independent sources. Where
reported by sources known to be generally reliable, such incidents have
been included. The source and status of information that has not been
independently verified has been indicated.
Previous reports on human rights abuses in Ethiopia which have been
compiled without visits to government-held areas have been subject to
criticism, chiefly from defenders of the previous government, that such
sources are wholly biased and unreliable. On these grounds a highly
critical report by Cultural Survival on the government's resettlement
program^ was dismissed by Professor Richard Pankhurst^ and Mr Kurt
Jansson, head of the UN famine relief operation in Ethiopia.^ As shown
in chapter 12, those dismissals were premature.
In compiling this report, Africa Watch has used as extensive a range
of sources as possible. Between 1978 and 1988, the Ethiopian
government denied the existence of the war altogether, and at no time did
*
W. Clay and Bonnie K. HolcombcX
Cultural Survival (Jason Politics and
the Ethiopian Famine 1984-1985, Cambridge, Mass., 1985.
* Richard Pankhurst, The Ethiopian Famine: Cultural Survival's Report

Assessed," Anthropology Today, 13, June 1986, pp. 4-5.
' Kurt Jansson, Michael Harris and Angela Penrose, The Eihiopim Famine,
London, 1987, p. 26.
it allow independent access to the war zones. There is virtually no
reliable information available about human rights abuses associated with
the war from official Ethiopian sources. The Ethiopian government
displayed an unhealthy obsession with statistics, and ostensibly-precise
numbers for damage to property and "affected populatkms* fonn the
greater part of its assessmeiit of the impact of flie wus and fBooBM,
Given that the govemmeBt cooastently ovedooked te Cixiflteiioe of a
million people in Tigray, and invented half a million retunriqg refiigees
who did not exist, such figures must be treated with cavtiOD.
Concerning famine, Africa Watch has made extensive use of official
documents, aid agency reports and the research undertaken by Ethiopian
and foreign scholars working in government-held areas. In many cases
it is necessary to "read between the lines" as these scholars were anxious
not to endanger their sources, careers, liberty or lives by tellii^ the truth
in plain wordb.
Asignificant part of the mfioimation contained In this re^rt originates
from the reports, newspaper aiticies, diaries and testimonies of raceigD
visitors to areas controlled by the rebel fronts, principally the Eritrean
People's Liberation Front (EPLF) and the Tigrayan
Peopled liberation
Front (TPLF, which after January 1989 was the leading member of the
Ethiopian People's Revolutionary Democratic Front, EPRDF). Though
these fronts gave greater access to the war zones, that access was never
unrestricted. Visitors were usually accompanied by armed guards,
primarily to protect them from government saboteurs, but which also
identified them with the relevant front. The information obtained is
therefore less tlian Ideally Independent. However, no vuilor — indndiqg
those who were
nnsynqwllietic to the fironts and snbaegnently wrote
critical reports of their activities —
has made a nbstantial criticism of
his or her access to the civilian popnlatlDn, or come away with belief
that the pêople he or she spoke to were influenced by the presence of
EPLF or TPLF-EPRDF* representatives. Consequently, some of this
information has been used, after careful scrutiny and cross-checking.
Much of the information obtained by visitors to rebel-held areas
consists of eye-witness accounts of atrocities and their aftermath; this is
not subject to the same problems of potential distortion.
Other mfofmatlaa origlBites fiom refugees. In refugee camps,
mdependent access to civilians is possible. Refugee testimony cannot be
regarded as distorted simply because the lefqgee has made a politiod
^ The designation TPLF-EPRDF is intended to encompass the TPLF from

its incôn up until the finmatiflo of the EPRDF, and the EPRDF thenafler.
11
statement by fleeing his or her country; neither of course can it be used
uncritically. In addition, contrary to the allegations of some defenders of
the government, many refugees (particularly in the late 1980s) fled to
neighboring countries precisely because they were unsympathetic to the
rebel fronts. For eiample, most of the lefiigees vifbo amved in Sudan
fitom Eritrea in 1989/9 were not suppoften of ttie EPLF: displaced
dvflians who supported the EPLF had remained behind in relief camps
run by the EPLF and the Eritrean Relief Association.
Hie EPLF and TPLF-EPRDF also displayed an abiding preoccupation
with numbers. These may or may not have been more accurate than
government figures. On the rare occasions when these figures have been
alluded to, their origin and our view of their reliability has been noted.
This report covers abuses by all sides. Where documented, abuses by
the rebel fronts have been included as well as those conunitted by the
government. However, the great majority of abuses against civilians* and
actions leading to funine, were committed bf the govenmient. The
fronts certainly had authoritarian political structures and tolenited litde
dissent in their own ranks, but — like the government's crackdown on
the institutions of dvil society — such abuses foil outside the scope of
this report.
The relative paucity of rebel abuses noted in these pages is not a
matter of the absence of reliable sources of critical information on the
activities of the fronts. All the fronts have their dissenters, who are
fiercely critical of certain of their actions and policies. These people
have provided information on some abuses by the fronts, but generally
agree that the treatment of chfilians and prisoners of war has been good,
even enmpUuy. These critics mdude refugees mteiviewed m
Sudan.
This report does not seek to justify or condemn the decision by rebel
fronts to engage in armed struggle, nor the decision by the government
to respond with military action. Africa Watch's mandate does not extend
to directly promoting peace. Instead the focus is on tibe manner in which
the wars were fought.
Statistics
This report contains dfacusskms of tfie available statistics for tiie

number cxf deaths attributable to the fBomnes and foiced idocations m
Ethiopia, and tiie controversy over the size of the population of Tigray.
The technical aqiects of statistical analysis and demographic modelling
have been kept to a minimum. The reason for the indusioo of this
material is that it is unpoctant to know how many people suffered and
died. It is is due to the victims to have the fact of
the least respect that
their living and dying taken seriously.
The United Nations and other concerned institutions have been
remarkably cavalier about the numbers of people who died, especially in
the 1983-5 famine. Usually the figure of one million famine deaths is
quoted for 1983-S. Hiis Ggm has absoliildy no scientific basis
whatsoever. Itisatrivialization aadddiamanizidon of lioman^^
such a figure to be produced without even a mfaiimal pretense at a
systematic investigation. If the UN
were to be equally cavalfer about
numbers of political detainees, or numbers of people killed when security
forces fire on protesters, it would be rightly condemned as ignorant and
irresponsible. Yet in Ethiopia and other countries which receive little
international attention, it appears able to quote wholly fictional figures for
famine deaths and remain unchallenged.
It is now too late for a proper demographic investigation into famine
mortality during 1983-5, and Aitica Watdh lacki the lesonrces to cany
out such an inquiry in any case. The analysis is theralDre veiy cmsoiy,
and based on existing surveys. Neverlfaele8S» it is the fiist tifloe that sndh
an analysis has been done, and it certainly provides a more accniale
assessment of the human impact of the famine than other figures
produced to date. Throughout, lower figures for deaths have been used,
so all estimates eii on the side of caution (or optimism).^
Geographical Terms
Ethiopia is beset by coofliciiqg teiritarial daims; the status of Eritrea

isonly the best-known instance. Eritrea is refeired to as a "tenitory", a
word that aspires to be neutral between the anirfiigtli^ claims Ifatf it is a
province and that it is by right an independent country. In addition, the
provinces that existed under Haile Selassie had their boundaries and
names revised on certain occasions, became "administrative regions" after
the revolution, and then were completely reconstituted in 1987, with the
introduction of a larger number of regions and some autonomous zones.
This report used the term "province" to refer to geographical and
administrative entities such as Harerghe and Shewa in the form in which
they existed up to 1987. "Oonder" is used for the province formerly
known as "Begemdir". Hiere are anrflictiiig daims as to
geographical extent of Hgiay. Purely tat coBwaiaaa^ the smaller
^ This is contrary to the general pnctice of citing only tiie upper lunits to
estimates of £unine mortality.
bf
government-defined Tigray is used in this report; the Tigrinya-speaking
areas of north Gonder and Wollo claimed by the TPLF are generally
referred to as the "borderlands" of Tigray. The post- 1987 names and
boundaries are not used, because they were instituted only very late in the
day, and only incompletely. "Tbe Ogadn" ii med to lefer to te lowhuid
area of Haierghe, Bale and eastern Sidamo inhabiled by etfanic Somalis,
moat of whom beloog to the Qgadeni dan.
Acknowicdgenicnti
This report was researched and written by Alex de Waal, Associate

Director of Africa Watch. It was made possible by the cooperation of
many people who have lived, worked or fought in Ethiopia, or who have
worked with refugees in neighboring countries. Many of these people are
credited in the tact, otfaen have had to remain anonymous.
This report is dedicated to the many unknown victims of thirty years
of war and funine in Ethiopia.
Glossary of Acronyms
ALF Afar Liberation Front

AMC Agricultural Marketing Corporation
Biir Ethiopian Dollar (approx. equal to US$0.47)
Deigue Provisional Military Administrative Committee (government
after 1974)
EDU Ethiopian Democratic Union
ELF Eritrean Liberation Front
EPDM Ethiopian People's Democratic Movement (a part of JBPRDF)
EPLF Eritrean People's Liberation Front
EPRDF Ethiopian People's Revolutionary Democratic Front
EPRP Ethiopian People's Revolutionary Party
ERA Eritrean Relief Association
ICRC Intanational Committee of the Red Cross
Kebele Urban DweUera' Association
MEISON All-Ethiopia Socialist Movement
MT Metric tons
OIF Oromo Islamic Front
OLF Oromo Liberation Front
PA Peasant Association
REST Relief Society of Tigray
RRC Relief and Rehabilitation Commission
SALF Somali Abo Liberation Front
Shifta Bandit
SNM Somali Naticmal Movement
SPLA Sudan People's Liberation Army
TPLF Tiâyan People's Liberation Front (a part of EPRDF)
UNHCR United Nations High Commissioner for Refugees
USAID United States Agency for International Development
WPE Workers' Party of Ethiopia
WSLF Western Somali Liberation Front
vii
MAP 4. EASTERN ETHIOPIA
INTRODUCTION
In May 1991 there was a dramatic change of government in Ethiopia,
in which the government of former President Mengistu Haile Mariam was
militarily defeated by the Ethiopian People's Revolutionary Democratic
Front (EPRDF) and the Eritrean People's Liberation Front (EPLF). If the
new govemmeiiC of Ethiopia and the adminifltnitioa hi Eritrea id>ide by
their promises of ensuring democracy and respect for human rights, there
is the veiy real ptospcCt that war and fEonme will be banished, for the
first time in a graeration.
This report is concerned with the thirty years of wars and famines in
Ethiopia between September 1961 and the overthrow of the Mengistu
government in May 1991. The starting date is inevitably somewhat
arbitrary: alternatives could have been chosen at various dates between
1960 and 1966. September 1, 1961, was the occasion of the first armed
clash between the newly-formed Eritrean Liberation Front and the army,
and is generally recognized by Ethiopians as the "official" outbreak of the
war in Eritrea. Other wars started at dates ranging from 1962 to 1975.
Why this Report Now?
The principal wars in Ethiopia are now

over. It is now too late to
influence the policies of the Mengistu
government, or those of the
international community towards that government. Yet there are a
number of reasons for continuing concern with the past. One is that
demands that many of those were responsible for perpetrating
justice
human rights abuses during the war should be brought to trial. Africa
Watch welcomes the promise by the new government headed by the
EPRDF that such trials will take place in accordance with mtemationally-
accepted norms of due process and in the presence of international obser-
vers. This report, while not intended to provide the specific evidence
needed to obtain criminal convictions against individuals for violations of
the humanitarian laws of war, documents the range and extent of abuses
that such trials should be concerned with.
A second reason is people need to
that the sufferings of the Ethiopian
be documented. The minimum duty of a human rights organization to
the hundreds of thousands of victims of war and famine is to record their
plight, so that they are not forgotten by history, and that history is not
rewritten to conceal or distort anbonasshig fMte. In aider to understand
the problems focmg the peoples oi Ethiopia at the present tune, it is also
necessary to understand the honors they have suffered.
A third reason is that the documentation of the abuses inflicted upon
the civilian population by successive Ethiopian governments leads directly
to an obligation on present and future governments not to repeat the same
crimes. By showing that the methods of counter-insurgency warfare
adopted in Ethiopia have led to a wide laiê of abuses, Africa Watch
hopes to discourage the adoption of such metiiods in future, not only in
Ethiopia but in other countries as well. When governments face
insurgencies, their response must be constrained by Imman rights con-
siderations.
In addition, the analysis of the genesis of famine contained in this
report has broad implications. The repeated famines that have struck
Ethiopia, and in particular the great famine of 1983-5, were in large part
created by government policies, especially counter-insurgency strategies.
It is possible that, at the time, neither government, army nor
international relief agencies were fully aware of the way in which these
actions were creatmg exceptionally severe famine. Members of these
institutions could thus plead ignorance when faced with the unacceptable
consequences of what they did and failed to do. Ignorance is the
flimsiest of excuses, especially when —
as in this instance —
there was
a notable lack of investigation into the causes of the famine. After the
examination of the evidence which we have relied on in compiling this
report, no such plea of ignorance should be acceptable in the future. This
report is intended to demonstrate conclusively, not only that war created
famine, but that particular strategies which the government adopted to
fight the wars created a particularly severe form of famine. Moreover,
international aid suppli^ to the government and to relief agencies
working alongside the government became part of the counter-insurgency
strategy of the government, and thus —
while meeting real and
immediate need — also served to further the government's war aims, and
prolong its life.
The findings of this report therefore have implications, not only for
the Ethiopian government and other governments faced with insurgencies,
but for relief agencies faced with the humanitarian needs resulting from
civil strife elsewhere in the world. This report raises disturbing questions
about the nature of the relationship between humanitarian agencies and
the host government. It calls into question the ethic of relieving actual
suffering wherever it is to be found, without preconditions other than safe
access and accountability for donations given. In tfie vndet context 6t a
counter-insurgency operation, the supply of such relief may actually
extend, intensify, or legitimize that counter-insurgency operation, whkh
is creating more suffering than is being relieved by the humanitarian
intervention. This is a tough ethical dilemma with no easy solutions, but
Copyrighted rnateriaLJ
one that must be faced and debated by the international community.
Putting conditions on the provision of humanitarian assistance is bound
to provoke the criticism of "playing politics with the hungry." But, when
more powerful actors are already playing politics with the hungry, for a
relief agency to ignore this fact is naive and may merely result in it
becoming their unwitting tool.
Main Findings
Violence against Civilians
The most characteristic feature of the war has been indiscriminate

violence against civilians by the Ethiopian army and air force. The army
deliberately killed and wounded tens of thousands of civilians and the air
force bombed civflians and civilian targets. It is not ponible to produce
an accunrte estimate for the number S
innocent people killed over the
decades, but it undoubtedly exceeds 150,000 (leaviiig aside those killed
by famine and the resettlement program). The atrocities discussed in this
report are but a fraction of the total number which occurred, although
most of the major massacres, especially in the north, are mentioned.
Every investigation uncovers more abuses, and many investigations into
massacres already documented discover more victims —
injured people
who died later and forgotten victims who had no relatives in the vicinity
to pronounce them missing. Very few incidents of killings turn out to be
less than has been reported.
Deaths from the war win continue even after the fightmg is over,
because many areas remam heavfly mmed, and unexploded munitions lie
buried in many marketplaces, fields and roads, waiting to claim victims.
Hiese gross violations have occurred consistently over thirty years.
Some years have been worse than others, but throughout the rule of the
Emperor Haile Selassie, his immediate successors, and Colonel (later
President) Mengistu Haile Mariam, there was no significant attempt to
curtail such abuses.
Counter -Insurgency Strategy
Throughout the war, the government followed a more-or-less

consistent set of counter-UBurgency strategies, with significant variations.
These consisted
Copyrighted matBrial
(1) The forcible relocation and control of much of the rural population,
using protected villages, forced resettlement from the insurgent
zone, and restrictions on movement.
(2) Hie systematic lestriction of food supplies (botfa ooBunacial and

relief) in insoigent areas, by catting key toads or by bombing
marketplaces and transport links.
(3) The use of indiscciminate violence and exemplary tenor against

civilians who remained outside the controlled ZQOes, usil^ both
ground forces and aerial bombardment.
(4) The fostering of divisions within the insurgent movements, and the
use of rebel forces opposed to neighboring governments to fight
against insurgents inade Ethiopia.
The strategies were implemented of the

differently in different parts
country, ^^fiagization was enforced the southeast, the west and
in
Eritrea, but not in Tigray, where restrictions on movement, forcible
resettlement (after late 1984) and the use of terror were used instead
This combination of strategies is familiar from many insurgencies
around the world. The Ethiopian case stands out as particularly
destructive because of the extraordinarily prolonged level of sustained
violence, and the frequent lack of any compensatory assistance to the
relocated and restricted population. Because of the fragile rural economy
and the dependence of rural people on mobility and a range of ecanomic
activities, this was patticnlvly damaging, and a major oontributor to
famme.
The Creation cf Famine
One consequence of the government's military policies, particularly

during the early 1980s, was the creation of famine. The great famine of
1983-5 is officially ascribed to drought. While climatic adversity and
related factors certainly played a part in the tragedy, closer investigation
shows that widespread drought occurred only some months after the
famine was already under way, and that information on food piodnction
and food prices gives an account which contradicts important dements of
the drottg^ hypothesis.
In recent years, "war and drought" has become the fovoced explanation
for famine. This is closer to the truth, but remains vague. It is not war
itself that creates famine, but war fought in particular ways.
The most important factors explaining the famine are the counter-
insurgency strategy adopted by the government, and restrictions and
burdens imposed on the population of non-insurgent areas in the name
of social transformation.
Repeated militaiy offensives destroyed the oops in soiplus-produdng
areas, and with them much rural employment. Hie bombing of
marketplaces restricted rural trade and exchange, impeding the
redistnbution of the surpluses that existed locally. In the areas where the
government retained some control, restrictions on migration, labor and
trade, and policies of forcible relocation in protected villages, served to
prevent hungry people from utilizing time-honored strategies for
obtaining food, by labor migration and petty trade.
These restrictions were enforced far beyond the areas of insurgent
activity, partly from fear that the insurgency would spread, and partly
from ambitions for socialist transformation.Other burdens on the
peasantry such as ponitively high delivery quotas of staple grains to the
Agricultural Marketmg Corporation and heavy taxation abo contributed
to the fEunine. Economic policies themselves are not within the mandate
of a human rights organization. However, when these policies are
implemented with a ferocity and single-mindedness that leaves no room
for dissent leading to possible revision, and when they do in the event
contribute to famine, at least the manner in which they are enforced
warrants consideration as an abuse of basic human rights.
Some responses to the famine only served to make matters worse.
The government's resettlement program was a disaster when considered
from ahnost any angle, and killed a mmimum of 50,000 people. Hie
largeHscale piovisioii of food idief to the government-controlled areas
of the north was a cantroversial initiative, whKh dlowed the government
to extend its control to contested areas» and maintain tfiat control longer
than would otherwise have been the case. Despite the fact that the
government had access to only a minority of the famine-stricken
population in the north, the great majority of relief was channelled
through the government side. This almost certainly prolonged the war:
the rebel fronts only regained the military position they held in 1983/4 (at
the beginning of the famine), some four to five years later.
The fsmnie of 1983-5 m nortfaon Ethiopia is estimated lo have killed
a nunhnum of 400»000 people (not connthig those killed by resettlement).
Somethu^ over half of this mortality can be attributed to human rights
abuses causing the famine to come earlier, strike harder, and extend
further than would otherwise have been the case. Famine also struck
other areas of Ethiqîa, for similar reasons, causmg tens of thousands of
deaths. The famines of 1973-5 killed at least 40,000 in Wollo and
55,000 in the Ogaden; the famine of 1965/6 also killed tens of thousands.
The Treatment of Soldiers
Soldiers were victuns of the war too. At least 100,000 soldieis were
killed. From the mid>1970s onwards, the govemment was oontiinially
increasing the size of its army, which numbered over 450,000 by 1990.
Conscripts were increasingly obtained by forceful and deoeitful means,
such as press-ganging or using food aid as a bait. Many were below the
age limit for military service according to international and domestic law.
In the army, they were subjected to a range of abuses, including summary
execution for disciplinary offenses. While not formally conscripted,
women were subjected to rape and kidnapping by soldiers. The EPLF
and Tigrayan People's Liberation Front (TPLF, a founder member of the
EPRDF in January 1989) treated prisoneis of war wdl, tfiongli the
International Committee of the Red Cross had disagreements with bofli
fronts concerning the tetter's access. Prisoneis of war were attacked by
govemment airplanes, and were subject to detentkm and torture or worse
by the govemment following their release. The Western Somali
Liberation Front (WSLF) and Somali army treated prisoners less well,
and frequently killed them. The rebel ti'onts also used methods of
conscription which on occasion were violent.
Inflaming Local and Neighboring Conflicts
Deliberately or inadvertently, the Ethiopian government inflamed other

conflicts. supported insurgent groups in Sudan and Somalia, wluch
It
committed human rights abuses. The destabilization of border areas and
the availability of modem weaponry through the market or through
militias and insurgents in neighboring countries helped to make local
disputes between ethnic groups more violent. Several small groups in
southwest Ethiopia have suffered severely from the depredations of their
well-armed neighbors, as a direct if unforseen consequence of
govemment policy.
The Record of the Rebel Fronts
The committed abuses against Chilians and prisoneis of

rebel fronts
war. Someof the rebel fronts have extremely poor human rights leoocds.
However, in the case of the EPLF and TPLF-EPRDF, these abuses woe
on a much smaller scale than those conunitted by the government focoes.
The of respecting civilians were the EPLFs and TPLF-
policies
EPRDFs weapon," and meant that with every atrocity committed
"secret
by the government, more civilians supported —
at least passively the —
rebels. Similarly, the released jvisoncis of war were the best
propagandists for the rebel cause witfaiii tfie govemmeiit army. The
history of the war is therefore, in a very real sense, a demonstration of
the fotflity of evil.
Hie war in the north was essentially not fought between the
government army and the fronts; it was fought between the army and
large sections of the people. It is this asymmetry between the contending
parties that was the major reason why one side behaved in a manner
radically unlike the other. It was an appreciation of this fact that enabled
the fronts to win.
The Impact on Civil Politics
With one exceptions, this report does not deal directly with civil
either in government- or rebel-held areas of the country.
institutions,
This exception is chapter 6, which deals with the Red Terror of 1977/8,
in which the urban insurgency of the Ethiopian People's Revolutionary
Party (EPRP) was crushed by the government, in the process killing a
minimum of 10,000 educated people and suppressing any form of dissent
for over a decade. This bloody episode was civil repression turned into
all out warfare, which is the principal reason why it is included in this
report.
More generally, tiie wats seriously stunted the growth of dvil society
in Ethiopia. Hie war was an excuse for the inddSnite continuation of
military rule — under a civilian mask during 1987-^
albeit and the —
creation and maintenance of security forces empowered to violate human
rights with impunity; it was an excuse for severe censorship and the
restriction of civil and political rights. Even after Marxism-Leninism
was formally abandoned by President Mcngistu in March 1990, there was
no progress towards the respect for civil and political rights, and even
regression, with the establishment of a de facto state of emergency
throughout the country, under military administrators.^ One casualty of
the war, not documented in this report, has been the stifling of any
initiatives towards democracy, the nde of law, and the development
^ See: "Ethiopia: Mengisln's Empty Democracy," News jum Africa Watch,

March 5, 1991.
civil society. These will be documented in future Africa Watch
publications.
The International Role
Hie international community has played a mixed ide in Ethiopia's

recent history. Until 1977, Ethiopia was a dose ally of the United States,
and received extensive military and economic assistance. Thereafter, it
was a close ally of the USSR, and was shunned by the US, which
provided no development nor military aid, and was a constant critic of
government actions. Other western countries, such as members of the
European Community, provided extensive economic assistance though not
military supplies, although the west's ally Israel broke with that policy in
1989, when it supplied armaments in return for the emigration of the
Ethiopian Jews.
Between 1977 and 1991 western aid to Ethiopia rose tenfold. Much
of this aid was humanitarian famine relief. Though not achieviqg all that
its proponents claim, this aid no doubt saved many lives and livdihoods.
However, there is a more questionable side to aid. In the crucial year
1985, about 90 per cent of the aid was given to the government and to
humanitarian agencies working on the government side, despite the fact
that they had access to only a minority of the famine-affected population.
Much assistance was given without scrutinizing the context in which it
would be and in fact served to support the counter-insurgency
utilized,
strategies of Mengistu government.
the Instances indttde the
"rehabilitation" assistance to the southeast durmg 1980-2» the Food for
the North Initiative of 198S--6, and the outpouring of aid in 1987-8.
More generally, international assistance undoubtedly prolonged the
Mengistu government's life.
The US government was consistently critical of the Mengistu
government's human rights record. However, the main US investigation
into government culpability in creating the famine was given a restricted
agenda which obliged it to find the Ethiopian government "not guilty" of
deliberately using starvation as a weapon of war. Had the investigation
been mandated to examine a slightly broader set of issues and a longer
time period, the verdict would have been much more likely to be "guilty."
The British government was also critical, but to a lesser degree. Most
other major donors did not include a significant consideration of human
rights concerns in determining their relations with the Ethiopian
government. The UN in particular was more eager to conceal evidence
of abuses, especially those associated with famine and famine relief, than
expose them, and the European Community gave substantial aid with no
questions asked.
The US was the only major donor to direct the majority of its
assistance to the rebel held areas in 1989-91, in accofdance with an
objective assessment of the nombocs of needy people to be found there.
Duing 1988-91, a major concem of Israel and the US was the
population of Ethiopian Jews, known as Falashas. Israel supplied arms
to the Ethiopian government in exchange for allowing the Falashas to
leave. The US lobbied for the Falashas to be allowed to emigrate but
objected to the arms supply.
Despite repeated approaches by the Mengistu government, the US
refused to give economic or military assistance. Apart from humanitarian
concerns and the Falasha issue, the US was active in trying to obtain a
negotiated peace. The US was ultimately successful in assisting the
handover dpowa from the Mengistu government to the EFRDF and
EPLF with less bloodshed than would otherwise have occuned. Hiiswas
achieved through persistent diplomatic efforts and through pragmatism,
notably the abandonment of the loqg-standing US objection to Eritrean
indepcaidence.
Structure
For the most part, this report is structured chronologically. Some

parts may appear repetitive —but the war was repetitive: there were
thirteen major offensives in Eritrea alone, ten in Tigray, and a dozen in
the southeast. Followiiig the historical background, it breaks naturally
into five sections, whidh cover the following periods, provinces and
topics:
(1) Insurgency and counter-insurgency under HaUe Selassie and

during the first years of the revolution, up to and induding the
Somali invasion (chapters 2-4).
(2) Counter-insurgency in the years between 1978 and 1984, when the
army was newly equipped and expanded with Soviet assistance,
leaduig to victory over the msuigenctes m
the southeast, the Red
Tenor m the cities, and the creation of fiamme in the north
(chapters S-9).
(3) The use of humanitarian reUef for war and programs of social
enigineering, mdudiqg pacification m Eritiea, the withholding of
relief from Tigray, resettlement and villagization, in the period
1985-8 (chapters 10-13).
(4) War and famine during the last years of the Mengistu government,
when the EPLF and TTLF-EPRDF had won llie military advantage
(chapters 14-16).
(5) Neglected issues and regions: conscription and the treatment of

soldiers and prisoners of war; the wars in the southwest, and the
wars in the cast including those involving Ethiopian refugees in
Somalia (chapters 17-19).
A final chapter deals with the policies of western governments to

Ethiopia, with a focus on the United States.
Wars and Famines, 1961-77
The "official" outbreak of the war in Eritrea was September 1961,

when the army first engaged the forces of the Eritrean Liberation Front
(ELF). Over the following decade the army acted in a brutal and
destructive manner towards civilians and their villages, farms and animal
herds. There were numerous massacres of civilians, hundreds of villages
were burned, and much of the population was forced to relocate in
fortified government villages. The worst atrocities occurred during 1967
and 1970-1. The ELF and the younger EPLF also engaged in abuses
such as kidnappings and assassinations, and fought against each other.
Shortly after the 1974 revohition, the ELF and EPLF combined in a
major attempt to defeat the army in Asmara. 1975 was to be tibe most
bloody year of the war to date. The government continued to use
indiscriminate violence against civilians, it instituted a food blockade of
the Eritrean highlands, and tried to mobilize a huge peasant aimy to
overwhelm the fronts by sheer weight of numbers.
From the start, the policies of scorched earth and the use of a food
blockade as a weapon meant that the war was fought at the cost of
creating hunger among the civilian population.
Haile Selassie also faced insurrections in the northern provinces. In
the 1960s the main such revolt was in Gojjam, due to d&content over
taxation and land measurement. In Wollo, inter--etfanic ^ghting in the
lowlands, coupled with government-promoted or sanctioaed processes of
land alienation and enforcement of crippling tenancy agreements in times
of drought all contributed to the creation of famine in 1972-4, a famine
made worse by the government's concealment of it and refusal to consider
10
assistance. After the revolution, there were insurgencies led by feudal
leaders in allthe northern provinces. The suppression of the Afar revolt
in to much loss of civilian life.
1975 was particularly bloody and led
In the southeastern Ethiopia there were two separate
part of
insurgencies during the 1960s and 19705. One insuigency was in the
Ogaden, where tfie WSLF, supported by the Somali government^ was
active. Hie second insugency involved a number of Oromo movements;
in the 1960s led by Wako Gutu in Bale and Sheikh Hussein in Harerghe,
in the 1970s led by the Oromo Liberation Front (OLF) and Somali Abo
Liberation Front (SALF). The government's military response included
indiscriminate violence against civilians and war against the economic
base —killing animals, poisoning wells, cutting food supplies, and
restricting movement. These military strategies were instnimental in
creating the famines which struck the area in 1973-4.
In 1977 the Somali army invaded tfie Ethiopian Ogaden, first secretly
and tfien openly, leading to a laige-scale conventional war. Both sides
m the war committed abuses agamst the dvflian population.
Counier-'Insurgency, 1978-84
In 1977 a number of simultaneous military changes occurred which

heralded a significant break with the past. These included: the Somali
invasion of the Ogaden, Ethiopia's break with the US and the turn to the
USSR, the massive enlargement and re-equipment of the army, the
prosecution of the Red Tenor, and the undisputed ascendancy of Colonel
Mengistu Haile Mariam.
The defeat of the Somali army in 1978 was followed by six years of
intense counter-insuigency warfare against the forces of the WSLF and
the OLF. The continuing war was waged largely in secret, and ended in
the defeat of the WSLF and the eclipse of the OLF. Other rebel fronts
such as the SALF and the Sidama Liberation Fioot were also defeated.
The government strategy included:
* Repeated military offensives, involving many abuses against civilians,

including indiscriminate aerial bombardment of villages.
* A policy of forcible relocation m protected villages.
* Hie use of Somali opposition movements to fight against the WSLF.

* Pressure on Somalia and the international community to obtain the
repatriation of refugees.
11
A particularly insidious element in the government strategy was the
utilization of international humanitarian aid for counter-insurgency
purposes.
1977/8 also was the year of the Red Terror, in which Colonel
Mengistu crushed the urban opposition of the Ethiopian PeopIe*s
Revolutionary Party (EPRP) by a massive campaign of IdlUng* detention
and torture, and intimidatioo. Later, the tenor was tnined on the All-
Ethiopia Socialist Movement, the government's eistwhile ally. Tens of
thousands of young peo|de were killed, and the government became
addicted to the use of terror as a weapon of war.
The Mengistu government was simultaneously fighting an intensified
war in Eritrea, seeking to crush the Eritrean insurgency with a vastly-
expanded and re-equipped army and air force. The years 1978-84 saw
an expansion of the scorched earth and forced relocation policies of
earlier years, together with satuiation bombiiig of areas of rebel streii^.
A series of large-scale offensives culminate in the huge '*Red Star"
offensive of 1982. These years witnessed the demise of the ELP, and the
retreat of the EPLF to the monntainous Sahel district close to Sudan.
Government policies led to wholesale ecological destruction, which,
together with the damage to livelihoods caused by the war and
restrictions, were major factors creating famine.
Meanwhile, in Tigray, the government faced the insurgency of the
TPLF, which was fighting a classic guerrilla war throughout the province.
The government's counter-insurgency strategy included:
(1) Military offensives into the TPLF heartlands, which were also the
richer, surplus-producing districts of the province.
(2) The bombing of nuuketplaces to disrupt commerce.
(3) The imposition of strict bans on the movement of petty traders and
migrant laborers.
As well as directly causing large-scale civilian suffering and death,

this cond>ination of policies meant that the normal processes of
redistribution of surplus grain to poorer areas and miigratioo of seasonal
laborers to richCT areas were no longer able to occnr, leadhig to mtense
famine in the food deficit areas. A
ck>se investigation of the evidence for
drought and production failures confirms that these cannot account for tfie
timing, severity or extent of the famine of 1983-5. Instead, an analysis
of the timing and location of the major offensives of 1980-1, 1983 and
1984, and the timing and scope of the restrictions, confirms that the
12
government's counter-insurgency strategy was the prime culprit for the
disaster.
Simultaneously, a set of restrictions and impositions were placed upon
rural people in government-controlled aieas» indndiqg punitive taxatiim,
lequisitioiuqg oi food for the Agricultural Marketing Corporation, forced
labor on government projects and state farms, and bans on labor,
migration and trade. Poor people were often forced to sell their reserves
of food to meet these demands, which were iMcked up by sanctions such
as imprisonment. These brought large sections of the population close to
the brink of famine. These policies were pursued partly for reasons of
counter-insurgency and partly from ambitions for socialist transformation.
The famine killed in excess of 400,000 people. The human rights
abuses made it come earlier, strike harder, and reach further. Most of
these deaths can be attributed, not to the weather, but to flie govenm^t's
gross violations of human rights.
Humanitarian Relief as a Weapon of War, 1985 -88
In late 1984 the famine reached the television screens of the west.
This caused a massive inflow of relief, which was utilized by the
government in its counter-insurgency strategy in the north. It coincided
with renewed offensives in Eritrea and Tigray and the launching of the
programs of resettlement and villagization.
The relief aid generously provided to the Ethiopian government and
the humanitarian agencies working alongside it was a boon to the
government's war plans. The politics of akl resulted in the government
side receiving a diqiroportionate share of the assistance compared to the
rebel-held areas. In Eritrea, aid was used as part of a military
pacification strategy, with aid agencies moving in behind the military to
secure newly-occupied areas. This allowed the government to score
significant military successes in 1985, and to control areas it had been
unable to hold on to before. Most aid agencies, in particular the United
Nations, preferred not to face the dilemmas of providing aid in such a
situation, and thus actively abetted the army's efforts. Some relief was
supplied duectiy to the army and militia.
When counter-attacking in 1987, the EPLF disrupted the relief
programs. These actions gained international condemnation, but the
government continued to object to plans to allow safe passage to
humanitarian supplies. The relief programs have also to be considered
within the wider context of the pacification strategy.
In Tigray, the government preferred to withhold aid from the province,
thus starving the people. Hie government attempted to conceal this fact,
13
and in doing so was abetted by the United Nations, which in August 1985
produced a mendacious report endorsing the government's claim that it
was feeding most of the famine victims in Tigray, at a time when it was
in fact feeding very few indeed. A
huge army offensive into Tigray in
1985, as well as being particularly bloody and destructive, was simed at
preventing clandestine relief supplies coming across the Sudanese border
to TPLF-held areas. Relief convoys, feeding centers and fefvgees were
aU attacked from the air. As in Eritrea, flie government was able to make
substantial military gains on account of the famine, but it also deeply
alienated the rural people, who realized that they would never be free
from famine while the Mengistu government remained in power. They
were thus prepared to undergo great hardships in order to fight against it.
The government's main response to the famine was a gargantuan
program of forced resettlement. This involved numerous abuses of
human rights,including the violent and mmmct in which
aifoitraiy
resettlers were taken, appalling conditions and on arrival, the
in transit
displacement of indigenous people in die resettlement areas, and violenoe
against resettlers who attempted to escape, including enslavement by
soldiers of the Sudan People's Liberation Army (SPLA). The resettlement
jM"ogram is estimated to have killed 50,000 people.
at least
The famine period also saw
beginnings of a comprehensive
the
villagization program. First implemented in the east as a counter-
insurgency measure, it was spread to other areas as an attempt at social
enguiccring. The implementation of this program also involved violence
and coercion, especially in the war zones.
The EPLF and TPLF-EPRDF Ascendancy, im-Ql

Early 1988 marked another watershed, in which the rebel fronts in
Eritrea and Tigray decisively gained the military initiative.The stalemate
in the war was broken, and their military victory duly followed three
years later. The government continued to fight the war with total
disregard for the rights of civilians, and the army and air force engaged
in reprisal killings of civilians. 1988 was a particularly bloody year in
the war, during which the government cast aside all restraint in its
repeated attacks on civilian targets. The massacre of 400 Eritrean
civUians by the army at Sheib and the death of 1,800 peofde when the
Tigrayan market town of Hausien was bombed are two of the most
notorious examples.
In 1989 the TPLF defeated the army in Tigray and occupied the whole
province. The government attempted to destroy what it could before
leaving, and then sent bombing missions against major towns includii^
14
Copyrighted material J
the capital, Meqele. In 1990, the EPLF captured the port of Massawa,
which the government then tried to destroy using aerial bombardment.
Meanwhile the war intensified in the provinces of Gonder and Wollo and
spcead to Gojjam and Shewi^ wfam the anny oontumed to commit
numennis abuses against civilians. Abuses continaed up until the last
days of the Meqgistn regime in May 1991 — for ezanuple, just a few
weeks before, over 100 Tlâyan prisoneis were summaiuy executed in
Gonder prison.
The years 1988-91 also saw continued food shortages, and belated
efforts by the international community to ensure that civilians on all sides
had access to relief. These plans were continually subject to political
interference, to a certain extent by the fronts but very largely so on the
part of the government. The besieged Eritrean capital, Asmara, witnessed
tiie worst extremes of hunger during 1990-91, where military requisition-
ing of food and the imposition of a blockade created severe suffering. In

Tigray, despite the excq^tionally severe drought, severe famine was
avoided because the tAJP did not impose the same burdens and
restrictions on the populace as the government had done earlier in the
decade.
Other Wars
From the 1960s, and increasing in intensity during the 1980s, there
has been a series of wars in western and southwestern Ethiopia. These
have included:
* An insurgency by the EPRP in westem Oojjam, against the Mengistu

government and (latterly) the EPRDF.
* An insurgency by the OLE in westem Wollega, against the

government. This and the war in Gojjam were also inflamed by the
resettlement and villagization programs, and saw large-scale violations
of human rights during 1989-91.
* Conflict between the OLF and the SPLA, which had a wing of its
own army inside Ethiopia, operating in alliance with the E&opian
army. The SPLA committed atrocities against local civilians, and the
OLF attacked a Sudanese refugee camp in 1990. The main exodus of
Sudanese refugees back to Sudan in May-June 1991 was not,
however, caused by attacks by the OLF or EPRDF.
15
* Conflict between the Anuak of Illubabor province and the government,
inflamed by resettlement, villagization and the presence of the SPLA.
* Local warfare by pastoralist groups in Gamu Gofa who are either

self-armed or armed by a neighbor. These wais aie lood disputes
intensified by the selective ivaiUbility of anns ta oofdi groups
following on alliaaoes with psstoialist militias fesoB^^ Ite
Kenyan govenunent has also been responsible for abuaes i^gainst fliese
pastoralSts.
In eastern Ethiopia, following the defeat of theWSLF and OLF in the

early 1980s,fragmented wars have continued on both sides of the
Ethiopia-Somalia border. Both governments supported insurgent groups
operating in each other's territory; Ethiopian refugees were armed by the
Somali govmment, sad the inaurgent groups daned widi each ofbet and
with the two national acmies. The war between iie Sonali NatioBal
Movement (backed by Ethiopia) and a Sonatt-badced Oramo front,
mside northern Somalia durhtg 1988 was a lesnlt off this process of
fragmentation and manipulation.
Following the EPRDF victory in May- June 1991, there was an
upsurge of local violence in southern and eastern Ethiopia, involving
banditry, inter-communal strife, and resistance to the imposition of any
form of central authority.
The Total Impact
The cost to Ethiopia, in human, economic and social terms of the last
thirty years of war and fomhie has been literally incalculable. Acocndiqg
to the very conservative estiouites oootained in this lepoil; the wan and
famines have cost a minimum of one million lives, and possibly as many
as 1.5 million. An even greater number have been wounded, traumatized,
forced to flee as refugees, or have been displaced within the country.
The economy is bankrupt, the natural environment irreparably damaged,
the country torn apart, and the growth of civil politics aborted. The last
thirty years have indeed been evil days.
16
Recommendations
Africa Watch's recommendations to the government of Ethiopia and

the administration in Eritrea include:
* That tiiose primarily lesponsible for gross abusea of human rights be

brought to trial, before an independent oourt» with due process of law.
* That a pennanent Human Rights Commission be set up with wide

nqgmg powers to investigate allegations of human rights abuse.
* Hiat an inquiry into the causes of famine be set up.
Africa Watch considers that the aid donors need to reconsider

carefully theiride in providiqg hunianitsrian asststanoe io sitnations of
armed conflict and hoinan rights abase. Africa Watcfat leoommendations
to the aid donors indnde:
* Hiat they should cooperate with the inquiry into the causes of famine,
and accept responsibihty for any instances in which they have abetted
or failed to prevent human rights abuses.
* That they adopt human rights conditions for the delivery of emergency
aid.
1. BACKGROUND TO WAR AND FAMINE IN ETHIOPIA
"The introduction of billeting into Gayint led to the death of a
peasant. The king's response to an appeal for justice was terse.
'Soldiers eat, peasants provide.'"^
Histories of Ethiopia
There is no impartial history of Ethiopia: every presentatioo of

historical facts is laden with modem-day political implications.
Ethiopia is a multi -ethnic and multi-religious state — Christians and
Moslems are present in approximately equal numbers, and there are also
followers traditional religions and until very recently—there was a —
community of Jews. Ethiopian nationalism, however, is largely based
upon and cultural symbols that derive from the Amhara-Tigray
political
tradition of the noitfieni highlands: Orthodox Christianity, an almost
unliroken tradition of Independence, liteiacy m
the ancient Ge'ez script,
and the use of ox-plough agriculture. One of the main reasons for ttie
last thirty years of waifiire has been the unwillingness of marginalized
people in Ethiopia to acocpi the noithon-highland definition of national
identity.
Ethiopia as we know it today is the product of the expansion of a state
centered in the northern highlands into adjoining areas, mostly to the
south. The northern highlands are inhabited by the Amhara and Tigray,
who are culturally and politically dominant, and a range of minorities,
notably the Agau. The Ethiopian Jews, known as Falasha, are ethnically
a sub--category of the Agan. The stales located this region daim a m
oontumous tradition gomg hack to the Idqgdom of Axnm in the early
middle-ages and beforehand, to the ofb|Ming of King Solomon and die
Queen of Sheba. Until 1974, the Eaqierais claimed to rule by virtue of
royal descent in the line of Solomon, and by being anointed by the
Patriarch of the Ethiopian Orthodox Church. Until the mid-twentieth
century, this state was known as "Abyssinia," and in the historical
context, will be called that in order to distinguish it from the larger area
of modem-day Ethiopia. In the region, highland Amhara and Tigray are
commonly called "habash," a word with the same derivation.
^
in: Donald Crummey, "Banditry and Resistance: Noble and Peasant
Quoted
in 19thCentury Ethiopia," in D. Crummey (ed.) Banditry, Rebellion and Protest
mAfiica, London, 1986^ p. 142. The kutt questUm m wm
the Bmpetat Teodros,
who rated ihe noitliwesteni part of modem-day Ethiopia from 1855 to 1868.
19
History lives in Ethiopia, The question of whether the ancient
Abyssinian state controlled all or part of modem-day Eritrea generates
acrimonious dispute among scholars and politicians. Contemporary
claims and counter-claims on the Eritrean issue are based on differing
readings of historical texts, which purport to show dttier that the territoiy
was traditionally indqpendent of nortfa-cential Ethiopia, or the "cradle A
Ethiopian dvUization," and inextricably linked to Ethiopia.
A simUar dispute rages over the origins of what today are the southern
provinces of Ethiopia. South of Gojjam and north Shewa, Ethiopia is
dominated by the Oromo (the largest ethnic groiq> in the country), with
significant representations of a range of other ethnic groups. Advocates
of "greater Ethiopia" claim that these areas —
and sometimes territories
even further afield in Sudan, Djibouti, Somalia, Kenya and Uganda —
were ruled by Abyssinian Emperors in the middle ages, until the
sixteenth-century "invasions" of the Moslems from the east and the
Oromo from the south. Odieis argue that tfie homelands of the
"invaders" fall within the boundaries of modem Ethiopia, and that a
reading of history that regards them as extraneous to EUiiopian histoiy
gives undue primacy to an Abyssinian version of events. They dispute
the territorial extent of the Abyssinian empire, and claim that western
historians have been seduced by the allure of the literate Christian legacy
of Abyssinia into regarding its people as somehow superior to their non-
literate Moslem and pagan neighbors, thereby endorsing the legitimacy
of Abyssinian imperial expansion —
and facilitating it through the supply
of firearms.
What is certain is that from the sixteenth to the mneteentfa century
there was no hegemony of a single group over modem-day Ethiopia. At
times, independent Amhara kingdoms appeared to be on the point of
being vanquished by the Moslems and the Oromo. In the mid-nineteoilfa
century, this began to change, as the northern kingdoms began a process
of political centralization, acquisition of European weaponry, and
conquest of their southern neighbors —
a process called "unification" by
its advocates and "colonization" by its detractors. This reached its climax
under Menelik, King of Shewa and Emperor (1889-1913). The Emperor

Menelik doubled fie size of the entire within a few decades and
established the boundaries of modem Ethiopia, and estab&hed
supremacy of the Shewan Amhara, not just over the Oromo and other
southern groups, but over the Oonder, Gojjam and Wollo Amhara and
Tigray as well.
20
Ethiopia and the West
To an older generation of people in the west, the name "Ethiopia" is

linked to the Italian fascist invasion of 1935; to a younger generation, it
is linlKd to the famine of 1984. There is a oommon thread to these two
momentons events: vicdations of the humanitarian bws of war.
The Italians invaded Ethiopia, deployed chemical weapons, bombed
Red Cross ambulances, and when in control massacred most of the
country's educated These abuses, against a fellow member of the
elite.
League of Nations, scandalized liberal public opinion in Europe and

America, and led to widespread sympathy and support for the exiled
Emperor Haile Selassie. Ironically, Haile Selassie was later to violate
international law in his annexation of Eritrea, and his army was to engage
in gross violations of human rights in combatting insurrection in different
parts of the empire — was still able to play upon the west's
but he
oonsdeooe and exploit his image as "victun." John Spencer, Ifoile
Selassie's longstanding advisor on international law, with unconscious
irony concluded his book Ethiopia at Bay with a quotation from
Herodotus: "thank the gods that tfaiey have not put it into the heart of the
sons of the Ethiops to covet countries which do not belong to them."^
Similar but intensified methods of warfare conducted against
subjugated peoples in Ethiopia by the Emperor's successor, Colonel (later
President) Mengistu Haile Mariam were instrumental in reducing much
of the rural population of Ethiopia from a state of poverty and hardship
to one of outright famine. International sympathy for Ethiopia's plight
led to the greatest outpouring of charitable donation in modem history —
a sympathy whkh a^dn deflected aHaiHon from the gross human rights
vicdatioos in the oounliy.
£tUopfai: Military and State
Army and state in Etliiopia are traditionally so dose as to be at times

indistinguishable.
The traditional Abyssinian state (or self-proclaimed empire) was
founded upon the principles of conquest of neighboring peoples, and the
settlement of soldiers on peasant cx)nimimities. The soldier-settlers were
entitled to administer the locality, laise taxes and requisition produce
firom the farmers. Their chief obligition hi return was to he prepared to
^ John D. Spencer, EAiapia at Bt^: A Penomd AccouM of Oie Haile

Sekasie Yean, Algnnie, Mich., 19H P- 3^-
21
fight for the king, and to raise a levy of peasants to do the same. One
Ethiopian historian has described the system of conquest and rule thus:
All the Christian provinces of the north were originally acquired by

wars of conquest ... [Tbe king] appropriated all the people and their
land, and reserved the right to dunoae of item aoooiding to his
wishes. He executed all resistanoe ngbterB who fell into his hands,
and reduced to slavery other cqitives of war ... These acts cmel d
repression were delibenrtdy committed ... to force the people to
surrender and to give them a terrible example of th& destructive feCCe
of the Christian army in case of further revolts.^
The twin themes of brutality in conquest —

including the use of
exemplary violence to instil fear and subservience —
and the military
mode of administration, which are here referred to in the context of the
15th century, recur hi moie recent history.
Traditional Forms of Warfere
There were two basic forms of traditional warfare in Ethiopia. One

form was that practiced by centralized states, including the Christian
Abyssinian kingdoms located in the northern highlands, and the Moslem
sultanates in the east. The second type of warfare was practiced by non-
centralized states, including both small-scale societies on the peripheries
of Ethiopia, and the large but decentralized confederacy of the Oromo.
Warfare by CeninUzed States
Abyssinian armies consisted of a group of pennanent soldiers attached

to the court,and a mass levy in times of emergenqr. Only the permanent
soldiers were trained. The levy method could be remadcably effective in
raising huge armies, though for relatively short campaigns. Armies of
over 100,000 men were not uncommon in the late 19th century. In 1935,
Haile Selassie raised an army of up to 350,000 to fight the Italians.
A central element in traditional warfare in Ethiopia was the
unremitting brutality of armies, in wars of conquest, rebellions and
counter-insurgency campaigns. "Whose face have you not disfiîred?
'Tadesse Tanuat, Church and StaU ht EOUopia, IZTO-ÎSI?, Qsfbid, 1972,

p. 99.
22
Copyrighled material
Whose wife and child have you not captured?" ran a soldier's song from
the 14th century.^
In the absence of the institutions of civil society, the principal way for
the ambitious to advance themselves was through warfare. Banditry was
a tiaditional mode ai a frustated local leader would
social mobility:
become a bandit (Amharic: a wfafle, obtaining wealth* a warlike
shifta) for
repolation, and a retinue of like-minded followers, until his ruler was
compelled to bestow a high office upon him. The most feonons example
of the successful use of banditry for political advancement was the case
of Ras Kassa of Quara, who rose to become the Emperor Teodros, ruling
Abyssinia from 1855 to 1868. Some such bandits, such as the Moszagi
brothers in Eritrea in the 1940s, have been described as "social bandits,"^
but undoubtedly the majority were predators rather than protectors of the
poor. The robbery, destruction and general insecurity created by banditry
rendered rural people vulnerable to famine.
Countor-insurgency strategies consisled In attonpting to buy off the
leados of the rebellion with promises of rank and riches, whfle wreaking
destruction on the rural people in the rebellious area. The inhabitants of
an insurgent area withui the empire were treated no differently to a
newly-conquered "enemy" population. One 19th century European
traveller was given an explanation for the scale of destruction: "if an
invasion did not completely ruin a country, the inhabitants would sooner
or later rebel and it would be necessary to send a great zemetia [military
expedition] and start all over again."** Examples of this will be given in
following chapters. The notion that wanton brutality towards ordinary
civilians might make rebel leaders less willing to accept a political
compromise does not seem to have figured In official military thinldqg.
As a result: "a constant enmity exists between fbc military and the
population in general" so that *coontiy peo|d6 sli^ remorselessly any
fugitives of either side from the field of battle."^
* Quoted in: Richard A. Caulk, "Armies as Predators: Soldiers and Peasants

in Ethiopia, c 1850-1935," buemaiUmal Jaurml of African mstarical Studies,
11, (1978X p. 46a
^ Eric Hobsbawm, Bandits, New York, 1981, p. 14.
^ Quoted in Bonnie K. Hoicombe and Sisai Ibssa,

The Invention of Ethiopia:
The Making of a Dependent Colonial State in Northeast Africa. Trenton, NJ,
1990, p. 108.
^ IK^lUam Flowden, Thtvds In Abyssinia, London, 1868.
73
Though most of the discussion has referred to the Christian kingdoms,
Moslem Ethiopian warlord Ahmed
the great al Iman Gran, who led a
jihad (Islamic holy war) against the highland states in the 16th century,
practiced similar fonns of warfare, and was renowned for his lack of
mercy.
Slaving and Pac^kaHon
One of the most OGnmion activities of Abysuiian armies was raiding

for slaves. Vast areas of the country were depopulated, and entire
peoples decimated by slaving raids. Some of those raided included
highland peoples such as the Agau; in the last century, as the highland
states expanded, the slaving frontier expanded towards the borders of
what are now Kenya and Sudan.
The age-old Abyssinian system of settling soldiers on a rural
population, which ww flien required to support tem, was dnmatkally
expanded m Hie late 19lfa oeatuiy. Thisoeamd whea the rtale» oeBieied
in the highlands of Shewa, caqpanded lo conquer the aiea known ai
soathem Ethiopia. The conquest of tiie southem lands, wliich aie taaMy
Qromo, was achieved by massive military r^wpaigng using firearms
provided by European powers. The pacification was achieved through the
so-called neftegna system. Neftegna is Amharic for "rifleman." The
Emperor Menelik paid his soldiers with grants of land — or, more
precisely, grants of the services of the indigenous people who were
required to till the land. An ordinary soldier was awarded a minimum of
two seife; lanldng officeis received tens or even hundreds. By tfiese
means, Menelik was aUe to reward his soldien and also oooliol the
newly-conquered southem regions. The serfs, however, were obliged to
pay half or more of their produce to thek newly-imposed landlords, in
times of hardship, this extra burden meant that the serfs descended into
famine. The practice of rewarding soldiers and other state servants with
grants of occupied land continued until the revolution.
Warfare among Non -Centralized Peoples
Traditional waifue among die Qtobbo fcSktmi a nther different

pattern. TraditionalQromo society was oiganized aocoidmg to a system
of age-grades, known as the gada system. The kadoshlp of the
community chaog^ eveiy eight years to a new age-set Each age set
was required to engage in a war, involving important ritual ekarents*
known as a butta war, before assuming leadership. The butta war could
be a hunting party directed against wild animals, oc a raiding eiîeditioa
24
targeted on a community that had not been raided previously. It was
preceded by a huge feast, in which many animals were slaughtered; one
of the aims of the war was to replenish livestock by raiding.
Such forms of warfare were central to the traditional Oromo religion.
All young men of a certain age grade participated, with women, children
and men of other age grades lemainlQg at home.
When confrooting the armies of Christian Abyssinia, the Oromo
employecl a version of guerrilla warfare. Relatively small and highly
momle bands would utilize night raids to weaken and demoralize
communities. Obtaining booty was also an important component of the
attacks. When the raided communities accepted Oromo domination, and
became part of the expanding Oromo political confederacy, the area
would become the base for guerrilla-style attacks on the adjoining
region.'
The Oromo rules of warfare required that when Oromo groups fought
among themselves, the level of videnoe was oonstiained, and captives
should be returned after a peace agreement was reached. When the
Oromo attacked non-Oromo gjmaps, the level of brutality was certainly
much greater. In general, in contrast to the huge and all-consuming
Amhara armies, the smaller and faster-moviqg Oromo bands would leave
less destruction in their wake.
In response to counter-attacks by the larger and better-armed
"conventional" Abyssinian armies, the Oromo would simply disperse.
The Jesuit Manoel de Almeida noted that the Amhara armies were unable
to invade Oromo territory because the Oromo pastoralists did not grow
food but instead relied on their cattle, which could be evacuated from the
path of an advancing army, wfaicfa could not thetefbie feed itself.^®
In the 18th and 19th centuries^ some Oromo states grew up in the
south-west of Ethiopia, and d6vek^)ed patterns of warfare that were more
akin to those of the northern highlands. An Italian missionary described
the behavior of these Oromo armies when attacking neighboring, stateless
peoples: "When 'foreign' sokUers enter a country, nothing is spared.
^ Mohamed Hassen, The Oromo cf Ethiopia: A History 1570-1860,

Cambridge, 1990, p. 12.
' Mohamed Hassen, 1990, pp. 22-4.
M. de Almeida, Historia de EAiapia a a!ta on Abas^^ quoted in R.

Fankhoist (ed.) Ihtvetters bt EMopia, London, 1965, pp. 42-3.
25
families, villages, agriculture, cattleand everything disappears in a few
days.""
Marginalized people in southern Ethiopia traditionally practiced forms
of warfare akin to mutual raiding for cattle and captives. While both
common and violent, a number ol foctois necessitated a limit to tiie level
of destruction. One such factor was the low level of military technology;
a second was the need for a negotiated peace at the end of the conflict,
so that both parties could resume basic economic activities such as cattle
herding. A
third was the ritual element in warfare. The development of
trading relations with more powerful states to the north, which demanded
slaves and ivory, and the introduction of firearms, upset this system in
historical times, so that the last two centuries have witnessed peripheral
warfare that has been just as bloody and destructive as that seen in the
central highlands, and has at times verged on the genocidal.
Armies and the Creation of Famine
The chronicles of the history of Ethiopia are filled with stories of

famines. Richard Pankhurst, a leading historian of Ethiopia, documents
that between the fifteenth and nineteenth centuries there was on average
at least one famine every decade. Prof. Pankhurst joins the majority
of scholars who study Ethiopian famine and develops the theme that
famine is a natural disaster, caused by drought, pestilence or other kinds
of visitation, working upon a society impoverished by thousands of years
of isolation and technological stagnation. However, in this history
(commissioned by the government of Colonel Meqgistu) Pkof. Pankhurst
makes little mention of the role of warfare, the state's forcible
procurement of produce, or the failure of the state to undertake significant
actions to prevent or ameliorate famine. Ihis is conveniently ckrae to the
official view of the causes of contemporary famines that was propounded
by that government.
The Ethiopian has always been autocratic and unresponsive to the
state
needs of its people. Many
of the sources used by historians derive from
documents written by officials at court, and so cannot be expected to give
an accurate account of the official response to famine. In fact, like some
modem publications on famme, their role is to glorify the ruler and
present him as pious, generous and enlightened.
" G. Massaja, quoted in Mohamed Hassan, 1990, p. 127.
^ Richard Pankhurst, The History of Famine and Epidemics in Ethiopia

Prior to the Twentieth Century, Addis Ababa, 1985. RRQ
26
Prof. Pankhurst disparages "the long-established Ethiopian tradition
of blaming natural calamities on the wickedness of the people."*' Both
historical and contemporary evidence suggests that the opposite is more
often true: famines brought about by the callous actions of powerful
people are cominonly blamed on natme —
espedaiXSy by court
chroniclers.
Requisitioning Food
The historical sources, despite their limitations, tell a story of centuries

of state-created famine in Ethiopia.^'* The manner in which wars were
conducted was instrumental in creating famine. One way in which this
occurred was the requisitioning of food by armies, which provisioned
themselves from the local inhabitants. The Portuguese Jesuit B. Tellez
visited Ethiopia in the 17tfa century, and wrote that famine was common
on account ik visitations of locusts and "the marcfaii^ of the soldiers ...
which is a ]dague worse than the locusts, because thc^ [the locusts]
devour only what they find in the fields, whereas the other [the soldiers]
spare not what is laid up in houses."^ Armies on campaign were
described as leaving almost as much ruin and devastation as if the
country had been invaded by the enemy. The mass armies raised for
discrete campaigns were huge, even by modem standards. Noting that
"an Abyssinian army often numbers 80,000 men, accompanied by 30,000
women, slaves and camp followers," the Englishman Colonel Berkeley
wrote: "it will be understood that it leaves a desert wherever it goes.***
He was writing from experience in the 1860s —
even larger armies were
fielded in the following decades. Even the staff of the Ethiopian Relief
and Rehabilitation Oommission (RRC), reviewii^ the historical evidence
" Fankhuist, 1985, p. 46.
^*
Many of these sources were utilized by Prof. Pankhurst in earlier
publications written under the auspices of independent academic institutions.
B. Teilez, quoted in Margery Perham, The Government of Ethiopia,

London, IS^, p. 163.
^ Quoted In Ht>m cf Africa, 4,1, 1981, p. 29.
27
for famines, were obliged to conclude that many famines were caused
directly by armies' food requirements."
In the 19th century, Emperors Teodros and Menelik both tried to
introduce strategic grain stores for the specific purpose of feeding the
army, but in neither instance was this consistently impkmciited and it is
also unclear how the grain was obtained. Haile Selassie was the first
ruler tomtroduoe a standing army under unified oential control, and to
undertake systematic measures in order to provide for it.
Access to food supplies was therefore central to military strategy. This
has already been noted with regard to Amhara-Oromo warfare. Lack of
food to feed armies played a critical role in the Tigray rebellion of
1913/14. Local Tigrayan nobles wrote to Ras Wolde Giorgis, head of the
Shewan army, imploring him not to invade Tigray to put down the revolt,
because a crop failure meant that the province could not support two
armies. When Wolde GiorsJs did inva&, many of his soldiers deserted
on account of lack of food.^
War against the Economic Base
A second manner in which warfare contributed to famine was the

economic nature of many wars. Wars between Abyssinian principalities
were in part designed to remove the power of the adversary to raise an
army — with the implication that the destruction of grain stores and
other essential items in the economy was a military objective. Thus the
Emperor Yohannes deliberately despoiled Gojjam in the 1880s so that
King Teklehaimanot could not support his soldiers —
instead the Kii^
invaded tfie independent principality of Keffia and quartered his army
there. Wars of conquest, fought mainly against people in southern
Ethiopia, were designed to subjugate the people so that they could be
brought under Amhara domination. Most of these people were Oromo,
and many were pastoralists. The empire required a class of serfs, tied to
the land. This was clearly impractical if the subject populations had their
own independent source of livelihood. Other reasons for confiscating
cattle and crops and destroying villages were desire to break the spirit of
resistance, the need to obtain food for the army, and the tradition of
* RRQ "Food Shortage Report on Tigrai, Annex: A Historical Perspective

on Famine in Tigrai/ Addis Ababa, 1979, p. 39.
National Records QfiBce, Khartoum, File bid 2/19A60.
Peiham, 1969, p. 161.
28
wanton brutality. These coincided with a strategic politico-economic
imperative destroying
of the independent source of sustenance of
conquered people.
Another aspect of the relationship between annies and famine is also
worthy of note. One is that annies weie cairien of disease, and
themselves disease-ridden. Many moie soldieis died from iUnesses
contnM^ on campaign than firam the wounds of battle. In Menelik*s
campaign against Gojjam m 1882, 3,000 soldieis were killed by disease,
as against 900 in battle (plus 50 killed by peasants while looting).^
Local populations would flee before approaching armies for fear of
contracting diseases. In 1913/14, for example, the Shewan army was
responsible for introducing a cholera epidemic to Tigray, at that time on
the verge of famine.
The Greai Famine cf ISSS-QZ
The great famine of 1888-92, popularly known as Kifu Qen, ("Evil

Days") was possibly the woist famine in Ethiopia's recorded history, and
some estimate that as much as one third of the population died. A
rinderpest epidemic decimated cattle herds, while a combination of
drought, locusts and army worms damaged crops. The actions of the
state, however, were also important hi deepening the famine for many
sections of Ethiopian society, while helping others to survive aiKi even
benefit.
The famine struck at a time when the Emperor Yohannes was fighting
the annies of the Sudanese Mahdi, and when King (after 1889, Emperor)
Menelik was waging large campaigns of destruction in the south.
Menelik was able to procure grain to help feed his new capital at Addis
Ababa, so that it was described as a "Noah's Ark*" amid general
devastation. It appears that much of this was obtained by the confiscation
of "concealed" grain in the rural areas.^^ A
common traditional storage
method uses underground pits —
the grain stored there is not a hoarded
surplus, but the family's diet. Sources do not reveal whether this was the
origin of Menelik's famine relief for Addis Ababa, but it is probable that
rural people suffered so that townspeople could eat. Magazines were set
up to feed aimy —
by levying a new tithe on the peasantry, which
contemporary dnomdeis optimiAcaDy itcofded as "popuhir." The
^ aulk. 1978, pp. 476-7.
^ Pankhuist, 1985, p. 101.
29
Emperor also made some symbolic
gestures such as no longer eating meat
at his court, and going with a hoe.
to a field to dig
However, Menelik's chief response to the famine was to try to export
it by invading his southern neighbors. Amhara lords reacted to the
famine by plundering the rich Ctomo province of Arsi (tfuee diffeient
governors participated) and raiding cattle in the Qgaden. Menelik
authorized the sending of northern garrisons lo sovtfaeani towns such as
Bure and Nekempte to be fed by the local population —
presumably
without the latter's consent.^^ The Gojjami army occupied Keffa to feed
itself there, and thousands of Shewans migrated to the highlands of
Harerghe.
Yet, if be believed, "the reaction of the Emperor
official sources are to
Menelik emergency was one of the few bright spots in an
to the
otherwise gloomy picture."" The despot was, we are told, "very
distressed to see his army famished**^ and ordered provincial governors
to supply famine relief to the troops —
the sources of supply of this
generous relief are not mentioned. Parallels with the late twentietih
century need not be stressed.
An Outline of VulnerabiUty to Famine in Etiiiopia
This section consists in a discussion of the socio-economy of rural

Ethiopia, with a focus upon Tigray and the adjoining areas. This analysis
is important because it provides the framework in which the
governmental actions of the 1980$ can be appreciated. Hie government's
counter-msurgency methods tore at the very sinews whidi kept the rural
economy together, turning a period of hardship into one of outright
famine. Some of those sinews are obvious, such as peasant farming,
some of them are less obvious. It is necessaiy to analyze how they held
the economy together in normal times.
The most common view of the socio-economy of Ethiopia centers on
the relationship between the soil, the peasant and those who tax the
peasant. The focus is upon the agricultural system. In the central and
Richard Pankhurst and Douglas Johnson, "The Great Drought and Famine
of 1888-92 in Northeast Africa," in D. Johnson and D. Anderson (eds.) The
Ecology of Survival: Case Studies From Northeast African History, London,
1988, pp. 56-7.
^ Pankhurst, 1985, p. 98.
^ Pankhurst, 1985, p. 101.
30
Copyrighted matsrial'
northern highlands, farmers grow a variety of crops in the spring {belg)
and summer {meher) growing seasons (Eritrea and much of Tigray
receive only the meher rains). Ox -plows are used to prepare the land,
using technology that has changed little over millennia. Ethiopia is one
of & origjiuil lod of oeceal cultigens, and llim Is a huge variety of
stiaiiis of each of the major crops: teff (the most piestigious staple),
wheal, bailcy, millet, sorghum and maize.
Whfle agriculture is central to the Ethiopian economy, the view that
focusses upon it to the exclusion of other aêcts of rural life can be
misleading. James McCann, a leading authority on the northern
highlands, considers that "the image of the insular, long-suffering
Ethiopian peasant" and the view that sees "highland farmers and highland
agriculture as static and self-contained ... [have] obstructed understanding
of the rural economy and social history of northern Ethiopia."
Ethiopian highland peasants do not survive just by growing things in the
fields — mignition, trade and animal reariog are impottant too.
Rather than concqitiializiiig rural Ethiopia as an agglomeration of
independent peasants each provisioning himself or herself from frummg,
it is more useful to see the region as a pattern of geographically-
specialized areas, with a set of links between them. Hiis is particularly

the case for the north.
The fundamental dichotomy is between areas which are normally
surplus-producing, and areas which are regularly in food deficit. In the
north, the surplus areas include: Gojjam province, southwestern Wollo,
central Gonder and Simien, Raya district in eastern Tigray and Shire
district in western Tigray. These areas are not particularly drought-
prone; even when drought hits other parts of the counliy, ttif noimally
continue to produce adequate crops or surpluses. When diew areas do
suffer a partial crop faflure, it is usually localized and may be due to
hailstorms, frost, infestations of pests sudi as locusts or anny-wonns, or
indeed too much rain.
The food deficit areas include most of the rest of northern Ethiopia,
especially some parts of eastern Gonder, northern and eastern Wollo,
most of Tigray, and almost all of Eritrea. They form an arc along the
eastern escarpment of the highlands, with an extension into the Tembicn-
Wag area, where there is a ram shadow on account of the Simien
mountams. Many of these areas have been in chronic deficit for a
hundred years (smce the great frunme of 1888-92). They produce a good
^ James McCann, From Poverty to Famine in Nartheost Et/Uopia: A Rural

History 1900-1935, Fhiladdphia, 1987, p. 69.
31
harvest perhaps once or twice a decade, and a nm
of six or seven years
of crop failure is not unusual.
While adequate rains and the absence of pests are important to
production in all areas, for many farmers
the critical factor causing them
to fail to grow adequate crops lack of Qxen to plow the laid. The
is
more times a field is plowed, the higher the yield. Investigations have
shown that in poorer areas of northern Ethiopia, approodmatdy one third
of the farmers have one ox, and one third have none —
only one third
have two or more. A
farmer with one ox is traditionally known as '*half
a man". He must team up with a neighbor in order to put together a
plow team, which then divides its time between the farms. A farmer with
no ox must try to hire a plow team, and can usually only afford to make
one or two passes with the plow, instead of an optimal higher number.
The single most important short-term constraint on higher food
production is shortage of plow oxen.
Oxen are expensive. Fanners must save in oider to buy one, and the
loss of an ox is a devastating blow to a poor hooaeiiold.
Treated as a whole, northern Ethiqna is very rarely m
overall food
deficit. Conceding that statistics are very unreliable, perhaps the only
year in recent times when there was such a deficit was 1984. In all other
years between 1975 and 1991 there has been a net surplus, though
sometimes a small one. The problem is not food availability, it is food
distribution.
This leads to the question: how do the people who live in the deficit
areas make good that deficit? Hie answer is, through four basic means:
* Working for money. Large numbers of rural Ethiopians undertake

paid labor, either dose to home, or migrating to surplus-producing or
coffee-producing areas, cities, or farming schemes. Migrant laborers
buy food where it is cheaper and more plentiful than in their home
areas, in effect redistributing it from surplus to deficit areas.
* Engaging in petty trade. In the seasons when agriculture is not

possible, many men engage donkeys and mules. One
in trading, using
researcher in the village of Adiet, near Axum in Tigray, found that
more than half of the luhih men in the village were part-time trados,
in salt, grain, animals and consumer goods. Some travelled as fur
afield as Gojjam to buy giain.^ For the pastoralists of the lowlands^
^ Richard Baker, interviewed by Alex de Waal, November 1990.
32
trade is even more important. By these means, food is moved from
surplus to deficit areas.
* Selling aniflials.Not only the specialist pastoialisls of the lowlands,

but many highland farmers, earn a considenible income from selling
animals. Animals are kept for sale and for milk. One (tf the few
investigations of the role of animals in the highlands was done by
Noel Cossins and Bekele Yemeron in the 1970s.^^ They concluded
that in many areas, animals played a role at least as important as
farming. One of these areas was the Tembien-Wag area in Tigray-
Wollo, which was the epicenter of the famine of 1983-5. In the
Eritrean highlands thisis also the case, and many farmers are in fact
semi-nomadic, migrating with their animals out of the highlands to

the eastern and western escarpments.
* Remittances from relatives woildng in towns and abroad are

important, paiticulaily in which has larger
Eritrea, cities, more
industry, and more ready access to neighboring counUies.
Other such as eating wild foods and borrowing, are also

strategies,
commonly Another strategy followed is going hungry: rural
followed.
people will reduce consumption in order to preserve essential assets such
as seed, farm tools and plough oxen. The meager diet upon which
Ethiopian peasants are able to survive appears to contradict the basic
tenets of nutritional science, but is remarked upon by all who have
studied famine survival strategies.
A Digression Into Statistics
There arc no reliable statistics for Ethiopia, especially the northern

regions. This is for three main reasons: (1) there have been no systematic
investigations of most aspects of the rural economy of the north, in part
because of prolonged warfare in the area, (2) even the most basic facts
are subject to political manipulation, and (3) peasants are reluctant to
divulge infoimation to outsiders.
Infonnation is power. The Imperial govemment withheld information
about famines, and news about the famines in the post-revolutionary
period was subject to both censorship and diôrtion. Similar
^ Noel J. Cossins and Bekele Yemeron, "Still Sleep the Highlands: A Study
of Farm and Livestock Systems in the Coitral HighUmdB of Ethiopia," Addht •
Ababa, Livestock and Meat Board, 1973.
33
considerations apply to basic facts about ruial production, rainfall, and the
human population.
For many years, even rainfall data were regarded as a state secret fthe
logic behind this will become evident in this report). In addition, due to
the war, many monitoring stations ceased to function.
rainfall
The size of the human population is not known for sure to within
several millions. In part this is due to problems with oountiiig people.
Some peasants remain totally unreached, hiding in the mountains: "since
the state provides very little that is beneficial to the peasant but siphons
off a good deal of the lattcr's harvest, a large number of these mountain
peasants will soon be driven to break their ties with the outside world,
and to retreat to their rugged fortresses to live a life of independence in
sublime isolation. "^^ When the first national census was conducted in
1984, the enumerators found 29 per cent more people than expected in
the areas in which they were able to survey — suddenly, an extra eight
million Ethiopians were "discovered."
The most contentious population issue in noithem Ethiopia is the
question of the number of people in Tigray. This was an issue of dispute
throughout the 1980s. On one side, the government claimed that Tigray
contained between two and 2.8 million people. On the other, the TPLF
claimed that figures collected from village committees indicated a
population of between 4.5 and 5 million. In 1989/90 the population
estimates were 2.73 million (government) and 4.82 million (TPLF)
respectively. This issue became important in 1985 when the government
claimed that most of the famine victims in Tigray were receiving rations
— a claim that could only be made if the government's 1984 population
figure of 2.41 million was accepted.
Demographic data from Ethiopia are extremely poor. Nobody knows
how many people there really are in Tigray. However, the evidence that
is available does allow a set of preliminary estimates to be made. These
show, that while the population may not be as high as claimed by the
TPLF, it is undoubtedly much higher than that claimed by the Mengistu
government. There are several elements to the revision of the population
estimate.
I TPLF-defined Tigray includes parts of Raya Kobo, IVwlemti, and

Wolkait, which are in government-defined Wollo and Gonder.
According to the TPLF in 1989, these extra areas contained
^ E>essaiegn Rahmato, "Famine and Survival Strategies, Case Study from

A
Northeast Ethiopia," Addis Ababa, Institute for Development Research, 1987, p.
51.
34
400,000 people. The following discussion will be confined to the
smaller government-defined Tigray.
II Hgray was not eoimientBd in the 1984 census, because of the war.
Hie population estimate given by the government was therefore
derived from the 1969 National Sanqile Survey, which estimated
a population of 1^6 million, with 2.7 per cent per annum added
on for population growth. However, the results of the 1969 survey
were suspect. The figure first announced was 1.36 million, down
from the estimate based on the 1965 sample survey, of 1.41
million; in January 1970 it was revised upwards. A
more detailed
survey in the early 1970s estimated the population at 2.04 million,
but internal evidence in the survey results indicate that is likely to
have been an underestimate.^
A major source of inaccuracy for the figures for Tigray was the
large number of temporary outmigrants from the province,
probably more tiian ten per cent of the total population. After the
revolution, most of this temporary outmigration ceased, boosting
the local population accordingly. Thus it is likely that in 1969 the
real population was at least 2.2 million.^ Hiis would imply a
1989 population of 3.65 million.
III In the areas where enumeration was possible, the 1984 census
found an additional 29 per cent of people compared to estimates
derived from the 1969 survey.'^ It is safe to assume that, had
enumerators actually visited Tigray, they would have found an
Hunting Technical Services, "Tigrai Rural Development Study, Annex 3:

Population," fiorehamwood, 1975. 'UiA survey found a relative sliortage of
young girls.
^ It is notable that Ac 1970 language micvey found 356 million Tigrinya

speakers, about 400,000 more than could be accounted for in the total estimates
for the settled Tigrinya-speaking populations of highland Eritrea, Tigiay and its
borderlands, Addis Ababa and elsewhere.
* The census was conducted shortly after the govemment began a program
of sêmatic conscription of young males into the army. The data show a
leUtive shortage of young males —
pcesnmahly concealed from the census
enumerators by themselves and their fionilics.
35
additional number of people. An additional 29 per cent on the
official figure would imply a 1989 population of 3.62 miUioa.
rV The results of the 1969 survey were manipulated for political

reasons. The most well-known example of that was the
undeiestimaticm of the mmiber of Oromo. Though the Qromo
actually cmtmmibeied the Amhaia, this poMtically-seBflitive fiict
was suppressed, and the results dained tfiat 7J^ miUkMi Amhaca
outnumbeied 6.8 million Oromo. There were also allegatkna
(never systematically investigated) that the total population of
Tigray was also reduced for political reasons, and that the district-
by-district totals added up to much more. As this possible
distortion cannot be proved or quantified, it will not be considered.
Factors II and III should be combined. Even if the corrected results

of the 1969 survey had been known, the 1984 census would still have
found more. Combinmg factors II and m lesnlls m a total 1989
population estimate of 4.70 millioa. A
more oonsefvative estimate would
assume that factor m operated at "half stieiigth", which would imply a
total 1989 population of 4.20 million (for government-defined T^pray)
and 4.60 million (for "greater" Tigray).
This shows that in 1989/90 the government of Ethiopia underestimated
the population of Tigray by at least 1.4 million people (i.e. one third).
Figures for rural production are highly suspect. Most estimates for
areas of farmland derive from tax records, which are arrived at by
processes of negotiation, influence and bribery. One scholar studying
land before the revohitioa noted that oo^ two per cent of ttub fnndand
in Tigray had ever been measured, which was better tfaam 6ojjam (0.1 per
cent) and Oonder (none at all). Ife oonchided: "on two fionta, ownerBhip
and boundaries^ these farmers have succeeded preventing thek
government from kaming the substantive details that would allow any
forceful application of the land taxes."'^ As a result, accurate
knowledge of farmed area was non-existent. After the revolution, as the
Agricultural Marketing Corporation (AMC) gradually extended its
procurement in the late 1970s, sample surveys were begun in selected
villages to assess the crops. In WoUo, the assessments of cropped area
and harvest, which had been approximately stable up to 1978/9, suddenly
more than doubled in 1979/B0» remaining at compand>le levds tfaereafker.
^ Peter Schwab, Decision-MMig in EOUapia:

Process, London, 1972, pp. 30-1.
A Study of the MBM
36
It is likely that pressure to identify surpluses available for procurement
by the AMC influenced the sudden increase in harvest estimate. In 1981,
the Food and Agriculture Organization of the UN (FAO) gave technical
assistance to the Ministry of Agriculture, enabling it to double the number
of surveys. The result was a huge increase io te esttnurted yield.
Ovenrnling objectkms fimm the slitisddaiis who wanted to double-check
the figures, the central government published the unchecked first estimate
— because the hi^ figure could be used to demooBlnte the ckdmed
"success" of the government's newly-launched campaign to increase
agricultural production.
Figures for amounts of grain in storage or numbers of domestic
animals are even more unreliable — the information must be obtained by
asking peasants, who are reluctant to divulge anything to an outsider who
is likely to be seen as wanting to assess taxes, recruit soldiers, confiscate
land, or otherwise interfere with the peasant's autonomy. Official

statistics are thus politically loaded facts based on varying degrees of
ignorance about the reality.
Figures for those "in need" on aoooont of famine are similarly suspect.
Figures from the government Relief and Rehabilitation Conmiission were
based on village reports (in government-held areas) and gross estimates
elsewhere. In 1983, the total for Tigray was exactly one million; for
Wollo, it was given down to the last individual. Examples of political
manipulation of the figures will occur in this report. Figures produced
by the FAO are based upon satellite imagery of vegetative growth and
ground surveys. Satellite imagery of vegetation is useful for assessing the
extent of drought, but cannot evaluate the impact of war^ trade disrup-
tions, taxation or pests, limiting its value for Ethiopia in the 1980s. In
addition, throughout the 1980s, UN teams could not travel to rural areas
of Eritrea, Tigray and north Wollo to coodnct ground checks.
Ethiopian peasants are notoriously miwiUing to divulge even the most
basic information to outsiders. This stems from the well-founded fear
that information is liable to be used against them. One of the peasants'
few weapons against excessive exactions by the government is the latter's
ignorance about their real condition, including how many of them there
are. Several rural revolts (notably in Gojjam) have been sparked by
government attempts to measure farm land. Peasants prefer the "sublime
tsolatioo" alluded to above.
The Ethiopian peasantry cannot be counted against its will. Hitherto,
that win has afanost ahvays been bckmg. Only with the advent of famine
relief and more democratic forms of government are reliable statistics
about the number of Ethiopians and their condition likely to be produced.
37
One tendency is certain: the closer an investigator probes into the rural
economy of northern Ethiopia, the more he or she finds —
whether it be
people, farmland, animals, or any other form of resources. One fact is
equally certain: any attempt to quantify any aspect of Ethiopian society
is a hazsudoiis enterprise; and moae vAo pot Qgnies an Ihefr dainw tot
things such as rural production, the extent of enWronmeotal dogndation,
or the mmibeis dead in a fiunme, are spealriqg eidier in tgnonmoe of the
truth, or with tlie intention of concealing ft*
38
2. SCORCHED EARTH IN ERITREA, 1961-77
Background to the Conflict in Eritrea
The political backgiound to the war in Eritrea has been studied many
timeshy scfaolaciaf ditfeientleaiiingB. Before conquest by Italian forces
in the late 1880s, the history of Eritrea is controversial — scholars disrate
whether significant areas ctf it were tributary to states in Ethiqîa proper,
or not.
For half a century Eritrea was an Italian colony, until occupation by
the British army during the Second World War. There then followed a
decade of political turmoil, as the post-war powers dithered about the
future of the territory, and the Eritreans organized themselves into a welter
of political parties which campaigned for nationalist, sectarian or unionist
causes using fair means and foiiL For the most part, the Mgbhmd Christian
population was sympathetic to union with Ethiopia, whereas the lowland
Moslem population wanted ittdependenoe. Some of tbe lowland peoples,
notably the Beni Amer tribe which straddles the Sudanese border, were
influenced by the growing nationalist movement in British-controlled
Sudan. After several false starts in trying to determine Eritrea's future,
the Alliedpowers turned the issue over to the United Nations, which
appointed a Commission of Inquiry to investigate the wishes of the
Eritrean people.
When the UN Commission of Inquiry made its investigation, there was
widespread rural unrest, which made it impossible to visit many lowland
areas in western Eritrea. The investigaticm was also cursory in the
extreme, and no referendum was held. Meanwhile, an Eritrean assembly
was elected, by indirect votii^, and was approxunately evenly-divided
on the crucial issue of independence or unity with Ethiopia. The Ethiopian
"liaison office" headed by Colonel Negga Haile Selassie financed the pro-
Ethiopian Unionist Party, mobilized the Orthodox church on his side, and
intimidated the opposition. These tactics served to alienate much of
Eritrea's elite.
Finally, US strategic interests proved the decisive factor. The US was
a dose ally of Haile Selassie, and wanted the use of the strategically -sited
air foiceaklcoiimiunicatioiis base at Kagnew, in Asmara. Inl952,Eritoea
was given aconstitution wfaidi mchided a democratically-elected assembly
and&ekey mstitutions of dvil society, but federated iindier the Ethiopian
crown.
99
Haile Selassie reportedly neither understood nor approved of the notion
of federation/ and at once set to work annexing the territory. The
Eritrean administration was stripj>ed of its powers and the Assembly was
undermined. The Emperor dismissed and appointed Eritrean ministers.
A staê by the Eritrean labor unioiis in 1958 was met with vioknce —
over 200 strflcers were detained and 60 injured by soldieR. Finally* in
1962, coerced and bribed, and with the building surrounded by soldiers,
the Assembly members voted to dissolve the Federation.
Throughout the 1940s there had been widespread shifta activity in the
lowlands, with up to 3,000 bandits active. This abated after the British
offered an amnesty in 1947, but by the mid 1950s many former shifta
returned to violence. When Haile Selassie's intentions were clear but the
act of annexation not yet consunmiated, leading pro-independence
Eritreans fled into exile and formed the Eritrean Liberation Front (ELF)
in 1961. They made contact with sh^ leaders, and armed revolt began
in September of that year. Most of the membership of the ELF was
Moslem, from the western lowlands, and in particular from the Beni Amer.
The government deepened Eritrean opposition by continuing to
dismantle Eritrean institutions and deprive Eritreans of basic civil and
political rights. There were also a number of political detentions and trials,
including one of 19 government officials accused of supporting the ELF,
which was transferred to Addis Ababa in 1963 because the government
believed that no convictions would be obtained in an Eritrean court. Even
those Eritreans who had previously been ardent supporters of the Union
became disillusioned. For example, Tedla Bairn, the first Eritrean prime
minister, defected to the ELF.
Eritrean Resistance 1961-74
Thefirst shots in the Eritrean war were fired on September 1, 1961

in the lowland district of Barka. The leader of the ELF detachment that
engaged the Ethiopian soldiers was a well-known QX-shifta called Idris
Awate. Most of his band of eleven followers had previous military
experience in the Sudan Defence Force.^ Over the following few years
the political and military leaders operated largely in mutual isolation. Idris
Awate was killed in 19iS2, but the ELF ooatinued to grow on the ground,
and in the mid 1960s it had over 2,000 fighters.
^ John D. Spencer, Ethiopia at Bay: A Personal Account of the Haile Sdassie

Years, Algpnai, Midi., 1984.
^ The British-run colonial araiy of Sudan.
40
Prominent political leaders of the ELF included Woldeab Woldemariam,
Osman Saleh Sabbe and Idris Mohamed Adam. In 1963 a ten-member
Supreme Council was established in Khartoum. Finance, armaments and
training came at different times from most Arab countries, including Syria
and Inq — under whose Baatliist ideology Eritrea was inoQipocated into
the Aiab world. In tfie late 1960s, the ELF received assistance firom Cuba
and other Marxist states.
Sudanese support was important but not unwavering. While public
opinion in Khartoum and Kassala was strongly pro-ELF, successive
Sudanese governments followed vacillating policies: support during 1961-
4; increasing coolness during 1965 (13 ELF fighters were extradited to
Ethiopia that year and some were subsequently executed); and coldness
from 1967-9, during which time ELF political activity was prohibited.
During the early radical phase of Colonel Jaafar Nimeiri's rule (May 1969-
July 1971X Sudanese support for the ELF was again strong; it was then
cut off m response to an Ethiopian withdrawal of support fat the Anyanya
rebels in southern Sudan. Relations between Sudan and Ethiopia then
deteriorated after the 1974 revolution, reaching a nadir when the ^hiopian
government backed an attempted coup in Sudan in July 1976. In response,
the Sudanese government followed policies strongly supportive of Eritrean
independence — for two years. The most important role played by the
Sudan government was allowing the Eritrean fronts to transit supplies,
including military equipment, through Port Sudan.
In 1965, the ELF reorganized itself into five operational zones along
the lines of the Algerian Front de Liberation National.
Ihe ELF was an amalgam of different elonenlB — indeed, the zonal
structure of 1965 was adopted partly in order to prevent conflict between
different regionally-based elements within the oiganization. Many
substantial conflicts within the organization were left unaddressed. In the
early years, the membership was almost entirely Moslem, and the
Mganization had a strong tinge of pan-Arabism. A position often adopted
was that the liberation of Eritrea should precede social or economic
transformation. In the late 1960s there was a growing radical element
within the ELF, leading to the adoption of Marxist ideas.
Much of the insurgent activity in the 1960s and early 1970s involved
ads of sabotage against government installations and Imdges, and
ambushes of convoys and trains. Ihe ELF also regularly engaged army
patrols, and attacked small garrisons and police posts. It quiddy made
most of the lowland countryside impassable to govenmient forces except
in military convoys.
The ELF was anxious to avoid the mistakes of the Palestine Liberation
Organization, and was reluctant to let people evacuate to Sudan in large
41
numbers. ELF leaders recognized that life in refugee camps would breed
a generation of young people bitter and addicted to violence, and argued
that this would undercut the revolutionary idealism necessary for their
membership.
The first of a series of sp^
in the ELF oocuned In 1968. Osman
Saleh Sal>be, a prominent exiled leader, departed and founded his own
organization. Ibrahim Tewdde (lator poisoaed in spspidoos
circumstances) and Isseyas Afeworki founded the Tripartite Union. This
led eventually to the formation of the EPLF in February 1972» led by
Isseyas and Ramadan Nur. The ELF, under the military command of
Abdalla Idris, denounced the new organization.
There then followed two years of internecine strife between the two
movements, which caused an estimated 3,000 casualties among the fighters
— more deaths among the combatants than the entire thirteen years of
government military actions to date. Purges within the two fronts also saw
a number of people murdered. A
battle between the two fronts at Wolki
in October 1974 left 600 fighters dead, and caused a spontaneous
demonstration by the citizens of Asmara, who marched to the battlefield
and demanded that the fronts settle their differences by negotiation.
Counter-Insurgency 1961-74
Idris Awate's personal history as a shifta leader, the background of the

other guerrillas, plus twenty years recent experience of intermittent banditry
in western and southern Eritrea, enabled the government to regard the
threat as one of straightforward brigandage utiliased for personal
advancement by Moslem sectarians.
Throughout the 1960s, the counter-insurgency strategy followed
consisted of punitive patrols, interspersed with large offensives and a
policy of forced relocation in fortified villages. This was a colonial-type
technique designed to impress the subjugated population with the firepower
and determination of the government. It led directly to large-scale human
rights abuses.
The actions of the army also led to the impoverishment of rural people,
food shortages, and famine. The soldiers requisitioned food, destroyed
crops, killed animals, prevented trade, drove people firom their land, and
(during 1975) blodcaded the entire highlands. From tfie very start of the
war, the Ethiopian army used hunger as a weapon.
Starting in September 1961, the day-to-day brutality of army patrols
caused many abuses of human rights. Dawit Wolde Giorgis, later head
of the RRC, served as a mflitary officer at the time. While fervently pro-
42
unity (i.e. anti-Eritrean independence), he was critical of the army's
operational methods:
The anny ... entered Eritrea in the 196Qs with flie mentality of a
conqueior. It belittled the small bands of Moslem separatists opeiatiqg
in die lowland areas and believed it conld command respect and loyalty
fimm the people by sheer show of foice.... The army made a crucial
error in this operation; it did not concentiate QO attacking the guerrillas
directly; instead it devastated the villages suspected of iiarbohog them.
Another disastrous decision was that the army would carry only two
or three days' rations at a time. After they ran out, they were expected
to live off the land, to take what they needed from the people. I
remember soldiers slaughtering cattle, eating what they wanted, and

then leaving the rest to rot. Sometimes soldieis would kill cattle just
to get the livers. It was as if they were in enemy territoiy.^
Haggai Erlich, an Israeli scholar, commented that "usually in Eritrea

a company would leave for a month-long routine counter-insurgency
operation carrying almost no supplies, which almost inevitably turned them
into shifta in uniform."^
A more effective aspect of the Ethiopian counter-insurgency operation
was the training of an elite commando force, with Israeli assistance.
Known as Force 101, this unit became operational in 1965.
Scorched Earth: The 1967 lOUings
The first large-scale abuses of human rights occurred during three amiy
offensives in 1967. The governor of Eritrea, Ras Asiate Kassa is reported
to have boasted he would leave Eritrea as bare as his bald head.
that
Certainly, throughout the year thearmy behaved as though the depopula-
tion of the Eritrean countryside was its aim.
The first offensive began at dawn on February 11, when a force of
about 5,000 soldiers began burning villages in Barka district. An Israeli
advisor wrote in his diary:
' Dawit Wolde Giorgis, Red Tears: War, Famine and Revolution in Ethiopia,
Tienton, NJ, 1989, p. 82.
^ Haggai Erlich, The Struggle over Eritrea 1962'197S: War and RevobttUm
in the Horn Africa, Stanford, 1983, p. 39.
43
The Second Division is very efficient in killing innocent people. They
are alienating the Eritreans and deepening the hatred that already exists.
Their commander took his aides to a spot near the Sudanese border and
ordered them: "from here to the north —
clear the area." Many
innocent people ivere massacied and nothing of substance was
achieved/
Between February and April the army binned 62 villages, including

Mogoraib, Zamla, Ad Ibrahim, Qetsei Gingiir, Adi Bera, Asir, Fori and
Ad Habab, while villagers were ordered to collect at army posts for
screening. The soldiers were assisted by artillery and aerial bombardment
using incendiaries. According to reports from local community leaders,
402 civilians were killed, and about 60,000 cattle and camels slaughtered
with machine guns and knives and by burning alive.^ In addition, 21
detainees, most of them teacfaeis and government employees, were
summarily executed in Tessenei prison on February 12. Traders were
singled out for detention or kiUmg, because they were believed to be
responsible for supplying food to the insurgents.
Between March and May, 25,500 Entrean refugees were registered in
Sudan, and an estimated 5,000 more crossed the border and stayed
unregistered in Beni Amer villages. The Ethiopian government objected
to humanitarian assistance, calling them "rebels not refugees."^ On March
15, the Sudanese village of Guba, close to the border, was attacked by the
Ethiopian army on the suspicion that ELF fighters were present. One
villager was shot dead and three were kidnapped. A
nearby village, Debre
Sultan, was also attacked. Inunediately before the offensive began, Haile
Selassie had visitedSudan to ensure that the Sudanese govenunent would
respond appropriately: it duly withdrew the refugees fitom the border area
and prohibited the activities of the ELF, but succeeded in obtaining
humanitarian assistance from the United Nations High Commission for
Refugees (UNHCR) and setting up a refugee camp.^ This was UNHCR's
^ Quoted in: Eriich, 1983, p. 58.
* Mohamed Hamed, The International Dimension of the Battle for Eritrea,

Beirut, 1974, pp. 126-7 (in Arabic).
^ Ahmad Karadawi, "Refugee Policy in the Sudan 1967-1984," DPhil Thesis,
Oxford, 1988, pp. 57-9.
' Karadawi, 1988, pp. 62-4.
44
first involvement in the Horn of Africa, and the first of many refugee
flows over the following 25 years.
After these atrocities, a large area of Barka was left uninhabited.
In July 1967, the army turned its attention to the eastern lowlands of
Semhar. The villages of EOet and Omnliot woe buioed oo July 11, and
30 young men weie tied np and burned alive inside a house. Five other
villages were burned over the foUowiqg days,' and 51 people killed.
6,000 domestic animals were lolled; according to reports, the soldiers
singled out camels for slaughter, because they were vital for transport.'"
The third offensive started in November 1967. 7,000 soldiers from the
Second Division began to bum villages in the vicinity of Keren. Almost
all the villages of Senhit — 174 in all — were destroyed. Some reliably
reported atrocities included:
* Kuhul and Amadi: the army ordered the people to collect in one place,
where they were bombed by air force plan^
* Asmat: the army q>ened fire on a weddiqg party, Idllmg an unknown

number.
* Melefso: thirty community leaders who met the soklieis and offered
them hospitality were killed.
The army burned crops, killed or confiscated livestock and poisoned

wells, with the clear intention of making the area uninhabitable.
After Senhit, the burning of villages spread to the highland districts
of Seraye and Akele Ouzai, which had been hitherto little afifeded by tiie
war. 86 villages were burned in these two districts, and at least 159 people
killed.
Forced ReloaUion in Fortified ViUages
In December 1966, the army began a policy of forcible relocation into

fortified villages among the pastoral population of Barka. After each of
the three 1967 offensives much of the population was relocated in new
villages or forced to live in nearby towns.
' One of those destroyed was Sheib, the scene of massacres in 1988 and 1989
(see diapter 14).
^ Mohamed, 1974, p. 13L
45
In the villages, curfews were imposed, together with
fortified
restrictions on daytime movements. This had a profound impact on the
livelihood of the herders. In normal times, pastoralists leave their village
for days or weeks during the dry season in order to search for grazing;
if they are confined to a radios of half a day^ walk firam te village, they
can only keep a soiall nuoiber of animals on tlie limited fptu and biowae
of this area. In addition, because of daytime heat duimg certain tunes
of year, herders prefer to graze and water animals at night TUs is better
for the animals' liealth and means they drink less water — water is a
scarce and expensive commodity in Eritrea. The policy of curfews made
this night-time grazing impossible. On occasions when the strategic
hamlets proved too difficult to maintain, the army forcibly relocated the
population in nearby urban centers, with even more drastic consequences
for their livelihoods. Strict controls on trade were established, with the
inspection of all food items brought in and out of villages and towns by
vehicle or pack animal.
However, as the ELF regained control of the countryside, the people
began to return to their previous homes.
The 1970-1 Massacres
In January 1970,army re-launched the policy of forcible

the
villagization, coinciding with a large militarysweep through Barka and
Sahel. Tens of villages were burned. In March, 16,700 refugees fled to
Sudan, mostly to the Tokar area along the Red SetL coast. The period from
April to July witnessed many atrocities as military activities spread to the
eastern lowlands. 32 dvflians were shot dead wliea tlie army binned
Arafali village. 88 people were executed when the people of Atdioma,
between Massawa and Ohinda» refused army orders to relocate in a
protected village.
Violence against the civilian population increased towards the end of
the year. In retaliation for an ambush in which a senior army commander
was killed, there was a series of massacres near Keren. One atrocity was
the killing of 112 people in a mosque at Basadare in November. The
people were collected in the mosque by soldiers who said they would be
safe therefrom a planned air strike; the soldiers then opened fito. Anollieff
was the destruction of the village of Ona on December 1, in whidi an
estimated 625 people were IdUed. This was apparently in retaliation fx
the villagers betraying the presence of an army patrol to flie ELF^ which
46
ambushed it and killed several soldiers. Acting contrary to official
orders, local boy scouts re-entered the village to bury the dead.
In December, the military governor imposed a state of emergency on
Eritrea, and assumed powers of arbitrary arrest, and the right to displace
the population. Ten-kilometer strips of land along the coast and the Sudan
bolder were dedaied "prohibiled ziODes»" wlieie tiie anny could open fite
on any target. On January 27, 1971» about 60 civflians, most of them
elderly people, were killed by soldiers in a mosque in the village of
Bafcored.
As well as being ruthless and violent, the administration was described
— even by those sympathetic to its cause —
as "inefficient, brutal and
corrupt".^^ Senior officers sold the supplies destined for their units, and
the governors (the civilian Ras Asratc Kassa and the soldier Lt-Gen
Debebe Haile Mariam) and their associates are alleged to have often
detained prominent citizens in order to obtain large bribes for their release.
Human Rights Abuses by the Eritrean Fronts
Both the ELF and EPLF committed abuses against human rights during
this period. At first, the ELFs Moslem leadership was intolerant of
Christians. Some of the first Christians to attempt to join were summarily
executed, as were members of a group of university students known as
Siriyat Addis. The defection of a large number of Christian fighters in
June 1967 presaged the splits that were to destroy the organization in the
1970s. In 1968, the Tripartite Union breakaway group was subject to
harassment from partisans of the mainstream ELF based in the western
lowlands, and two members were arrested and killed.
After the army atrocities of 1967, there was a widespread demand in
the ELF rank and fOe for retributive attacks on civilian targets. In
response, the leadership set up a special unit to engage in hijadcings of
airplanes. Civilian airplanes were hijacked at Fraidcfiirt in March 1969
and Karachi in June 1969. In neither incident were any passengers injured.
The ELF also set up a special unit known as "Quattro Cento" after
the death penalty in force during the Italian occupation. Its task was the
assassination of civilians associated with the government. Gwynne
Roberts, a journalist who travelled with the ELF in mid-1975, estimated
that at that time the ELF were carrying out about 15 assassinations each
" Godfrey Morrison, Eritrea and the Southern Sudan: Aspects of Some Wider
African Problems, London, Minority Rights Group, 1971, p. 29.
^ Erlich. 1983. p. 39.
47
week. Those accused of collaboration were sent two warnings, and if they
ignored both, an assassination squad was dispatched. Roberts also saw
numerous civilians who were detained by the ELF for having failed to give
them assistance, or on suspicion of collaborating with the government.^^
The Tripartite Union and the EPLF also carried out assassinations. On
April 14, 1970, two judges who had earlier senleiioed rebel fighteiB to
d^Kth were shot dead in a bar m
Asmara together with fimr oilier peo[de
presrat.
The ELF taxes and enacted reprisals against villages and
levied
individuals who
refused to cooperate. In. one of several similar incidents
in 1971, 52 Christian villagers were burned to death in their huts in a
village named Debre Sila for non-payment of ELF demands.^'' On March
7, 1971, the ELF plundered the village of Halib Menal, stealing many
cattle, after the villagers had killed two ELF fighters while resisting an
ELF attempt to occupy the village two days before. Individuals who failed
The organization also repeatedly demanded cash payn

Christian missions ui its operational areas, andoe occasions confiscated
medicines from hoîtals and pharmacies.
There were several instances in which the fronts took European and
American hostages. For example, on May 24, 1974, the ELF attacked the
American Mission Hospital at Ghinda, and seized an Eritrean nurse and
an American missionary as hostages. The nurse was summarily killed the
same day because she could not keep up with her captors' walking pace;
the American was released unharmed after three weeks. In 1975 and 1976,
both the ELF and EPLF took US servicemen at Kagnew base as hostages;
at one point the ELF tfueatened to try two servioenen fat the damage and
deaths caused by US-snpplied munitions. The ELF also kidnapped three
British tourists* whfle the EPLF took Mr B. H. Burwood-Taylor, the
British honorary consul, from his office in central Asmara and kept him
for five months. All were released unharmed, though Mr Burwood-Taylor
was kept in solitary confinement for extended periods.
The Eritrean War and the Revolution
The war in Eritrea was a principal cause of the revolution of 1974

which overthrew the Emperor Haile Selassie and broi^ to power the
^ The Financial Times, London, October 24, 1975.
^ j4Jrica Contemporary Record, 1971-72, p. B113.
48
Provisional Military Administrative Committee, known by the Amharic
word for "committee", the Dergue. Senior army officers came to believe
that the war could not be won by military means, and that a negotiated
settlement was needed.
Inmically, the revolution foUowed two years in which there were fewer
army offensives, and modi of the military activity consisted in the war
launched by the ELF against the EPLF to enforce unity. The government
tried to sow dissension in the Eritrean ranks; attempting to inflame the civil
war between ELF and EPLF. Over half of the Ethiopian regular army was
stationed in Eritrea, a total of 25,000 men. Regular patrols continued,
together with the attempted enforcement of the protected villages strategy,
and punitive missions after successful ELF or EPLF guerrilla raids.
After a spate of assassinations by the rebels June 1974
in —
including
the killing of an army colonel — the army burned the village of Om
Hager, near the Sudanese border, in July. The villagers were ordered to
collect in the local stadium. After waiting for two hours, soldiers opened
fire with machine guns. At least 54 civilians were shot dead and a further
73 drowned while trying to swim a flooded river. Some reports indicate
that 17 elders were burned to death in a hut. About 4,000 refugees fled
to Sudan. The new military head of state, Gêral Aman Andom, himself
an Eritrean, apologized for the massacre and promised compensation. This
was never paid.
General Aman himself was killed in November in a shoot-out with an
army unit sent to arrest him by Colonel Mengistu Haile Mariam, who was
emerging as the strongman of the Dergue. Aman had advocated
compromise with the Eritrean fronts; Mengistu wanted to settle the
problem by force of arms. Starting in 1974, Mengistu made every effort
to increase the size of the aimed forces. During the years 1974-6, military
assistance from the US was actually increased. However, the amounts
provided fell short of Mengistu's ambitions.
The intervention of the citizens of Asmara after the battle of WoUd in
October 1974 forced the ELF and the EPLF to try to settle their differences
by negotiation. An agreement was reached at Koazien in January 1975
and the two fronts then began a concerted war on the government. In May
1975, at its second congress, the ELF voted out a number of exiled leaders
and emphasized unity with the EPLF. Ahmad Nasir became the new
chairman. The EPLF broke with Osman Saleh Sabbe (then head of its
foreign mission) and adopted a radical socialist program at its first
congress in January 1977. Sabbe then founded a third front —
the ELF-
Popular Liberation Forces (ELF-PLF) —
and a series of triangular
negotiations ensued; unity was not, however, achieved.
49
Between 1975 and 1977 the fronts succeeded in overrunning almost
the entire territory, save Asmara, Massawa and town of Barentu
the small
in the west. At this time, the fighters of the ELF
and EPLF outnumbered
the Ethiopian forces, and their victory appeared to be inevitable. This was
the first significant military threat faced by the new government of the
Dergue.
The Siege of Asmara, 1975
Following Aman's death, the level of violence increased in Eritrea, with

the ELF and EPLF closing in on Asmara. Some guerrillas infiltrated the
city on December 22. In retaliation, at least 50 Eritrean youths were killed
by security forces, 18 of them strangled with piano wire on the night of
December 23. ELF and EPLF commando units attacked installations in
the city, including civilian targets. On December 22, a grenade attack
killed six civilians. In mid-January, the US Infonnation Service building
and the main post office were atCadked with grenades.
On January 31, 1975, the Eritrean fronts lannclied an attack on Asmara
city. The attack started with a rocket attack on an army post, and there
were a number of battles in nearby villages and oo the outskirts 6f the city.
Despite fierce fighting, the city remained under government control.
Over the following four days, government soldiers went on the rampage
through the city. Civilians were dragged from their houses and executed.
Soldiers opened fire on passing cars —
in one incident, six people were
killed in a minibus. Passers by were stopped and shot dead. The soldiers
looted and pillaged, in some instances cutting away women's ears to seize
their eanings. According to eye-witness reports obtained by journalists
who arrived in the city a week later, the soldiers used their bayonets when
they ran out of bullets.^ There are reliable reports that exactly 100
civilians were killed at the village of Woki Debra, just outside the city.
About 50 people were killed at Adi Sogdo.
The government claimed that 124 civilians were killed by "^ray
bullets," during the four days of killing, but in a rare admission of a
greater degree of responsibility, executed 13 soldiers for "excesses" and
transferred 90 others in mid February. Journalists obtained reliable
accounts of 331 civilian deaths, and other sources claimed that up to 3,000
were killed. The number of combatants killed by on all sides was about
3,000.
Dial Torgerson, The Guardian^ London, February 15, 1975.
SO
After the failure of the ELF-EPLF attempt to capture Asmara, the war
intensified in the countryside as both sides prepared for a prolonged siege.
The EPLF estimated that 110 villages were partly or wholly destroyed by
the army in 1975. During February-April there were a number of
inassacins in diffeieiit parts S Eritrea, witii te^
occasion: Haddko in die eastern lowlands^ Adi Worehi Sub, Wanlki and
Gnilla in the west. In one incident at the village of Wold Debah on March
13, a gicoap of soldiers travelling between Asmara and Keren rounded up
a group of villagers early in the morning and shot dead 37. Some of the
survivors fled; others remained behind. The following day the soldiers
returned and burned the entire village, reportedly killing 500 people by
shooting and stabbing with bayonets. Many livestock were also
slaughtered.
After a lull, later in the year there were more killings of civilians,
usually in reprisals for rebel attacks. Three families were shot dead in
Asmara in August. On August 22, at least six boys were killed in Asmara.
A number of village were destroyed in air attacks.
As the ELF and EPLF
besieged Asmara, the government used a
blockade of food in order to try to sap their strength. Qieckpoints around
the city prevented food from being smuggled by sympathizers to the
guerrillas. Hie import of food from Tigray was also severely curtailed,
and the Ethiopian Red Cross was prevented from bringing in food relief.
Brig. -Gen. Getachew Nadew, the military administrator, explained the
blockade in these terms, made familiar by Mao Zedong: "if you wish to
kill the fish, first you must dry the sea."^^ These measures brought
considerable hardship to the rebel fighters, who were compelled to rely
on food brought by camel trains from the Sudan border. Tliese supplies
were subject to ambush and aerial bombardment, and were inadequate.
Journalist Roberts described the ELF fighters as running short
Owynne
of food, and surviving on reduced diets.^^ As some of the villages near
Asmara ran out of food, ELF units were forced to withdraw. In and
around Asmara, the price of food rose more than twenty-fold, to highs
surpassing those reached in the 1984/5 famine. Despite the food crisis -
- indeed famine —
the government maintained its blockade, and prevented
significant food imports until the military situation liad improved.
" Quoted in: Sunday Times, London, September 7, 1975.
Financial Tunes, London, October 22, 1975.
51
The Peasants' March
1976 and 1977 saw relatively fewer atrocities in Eritrea. This was
related toa number of factors, including several rounds of peace
negotiations and the govemmeiifs experience in 1975 that each massaore
merely drove people into tihe arms of fhe rebel fronts. Tbe most important
reason, however, was that the two main offensives planned into Eritrea
failed to reach the territory.
In 1976, Mei^tu planned to overcome the problem of a relative
shortage of armaments by resorting to a traditional Ethiopian tactic — a
mass levy of peasant soldiers, mobilized with the promise of booty and
land in the enemy territory. The soldiers were told that they could take
any land for themselves, after first driving the Eritreans from it.
The "Peasants March" on Eritrea was planned and implemented starting
in March 1976. Major Atnafu Abate, a close friend of Mengistu, was
responsible for die march. About 50,000 peasants, most of fton from
Wollo, were recruited. Some were volimleciSi templed by the
government's promto» others were forcibly conscripted. Most were given
antiquated surplus rifles from the armory, some were unarmed —
told they
would be able to obtain weapons from dead rebels. Without training, the
peasants began to march northwards. Meanwhile, in April the
government summarily ordered out all foreigners engaged in evangelical
or humanitarian work in Eritrea, closed all mission hospital and
confiscated most of the equipment.
Despite intense diplomatic pressure from the US, which objected to
this "medieval" manner of oonductiqg warfare, tte march went ahead.
However, the untrained peasant army was no inatdi for the Eritreans —
or indeed the newly-formed Tigrayan People's liberation Front (TnUP).
On the night of June 1/2, a surprise attack on the marchers was made at
Zalenbessa, in Tigray, before they even entered Eritrea, and over 1 ,200
were killed and the remainder dispersed. Much light weaponry fell into
the hands of the TPLF.
Offensive and Counter -Offensives, 1977
Atnafu and Mengistu then proceeded with a second plan for an

offensive in Eritrea. Atnafu planned to create over 55,000 "Defense of
the Revolution" squads, coBtaimqg about 300,000 miltriaimtn wi^ three
months' military traimng. (At diffmit times, government mraibers spoke
of a force numbering between 200,000 and 500,000; probably 150,000
were actually recruited by mid-1977.) The militiamen were to march first
on rebel forces in Gonder and Tigray, and then on Eritrea.
52
Oopyiigriici
Implementation of this plan started in December 1976. In February
1977, Mengistu eliminated senior members of the Dcrgue who advocated
negotiation with the Eritrean fronts, including the Dergue chairman,
General Teferi Bante. In April 1977, the first contingents were sent to
Gender (see chapter 3). ELF forces crossed into Gonder to engage the
militia in June, but the planned offensive into Eritrea was overtaken by
events —Somalia invaded the Ogaden, and die mflitia were diverted to
ttie southeast.
Regular military units remained active in Eritrea, and there were
numerous instances of violence against civilians, though not on the scale
of 1975. In early March 1977, a naval unit killed between 100 and 160
civilians at the Red Sea village of Imberemi. On March 31, 1977, army
units reportedly killed 42 civilians in reprisal for actions by the fronts.
During 1977, the Eritrean fronts remained on the offensive. They were
able to enter Asmara at will — in June the ELF displayed its confidence
by taking a British journalist into the city at night. Starting in March, both
fronts began to capture provincial towns — Nacfo in Mardi, Tessenei and
rain
•
They appeared poised for an assault on Asmara. The main
IMMIII factor deterring
the attack was a fear of renewed conflict between the ELF and EPLF —
the "Angola-ization" of Eritrea.
In early 1978, the Ethiopian administration in Eritrea made an estimate
for the cost of the war over the previous sixteen and a half years.
According to the estimate, 13,000 soldiers and between 30,000 and 50,000
civilianshad been killed or wounded. (No figure was given for casualties
among ELF and EPLF.) There were more than 200,000 Eritreans
the
forced into exile over half of them in Sudan. The financial cost in terms
of damage to property and the expenses dporsning the war amounted to
US$1.2 biUion.''
^ Quoted in UN Oammission on Human Rights, '*A Study on Human Rights

and Mass Exoduses," by Prince Sadmddin Aga Khan, Annex II, 1981, para. 32.
S3
3- REBELLION AND FAMINE IN THE NORTH UNDER
HAILE SELASSIE
Northern Marginalization under Shewan Rule
Hie norlfaem provinces of Gender, Gojjam, Wollo and Tigray are the
heartland of the "core" culture of Ethiopia —
the Ethiopian Orthodox
dnuch, the Amharic language and script, plow-based agriculture, and many
elements of the social system of the country derive from this historic region.
Most of the Emperors also came from here.
At the end of the 19th century, the center of power in Ethiopia decisively
shifted from the north to Shewa, with the assumption of the title of Emperor
by Menelik, King of Shewa. Menelik was an Amhara, from the dynasty
that ruled Manz, at the northern tip of the modern province of Shewa.
The majority of the inhabitants of the rest of Shewa were Oromo as —
is the case today. In terms of descent, the group that became politically
dominant in Shewa (and subsequently in Etiiiopia) was a misctuie of Amhaia
and Oromo; m tenns of language, religion and cultural practices, it was
Amhara. The northern Anihara regarded the Shewans as "Oalla" (the
pejorative tenn for Oromo),* and together with the Tigrayans and some
of the Agau and Oromo people in Wollo, resisted the new Shewan
domination, which led to their economic and political matginalization.
Revolt in Wollo
Between 1928 and 1930 there was a rebellion —

or a series of rebellions
— in northern Wollo against Shewan domination. The specific political
cause was support for Ras Gugsa Wale, a northern Amhara lord with a
strong claim on the throne, against the Shewan Ras Teferi (who crowned
himself the Emperor Haile Sekssie after defeating the revolt). The
government suppression of the revolt led to quartering soldiers with local
people, interrupting the salt trade, and involved massive looting and
confiscation of cattle. Combined with drought and locusts, the result was
famine.^ Haile Selassie ordered the importation of grain from India to
supply Addis Ababa, but there was no relief for north Wollo. Political
measures were taken after the revolt, including the replacement of much
^
Gerry Salole, "Who are the Shoans?" Horn of Africa, 2, (1978). pp. 20-9.
^ James McCann, From Poverty to Famine in Northeast Ethiopia: A Rural

Hislory, 1900-1935, FhUadelphia, 1987.
55
of the administration, which formerly had local roots, with appointees from
Shewa; and the joining of the rebellious districts to the province of southern
Wollo, which was ruled with harshness and venality by the crown prince.
These helped to contribute to the further marginalization of the area, and
the series of famines which plagued the area up to the fall of the Emperor.
The cumulative impact of imperial misrule and the petty tyrannies of
local lancDocds created an atmoapheie in wlikli development was extremely
difficult, as described by two consultants investigating the possibility of
starting livestodc projects:
Wollo is ... there is such an obscuring weight of

virtually impossible
disbelief, suspected innuendo and antagonisms; such a mess of mis-
government at petty levels, and such a lading of landlords that there
is almost nothing to start with and nowhere to start that will not go
wrong or sour ... [there is] the smothering welter of the weeds of an
entrenched and stagnant society.^
The Weyane in Tigray
Following the restoration of Haile Selassie after the defeat of the Italians
in 1941, there was a revolt in Tigray. Known as the Weyane, this was the
most serious internal threat that Haile Selassie faced. An alliance of the
Oromo semi -pastoral ists Raya Azebo, disgruntled peasants, and some
of
local feudal lords, under the military leadership of a famous shifta^ Haile
Mariam Redda, the rebels nearly succeeded in overrunning the whole
province."* British aircraft had to be called in from Aden in order to bomb
the rebels to ensure their defeat. While some of the aristocratic leaders,
such as Ras Seyoum Mengesha, were treated gently and ultimately allowed
to return and administer the vecaldtrBnt province, theie were reprisids
against the ordinary people. Most notably, the Raya and Azebo Oromo
were subjected to wholesale land alienation, and nmch of tiieir territory
was transferred to the province of Wollo. Tliis area was badly hit in
subsequent famines, partly as a consequence.
^ Noel J. Cossins and Bekele Yemeron, "Still Sleep the Highlands:

A
Study
of Farm and Livestock Systems in the Cential HigbLands of Ethiopia,'' Addis AbatM,
1973, p. 2.10.
* Gebru Tareke, "Peasant Resistance in Ethiopia: The Case of Weyane," Journal

of African History y 25, (1984), pp. 77-92.
56
Copyrighted matarial
Tax Revolts in Gojjam
Gojjam treasured its independence for centuries, and did not submit
willingly to Shewan rule. The issue around iîkh opposition repeatedly
coalesced was any attempt by the central govemment to measme land and
tax it. Taxation was not only resented as the imposition of unjust exactions
by government, but was feared as the means whereby the traditional land
tenure system would be undennined, and the fanners* independence
destroyed.
In the 1940s and '50s there was a series of attempts to measure land
in Gojjam, prior to taxation. In the face of peasant resistance, including
violence, all attempts failed. In the early 1960s, only 0.1 per cent of the
land had been measured, and Gojjam, one of the richest and most populous
provinces, paid less land tax than the poor and thinly populated province
of Bale.^ In 1950/1 there was armed resistance, including a plot to
assassinate Haile Selassie. Howevor the most serious revolt occurred in
1968, in re^Kmse to the most systenuitic attempt to levy an agricultural
income tax to date.
In February 1968, in reaction to the airival of parties of government
officials accompanied by armed police, the peasants of Mota and Bichena
districts resorted to armed resistance. After months of stalemate while much
of the province remained out of government control, Haile Selassie sent
troops to Gojjam in July and August. The air force bombed several villages;
it burned houses but its main task was probably intimidating the resistance.
Several hundred people died, according to contemporary accounts, but the

GoJaamis remained defiant. Finally, in December, Haile Selassie backed
down. He visited Gojjam in 1969, canceled all tax arrears, and made no
serious attempt to collect the new taxes.
Famines in Wollo and TIgray
In 1974, the Emperor Haile Selassie became notorious for his attempts
toconceal the existence of the famine of 1972-3 in Wollo. This, however,
was only one in a succession of such incidents. Prof. Mcsfin Woldc
Mariam of Addis Ababa University has documented how the famines of
1958 and 1966 in Tigray and Wollo were treated with official indifference,
bordering on hostility towards the peasants who were consideied sufBdently
^ Peter Schwab, Decision-Making in EOUapia:

A Stiufy afOie Pdlitical Process,
London, 1972, pp. 4» 159.
57
ungrateful for the divinely-sanctioned rule of Haile Selassie as to allow
themselves to defame his reputation by dying of famine.
There was severe famine in Tigray in 1958 which went without
significant government relief. In 1965/6, repc^rts of famine from Were Ilu
awraja in WoUo arrived at the Ministry of the Interior in November 1965,
one month after the situation became clear to the local police, but no action
was taken. The information took a further 302 days to reach the Emperor,
who then requested the Ministry of the Interior to act —
which it did by
asking officials in Wollo to send a list of the names of the people who
had died.^ A
small relief distribution was then authorized, llie only
consistent response to famine was to regard it as a security problem —
famine created destitute migrants, who neêd to be prevented from entering
towns, particularly Addis Ababa.
Both the 1958 and 1965/6 famines killed tens of thousands of people.
The famine that struck Wollo during 1972-3 played a crucial role in
Ethiopian history: the revelation of that famine by the British television
journalist Jonathan Dimbleby played a key role in precipitating the downfall
of the rule of Haile Selassie. Between 40,000 and 80,000 people died.^
The famine also led directly to the creation of the Relief and Rehid>ilitation
Commission (RRC), the powerful government de|Mrtment mandated to
prevent and ameliorate future famines, and to coordinate international
assistance. The 1972-3 famine was the last one in which there were no
functioning mechanisms for the delivery of large-scale humanitarian relief.
The Wollo famine was px)pularly blamed on drought, a backward and
impoverished social system, and the cover-up attempted by the imperial
government.^ These factors were all important —
though it must be
remembered that specific actions by the government, especially after the
Ras Gugsa and Weyane revolts, were instnmiental in creating the absence
^ Mcsfin Wolde Mariam, 1958"

Rural Vulnerability to Famine in Ethiopia ->
1977, London, 1986, p. 106.
^ John Seaman and Julius Holt, "The Ethiopian Famine of 1973-4: I. Wollo
Province," Proceedings of the Nutrition Society, 34, 1975, p. 114A. The Ethiopian
Nutrition Institute calculated a figure of 200,000 deaths but this appears to have
been based on over-pessimistic assumptions about the size of the fomine-affected
population.
* Failure to relieve a famine already under way

can scarcely be called a cause
of that famine, but popular coverage of "the politics of famine" continues to focus
on the politics of famine relief, failing to draw the necessary distinction between
the two.
58
of development. In addition, forcible alienation of resources and violence
also played an important role.
The group that suffered most from the famine were the Afar pastoral
nomads of the Danakil desert. Famine had already gripped them in early
1972. Hie Afar inhabit an arid semi-wilderness, utilizing pastures over
a large area to support their herds. In times of drought, they are forced
to move to areas which they do not normally exploit. Traditional drought
reserves included the Tcheffa Valley, on the rift valley escarpment, and
pastures along the inland delta of the Awash river where the waters dissipate
into the desert. In the 1960s the Tcheffa Valley became the location of
commercial sorghum farms, and small farmers from nearby also began to
use much of the land. Meanwhile, large cotton plantations were developed
along the Awash. By 1972, 50,000 hectares of irrigated land had displaced
20,000 Afar pastoralists.^
During the years of good rainfall, the loss of the drought reserves was
not noticed by the Afar, but when repeated drought struck, they found that
a necessary resource they had utilized sporadically for generations had been
alienated, without compensation. Famine among the Afar was certainly
caused by drought —
but by drought acting on a society that had been
deprived of the means of responding to that threat.
Official indifference to the plight of the Afar is illustrated by an incident
in 1974, when the flood waters of the Awash river were directed to the
Dubti valley in order to irrigate cotton plantations. The resident Afar
population was not informed, and 3,000 lost their homes, while 100 were
"missing."*"
Mobility is crucial to survival among the Afar. Nomadic in normal
times^ the ability to move freely over large distances becomes a vital
concern when resources are short. In the early 1970s, the Afor's mobility
was further restricted by the flow of weaponry to their nomadic neighbors
and competitors, the Issa (who are ethnic Somali). The Issa themselves
were suffering from the alienation of much of their pasture and restrictions
on their movement. The result was an attempt by the Afar to appropriate
wells formerly used by the Issa. This led to widespread armed clashes,
especially in 1972. One Afar reported "Many people die. Disease is the
' Maknun Gamalcdin, "State policy and famine in the Awash Valley of
Ethiopia," in D. Anderson and R. Grove (eds.). Conservation in Africa, Cambridge,
1987, p. 335.
^^RRQ "Food Shortage Report on Wello, 1979; Annex: A Pkofile of Disasters

in WeUo," Addis Ababa, 1979, p. 53.
59
first cause but the Issa are the second. "^^ Meanwhile, a survey done
among the Issa reported that homicide by the Afar was a major cause of
death. The famine also resulted in large-scale aimed clashes between
the Afar and their Oromo neighbors in Wollo.
The second group which suffered severely from the famine included
farmers in a narrow strip of middle-altitude areas of northern and central
Wollo. Those who suffered most were tenants. The Raya and Azebo
Oromo bad been reduced to that state by massive land alienation after they
participated in the Weyane revolt against Haile Selassie in 1943. Otben
were forced to mortgage or sell their land by the stresses of repeated harvest
failures in the early 1970s. Landlords took advantage of their tenants'
penury by insisting on the payment of large rents, often in kind. This
demand could be backed up by force, as most influential landlords had
a retinue of armed guards. The enforcement of crippling tenancy contracts
in time of shortage had the effect of taking food from the hungry. Thus,
during 1973, the famine area exported grain to the provincial capital, Dessie,
and to Addis Ababa.
The famine was much less severe in Tigray province, despite the drought
affecting both provinces. The difference can be bigely accounted for by
the different modes of land tenure —
in Tigray, moat formers owned their
own most woe tenants.
land; in middle-land Wollo,
Finally, the Emperor Haile Selassie considered that the peasants and
nomads of Wollo were shaming His reputation by starving, and resolved
to ignore them. Reports of famine were consistently ignored or denied.
In response to a report by UNICEF documenting famine conditions in July
1973, the Vice-Ministcr of Planning retorted: "If we have to describe the
situation in theway you have in order to generate international assistance,
then we don't want that assistance. The embarrassment to the government
isn't worth it. Is that perfectly dear?"^^
Quoted in: Noel J. Co.ssins, "No Way to Live: A Study of the Afar Clans
of the North-Easl Rangelands," Addis Ababa, Livestock and Meat Board, 1972,
p. 51.
J. Seaman, J, Holt and J. Rivers, "The Effects of Drought on Human

Nutrition in an Ethiopian Province," International Journal cf Epidemiology, 7,
1978, p. 37.
K. J. Lundstrom, North-East Ethiopia: Society in famine, Uppsala, 1976.
^ Quoted in: Paul H. Brietzke, Law, Development and the Ethiopian Revolution,
Lewisburg, 1982, p. 127.
60
Thougli the governor of Wollo, Grown Prince As£a Wossen, was both
greedy and incompetent (at the time of the fiunine he forced the closure
of commercial sorghum farms in the Tcheffa Valley by engaging in
litigation, claiming their ownership), Haile Selassie was never in ignorance
of the conditions in Wollo. A
UN official visited him in early 1973 and
found him well-informed —
his attitude was that peasants always starve
and nothing can be done, and that in any case it was not the Shewan
Amhara who were dying.^^ On belatedly visiting the province in
November 1973, his one remedial action was to announce that all who had
sold or mortgaged their land in the previous year could return and plow
it during the comii^ season, only leaving it to their creditors aflowards.^^
Even this minimal and tardy gesture was not enforced.
The 1975 Northern Rebellions
The Wollo famine contributed to the downfall of Haile Selassie, not

because the hungry peasants and nomads revolted and forced him out, but
because the issue gained political currency among the students and middle
classes of Addis Ababa. However, that is not to say that the famine, and
more generally the eight decades of political marginalization and economic
stagnation that preceded it, did not have serious consequences at the time
of the 1974 revolution and the years following.
In the eariy 1970s, "peasant rismg$ in various provinces [were] an even
more closely guarded secret than the famine".^^ Hiese revolts intenafied
in during the revolution, witii a series of rebellions led by feudal leaders
in each of the nonhem provinces. In Wollo, there was a revolt by a feudal
lord, Dejazmatch Berhane Maskal. In March 1975, he destroyed an
Ethiopian airlines DC3 at Lalibella. In October, he rallied supporters after
a spree of killings of former landlords by peasants and government security
officers. Dej. Berhane's ill-armed force of 5,000 was defeated by
government militia and air force attacks near Woldiya in December 1975,
but he continued to cause problems for the government for years. Another
feudal leader* Ougsa Andww, had brief military successes in northern
Wollo, before the army foiled an attempt to capture Kbrem mmid-1976,
" Brietidce, 1982, pp. 126-7.
^ Lundstioiii, 1976, pp. 52-3.
" Lionel Cliffe, "Capitalism or Feudalism? The Famine in Ethiopia," Review

of African Political Economy, 1, 1974, pp. 34-40.
61
reportedly causing 1,200 fatalities among Gugsa's peasant army and local
villagers.^* Other smaller revolts occurred in Gojjam and Shewa.
The most significant rebellion started in Tigray. This was an insurrection
led by the former governor, Ras Mengesha Scyoum (son of the governor
at the time of the 1943 Weyane). Ras Mengesha fiksd to fbt hills with about
600 followers in November 1984, when tiie Decgue executed 60 officials
of the previous regime. Ras Mengesha combined wiA other membeis of
the aristocracy, notably General Negga Tegegne (former governor q£
Gender) and formed the Ethiopian Democratic Union (EDU) in 1976. They
obtained encouragement from western countries. With Sudanese military
assistance, the EDU occupied the towns of Metema, Humera and Dabat
(all in Gondcr province) between February and April 1977,^^ but was
defeated by the militia force sent to the province in June-July.

The EDU remained active in Tigray, where two other rebel groups were
also operational. The Tigrayan People's Liberation Front (TPLF) was set
up in February 1975 by a group of left-wing students and peasants,
incorporating the Tigray National Organization, created three years earlier.
Prominent among its early leaders was Berihu Aregawi; later the hoot was
headed by Meles Zenawi. In 1978, the TPLF set up the Relief Society
of Tigray (REST), headed by Abadi Zcmo. It espoused a mix of Tigrayan
nationalism and socialist transformation. The Ethiopian People's
Revolutionary Party (EPRP), after defeat in the urban Red Terror (see
chapter 6), retreated to a base in rural eastern Tigray in mid-1977.
The EDU was rent by divisions between its leaders, and its aristocratic
leaders failed to gain popular support among their erstwhile tenants.
Crucially, it suffered defeat at the hands of the TTLF.^ The EPRP was
also defeated by the TPLF and driven into Gonder, creating lasting bitterness
between the two organization.
^'
Michael Dobbs, "A bloody ferewell to feudalism," Ihe Guardian, London,
November 24, 1976.
^'
Sudanese military support for tibe EDU was given in retaliation for the
Ethiopian government's backing for the Sudanese opposition National Front, headed
by Sadiq el Mahdi, which had staged an unsuccessful armed insurrection to
overthrow the government of Jaafai Nimeiri in July 1976.
^ The two organizations later fought in the refugee camps in Sudan, for
example in May 1979 and July 1982. The latter attack, in which eleven refugees
died, was instigated by the EDU. Ahmad Karadawi, "Refugee FUlicy in the Sudan,
1967-1984," DPhii thesis, Oxford, 1988, pp. 181, 193.
62
After the ill-fated Peasants' March of 1976, the government launched
a series of five military offensives in Tigray: November 1976, June 1978,
October-November 1978, March-April 1979 and May- June 1979. Small
towns such as Abi Adi changed hands several times. By 1979, REST
estimated that 50,000 people in Tigray were displaced on account of war.
Refugees from Tigray and Gonder began to arrive in Sudan in early 1975.
By May there were 34,000; by 1978 there were 70,000. In February 1979,
the Ethiopian army invaded Sudanese territory at Jebel Ludgi, forcing the
eviKûation of the nearby refugee camp of Wad el Hileui.^^
Fighting against the Afar
Another serious though short-lived rebellion occurred among the Afar.

The Afar leader and Sultan of Awsa, Ali Mirrah, had been accorded a high
degree of autonomy by Haile Selassie, and the well-armed Afar had never
come fully under tte administrative or mflitary contiol of the government.
Ali Mirrah was also a large landowner and feared the confiscation of his
cotton plantations in the land reform. This brought him into conflict with
the government, and on June 1, 1975, a military force was dispatched to
arrest the Sultan at his headquarters. Ali Mirrah escaped to Djibouti, but
his followers launched a coordinated series of attacks on military outposts
on June 3, and claimed to have killed several hundred soldiers. Many
unarmed Amhara laborers in the Awash cotton plantations were also
massacred.
Government reprisals were swift. Starting at Assaita in the Awash valley
on June 3, and expanding to an arc stretching from Mille in the west,
pacing through Awsa to Serdo mthe east, tiie army attadced towns,
agricultural schemes, and Afar nomads' encampments.
The cotton plantations had a large hiboT force, including Moslem
Eritreans, local Afars and highland Amharas. The army selectively killed
Eritreans and Afars. Refugees in Djibouti reported that 221 Afar workers
were killed. Women and children were gunned down as they tried to flee
on tractors. In one incident in mid-June, 18 men were vShot dead on a
bridge. The killing then spread to the towns. An estimated 300 died in
Awsa and 100 is Assaita, including the Imam, killed in his mosque. The
soldiers also spread out into the countryside and attacked small Afar villages
and cattle camps. While many of the casualties were armed Afar men (the
distinction between a civflian and combatant is a fine one among the Afar)
women and children were also killed when tanks and artillery tombarded
» Karadawi, 1988, p 117.
63
cattle camps and troops opened fire. The killing lasted six weeks.
Estimates for the total number of civilian casualties amount to more than
1,000; some run as high as 4,000.
One of those killed was a young British social anthropologist, Glynn
Flood, who was arrested by the army and detained. According to other
people detained in the same prison, after two weeks he was talsn oat by
four soldiers, and then ttey heard a scream. It is —"wy^ tiiat he was
bayonetted to death and his body thrown into the river Awadi. Oovemment
officials had tried to stop Mr Flood bam going to the area a week earlier,
and it is probable that he was murdered in order to prevent him from
producing evidence of the killings.
In exile, Ali Mirrah founded the Afar Liberation Front (ALF). His son,
Hanafari conducted military operations and succeeded in closing the
strategic Assab-Addis Ababa highway. The government responded by
some leading Afar
the twin tactics of another military campaign, and giving
positions within the administration and a measure of autonomy. A
faction
led by Habib Mohamed Yahyo was given a large quantity of arms, and
proved to be a loyal supporter of the government.'' These tactics
prevented the ALF from posing a major military threat, though it was able
to mount occasional attacks throughout the 1970b and 1980b.
^ There is a dispute over the Sultanate of Awsa between the families of Ali
Mirrah and Habib Yahyo going back several generatkma.
64
4. INSURRECTION AND INVASION IN THE
SOUTHEAST, 1962-78
Introduction
Wais have been fouglit contnmally in soutlieast Ethiopia for as long

as in Eritrea. The two principal subjugated peofdes of the region the —
Oromo and the ethnic Somali inhabitants tlie Ogaden have been —
staging insurrections more or less continually since the early 1960s. The
firstphase of the wars lasted until 1977, with the government utilizing a
set ofcounter-insurgency strategies familiar from Eritrea, though arguably
more successfully. The second phase was a large-scale conventional war
which began when the Somali army invaded the Ogaden in July 1977,
concluding when the Ethiopian army defeated that invasion in March 1978.
A pcotiacted tiiird phased began ate that victory, with a return to counter-
insurgency waÊare; that is the subject of the followii^ chapter.
Southeast Ethiopia consists of well-watered highland areas, largely
inhabited by Oromo people, who practice a mix of agriculture and livestock
raising, and drier lowlands, known as the Ogaden. HieQgaden is inhabited
by ethnic Somalis belonging to 12 different clans, among whom the Ogaden
clan is numerically dominant. Other important clans include the Issa, to
the north of the Harerghe highlands, and the Isaaq, who inhabit the Haud
reserve, an area adjacent to northern Somalia.^ The Somalis are largely
pastoral nomads. The population of the Ogaden itself is about one million;
the surrounding Oromo areas contain a much larger number of people.
At Independence in 1960, the Somali government laid daim to the
Ogaden and adjoining Oromo areas in the Bale and Harerghe highlands.
The Conquest of the Oromo and the Ogaden Somali
Between 1886 and 1889, the Emperor Menelik conquered the

independent Oromo states of Arsi and Bale, and occupied the trading city
of Harer, which is a holy city for Moslems. Menelik then laid claim to
the vast desert area of the Ogaden.^ This claim was recognized by the
^
Until 1960 known as British Somaliiand; in 1991 the Somali National
Movement, controlling the area, unilaterally declared independence and named
it the Republic of Somaliiand.
^ This claim was made wifli Italian encoiuaêment, partly in order to prevent
the British annexing the area to British-occupied northern Somaliiand.
65
European powers in 1910, though effective occupation of the aiea was not
attempted until after the Second World War.
Amhara domination followed. In the context of southern Ethiopia, the
term "Amhara" needs to be treated with care. While the Amhara who came
to the south as conquerors originated from all parts of the northern
highlands, all came as vassals cxf the specifically Shewan Amhara state.
Local people, whatever their origins, were also able to assimilate into the
Amhara class, by vutue of marriage, or adopting the religion, language
and cultural traits of the Amhara. A social anthropologist working in the
neighboring province of Aisi noted that for the indigenous Oromo "'Amhara'
and 'self-satisfied dominant elite' have become convergent categories."'
In the highland areas of the southeast, Amhara neftegna were given
grants of land, with accompanying rights to extract produce from the local
population. The indigenous peoples were unhappy with the loss of their
independence and with the new burdens imposed upon them by their
Amhara overlords, and armed resistance was frequent. The Italian conquest
of 1935 came as a liberation from Amhara rule for many inhabitants of
the southeast; and after Haile Selassie was restored there was intomittent
armed resistance against the re-imposition of the hated land tenure and
taxation systems, notably in Harerghe in 1942, 1947/8 aind 1955.
When Somalia gained its independence in 1960, there was agitation
in the Ogadcn ("western Somalia") for independence, or for separation from
Ethiopia to join the Somali state. The Somali government set up the
Western Somali Liberation Front (WSLF) in that year. There was a revolt
in the Ogaden in 1963-4, which was put down with customary brutality
(see below).
Hie late 1960s and early 1970s saw the growth of an Oromo nationalist
movement. Hiis was first expressed through traditional-styte shifta
rebellion in Bale and Harerghe, and in the oeatioo of Ovomo community
associations among groups in Shewa, WoUega and ArsL
The 1960s Rebellion in Highland Bale
The rebellion in Bale was the outcome of the systematically brutal

subjugation of an indigenous population by a ruling class of armed settlers
drawn largely from the Shewan Amhara. John Markakis has written of
the conditions preceding the revolt:
' Paul Baxter, "Ethiopia's Unacknowledged Plrobiem: The Oramo," African

Affairs, 77, (1978X pp. 283-96.
66
The legal exactions of the state and the landlords were compounded
by a host of illegal impositions levied by the ruling class on the
peasantry, usually associated with matters related to land. Land
measurement, classification, registration, inheritance, litigation and so
on were matters that could be concluded only through the payment of
enforced bribes to a series of officials, and were subject to the risk of
firaud in the process. Tax payment itself required the running of a
gauntlet manned by officials who had to be bribed to conclude the
transaction properly. Venality, the hallmark of Ethiopian officialdom
throughout the empire, reachied its apogee in the conquered areas of
the south, where the hapless peasantry had no recourse against it.
Northern officials serving in the south hoped to amass a small fortune
during their tour of duty, and to acquire land through grant, purchase
or other means. The scale of their exploits in Bale affronted even some
of their colleagues ... There was precious little return for such
impositions.'*
Armed rebellion started in highland Bale in 1962, fanned by both

encouragement from the newly-independent Somali government, and the
heavy handed response of the governor. The rebels were led by a minor
chief named Wako Gutu, and used traditional shifta-stylc tactics, with no
central command. Fighting gradually intensified until late 1966, when it
was clear that the provincial police and militia could not contain the revolt,
whereupon the government declared a state of emergency and called in
the army.^ There were ground assaults and aerial bombardments in both
highland and lowland Bale in the early months of 1967, involving much
indiscriminate violence against civilians. There were also punitive measures
such as land confiscation, restrictions on nomads' seasonal migrations and
heavy fines levied on uncooperative communities. By 1968 more than
a quarter of the land was classified as confiscated.^ In order to recover
their land, farmers needed to pay their tax arrears —a dauntuig prospect
in view of the epidemic corruption of the administration. However the
rebels could not be dislodged from their mountauious, forested base.
^ John Markakis, National and

Class Outlet in the Ham ofAfrica, Cambridge,
1987, p. 194.
^ Patrick Gilkes, The Dying

Lion: feudalism and Modernization in Ethiopia,
London, 1975, pp. 214-18.
* Markakis, 1987, p. 200.
67
Military tactics then changed to a pacification approach, avoiding direct
military confrontation on the ground. Roads were built into the rebel
heartland, with the assistance of British military engineers.Combined with
restrictions on movement and military surveillance of the lowlands, this
helped to cut off assistance from Somalia. Air strikes continued, with US
technical assistance, aimed at intimidating the rebels and destroying their
food supplies (and tfaerôie also the food supplies of the local population).
These strategies were combined with leniency townds flie rebel leaden,
who were allowed to gp free or rewarded with lucrative positions.
The final demise St Wako Gutu's forces came in March 1970, after
military assistance was cut off by the Somali government, and the tightening
noose of government troops ensured that they ran out of food supplies.
The government then granted a general amnesty and made various promises
to the general population, which it failed to keep. Conditions in Bale at
the time of the revolution were almost exactly as they had been a decade
earlier.
The Destruction of Oromo PoBtkal MofcmcBls
The 1960s saw a growth in Oromo cultural, social and political

movements. In part this was related to the achievement of independence
by African peoples which had been colonized by the European powers —
many educated Oromo aspired to a similar "liberation." An article by an
Oromo student leader, Wallelign Mekonnen, in a 1969 student publication
expresses well the feelings of subjugation:
Ask anybody what Ethiopian culture is? Ask anybody what Ethiopian
language is? Ask anybody what Ethiopian music is? Ask about what
Ethiopian religion is? Ask about what the natkmal dress is? ft is either
Amhara or Amhaia-Tigiell To be a "genuine" Ethiopian one has to
listen to Amharic music, to accept the Amhara-Hgre religion. Orthodox
Christianity, to wear the Amhara-Tigre shamma in international
conferences. In some cases, to be an "Ethiopian" you will even have
to change your name. In short, to be an Ethiopian, you will have to
wear an Amhara mask (to use Fanon's expression).
Quoted in: Randi Ronmng Balsvik,

^ He^
SdasaM SmdaUK The hodkciuai
and Social Background to Revolution, 1952-1977, E
iMBâg, Mich., 1985, p.
277. Wallelign was killed by security forces during an attempted hijack of an
Ethiopian Airways airplane. The EPRDF offensive of May 1991 which overran
the govemmcni's northeast front and led directly to the flight of President Mengistu
Haile Mariam was named "Opeiatian Wallelign" in honor of this sUident.
68
The Mecha-Tulema Sclf-Hclp Association, founded in 1962, was the
most prominent attempt by the Oromo to organize legally. The association
was only legally registered after overcoming considerable opposition from
the government. It sponsored specific self-help projects, but had the
broader aim of developing an Oromo "national consciousness." In 1966,

nsiqg the pretext oi a hoab ejcplosion in a dnema in Addis Ababa, Haile
Selaê cradced down on the Mecha-Tulema Association. Over 100
leading Oramo community leadeis were arrested and brought to court, in
a trial that was a parody of due process. Based on confessions obtained
under torture and other dubious evidence, the two leading defendants were
sentenced to death,* and others to long prison sentences. The organization
continued an undc»:ground existence for several years thereafter.
Revolt in Highland Harerghe
In the mid-1960s, there was also a small insurrection in the highlands

of Harerghe. This was led by Sheikh Hussein, a farmer associate of the
Oromo politician Tadesse Birru. The organization was called the Oromo
liberation Front (not to be confiised with the second Oromo liberation
Front, founded in 1974), and carried out small-scale guerrilla activities.
In August 1971 Sheikh Hussein changed the name of his organization
to the Ethiopian National Liberation Front (ENLF), and incorporated some
of Wako Gutu's followers, but thereby also split his movement. By 1973
the ENLF controlled sigoificant areas of the highlands of Harerghe and
Bale.
The Revolution and the Oromo Movement
The revolution of 1974 split the Oromo movement. Many of the

members of the Deigue were themselves Oromo —
including General Teferi
Bante, Chairman from November 1974 imtU his execution by Mengistu
in 1977.' Many of the Dergue's initial programs, notably the land reform
of 1975, the change in official designation from the derogatory "Galla"
to the more acceptable "Oromo," and the legalization of the use of the
Oromo language for public speaking, were welcomed. As a result, the All-
' Lt. Mamo Mazamir

was executed for treason. Gen. Tadesse Birru was
reprieved by Haile Selassie, and released at the time of the revolution. In 1975
he was rearrested by the Dergue and executed along with a colleague.
' Gen Teferi's father's name was actually Benti (an Oromo name) but he
changed it to Bante (an Amhara name) to be mme acceptable to the Amhara elite.
69
Ethiopia Socialist Movement (MEISON), which was led by several
prominent Oromos, allied itself with the Dergue from 1975 until it was
purged in 1977 (see chapter 6).
The land reform of 1975 gave the Dergue great political capital in the
Oromo areas — which it promptly began to deplete by heavily taxing the
peasants and requisitioning food from than for the army and the towns.
The land reform also set up Peasant Associations (PAs), witii tiie Initial
aim of re~distributing land. PAs were given wider-ranging powers shortly
afterwards. In the south, most of the PA leadership originally consisted
of local people elected with much popular support. This began to change
in 1978. The purge of MEISON coincided with a slightly less violent purge
of the leadership of the PAs, and the formation of the All-Ethiopia Peasant
Association. From this point onwards, PA leaders were all appointed by
the government.
Many Oromo leaders went into armed opposition in 1974. They joined
defectors from the ENLF Oromo Liberation Front, and founded
aiiu the first
the (second) Oromo liberation Front (OLF). This was initially active in
the highlands of Bale and Harerghe, aind had its first meeting to publish
a political program in October iS that year. The initial insunectioa was
a decentralized revolt which encompassed a number of different groups.
The government launched two offensives, the first in 1974, and the second
in early 1976, using locally-recruited militia. These succeeded in scattering
but not suppressing the nascent OLF resistance. Influenced by MEISON,
the Dergue entertained hopes of negotiating a compromise with the OLF,
and several meetings were held, but without result. By early 1977, the
OLF had set up an administration in parts of the Cheicher highlands of
Harerghe, and was active in Bale, Arsi and Sidamo.^"
In 1976, the Somali goveroment set up a guerrilla force to fight in
Oromo areas, as a counterpart of the WSLF, calling it the Somali Abo
Liberation Front (SALF).^^ Wako Guta and Sheikh Hussein were among
the prominent Oromo nationalists who joined the SALE, which formally
superseded the ENLF. The pre-existing split between the ENLF and the
OLF, and fears that Somalia harbored irredentist ambitions to annex Oromo
areas, using the SALF as a vehicle, led to distrust and at times conflict
between the OLF and the SALF.
From 1974 to 1977 insurrection spread through much of the Oromo
highlands of southeast Ethiopia.
10
Africa Confidential, 19.11, May 26, 1978, p. 6.
" "Abo" is a form of greeting common to many peoples in southeastern

Ethiopia. The Somali govenunent was eager to avoid use of the name "Oromo."
70
Ethiopian Rule and Famine in the Ogaden
The inhabitants of the Ogaden received little from the Ethiopian

government during the first decades of the occupation except raiding parties.
Following independence in 1960, the Somali government set up the WSLF
and there was increasing agitation for the Ogaden to secede firom Ethiopia.
The Ethiopian army immediately moved to set up mflitary bases in the area.
Up to 500 civilians were killed when the village of Aisha was destroyed
to make way for a military post in August 1960.
In 1963, following the first systematic attempt by the Ethiopian
administration to collect taxes, there was widespread insurgency in both
lowland Harerghe and Bale. The guerrillas relied heavily on the Somali
government for support, and while they grew in numbers to about 3,000,
they never posed a serious military threat to the central government. Their
guerrilla tactics were unsophisticated and the army was able to engage and
disperse them on several occasions. The Ethiopian government also put
prc^re on Somalia by incursions into Somali territory and threats of a
largerHscale invasion. Following government militaiy offensives in late
1963 and an agreement between the Somali and Ethiopian governments
in March 1964, the insurrection was largely over.
More serious for the civilian population of the area was the government's
policy of mounting punitive expeditions, which killed or confiscated large
numbers of animals, depriving the pastoral communities of the basis for
their survival.
Military administration remained in the Ogaden after the insurrection.
Most major towns had curfews for at least a year. Ogaden clan leaders
documented a number of incidents in May 1964, when 75 people were
reported killed by the army, together with more than 14,000 domestic
animals killed or confiscated, and July 1964, when 22 people were killed
and over 8,000 anunals killed or confiscated." Ihis "economic war"
against the Ogaden was supplemented by a policy of encouraging Amhara
farmers to settle in the more fertile areas, especially in the Jijiga area. The
process of land registration became a vehicle for settler farmers claiming
land rights, depriving pastoralists of use rights.^"' The lack of access to
these pastures became critical when drought struck in 1973-4.
" Ismail Wais, "An Account of the Colonial Experience of the Western
Somalis," Horn of Africa, 4,4, (1981/2), p. 28.
Noel J.Cossins, "Pastoralism under Pressure: A Study of the Somali Clans

in the Jijiga Area of Ethiopia," Addis Ababa, Livestock and Meat Board, 1971,
p. 82.
71
The introduction of administration also led to attempts to regulate the
livestock trade. Selling animals is critical to survival for the Ogadeni
pastoralists. Hitherto, most Ogadeni animals had been sold to Hargeisa
and Bcrbcra in Somalia, along an age-old trade route which was now
technically regarded as "smuggling. **
The new administration confiscated
many "smugged" animals. TogcdKr wilb llie hansanent off heiden
attempting to sell animals in Ethiopian towns, this acted as a poweifid
obstacle to trade, leading diiecdy to the impoverislunient off the faeiden.
A final and key element to the pacification canâign in the Ogaden
was the government control of water points. A w
BCt oik of functioning
wells is crucial to the mobility which herders need in order to seek out
seasonal pastures. There are many reports of the wells dug by the Ogaden
people themselves being poisoned. New reservoirs (birkas) were built by
the government, but primarily to serve the interests of settler fanners and
townspeople.
In 1967 there were further military actions, chiefly in lowland Bale,
aimed at WSLF groups which were acting in concert with Wako Gutu,
and their civilian supporters.
In 1969 Maj.-Gen. Siad Bane seized power in a militaiy coup in
Somalia. He acted fast to consolidate his power, and one of his actions
was placating the Ethiopian govemment by formally disbanding the WSLF -
- though not renouncing Somalia's longstanding claim to the Ogaden.
During 1971-2 there was another round of atrocities by the army against
Issa and Ogadeni pastoralists. The conflict was based upon two factors.
One was Issa-Afar competition for political control of Djibouti, which was
moving towards independence from France. The Issa, who are the majority
in Djibouti, were in favor of inunediate independence; the minority Afar,
supported by Haile Selassie and favored by France, wanted independence
postponed. The second fiictor was drought. Coaflict was spailDed by
occupation of a series of unjKntant wdls aioiuid Banoti by the nei^^
Afar in the late 1960s. When there was poor lainfall in 1971 and 197%
the Issa tried to leoccupy the wells, there were aimed dashes lietween the
two groups.
On the pretext of the dispute over the wells, the Ethiopian army
intervened against the Issa. According to a letter of complaint written by
the chiefs of the region to the Ethiopian parliament, between April 1971
and May 1972, the army killed 794 people, as well as confiscating nearly
200,000 head of livestock.^'*
^ Quoted in: Wais, 1981/2, pp. 25-8.
72
In 1974, drought added to the Ogadenis' problems, and the area was
struck by famine. The loss of pastures to iimnigrant fanners, restrictioiis
on movement, and contimied amed dashes with the Afar all contributed
to tibe famine.
Accotding to a survey done in Haieighe in May-June 1974, death rates
among the lowland pastoralists were about three times normal.^^ Assuming
"normal** to be 20 per thousand, and the affected population to be 700,000,
this implies 28,000 famine deaths over the previous year. Another survey
done a year later found that death rates bad risen slightly, implying a similar
number of famine deaths in 1974/5.^^
Responding to the famine became one of the first tasks of the newly
constituted Relief and Rehabilitation Conmiission (RRC), which became
active in delivering large quantities of food relief and setting up feeding
centers.
The RRCs activities dunng 1974-5 had a large humanitarian component
However, whether through an ethnocentric view of the superiority of a
setded over a nomadic lifestyle, or through a deliberate policy of using
the drought as an opportunity to extend government control over the
recalcitrant population, the famine relief program served to undermine key
aspects of the Ogadeni way of life.
By early 1975, more than 80,000 Ogadenis were living in 18 relief
shelters. The shelters were run on military lines, with strict curfews
enforced at 8.00 p.m. Movement in and out was severely restricted —
making it impossible for each family to keep more than a handful of small
animals. Traditional festivities were reportedly banned in some camps.
The government had the e3q>Iictt infentiaa of turning the camp populations
into settled farmers, rather dum aUowtng them to return to a pastoral way
of life. Another intention was to relocate camps well away from the Somali
border.
As a result of these restrictions, the great majority of the Ogadeni men
stayed outside the camps, moving with their animals, unwilling to risk
approaching their families within the camps. Fear that the Ethiopian
government was intent on undermining their traditional way of life was
John Seaman, Julius Holt and John Rivers, "Hararghe under Drought: A
Survey of the Effects of Drought upon Human Nutrition in Haraigbe Province,
Ethiopia,** Addis Ababa, RRQ 1974, pp. 39-41.
M. Gebre Medhin, R. Hay, Y. Licke and M. Maffi, "Initial Experience of

a Consolidated Food and Nutrition System: Analysis of Data from die Ogaden
No. 48, 1977, pp. 33-4.
73
one factor that spurred many Ogadeni men into aimed opposition to the
government.
The Secret Invasion: March 1976-May 1977
After the revolution, Ethiopiawas in turmoil, and Hs aimy was bogged

down in the war and fighting insurgencies in sevml provinces.
in Eritrea
The Somali government had meanwhile greatty expanded its aimed fcMces,
with Soviet support, so that in certain areas —
such as artillery, tanks and
mobile battalions —
they outnumbered the Ethiopian aimy, and in most
other areas they matched them. There was, however, political discontent
in Somalia, as President Siad Barre's authoritarian rule was antagonizing
certain groups. Although President Siad had pledged to abolish "tribalism",
he came to rely more and more on support from certain clans. At the
beginning of 1976, President Siad turned to the elders of the Ogaden clan
for a political alliance, and concluded a deal whereby the government would
provide support for the cause of die "liberation" of the Ogaden from
Ethiopia, including military assistance, in retum for political loyalty to the
regime. It is important to note that the deal was struck with representatives
of the Ogaden clan, not representatives of the population of the Ogaden,
which includes many clans. The WSLF was re-founded, and members
of the Ogaden began to receive preferential treatment in matters such as
government and army posts and education.
In the first months of 1976, the WSLF became active once again inside
the Ethiopian Ogaden. With a leadership based in Mogadishu and close
to President Siad, it drew recruits from the frustrated and alienated pas-
toralists. main base was Hargeisa, infiltrating through the Haud reserve
Its
area into Ethiopia to make guerrilla raids. Hie Somali media trumpeted
its successes, but the claims made were out of all proportion to the reality:
in fact, the guerrillas were making little real military headway.

Officers in the Somali army became impatient with the slow progress
of the WSLF, and argued that if the Somali government's stated intention
to annex the Ogaden was genuine, it would be necessary to use the regular
army. In early 1977, President Siad responded to these complaints with
a compromise: soldiers from the regular army would henceforth fight with
the WSLF. About one fifth of the Somali army, numbering about 3,000
men and consisting mainly but not entirely of members of the Ogaden clan,
was deputed to become the principal force of the WSLF. The soldiers took
off their uniforms and put on the ragged clothes of guerrillas; they
74
abandoned their annor and heavy weapons foi light guns and hand
grenades."
Under the command
of senior military officers, the "army" units of the
WSLF engaged on Ethiopian military positions, while the pre-
in attacks
existing "guerrilla" units of the WSLF undertook activities such as
ambttshes» sabotage and laying land mines. Tlie plan did not succeed.
Tbe soldiers were not trained for gueirilla warfare, and the officers did
not like a method of warfiire which conflicted with then: conventional
training. When an attack on an Ethiopian garrison at Godc (s(ûthem
Ogaden) in May 1977 was repulsed with the loss of over 300 dead,
including 14 middle- and high-ranking officers, dissent in the army became
vocal. In June the decision was made to commit the Somali army, in
uniform and with full armor and support, to the Ogaden.
The Regular Invasion, July 1977-March 1978
On army invaded tbe Ogaden. The Somali

July 23, 1977, the Somali
government still dedare war and insisted in the media that all
refiKsed to
military actions were the responsibility of the WSLF, but the truth was
evident.
The first assault was made in the central-southern Ogaden, and Code
and Gabridaharey fell to the invaders within a week. Half of the Ethiopian
Third Division was put out of action, and the Somali force moved rapidly
north. This was followed by an invasion from the vicinity of Hargcisa
directed towards Dire Dawa; the attack on the town started on August 10.
Three attempts to take the town failed within two weeks, and instead the
Somali army turned its attention to Jijiga, which was evacuated by the
Ethiopian army on September 10. The Somali army then concentrated on
attacking Harer, advancing from Jijiga in October. A
brigade that was
originally directed to the south of Harer in a diversionary move actually
succeeded in occujîng a section of the town for several days in November
before being pushed back. Harer was besieged for two months.
WSLF units engaged in sabotage action, unpairing the mobility of the
Ethiopian forces by destroying communications.
The conventional fighting was confined to the north. In the southern
Ogaden, the Ethiopian garrison at Dolo (near the Ethiopia-Somalia-Kenya
triangle) withdrew to northern Sidamo, and three Somali brigades crossed
into Ethiopia, encountering no resistance. The iiivading force stopped short
The following account of the Somali invasion and abuses associated with
itdraws heavily on material provided by Abdi Razaq "Aqli" Ahmed, formerly
a Major in the Somali army.
75
of Negele (southeast Sidamo) when two of the brigades were re-assigned
to the northern front, and this area remained quiet. In Bale, the SALF was
active on a small scale, and there was no conventional military action.
A single battalion was assigned to El Kere, and then moved to Fiiq in
central Harerghe, meeting no resistance on the way.
In late December, the responded to repeated appeals bom the
USSR
Ethiopian government and switched sides. It airlifted several billlcMi doUais
worth of military equipment to the embattled Dogne, indndiqg over 600
battle tanks and 67 MiG fighter-bomber airplsuies.^* Approximately
16,000 Cuban combat troops were also flown to Ethiopia ti^ther witii
modem armor. The government had earlier launched a program of mass
mobilization, and was expanding the army from 60,000 regulars and 75,000
militia to 75,000 regulars and 150,000 militia. This led to a dramatic
change in the make-up of the Ethiopian army. Its firepower and mobility
became immediately greater than those of the Somali army. It was now
advised by the same Soviet strategists who had trained the Somalis.
In late January, the Ethiopian counter-offensive began, directed by
Soviet advisors and spearheaded by Cuban troops. The Sanali army was
pushed back from Dire Dawa and Harer and outflanked by mobile and
airborne units. Counter-attacks were repulsed, and dissent within the
Somali army escalated. In early March the Somali command gave the order
to retreat, and the Ogaden was evacuated and reoccopied by the Ethiopian
army without a fight.
Abuses by the Somali Army
The Somali army was regularly violent and abusive to the inhabitants
of the areas it occupied. Its treatment of Qromo dvflians was markedly
worse than ethnic Somalis —many of the troops came from the same dans
as the local Somalis, and therefore treated them with moce respect, not least
because clan loyalty demands vengeance on those who commit an oCfenae
against a clan member.
The
pattern of abuses consisted mostly of small giDttps of soldiers
committing the following types of violations:
* Soldiers destroying or looting property, and on occasion killing the

owners who protested.
" NOVIB, "War and Famine in Ethiopia and Eritrea, An Investigation into
the Arms Deliveries to the Struggling Parties in Eritrea and Tigray," Zeist, The
Netherlands, 1991.
76
* Soldiers raping women, and killing brothers, husbands or fathers who
objected.
* Soldiers taking food or livestock to eat, and sometimes killing the

owners who objected.
* Sddieis opening fiie on dvOians who returned to investigate the

condition of their houses or forms which had been occupied by soldiers.
The extent of these abuses and whether the soldiers responsible were
held accountable and punished depended entirely on the local conmiander.
On several occasions there were larger-scale violations which had been
authorized by senior officers. These included the mining of buildings in
Jijiga and other towns during the retreat in February-March 1978. In
addition, in November 1977, the commander of the force which had
occupied part of Harer town was instructed by a senior officer to destroy
as much of the town as he could before retreating. Harer is a holy city
to Moslems and the local commander, as a devout M
osinn, refused to carry
out the order.
Attacks by the Somali air force in late July resulted in dvflian casualties
at Aware and Degahabur. In mid August the Ethiopian govemment reported
that a Somali air attack on the aufield at Jijiga had only narrowly avoided
causing a large number of civilian casualties because an airplane on the
ground had been evacuated just minutes beforehand.
There were also violations of the rights of combatants by members of
the Somali anny and WSLF. These included:
* Summary execution of officers and men in order to maintain discipline.

In February 1978, the commander of the Jijiga front requested each
battalion commander to send him twelve soldiers, implying that they
would be considered for promotion or another foim of award or benefit.
All of them — numbering about 80 —were summarily shot. The
commander explained that his junior officers had been insubordinate,
and that in future all who disputed his orders would be similarly dealt
with. In January, five middle-rankii^ officers were shot separately
in suspicious circumstances.
* Retreating Ethiopian soldiers were set upon by WSLF fighters and armed
WSLF sympathizers and killed.
* Abuses against prisoners of war (see chapter 17).
77
Abuses by the Ethiopian Army
The Ethiopian army was also responsible for abuses against the civilian
population during the war. Before abandoning the towns of Jijiga,
Dcgahabur, Aware and Gabridaharey the army sunmiarily executed civilians.
In the case of Jijiga, nearly 100 were reported killed.
Individual acts of violence by Ethiopian soldiers agpinst civilians have
been reported. Aserious instance of an abuse used as part of a Biilitary
tacticoccurred just south of Harer in late December, when the Ethiopian
troops forced a line of ethnic Somali women to walk in front of their
advancing soldiers, using them as a human shield. The Somali soldiers
were faced with the alternatives of retreating under fire or opening fire
themselves — they chose the latter, and about 20 women were killed.
The worst and most systematic abuses by the Ethiopian army occurred
during the rcoccupation of the Ogaden in March 1978. Journalist Norman
Kirkham described how the Ethiopian and Cuban troops swept through
the Ogaden after the retreating Somali army, virtually unopposed:
I me that civflians had been shot

met numbers of the survivors who told
had been executed sunmiarily in house to house sweeps.
in the streets or
Sometimes whole crowds were machine-gunned; villages were burned
to the ground.
Some of the worst incidents followed the fierce battles for the town
of Jijiga where thousands of refugees had fled. One of them, Hassan
Khaireh Wabari, a 31 year old merchant, told me "Artillery, bombing
and tank fire devastated many of the buildings before the Cubans and
Ethiopians moved in at daybreak. Sick people and others trying to
protect their homes were shot, and later I saw people beiqg rounded
up and executed with machine guns. At first the women were saved
so that they could be raped. Then they were killed."
Sheikh Ali Nur, a Koranic teacher from Fiiq, near Harer, said that he
had walked many miles to tell me of similar attacks in his area. "They
shelled and bombed us. Then they shot the men, raped the women,
and destroyed the houses. I know that about 130 were exterminated
in my village and about 800 more died in the same district. Even the
animals were shot."^'
Sunday Telegraph, London. April 8, 1979.
78
Copy rightedm atsrial

Kirkham reported that neither side in the conflict gave quarter or took
prisoners. of the major abuses he witnessed was indiscriminate
One
bombing of civilian targets, which had forced thousands of people to
abandon towns and villages and set up temporary grass shelters concealed
in the bush.
We travelled for 120 miles to the bombed oat areas of Malako, a ghost
town deserted and to Oaibo» where the
after bitter fighting last year,
people had scattered into the nearby hills aStci an air raid had wiped
out their village a few weeks earlier. We walked across an acre of
charred ruins and ashes and I was shown cannon cartridges and a three-
foot rocket container as the villagers described what had happened.
The attack had begun at breakfast time when an American F-5 jet of
the Ethiopian air force suddenly swept out of the sky, roaring low over
the huts. The climbed again swiftly without firing and the people
pilot
sighed with but too soon.
relief, Slowly, tiie gêen-and-brown
camouflaged jet tuned and began to descend again, this time followed
by a MiO-21 loaded with napahn. The F-5 made foor nms, sprayii^
American cannon shells and rockets, while the MiG dived on the four
comers of the village* droppmg its deadly napalm in a neat rectangle.
Within ten minutes, Garbo had been turned into an inferno. The people
ran for their lives but in spite of the preliminary warning pass by the
F-5, more than 90 died in the flames or were killed by the strafing.
Others were hideously burned and are beii^ treated in a hospital across
the border in Somalia.^
Followii^ the coonter-oCfensive, SOOÔOO people were hilemally

displacedwithm Ethiopia, and refugees streamed into Somalia.
^ Sunday Telegraph, London, April 8, 1979.
79
5. THE SECRET WARS TO CRUSH THE SOUTHEAST,
1978-84
For most rural people in Harerghe, Bale and parts of Sidamo, the end
of the "official" Ogaden war did not represent the end of suffering and
human rights abuse —
rather, the end of the beginning.
The Ethiopia-Somalia war had profound consequences. Colonel
Mengistn HidteMaiiam wasfmmeasuiab^^ stieqgtheiiea — he gained t)o1h
prestige and a vast new annoiy. He benefited fiom a surge of nationalist
fcelii^ in Etfaiqna and also from international condenmatioa of Somalia's
aggression. President Siad Barre was famniliated, and processes of political
decay and fragmentation in Somalia were greatly accelerated.
The Somali army was gone, but internal conditions in southeast Ethiopia
had not improved, and the Oromo and Ogadeni insurgencies continued.
Large-scale human rights abuses by the Ethiopian army against the local
population increased. Six years of secret wars, that were both more
widespread and more bloody than the official war, were to end in the defeat
of the insurgents, and the creation of widespread famine conditions. Many
of die inhabitants fled to Somalia, where the refugees became pawns in
another political struggle that was slowly degeneiatiog mto dvil war, and
"where hunger and human rights abuse were common.
In early 1978 the Ethiopian government had acquired a new arsenal
from the USSR and had built a greatly expanded army, apeaiheaded by
Cuban combat troops. Though established to combat a oonventionid
invasion, that force was now to be used for counter-insurgency only. The
government could now contemplate cnishii)g an insurgency by brute force
alone.
The victory of the Ethiopian government was aided by dissension among
its adversaries. The WSLF was very strong in mid-1978. However, it
continued to be subject to manipulation by President Siad Barre, who used
itto bolster his position in domestic Somali politics, espedally after
diaoontenled army ofGcen staged an abortive coup in 1978. Hiis led to
disillnsion among the WSLFs erstwhile supporters, and resistance to it
from other Somali groups, notably members of the Isaaq clan.
In Ethiopia, a series of events in 1977-8 conspired to increase popular
support for the OLF. These included the purge of MEISON, which brought
an end to hopes of a negotiated compromise with the government, the purge
of the Peasant Association (PA) leadership, government declarations of
intent to collectivize agriculture, the resettlement of Amhara farmers in
Oromo areas, the enforced use of the Amharic alphabet in the literacy
81
campaign, and the brutality of the 1978 counter-offensive by the army.
Militarily, however, the OLF was in a weaker position due to the huge
build up of the army in Harerghe. The "liberated area" in the Chercher
highlands had to be largely abandoned in 1978-9. Tactics changed towards
a more classic guerrilla campaign. In 1981, the OLF also started to open
a new front in western Ethiopia, in Wollega.
The OLF also gained firom Ifae dedine cSthe SALF, wiiidi was dosdy
associated with the Somali government. Many Oromos had been
antagonized by the evident ambitions of Siad Bane to amiex Oromo areas
and the abuses conmiitted by the Somali army when occupying these areas.
Much of the leadership of the SALF joined the OLF in August 1980,
following large-scale rank and file defections. The SALF suffered further
defections to the newly-founded Oromo Islamic Front (also Somali-
backed), but maintained a small operational presence throughout the 1980s.
Another group, the Sidama Liberation Front, was formed in 1978 and was
active up until 1984.
The OLF remained suspicious of the Somali government, and hence
the WSLF, SALF and Oromo Islamic Front, and tSen was no cooperation
between the different insurgent groups.
Counter-Insurgency after the Somali Defeat
The Ethiopian army occupied the Ogaden for only a brief period after
defeating the Somali army. Six months later, rebel attacks were increasing
in frequency. WSLF was back in control of most of
Within a year, the
the countryside, and thearmy was confined to the towns, the main roads,
and the air. According to journalist William Campbell, 90 per cent of the
lowlands were in rebel hands.^ The OLF was also able to operate freely
in much of the highlands, and held its first congress at Bookhee in the
Giercher highlanck of Harerghe in April 1978.
Conventional battlefield tactics met with limited success against the
WSLF and the OLF. Sweeps and patrols throughout 1979 in tiie lowlands
temporarily reduced the insurgent activity, but failed to engage most of
the rebel forces, and became instead more akin to punitive expeditions,
attacking villages and herds, and forcing another wave of refugees to flee
to Somalia.
^
Africa Confidential, 25.15, July 18, 1984, p. 1.
' The Guardian, London, March 7, 1979.
82
An Ogadeni woman later recalled some of the violence that occurred
when she was driven from her home:
One morning, before sunrise, in July [1979], a woman in my quarters

who was on the road, heard over her ears the noisy click-dang of a
heavy convoy from a long distance, but maiGhing full-speed towards
our home. Running back home, fijgfateoed and crying out, she signaUed
us: '*Wake up! Wake up! They are Abyssinian."
Most of us came out of the house. Unfortunately we found ourselves

surrounded by armed soldiers who inmiediately began to shower bullets
on us before we had a chance of drivii^ our livestock out of the pens.
Most of us were unable to escape together in family groupings. I

remember that my husband and two sons jumped out of the hut together,
but inmiediately rushed in different directions.
After running for a few minutes I saw with my own eyes my nine-years'
old son cau^ by an Abyssinian soldier who mercilessly gâeged hun
by the hair and smashed him to die ground. The young boy was crying
out for mercy, saying: "Oh! Mamma! PappaP and sometimes calling
to the soldier: "Uncle, don't kill me, I am young!" While he was on
the ground at the feet of the soldier asking for clemency, a second
soldier standing by jumped out and bayonetted the boy with a push-
and-twist in the stomach several times so he was dead.^
The village was burned in the attack, and nine people killed: a mother
and her newly-delivered baby, four other children, an older gkl and an
did blmd man. As the group fled towards Somalia, they were again
intercepted by soldiers and two children were lolled. A
baby also died
of hunger.
In late 1979, the government changed its counter-insurgency strategy.
It adopted a four-pronged approach, consisting of:
(1) The forcible displacement of much of the population into shelters

and protected villages;
* Somali, Tigray and Oromo Resistance Monitor (STORM), 2.3, September

1982, pp. 2-3.
83
(2) Military offensives which attacked all people and economic assets
remainiiig outside the shelters and protected villages;
(3) Sponsoring insurgent groups against the WSLF and the Somali
government; and
(3) Attempts to promote the repatriation of refugees.
The counter-insurgency campaigns of 1979-84 wne Uog^ successful

in though small-scale armed resistance was never entirely
their aim,
ehminated. However, this "success" was gained at the cost of much
suffering by non-combatant civilians, and the reduction of much of the
population to a state of famine.
Numbers AffecUd
In mid-1978» when the "official" war was over, there were an estimated
500,000 displaced peisoas. Ttae were almost 200,000 RRC sfadtenm
in Harerghe, 66,000 in Bale, and 20,000 in Sidama^ By October, te
number "cared for" by the RRC in Bale had risen to 350,008; by 1979 it
was 586,000. Tbme were an additional 230,000 in Sidamo. Bale and
Sidamo had been scarcely affected by the Somali army. In 1980, the RRC
claimed that one million people in Harerghe were affected by drought and
war.
By number of "war affected" people who had been relocated
1981, the
in villages amounted to 880,000 in Bale alone, including 750,000 in the
northern part of the province, where the Somali army had never reached.
A further 1.5 million were liviqg in relief shellers.
Meanwhile, refugees streamed across the inleraatiaaal border into
Somalia. In inid-1^8 there were 80,000-^85,000 in camps in Somalia.
A year later there were 220,000; by the end of 1979 between 440,000-
470,000; and by the end of 1980 about 800,000. By 1983, the Somali
government was claiming a total of 1.3 million refugees, though this number
was hotly disputed by the aid donors, who argued that the true number
was perhaps 700,000-800,000. Many of the refugees were not ethnic
Somali but Oromo.
* RRC, The Challenges

of Drought: Ethiopia's Decade of Struggle in Relief
and RehabUUaHam, Addis Ababa, 1985, p. 135.
84
The war had left 600,000 displaced and refugees at the time of its
"official" conclusion in March 1978. After three years of "peace," the

affected population had risen by more than five times.
The fact is that the great majority of the war affected population of
southeast Ethiopia from late 1979 onwards was affected not by the fighting
between the Somali and Etfaiopiaii armies in 1977/8, but by the oonnter-
insnrgency stiat^ of the Ethiopian govemment which was implemented
tarn December 1979 onwards. Many of the people were affected by the
militaiy operations of the army, otheis weie affected by forced relocations.
Military Action during 1979-84
Starting in December 1979, the govemment launched a second military

offensive. Soviet advisors and Cuban troops participated. Tliis differed
from the counter-offensive of 1978 in that:
(1) It was more specifically directed against the population's means of

survival, including poisooinff and bombing water holes and machine
gunning herds of cattle; and
(2) It covered Oromo areas as well as the Ogaden.
At the outset of the 1979/80 offensive, the WSLF

was estimated to
control 60-70 per cent of the Ogaden.^ Hie OLF controlled large areas
of the highlands. The first govemment offensive lasted several months,
followed by a counter-attack by WSLF forces based in Somalia in March
1980, and stepped-up guerrilla action by the OLF. Ethiopian forces then
mounted five major attacks between May and July, which coincided with
counter-attacks by a joint WSLF-SomaU anqr fbice.^
A
new wave of refugees fled to Somalia. Some were interviewed by
journalist Victoria Britttiin:^
I had sixty camels. The Ethiopians waited at the water point and
machine-gunned my two eldest sons and all my camels. I brought my
six young children out on two donkeys.
Âfrica Contemporary Record^ 1979-80, p. B199.
Âfrica Contemporary Record, 1980-81, p. B177.
^ Tke Ouardkm^ London, May 21, 1980.
85
The Ethiopians came twice to my farm in Sidamo, once with white men
[i.e. Cubans]. They took stores of maize, pulled it from the field, beat
everyone in the house. They have taken young men from us to fight
in Eritrea. Nobody is left in my area.
In February there wn
bombing whidi made fires as teas the eye coidd
see. My
camels were bmned and many people ia our fmfly. WUSi
two camels carrying our house I have wdked siiioe then ....
Tbe air focce was deployed to attack villages, animal herds and fleeing
refugees. Reports indicate that napalm or phosphorous was used fteqiiently.
There were also several raids up to 20 miles inside Somalia.
The Ogaden alone during the year following
civilian casualties in the
the Somali defeat were estimated at 25,000.^ Combined with the flight
of several hundred thousand refugees to Somalia, this represented an attempt
to break the WSLF resistance by brute force. Perhaps half of the Ogadeni
population was in Somalia, and half of die remamder in Ethiopian camps
and settlements. Diplomals talked of the depopulatiim of the Ogaden as
the "final solution'*.'
The government offensives nqged well beyond the Qgaden, as witnessed
by Victoria Brittain's interviewees. There was also much military action
in the highlands of Harerghe and in Bale and Sidamo. Many areas which
had been affected little or not at all by the war of 1977/8 were devastated
by these offensives.
In October 1980, there was fighting in the lowlands of southern Bale.
During 1980, the OLF claimed to have engaged the Ethiopian army in 40
major battles, in its operational area of tfie highlands of Harerghe and
Bale.^ As late as December 1980, journalist Qiipg WHesmith was aUe
to travel more dian 100 Uknneteis niside the Ediiopian Qgaden widi WSLF
forces, and testified that most of the countryside was tmder rebel
administration.^^ However, by then die tide had turned; the govomment
counter-insuigency wp««flF* was meetu^ with snooess.
* David Lamb, In tenta tional Herald Tribune, May 14, 1980.
' The Guardian, London, May 20, 1980.
" Greg Wilesmidi, The Observer, London, Deoenber 7, 1980.
86
In early 1981, fierce fighting continued in Bale and Sidamo, and also
in Arba Guuga district of Arsi. The southeast was described as "the most
active trouble spot in the country."^^ The Sidama Liberation Front (SLF)
was becoming more active, largely in response to pre-emptive government
counter-insurgency policies.
The war in Sidamo in 1981 was one of the Deigue's best-kept seciets.^^
In January, 200 people were reported killed by an army patrol at Godaboke
Mito and Chire villages in Sidamo. Between March 19-21, helicopter and
airplane attacks at Gata Warrancha in Sidamo caused at kast 20,000 people
in one valley to flee» and over 1,000 (and possibly more than 2»000) were
reported killed when a "wall of flames" was ignited by bombing using either
phosphorous or ethylene.^'* Ethylene is a heavier-than-air gas which can
be sprayed from the air, whereupon it spreads out, hugging the ground,
and can be ignited by an incendiary to create instantaneous combustion
over a large area. Its use in this attack has not been confirmed by other
independent sources.
The government ordered the evacnatioa of a Norwîan misston staticm
and hoîtal, leavmg the wounded without medical care. In July* 615 were
reported killed at a meetiqg called by local adndtttstratori at Alo. Awell-
documented killing took place on December 1, 1981, when a defense squad
killed at least 48 people, including several entire families.
Throughout the southeast, the army took frequent reprisals against
civilians in localities close to where guerrilla attacks had occurred. In one
credible reported incident between Shilabo and Warder in the Ogaden in
August 1981, houses were burned and 12 villagers were taken hostage and
subsequently disappeared.
One aspect of the offensives which had far-reaching implications for
Ogadeni society was a government policy of poisoning wells, in order to
impoverish nomads and restrict then: movemenls.
Large scale war was effectively over in most of the lowland south east
by 1982, though sporadic guenilla activity oontintted into the following
year. The WSLF was able to make dramatic raids such as stonnii^ die
Ctmtemporary Record^ 1980-81, p. B178.
m Sidamo during 1980-2 are documented

Some of the atrocities committed
in STORM March 1983. The prindpai informant, a reftagee named Tadesse
, 3.1,
Baisamo, was later murdeied by Ethwpian security forces.
**
Reuters, April 27, 1981; St Paul Pioneer Press, April 12, 1981, reprinted
in: J. W. Clay, S. Steingraber and P. NiggU, The Spoils of Famine: Ethiopian
Famine Policy and Peasam ^gricMllKfv, C!nnbridge, Mass., 1988, pp. 224-5.
87
prison in Jijiga on August 12, 1983. In reprisal for that action, the
Ethiopian anny destroyed the villages of Birgot, Midha and Burey and killed
300 civilians."
In the highlands of Sidamo and Harerghe, wideqxead vicdenoe by
government foices continued tfaroughoat 19ffiL IWagt by defeiiae squads
took place in Sidamo in January; on several occasions, &e victims were
decapitated and their severed heads were di^layed in prominent places,
to warn their fellow villagers. In a military sweep than began on November
26 and lasted into January 1983 (i.e. during harvest time), the army made
numerous attacks on villages accompanied by the burning of crops and
confiscation of livestock. Villagers who could not escape were killed.
The survivors languished in relief shelters, suffering disease, malnutrition
and high death rates, or tried to flee to Sk)malia —
though some columns
of would-be lefogees were reportedly interoepted and te detanicca
imprisoned or lolled.
m
On April 1, 1983, a government reprisal for SLF activities dniit^
the previous two months, soldiers lolled 100 civilians in viOige of
Halile, Sidamo.^^ In 1984, the government was able to recapture most
of the areas previously held by the SLF, and forcibly relocated the
population in relief shelters. In Chire camp 3,000 people died, mainly
children, before relief agencies were allowed to provide services in 1984.
The war in the highlands of Harerghe continued in 1984, leading to
the forcible implementation of a large-scale villagization program (see
chapter 13, below).
The war in the southeast was largely a secret war, especially after the
WSLF ceased to take journalists mto the area after niid-1960^ on account
of lack of control of rural areas and pressure exerted ob Somalia by the
Ethiopian government. The incidents referred to above are but a few details
from a much larger story of routine brutality and indiscriminate killing of
civilians by the army. The figures for the numbers of people îq^lacgd
by the war also indicate the scale of human suffering inflicted.
Population Displacement in Counter-Insurgency
A major part of the counter-insurgency strategy adopted by the

government m
the southeast was the forcible rdocaliaii of the populatfcm
mto protected village where their movements could be oootrolled, and
" Afiica Contemporary Record, 1983-84, p. B139.
Africa Caaimporary Record, 1982-^, p. B152.
88
whereby the guerrillas could be denied access to essential supplies. This
program was implemented in Bale between 1979 and 1982, where almost
the entire population was villagized during this period. There was also
widespread villagization in certain areas of Sidamo. More than two million
people were forcibly relocated during the period 1978-82. In Harerghe,
universal villagization began in October 1984, ooinddiiig with intensified
military activity against tiie OLF.^^
The security aspect to tiie pcogiam was ofBdally lecogmzed
vOli^gizatioa
from the while the OLF opposed villagization because it saw it as
start,
an mstrument of government control, Teshale Tessema, of Addis Ababa
University, basing his information on the guidelines of the Central
Villagization Coordinating Committee, wrote:
the villagizationis a direct blow at them [the OLF], by depriving them
of any base from which they could carry out their banditry and anti-
revolutionary activities. Thus the efforts of these groups ranged from
counter-agitation to the burning of houses in new villages ... As some
bandits who submitted said: "the villagization is the Mgliest artillery
blow directed [against the] bandits. With this launching the possibility
ci obtaining fresh food bandits is over."^*
The government drew up ambitious plans for villagization in the south

east. By
1979, nearly 560,000 people had been villagized in northern Bale.
Two years later, this had risen to 750,000 in 280 villages. Villagization
proceeded more slowly in southern Bale —
the program to villagize
130,000 semi-nomads got under way only in 1981, and was completed
in 1984.*' (Southern Bale was affected by the war of 1977/8, northern
Bale was not —
but the latter was the locus of SALF and OLF activity).
In Sidamo, 40,000 were viUagized in 1979, and a further 190,000 gathered
" At the time, the program was generally called "resettlement," but in this
report that tenn is used exclusively to refer to the movement and senlement of
people from the northern regions in the south.
" Quoted in Survival hitemationaU For Oteir Own Chad ... EOtiopia's
Vinagizatkm Programme, London, 1988, pi 19.
UN Coordinating Committee for Relief and Rehabilitation and RRC of

Ethiopia, "Short-term Relief and Rehabilitation Needs in Ethiopia,** March 1981,
p. 13.
89
in sheltefs." FoUowiqg the fightiqg in Aiba Onoga distiict of Aiii»
villagization was impkmoiled tliae in 1962. PkoB weie flcMled to viU^âge
a further 2.4 million.
Commonly, the government would instruct people to relocate at a certain
place within a certain time. people remained behind, punitive
If the
measures would be used. Sometimes, no warning would be given, and
existing villages and homesteads were simply destroyed.
The following testimony of an Oromo refugee who had been subjected
to villagization is one of the very few pieces of direct evidence that is
available about the human impact of the program:
Tbe anny came and started bnming everything. We lan into the forest
with nothing. Some soldieis came and some men in white trucks, and
they told us to go back to our vilk^ and get the others. Then Aey
took us to a place fat from our homes and lokl us to make houses.
They gave us food every day, but there was never enough to save some.
We worked five kilometers from our homes, but if we complained, they
beat us. Also have any doctors and only dirty water, but we
wc didn't
couldn't say anything. told us the Somalis did it to us, but 1 knew
They
it was them. They kept saying it though, and tliey told us Aey were
helping us. Every time we harvested our crops, we had to give them
to tiie government, and tiiey gave us our rations.
I knew was them who burned us because they screamed at us and

it
called us names. They even said they hated us. They had men with
guns around all of the walls —
you couldn't move outside. If your
brother died in the next village, you couldn't go to buiy him. Just work,
they said.^^
A particularly insidious element in the government's policy of relocation

was its repeated attempts to obtain finance fiom the international community
to carry it out. These attempts were partly sncoesBfid —
had they been
more so, doubtiess villagization would have proceeded more quiddy in
Harerghe and Sidamo.
In the aftermath of the official war of 1977/8, it was not difficult for
Ethiopia to obtain assistance to assist the local population and repair some
^ RRQ 1985, pp. 136-7.
^ Quoted in Horn of Africa, "Refugees in the Hom\ 4J, (1981), p. 39.
90
of the damaged infrastructure in Harerghe. Somalia had been the aggressor
in the war, which had led directly to an estimated $1 billion in damage.
Later on, however, the government was not eager to draw attention to the
ongoing war. Displaced people were blamed in vague terms on the war
of 1977/8, and on dnmgbt.
In June 1980, the Ethiopian government started appealing for relief aid
for tiie "victims of drought." QfQdals from the RRC claimed that Haieigfae,
Bale and Sidamo as well as the northern provinces were stricken by an
eight-month drought.^ A reconnaissance team from the RRC had visited
Harerghe in February-March 1980, but delayed releasing its findings for
three months. The published findings indicated an urgent humanitarian
disaster — poor rains had affected one million people out of a population
of three million, mostly in lowlands. The report goes on to say that this
was made worse by the destruction of water facilities in the war of 1977/8 -
- it claims that pumping machines had been taken away by Somalis, and
40 supply points destroyed. Destruction of infrastructure and wells by the
Ethiopian army is not mentioned.^
Several tets about the Jmie 1980 appeal are odd. One is the daim
tiiat a dnmgjit of eight months had caused a major humanitarian disaster.
Such a drought indicates merely the faflure of one of the two annual rainy
seasons in the area —
a conmion occurrence and an indicator of hardship,
but no cause for serious alarm. Moreover, the RRC team had visited the
area before that rainy season was fully under way —
so the distress it found
could not be blamed on the alleged drought. The three month delay in
releasing the findings is itself suspicious, especially in view of the urgency
with which the matter was presented to the western donors. Recalling the
timing of the military operations in the area (i.e. the launch of the principal
counter-insurgency campaign in December 1979), the ffudings of the
reconnaissance team are less surprising, as is the delay in publication until
the security of the area was more assured some time later.
In May/June 1980 a UN
team visited Ethiopia and travelled to some
accessible areas of the southeast. Hie team recommended "the govenunenfs
In the lowlands of southeast Ethiopia there are two dry seasons (December
to March and May to September) and two wet seasons (April-May and October-
November).
^ RRQ Early Waming and Planning Service, "Report on a Reconnaissance

Trip in Hararghe Administrative Region," June 1980.
91
resettlement [i.e. villagization] shouldbe given all possible support."^^
The team recommended that183,000 metric tomies (MT) of food plus
transport be donated to the RRC. 158,000 MT
plus transport were actually
pledged, including $8.1 million from UNHCR.
The rationale for the mission's recommendation was in part that much
of the population had lost its economic assets and was collecting in relief
shelters. Unable to return to an economically active life without assistance,
the population could be better helped by social engineeruig in govenuBent
villages.
Following the RRC appeal in June, a second UN mission visited Ethiopia
from July 6-15, 1980. After consulting with the government, this mission
made the much stronger recommendation that 812,000 MT of food be
pledged, and the funds be provided direct to the RRC for "internal
handling."
If, as the government claimed, drought was now the main problem, the
end of the year saw a return to normal. The main summer rains in 1980
were good. In November 1980, the RRC reported that the food supply
situation in Bale, Harerghe and Sidamo was "normal."^ This did not,
however, stop an increasing flow of demands to intematioiial assistance
for victims of drought and war, and for villagiziitioa.
In early 1981, a mission from the International Council of Voluntary
Agencies (ICVA) visited the southeast as guests of the RRC. Ibongh not
as uncritical as the preceding UN missions, the ICVA team did reconmiend
support for the villagization program. The team noted "tight security
dispositions prevailing" in the villages it visited,^ but did not question
the official explanation that this was to protect the inhabitants from
"bandits." Others believe that the military presence was to keep the
population under tight control.^^
The Ethiopian government failed to obtain all the assistance it asked
for. However, it obtained enough to relocate almost tfie entire population
UN Coordinating Committee for Relief and Rehabilitation and RRC of

Ethiopia, "Short Term Relief and Rehabilitation Needs in Ethiopia," March 1981,
p. 5.
^ RRC, Early Warning and Plannmg Service, "Food Supply Status and Forecast
by Administrative Region," November 1980.
^ ICVA, "ICVA Mission to Assess the Situation with Regard to Displaced

Persons and Returnees, January 16-30, 1982," Geneva, 1982, p. 2.
^ STORM, Z3, September 1982, pp. 9-12.
92
Copyrighted
of Bale, plus substantial numbers of people in Sidamo and lowland
Harerghe, and some in Arsi. UNHCR
assistance rose from $100,000 in
1979 to $2.7 million in 1980 and $7.3 million in 1981; UN
Development
Program assistance also rose. 1981, the Babile shelter was
In early
established near Harer for "drought-affected" Hawiye nomads. Later in
the year, the Bisidiiiio scheme, lesettlii^ nomads for i^cdtnialwoik was
set vp. Voluntary agencies such as Lu&enn World Federation supported
some projects. The ironies of fanmanitaiian agencies woridog within a
counter-insuigency framework were not evident to the staff: one report
noted a large number of widows in the villages, explaining that "the
husbands have been killed or got lost during the Somali invasion."^
Despite the emphasis on drought given by the RRC and repeated UN
missions, all the refugees interviewed by Victoria Brittain in Somalia in
May 1980 — a month before the RRCs major drought appeal —
denied
that drought was the reason for their flight. Instead they mentioned violence
and destruction by the Ethiopian army.
A second element in the relocation strategy was a series of attempts
to obtain the letnm of refugees fiom Somalia. That will be discussed
below.
A final element in the population displacement strafpgy was the
introduction of settler populations from die north, in a small-scale
fwerunner to the resettlement program that was to attract much attention
in the later 1980s. Hie lesettters took land from the locals, who were
thereby displaced.
Two settlements were set up in Bale in 1979: Melka Oda and Harawa.
Harawa was highly mechanized, and was planned to have a capacity of
7,000 families. "The Amharas have given our land to others" complained
refugees in Somalia.^^ Many settlers were given military training and
arms.
Tbe use of relocation as a oounter-insuigency measure is common.
Under mtemational humanitarian bw
it is legitimate only if required by
the security needs of those to be relocated or by inq^^rative military

necessity, if the government provides the relocated population with sufHcient
resources to attain a reasonable standard of living, and if the relocation
avoids unneces5«iry suffering. Aside from the violations of international
law, the relocation policy followed in southeast Ethiopia was objectionable
on several grounds. One is that it was achieved through indiscriminate
* Quoted in: Jason W. day, "The Case of Bale," in Clay et al, 1988, p. 148.
® Victoria Brittain, The Guardian, London, May 21, 1980.
93
violence in areas of the countryside not controlled by the government and
the threat of such violence. A second is that it was achieved by hunger -
- people were obliged to congregate in relief shelters because of the
destruction of the basis for their way of life. A third is that, while
assistance was provided to the people in shelters and government villages,
this assistance was obtained under false pretenses from the international
community. Fortunately the mass human rights abuses that would almost
certainly have followed the large-scale refotilement of refugees in Somalia
did not occur (see below).
Sponsoring Insurgents against the Somali Goi?cnimait and WSLF

The Ethiopian government brought strong pressure on the Somali
government to withdraw assistance to the rebel groups. The pressure
included direct attacks into Somalia, usually by aircraft, notably in 1982,
and sponsoring rebel Somali groups.
The SSDF
In May 1978, immediately after defeat in the Ogaden, officers of the

Somali army staged a bloody but unsuccessful coop attenipC. The suiviviqg
coup leaders, most of whom were members of the Maj^rteen clan, fled
and founded the Somali Salvation Front. This soon amalgamated with two
other organizations to form the Somali Salvation Democratic Front (SSDF),
which gained support from the Ethiopian government. Though the SSDF
initially had a large armed force it failed to have a significant military
impact. Its major military success was a joint operation with the Ethiopian
army to capture the two Somali villages of Balambale and Galgudud in
1983. The leadership was divided and many fighters deserted back to the
Somali government. The SSDF came into oonflict witfi tte Ethiopian
government, reportedly for refusing to mount military operations against
the WSLF. In 1982, the Ethiopian government oonfiacaled annoied veliides
belonging to the SSDF. In 1985, after a spate of assassinations within the
SSDF, Ethiopian security forces detained the SSDF leader Col. Abdullahi
Yusuf and 12 others. Chie of them, Abdullahi Mohamed "Fash** died in
custody in 1986.
In Somalia, one of the consequences of the attempted coup was that
President Siad Barre purged the army, promoting his kinsmen from the
Marehan clan, and also bringing members of the Ogaden clan into more
powerful positions. This led to political conflict between the Ogaden clan
and other Somali dans and q>position movements.
94
TheSNM
Mengistu government began to support the Somali National
In 1981, the
Movement (SNM). The resulting war by the SNM against the WSLF was
an important element in Ac Bmofto stiategy, and it was leaoundingly
successfiil.
The foimation of die SNM was the outcome of systematic discrimination
and human riglits abuse against members of the Isaaq dan, whidi
predominates in northern Somalia, by the Siad Barre government.^
Many abuses against the civilian population of northern Somalia,
especially those living in the border area, were committed by the WSLF.
Killing, looting and rape were common from 1978 onwards. In late 1978,
Isaaq elders petitioned President Siad to form an Isaaq wing of the WSLF,
which would be able to protect local civilians. This organization, known
as Afraad, the "fourth unit," became operational in 1979. It inmiediately
came into armed conflict with the main (Ogaden clan) forces of the WSLF.
Shortly afterwards, an Isaaq army officer arrested 14 leading WSLF
fighters
at Oobyar who had been harassing and abusing the local population; they
were taken to Gebiley sad executed. The army command in Haigeisa was
tiien transferred to General Gani, a Marehan and a clansman of the
president; one of the changes that followed was the forcible transfer of
the Afraad away from the border zone. However, many members of the
Afraad became guerrilla fighters in their own right and continued the inter-
clan conflict, which intensified in 1981.
Other grievances felt by the Isaaq included the preferential treatment
of Ogadeni refugees compared to the local population, in terms of access
to education, health care and services, and discrimination against Isaaqs
in government and army posts and in business. A dispute over access to
the grazing in the Haud reserve was also flariqg.
After prolonged talks, leading members of the Isaaq dan met in London
in April 1981, to form the SNM. In January 1982, fliey negotiated with
the Ethiopian government to obtain a base and arms. The SNMsoon
became active in the border area, supporting the Isaaq clan in its ongoing
conflict with the Ogaden clan. In October 1982, there was fierce fitting
in the Gashaamo area.
The war between the Isaaq-SNM and the Ogaden-WSLF involved
violence against civilians, by both forces and on both sides of the border.
At first, the abuses were almost entirely by the WSLF, because it had a
* See Africa Watch Report, Somalia: A Government at War with its own
Pe(^le, January 1990.
95
near-monopoly on arms. Lorries were ambushed, traders stopped and
robbed, houses looted, animals stolen, women raped and civilians killed.
In early 1979, between Wajale and Alleybadey, two WSLF fighters raped
a woman, whose teenage brother then retaliated by shooting the fighters.
The commands of the WSLF unit then arrived and sammirily executed
the boy and two other family members.
One SNM abuse occurred m December 1981 when Isaaq fighters stopped
a truck at DhabCTooble, between Warder and Degahabur, and killed six
WSLF fighters and 13 civilians, all members of the Ogaden clan.^^
Clashes between the fronts were intense during late 1982 and 1983.
The SNM succeeded in cutting the WSLF off from its rear bases in northern
Somalia. Together with the Ethiopian army offensives, this was a fatal
blow to the WSLF, which never recovered. There was a final round of
fighting in December 1984-Janiiary 1985, but by this time the WSLF was
eS^Mveiy finished.
The Refugee Issue
Hie Ethiopian government engaged in a concerted campaign to make

refugees in Somalia return to Ethiopian territory. The refugees' destination
was to be government-controlled reception centers. On the whole, Oromo
returnees were then villagized, and Ogadenis were given animals.
The returnee issue over this period is complex. Tens of thousands of
refugees returned and were assisted. But the Ethiopian government
consistoitly exaggerated thenumber of letumees, sobjected many to abuses,
and used tfie returnee program for counter-insurgency purposes.
The Somali government was also at fault. Tliroughout the 1980b, as
unportant factor determining what refugees decided to do was die
increasingly unpleasant quality life in the Somali refugee camps. This was
because of the policies and corruption of the Somali government, which
diverted much international aid intended for the refugees, and increasing
levels of violence in Somalia, which the refugees were caught up in and
contributed to. Most seriously, the Somali government recruited refugees
into the WSLF, and after 1983, into the Somali army (see chapter 19).
The Somali government also consistently inflated the estimates of refugee
numbecSy and denied tfie existciioe of genuine returnee movenienbL Cfvifian
refugees were caught between two eiils, and then: letum to Etiuopia often
^ Aocordu^ to one report, a Majerteen woman was also killed, because she
was pregnant by her Qgadeni husband and the unbom child was tliDS an Qgadeni.
96
indicated merely that conditions in Somalia had deteriorated; not that those
in Ethiopia had improved.
The attempts to obtain the repatriation of the refugees involved the
international humanitarian community, especially UNHCR, which often
appeared to be ignorant of the realities of the situation.
The first attempts at lepatiiatkm occnned in May 1980 and coincided
with a visit by senior UN officials, invited by the government. One official
wrote:
Hie UNHCR representative broadly agrees with the [Ethiopian]

government perception of the situation, and feels that on both
humanitarian and pragmatic grounds a comprehensive approach is
needed; this would include assistance for both the displaced and affected
population in Ethiopia thus reducing the incentive to swell the number
of refugees in neighbouring countries. ... [The Ethiopian government]
feels that the UN system is taking a one-sided view of the situation
by launching a large-scale assistance programme in Somalia and doing
almost nothing in Ethiopia. They feel tfai^ this will only aggravate the
situation in attracting a laige munber of people to cross the border.^^
The UN andICVA missions referred to above followed from this

initiative, as did attempts to initiate a repatriation program. The opinion
(never tested) that the presence of international assistance on the far side
of a border "pulled" refugees there —
enticing them to abandon their homes
and trek through a wilderness to a strange country —
continued to have
a substantial and dangerous influence on assistance and protection policies
to refugees in the Horn for years afterwards.
In March 1981, a UN
mission asked for funds for 300,000 families (1.5
miUion people) to be viUagized over the next 18 months. Many of these
were anticipated to be returning refugees. The following month, at the
first International Cooferenoe for Assiûice to Refugees in Africa (ICARA
IX heM in Ocneva, the EtMopian Commissioner for Rdief and

claimed that "as a result of a general amnesty, more than 151,000 Ethiopian
refugees have returned," including 129»000 from Somalia.
However, according to one source, UNHCR officials could provide
details of merely "less than three hundred" retumees,^^ and the Somali
government denied that any had left. One part explanation for this
^ Dr Zaki Hassan, UNICEF Executive Board, ''Report on a Visit to Ethiopia,

13-17 May 1980," Addis Ababa.
" STORM, 1.3, June 1981, pp. 8-9.
97
discrepancy is that the returning refugees did not need to surrender their
ration cards from the Somali camps in order to receive assistance (chiefly
sheep and goats) from the Ethiopian RRC —
they merely needed to prove
that they had been in a refugee camp. It was therefore possible for a
refugee to collect the assistance on the Ethiopian side, and then return to
Somalia to continue drawing rations.
At ICARA I, on the prompting of the Ethiopian govemment, the UN
submitted projects anticipating the need to assist 268,000 letnini]^ refugees
from Sudan, Djibouti aiid Sconalia over flie next three years despite —
the absence of an agreement with either Sudan or Somalia for the voluntary
return of refugees. "It is anticipated that Ethiopians living in Somalia will
return as security and basic living conditions improve in the southeastern
part of Ethiopia" the proposal asserted.^ The UN document asked for
a total of $27 million plus food aid in international assistance for Ethiopia.
In September the government became more ambitious and asked for
aid for an anticipated 542,000 retumees.^^ In January 1982, it claimed
that 567,000 refugees had returned home.
However, these fignres were gross exaggeratioiis. Aoooiding to UNHCR
the following August, "over 10,000^ had returned. In 1981, the ICVA team
had met uidividual returnees, and was assured that several thousand were
living in the settlements it visited, but made no independent investigation
of the total numbers — the figures in its report had been provided by
all
the RRC. However, the "returnees" in el Kere (Bale) turned out to "have
returned from the bush and from Somalia. "^^ In 1982 a team from the
League of Red Cross societies also visited, spending six days on a guided
tour of returnee camps. Team members were told by their RRC hosts that
one of the camps, Degahabur (Harerghe), had held 10,000 returnees some
time previously, but the mhabitants present at the time had never left
Ethiopia."
^ UN Coordinating Committee for Relief and Rehabilitation and RRC of

Ethiopia, "Short-Term Relief and Rehabilitation Needs in Ethiopia," March 1981,
Annex D.
^ RRC, "The Returnee Problem in Ethiopia and Assistance Requirements,"

Addis Ababa, September 1981.
^ ICVA, 1982, Annex III, p. 3, emphasis added.
^League of Red Cross Societies, "UNHCR/RRC/League Programme of

Assistance to Returnees in Ethiopia," QicuUr, September 17, 1982.
96
Despite the absence of an impartial assessment of the situation, in 1981
the UNHCR initiated a small pilot program for returnees, which was
substantially enlarged in 1982, when it anticipated spending $26 million
to feed 200,000 returnees and set up three rehabilitation centers (two of
them in Harerghe), 25 reception centers, and various smaller projects for
letuinees.^ Ifowever, the leactioa of the mtematioiial community was
mixed. While Australia provided 25»000 NfT of food aid, the US refused
to participate.
The Ethiopian government faUed to obtain all the assistance it wanted
for this program. This was related to the lack of a "tripartite agreement"
between Ethiopia, Somalia and UNHCR to repatriate the refugees. Despite
the optimism expressed in the UN submission to ICARA I, relatively few
refugees did return home (in the tens ofthousands at the most), and the
Somali government resisted pressure from the Ethiopian government and
the UNHCR to assent to a program of "voluntary" repatriation.
A major reason why UNHCR promoted the returnee program was
impatience with the Somali gQVonment, which was also "playing the
mimbers gune" and trying to retain its refugee population and exaggerate
its size in order to obtain international assistance. The refugees were pawns
as all sides played politics with humanitarian assistance. However, the
exploitation and abuse of the refugees in Somalia did not justify promoting
the repatriation program.
Skepticism about how voluntary such a program would have been is
warranted, as can be shown by the case of Djibouti. In June 1980, the
Ethiopian government declared its intention of receiving the refugees back,
and shortly afterwards a tripartite commission of the governments of
Ethiopia and Djibouti together with UNHCR was formed to oversee the
repatriation.'^ In July 1981, reports indicate that 20 refugees were forcibly
rq)atriated, of whom 14 were sununarily executed on arrival. The following
two years saw numerous incidents of intimidation and harassment of the
refugees by the Djibouti authorities, and coercion to repatriate. Hiere were
no further reports of executions of returnees, but a number were detained
and sentenced to prison t^ms, despite promises of an amnesty.
Meanwhile, the Ethiopian government legislated against refugees. Under
Article 12 of the Revised Penal Code of 1981, attempting to leave the
country without official permission is a "counter-revolutionary act"
^ Letter from Poul Hartiing. head of UNHCR, to donors, dated April 30, 1982.
* Jeff Crisp, "Voluntary Repatriation Programmes for African Refugees, A

Critical Examination," British Refugee Council, London, 1984.
99
equivalent to treason, and punishable by between five and 25 yeais
imprisonment.
By mid- 1983, the repatriation campaign from Djibouti liad led to the
return of 13,500 refugees, about half through the tripartite program, and
half independently of it.
The Ethiopian government often used the term "returnees" to refer to

both prisoners of war released by the rebel fronts and ictundqg refugees.
Returning prisoners of war were placed in "recq|itkni oeatm" in Oonder,
M eqele and Harerghe, where thm were frequent reports of beating and
execution.
During government of Somalia consistently refused to
this period the
contemplate a agreement for repatriation. When such an agreement
tripartite
was made in 1986, only a small minority of refugees elected to use it —
about 7,200 over four years. However, during the late 1980s, perhaps
500,000 refugees did return spontaneously, as conditions deteriorated in
Somalia and marginally improved in southeast Ethiopia. In 1991, with
the outbreak of widespread violence in Somalia, that return flow accelerated.
The Creation of Famine
During 1978-84 the government of Mengistu Haile Mariam resqpQoded

to the insurgencies in the southeast with brutality, attacks on the economic
base of the population, restricting movement, and creating and exploiting
divisions within the society. This was the bloodiest period in the modem
history of the region. The government's military strategy was instrumental
in impoverishing the people, restricting their mobility and economic ac-
tivities, and creating famine. The activities of the Ethiopian army, the SNM,
WSLF and the Somali government combined to prevent the Ogadeni herders
from freely migrating, trading or cultivating. Meanwliiie, combination
of military offensives and forced relocation left mudi of the Oromo
population destitute.
There was chronic famine in much of the southeast during the whole
period 1979-84, and humanitarian assistance was used as an instrument
for the further extension of state control. When drought also occurred in
1984, the famine became more widespread and severe.
The true story of these campaigns and the related famine remains largely
unresearched and untold. The account given above is merely an outline
based on the few available sources —many of which were produced with
the clear intention of concealing what was actually going on.
100
6. THE R£D TERROR
The Red Terror was a campaign of urban counter-insurgency waged
in the main cities of Ethiopia, notably Addis Ababa, between 1976 and
1978. The name "Red Terror" was officially used by the government, and
itaccurately reflects the way in which excessive violeoce was used to teirify
the population and eliminale dissent. It was one of the most systematic
uses of mass murder by the state ever witnessed in Afiica.
The number who died in the Red Tenor is not known —
it is certainly
well in excess of 10,000. A

full treatment of the Red Terror would require
a separate and extremely lengthy report. What follows here is merely a
cursory examination, in order to place the episode in the context of the
Ethiopian government's counter-insurgency methods.
The Urban Opposition
In the 1960s and 1970s, opposition to the rule of Haile Selassie

crystallized among the educated, particularly university students. These
students were attracted by left wing political philosophy and nursed
grievances over their living and studying conditions* the lade of a student
union and student publications, and the shortage of career opportunities
for them following graduation. This led dire^y to support for violent
methods to overthrow the government. Student songs praised Ho Chi Minh
and Che Guevara, and a popular slogan was "Through Bale not Bole,"
referring to the expectation that revolutionary change would occur through
rural insurrection (as in Bale) and not through returning exiles (who arrive
in Addis Ababa through Bole International Airport).
Reflecting pre-existing divisions in the student movement, after the
revolution, splits soon appeared between different radical elements. The
Ethiopian People's Revolutionary Party (EPRP) was one group, the All-
Ethiopia Soddist Movement (known by its Amharic acronym, MEISON)
was another. Ibcre were some important tacticd differences between EPRP
and MEISON, notably over the Eritrean question,^ but their political
programs both espoused an almost indistinguishable biand of Marxism.
By 1976, the chief difference was that MEISON was prepared to cooperate
^ Randi
Ronning Balsvik, Haile Selassie's Students: The IntcUectiud and Sockd
Background to Revolution, 1952-1977, E. Lansing, Mich., 19&5.
^ EPRP was prepared to allow the Eiitreans to exercise a greater degree of

self-determination.
101
with the military government to achieve communism, whereas the EPRP
was not — indeed it was ready to wage urban guerrilla warfare in order
to bring down the Dcrguc. In the popular perception, MEISON also came
to be identified as a predominantly Oromo organization, and EPRP as
predominantly Amhara — perceptions that became self-fulfilling.
In mid-1976, responding to a government cndEdown on student
members of the opposition, the EPRP began to assassinate kadiqg membera
of the Deigue and its dient institutioiis, notably the mban dwdlei^
associations {kebeles)? TbeEPRP wa8 8ii^)ected of oomplidty inaMed
coup attempt in July 1976. 21 coup plotters were executed, and arrests
of EPRP members began in August. On September 23, there was the first
of nine officially listed assassination attempts on Mengistu. On October
2, the EPRP assassinated Fikrc Merid, a leading MEISON and government
cadre. Ten senior government officials and 15 members of the secret
service were killed in the next two months. The public assassinations
continued into 1977; several hundred were probably killed in this way,
though some of the murders attributed to EPRP may not m
fsu:t have been
carried out by that organization, but either by private individuals or by

government agents.
The Red Terror: The First Wave

The killing of people suspected to be members of the EPRP began in
September 1976. 21 were executed on October 21 and the deaths of a
further 17 were announced on November 18. However, it was not until
the killing of Gen. Teferi Bante by Mengistu in February 1977, and the
latter's assumption of supreme power that the Red Terror was officially
declared, and the mass killings began. Mengistu labelled the EPRP's
sporadic campaign of assassination &
"White Tenor" and Lt-GoL Atnafii
Abate promised "for every revohitioiiary lolled, a thousand counter-
revolutionaries executed.**^ The promised ratio was not to be much of
an exaggeration. Atnafii began organizing "Defense of the Revolutian
Squads," distributing arms to Addis Ababa kebele members who were
considered to be loyal. In a public speech on April 17, Mengistu called
upon the people to fight against the "enemies of the revolution" and
smashed three bottles filled with blood (or something resembling blood)
^ The following account relies heavily

oo: Rene LeFact,£^;UcyMa: An Heretical
Revolution? London, 1983.
* LeFort, 1983, p. 199. In November 1977, Atnafu was one of the

xevolutionaries to be killed — on die ordeni of his loqgstandiog friend, Meogism.
102
to signify the impending destruction of imperialism, feudalism and
bureaucratic capitalism.
On February 26, 44 prisoners were taken to a place outside Addis Ababa
and executed. On March 2, 1977, several people were shot dead by
government forces for distributing EPRP literature during a pro-government
demonstration.^ Over a thousand were detained during searches of the
city on Maich 23-27. May Day had become the occasion for levolutionary
groups to demonstrate their popular support, and EPRP planned to stage
large rallies. Starting on the night of ^pnl 29, rural Defense Squads arrived
in Sie capital, and, together with local kebele officials and soldiers, began
a massacre of suspected EPRP supporters. The official government estimate
is that 732 were killed over the next few days. Others believe the figure
was in fact 2,000 or more.^ On May 7-8, a daytime curfew was instituted
and house-to-house searches were conducted, with thousands being
arbitrarily detained by Defense Squads and soldiers. On May 17, the
Secretary-General of Swedish Save the Children Fund stated that "one
thousand children have been massacred m Addis Ababa and their bodies,
lying in the streets, are ravaged by roving hyenas." He estimated tiiat 100-
150 young people — some as young as 12 —
ime beiQg killed every
night^ On the night of June 4/5, about 400 students were lêd. In total,
at least 2,500 were killed in this first phase of the terror.
Bodies were left on the roadside to advertise the killings of the previous
night — those who inspected the piles of bodies to see if their friends or
relatives were among the corpses were targeted for execution or
imprisonment themselves. Relatives were forbidden to mourn. In other
cases, relatives had to pay one Ethiopian dollar for each "wasted bullet"
in order to have the body returned.
There were also mass arrests of suspected EPRP supporters. Many of
those arrested were subjected to torture, and many "disappeared" after
qênding some time in detention. Relatives were usually allowed to bring
food and clothing to detainees, and learned of flie detainees* trarsfer from
one prison to another when the prison guards instructed than to take their
food elsewhere. Similarly, they teamed of the death or disappearance of
their detained relative when the guards told them that it was no longer
necessary to briqg food. In some instances, the prison authorities deceived
p. 14.
* Amnesty International, 1978, p. 14; LeFort, 1983, p. 201.
^ Amnesty International, 1978, pp. 14-15.
103
the relatives, and continued to accept food for weeks or months after the
detainee had died or been executed.
Rene LeFort described the typical profile of the victim: "Simply knowing
how to read and write and being aged about 20 or less were enough to
define the potential or actual 'counter-revolutionary The authorities were
'
even able to institute a law authorizing the anest of children between eight
and twelve years."*
The Red Terron The Second and Thhrd Waves
The EPRP was largely crushed in Addis Ababa by the first wave of
the Red Terror, and retreated to a rural base in Tigray. However, the
killings and arrests continued. Though the ostensible target remained the
EPRP, the Dergue was now turning on kehele members suspected to be
more loyal to MEISON than the Dergue, and on MEISON itself. Haile
Fida, the leader of MEISON and confidante and ideologue of Mengistu,
was detained in August 1977. After spending several months in prison,
he disappeared. Many other MEISON cadres wm
accested shoitly
afterwards. In October the second wave of arrests and executions todc
place, during which tune an estunated 3,000-4,000 people were lolled.
Much of the killing in October was conducted not by the Defense Squads
and army, but in the course of a civil war between MEISON and the
remnants of the EPRP. Both organizations had been thoroughly infiltrated
by security agents, who were able to assassinate cadres of the opposing
organization, while disclaiming government responsibility for the act. The
continued killings by ostensible members of EPRP also created a justifica-
tion for the Dergue's continuing repression.
By the end of 1977, MEISON members had been thoroughly purged
from the ranks of government and the Uglber offices of the kebdes.
However, many remained at the lower levels, especially in the piovinoes.
The third wave of the Red Tenor tock place between December 1977
and February 1978. 300 were killed on the night of December 16. On
December 21, Defense Squad members (^ned fire with machine guns on
people praying in a mosque. One Ethiopian estimated that during this phase
25-30 people were killed in an "ordinary" day.' By the end of the year,
Amnesty International estimated that 30,000 political detainees were held
in the central prisons and the detention centers of the 291 kebeles of Addis
• LeFort, 1983, p. 202.
'Babile Tola, To Kill a Generation: The Red Terror in Ethiopia, Washington

D.C., 1989, p. 144.
104
Ababa. This figure must be considered a very caulioiis estimate because
of the large numbers held in provincial towns.
By this time, the killings were less public. Most were executed in
prison, and few bodies were left on the streets. However, the killings were
at least as frequent as before. Perhaps 5,000 were killed in Addis Ababa
in these months, and many more in provincial towns.
Though most of the killings —
at least in Addis Ababa —
were over
by Macch 1978, detentioiis and execntioiis cootiaaed tfaiougbout ibit year.
The Campaign Against tlie Mcrdianti
A significant minority of those killed or detained during the mid 1970s

were traders and shopkeepers. This group, and in particular grain traders,
were targets of the new government from 1975. While the campaign against
the merchants — like the corresponding campaign against landlords —
was not strictly part of the Red Terror, it deserves discussion, especially
in the light of the 1980s famines.
Grain merchants were blamed in part for the famines of 1972-4, and
seen as class enemies of the revoluticm. Many large merchants fled abroad
when tiie radical political leanings of the Dergue became dear, or joined
one of tlie conservative poUtical-milita^ resistance movements, such as
the Ethiopian Democratic Union (EDl^ which was militarfly active in
Hgray and Gonder.
In 1973, 90 per cent of all marketed grain was sold through an estimated
20,000-30,000 grain merchants. A small minority of 25 dominated the
supply to Addis Ababa, owning a storage capacity of 100,000 tonnes
between them.^^ This latter group was certainly able to engineer shortages
in the city, though in 1973 their chief contribution to famine appears to
have been to facilitate the export of grain from famine-stricken Wollo to
more prosperous Addis Ababa, where food prices rose a mere 20 per cent
during the scarcity. Over 75 per cent of grain laden were tmXIy based
and operated using pack animals. While fiesetradera drove hard baigains
and occasionaUy reaped windfisU profits, they openiled in a h(^
commercial environment, and perfoimed an essential service in the
redistribution of food.
^ Amnesty brtemational, 1978, p 8.

" J. Holmberg, Grain Marketing and Land Reform in Ethiopia: An Analysis
of the Marketing and Pricing of Food Grains in 1976 after the Land Reform,
Uppsala, 1977, p. 9.
105
The Special Penal Code of November 1974 included an article
prohibiting economic actions which might create or aggravate famine,
implicitly equating them with an attack on the state itself. Article 27 was
drafted in a vague and ambiguous manner, which was open to a variety
of interpretations. The Special Court Martial, instituted at the same time,
implemented the law in a draconian manner, and on the occasions when
a conviction couU not be guaianteed» the Deigne was leady to bypaai the
courts altogether. Artide 27 tfaerefoie acted as a powofiil detorent to
Intimate as well as illegal ecoaomic activity. It is worth quoting in full,
to iUnstiate the vague but iatenscly tfaieateouig euvirouBieat in ^^kfa tiadefs
were compelled to opeiate.
Article 27: Jeopardising Defensive Power <4 the State, Distress or

Famine.
(1) Whosoever intentionallyby commission or omission directly or

indirectly with culpable negligence conunits any prejudicial act
leading to the ocosequenoe ctf weakeaiqg the deiensive power
of the State or beiqg aware of such a CkI fiuls to do whatever
in his capability or creates within the country a grave stale of
misery, want or funine, epidemic or epizootic disease or distress,
especially by improperly hiding or hoarding, destroying or
preventing the transport or distribution of grain, foodstuffs or
provisions, or remedies or products necessary to the life and
health of man or domestic animals, or where the occurrence of
any imminent danger of distress or famine having shown a sign,
fails to do whatever in his power to control it, is punishable with
rigorous imi»isonment from ten years to Ufe, and where the
offence was intentional and where death has occurred or many
lives have perished the penalty may be death.
(2) Whosoever, in time of such distress, fails to carry out or carries

out improperly, except in the case of force majeure,
obligations or liabilities incumbent upon him, whether as a
purveyor, middleman, sub-contractor, carrier or agent, or in any
other capacity, in respect to the delivery or handing over of
provisions, remedies or any other products to be used to prevent,
limit or arrest the distress, is liable to the same punishments.
(3) Where the offender has acted for gain, a fine not emMog
twenty thousand [Ethiopian] dollars shall be impoBed m addition
to the penalty prescribed m sub-article (2) hoeof .
106
Many grain merchants were detained and their goods and vehicles
confiscated. Others withdrew from the grain trade, fearing the same actions.
With no state-run alternative distribution in place, the decline in the private
grain trade contributed to rapid food price inflation and shortages in Addis
Ababa.
The Dergue did not consider these legal prohibitions comprehensive
enough. In July 1976, the Special Penal Code was revised and leêd.
The same pfovisaons remained, but an additional one was added. Article
17(A) concerned "economic sabotage" and prescribed the death penalty
for actions leading to the destruction or withholding of grain, interruption
of work or transport, or "any other similar act."
Immediately after the revised Special Penal Code was promulgated,
seven Addis Ababa grain merchants were charged under this new article,
and sentenced to long prison terms. General Teferi Bante, then head of
state, intervened and changed several of the sentences to death. One of
those executed for "economic crimes" had been caught with 20 tons of
grain in stock — four trucks full, and scarcely enough to influence the
pncQ of grain in a city of ov^ one million residents. The others had been
found with stores of the s^Hce berbere.
An additional motivation for the increased pressure on private grain
merchants in 1976 was that in that year the government set up the
Agricultural Maiketing Coipoiation (AMC), which was to have a monopoly
on large-scale grain transactions. Coercion was needed to enforce the
monopoly.
In the provinces a large number of merchants were executed in 1976-8.
Every small town has stories of traders being killed by firing squads, thrown
into trenches, doused with petrol and burned, or disappearing while in
detention.
A number of other measures were implemented to humiliate and punish
merchants. Many of these consisted of macabre dramas orchestrated by
kebele officials, in whidi the poor exacted revenge on their previous
oppressors. Merchants weie required to participate in auctions, at which
an ordmary object — an egg, a cup of coffee, or a framed photograph of
Meqgistu —would be bid for. Each participant would be obliged to outbid
the others, from fear of a severe punishment. The price might reach five
thousand Ethiopian dollars (or, after the 1976 currency change, Birr) before
the auctioneers were satisfied. In some instances, the object of the auction
was the right to administer strokes with a whip on the back of another
merchant.
This campaign against traders, which continued well into the 1980s (see
subsequent diaptei^ was to have a profound negative impact on rural
people's ability to withstand adversity.
107
The Red Terror in the Provinces
While and detentions were most numerous and most

the killings
publicized in Addis Ababa, the campaign was also conducted throughout
the country, mostly during 1978. Young people in towns such as Asmara,
Gender, Balur Dar and Jinuna are known to have nffoied sevoely. One
document detafling the Red Tenor gives mfonnation oo the foUowu^
incidents:^
* Debre Markos, Gojjam: massacres in October 1976, Febmaiy, June and

August 1977.
* Dessie and Kombolcha, Wollo: batches of prisoners executed.
* Gondcr, Harerghe, Sidamo, Bale, May 1-15, 1977: 1,713 executed,

according to the government.
* Gonder, November 1977: 54 killed when security fotces opened fire

on an EPRP demonstration.
* Before December 1977: 56 prisoners killed in Tigray, 74 in Wollo, 32

in Chebo and Gurage (Shewa), and 56 in Gond^.
During 1978, the government instituted apurge of the leadership of

Peasant Associations, replacing the popularly-elected leaders with
government appointees. Though not as bloody as the urban terror, this
purge was accompanied by many detentions and some executions.
Next to Addis Ababa, the Red Tenor was the worst in Tigray. Three
differentaimed opposition movements were active Tigray at the time -
m
" the EDU, with support firom feudal lords and large tiados, the EPRP,
and the TPLF, which succeeded in capturing the headquarters of Tembien
district,Abi Adi, in early 1977. In Meqele, the pfovincial capital, students,
merchants and rural people suspected of supporting the TPLF were all
equally targets. The following case is an instance of a student suspected
of supporting all three rebel organizations, who was examined by Di Bent
Juel- Jensen, a physician, in Sudan in January 1986:
" BabUe Tola, 1989, pp. 143-4.
108
T.H. was one of 300 plus young men who had been educated by Ras
Mengesha/^ a Tigrayan from Tembien; he had worked in Addis as
an agricultural expert. In 1978 he was imprisoned in the house which
the Emperor had given to [a British professor] together with 470 other
unfortunates. He was accused of belonging to the EPRP. He was
tortured: electrical prods and bums to the soles of his feet and flogging
of the back, bolfa of wiiidi left tenible scats when I saw him *86. m
When they foond no evidence, they changed tack and accused him of
beuqg a member of the TPLF, because he was fiom Hgiay. Moce
torture. He was released after six months, and after an adventurous
flight via Meqele, from where he exapcd dressed as a poor shepherd,
he got to the Sudan. He still has problems waUdqg.
Smaller towns in Tigray suffered too. There, the atrocities of the Red
Terror followed on without a break from the violence against civilians that
was a normal part of counter-insurgency in the preceding years. The
following account of killings was given by Woreda Teka, a farmer and
trader and member of Abi Adi baito (council), to visitDis hi 1988.^ It
is indicative of the many atrocities that occurred in small towns without
gaining any publicity.
Well, to go bade to the beginning, the first incident was a massacre

by government troops. Hiis was in June 1976: 180 people were lined
up in the square (it was market day) and machine-gunned.^^ I
remember it was about 11 a.m. There had been a battle with the TPLF
a few days before this, and the soldiers said they were executing bandits
[wmbede]. In fact they just came into the market and rounded up
anyone they could find. About a quarter of those shot were women:
one had a baby who survived, and we found
it alive and still feeding
half an hour after she had been Hiere were about 5-600 soldiers
killed.
m the town that year. Hiey arrived m
April and stayed until October
when there was another battle. Ibey did have food of then: own but
kept coming round for extra money.
^ Former governor of Tigray and leader of the EDU.
Interview conducted by Sarah Vaugban and Geiry McCann, two visitors

to Tigray, November 1988.
^ Other sources give slightly different numbers: 160 total, or 149 peasants
and 19 students (total 168).
109
The soldiers returned in April [1977] and in August they killed four
people at the town clinic. They said they were TPLF leaders, and were
organizing resistance, but we all knew who they were —
one was a
trader and the other three were peasants.
The next year was the time of the Red Terror, and there was a whole
brigade [1,000-1,500 soldieis] in the area: there were banda [locally-
leoniited mercenaries] and aiilitia as well as regular aimy, and they held
the mountains aU aioimd here. On 29 Januaiy 1978 th^ killed seven
peofde in dieir homes with no ez|ilanation. They dida*t even allow the
families to touch the bodies for a whole day —
they were just left
outside the houses as a warning. Anyone caQght moBming those was
put in prison. Hie victims were:
* Techane Hagos, 28, who worked in the local gpvemment finance

office.
* Lowul Hagos, 25, his brother and a peasant fanner.
* Azanua, 32, originally from Ooader, working in the sanitation
dqMurtmrat.
* Abdel Hakim, 18, a nintfa grade student.
* Three other students from Hagerai Selam whose names I do not
know.
The residents of Abi Adi erected a monument to commemorate those

massacre of June 1976. Each time the army occupied the town,
killed in the
the monument was destroyed. Each time the garrison left, the residents
rebuilt it.
Impact of the Red Terror
Nobody knows how many people were killed, imprisoned, or forced

to flee abroad on account of the Red Terror. A minimum of 10^000 were
killed in Addis Ababa alone in 1977, and probably a OQaq>arable number
in the provinces in 1977 and 1978. A
larger number were detained, and
subjected to appalling prison conditions and torture. An even larger number
became refugees.
The main target of the Red Terror was a generation of urban people
with at leastminimal education. That generation was lost many —
physically removed, with the remainder so cowed and terrified that any
expression of dissent in Addis Ababa was unthinkable for a decade. EPRP
memben and sympatfaizeis^ and odiera with a sunilar social or eduGi^m
profile, were left with a bitter hatred of the Deigne. However, the EPRP
110
and MEISON were essentially crushed. Over the following 13 years the
EPRP engaged in armed opposition in rural areas but never regained its
preeminence, and MEISON became almost completely defunct.
Rene LeFort commented:
History offers few examples of revolutions that have devomed their

own chikben with such vidousiiess and so much cruelty. It can be
esthnated that, of ten civilians who had actively worked for a radical
transformation of Ediiopia, only one escaped arrest, imprisonment,
torture, execution or assassination. The revolution swallowed the whole
of the young generation of Ethiopian intellectuals, that is literates.^^
A was also lost in the Red Terror and the year

class of merchants
preceding Most of those who survived were either forced out of business
it.
or withdrew from fear of reprisals. While the numbers of merchants killed

or detained does not approach the tens of thousands of young educated
victims, this class has a special significance, because its absence contributed
to the famines of the 1980s.
In Tigray, tfie Red Terror encompassed groups not included in these
twodaues. Peasants and uneducated townspeople sufiSered too. Like the
EPRP sympathizers, ahnost all Hgrayans were left with a deep hatred of
the government. Unlike the EPRP, however, their resistance was not
crushed. The TPLF was battered by the events of 1976-8, but survived
and gained popular support. The Dergue was to find that peasant resistance
was harder to crush than urban insurrection.
A final consequence of the Red Terror was that it led the Dergue directly
to an addiction to rule by terror.In terms of crushing the threat from the
EPRP, the Red Terror was a complete success. It was an apparent
vindication of the use of indiscriminate and exemplary violence as a

counter-insurgency method.
" LeFort, 1983, p. 257.
Ill
7. TOTAL WAR IN £RITR£A, 1978-84
In August 1977, Colonel Mengistu Haile Mariam called for "total

people's war" against the "aggressors and secessionists."^ His first target
was the Somali invasion: his second was the Eritrean fronts, which were
in control of 90 per cent of Eritrea.
In December 1977, the EPLF stormed the port of Massawa, but the
attack was repulsed with heavy casualties. was
the turning point of flie
It
war in the 19708. Six numtfas later, the Ethiopian government was able
to redeploy its focces from the Ogaden, and continue to mobflize its still-
expanding army and air force, and counter-attack in Eritrea. Within a year,
the ELF was oo ttie verge of collapse and the EPLF had retreated to the
remote and mountainous district of Sahel. Confined to a few barren valleys
close to the Sudan border, the Eritrean rebels' final defeat seemed only
a matter of time.
In fact, the EPLF was able to withstand everything the Ethiopian
government could throw at it. The Ethiopian army continued to expand
and acquire more sophisticated weaponry, and employ more brutal
techniques. Between 1978 and 1984 the war was waged on an unparalleled
scale —
the numbers of offensives and boBibiqg attacks equalled and then
surpassed the levels ol the southeast Hie cost in human terms, both to
combatants and civilians, was huge.
Whfle the overriding reality of Eritrea in these years was all out warfare,
the government also attempted to employ counter-insurgency methods
similar to those used in the southeast, including population displacement
and control, economic reconstruction in government-controlled areas, and
the return of refugees. It tried to obtain humanitarian funds for this, but
met with little success. However, the input of economic resources, including
food aid was substantial. Thus, while the counter-insurgency strategy

directly created famine conditions, there was some compensatory assistance
to the population.
Army Offensives, 1978-81
In May
1978, using a newly-compleled aiifield in Meqele in neighboring
Hgray, the Ethiopian air force began a campaign of saturation bombing
of positions in Eritrea held by the ELF and EPLF. While many of the
targets bit were military, the bombers also attacked towns, villages and
^ Quoted in: Financial Times, London, August 25, 1977.
113
animal herds. In June, in response to peace offers from the ELF and EPLF
and diplomatic pressure from the USSR and Cuba, the government held
a political conference on the future of Eritrea, but failed to make any
significant concessions to the rebels.
The ground offensive started in July, and in a few weeks captured all
the towns that the ELF and EPLF had held in southern and central Eritrea.
Dunng the offensive, the army handed out basic oaaniodities tet had been
in short supply, such as sugar and soap, to the civflian popukrtioii of the
towns. Any benefits this rather obvious attempt to win fmrmay have
had were negated by a policy of mass detention of people who had
cooperated with the rebel administration.
The second offensive began in November 1978, aimed at the relief of
Massawa and the recapture of Keren. An even larger army was deployed,
including large contingents of armor. On November 25-26, there was a
huge tvr'o-day battle with the EPLF at Elabored, which ended
inconclusively. However, the EPLF was badly mauled and decided to
abandon Keren and the nearby towns, and withdraw to the mountains of
Sahel, whm the temin was appropriate for a last stand. This was called
the "strategic withdrawal."
The ELF, which had taken the brunt of the first offensive, was already
buckling as a military force. By continuing to engage tfie Ettiiopian army,
rather than retreat, it ensured its military defeat.
The Ethiopian attack included a number of incidents of the indiscriminate
bombing of refugees. The journalist Dan Conneli witnessed people leavii^
Keren just before its occupation by the army:
Over 20,000 people streamed northward toward the Sudan border. Some
carried small bundles in their arms, occasionally a battered leather
suitcase on tiieir heads.Aflatbed tmck cruised back into Keren carrying
seven women who had lost their children aloqg the way, their teantained
faces belying their stoical sOenoe ...
On the following day the Ethiopians began an indiscriminate

bombardment of the area with long range artillery, Stalin Organ rocket
launchers and MiG aircraft. Late in the afternoon, three MiG-23s hit
one makeshift refugee camp of 2,000 to 3,000 people some 40
kilometers north of the battle lines. Paramedics carried the 65 wounded
to the edge of the road and tended them there while awaiting EPLF
trucks to take them out after dark ... Among the injured was one family
of five. Berhane Gebreyesus lay on a canvas stretcher while his wife
and three children, also wounded, huddled aiDund hhn. His oneHmd-ft-
half-year-K>ld baby shivered with shock from a head wound that was
114
to claim his young the next morning. Ten had been killed outright
life
in the raid. Thirty more would not survive the next 24 hours.^
The EPLF's "strategic withdrawal" involved removing anything of use

in Keren and the surrounding area. "Not even a nail was left to the enemy.
Everything was dismantled and taken away, piece by piece," said one
eyewitness.' In some instances, particiilarly in the early days of the
witlidrBwal, wlien Keren was abandoned, this descended into somefliing
moie akin to panic looting.
The third oflScnsive took place
in January-February 1979, and consisted
in athree-pronged attack on Nacfa, the headquarters of Sahel district, where
the EPLF had set up its "liberated area" and was beginning to construct
defensive lines. More areas were evacuated in the face of the assault, and
the EPLF was able to dismantle and remove the infrastructure more
systematically.
A fourth offensive was launched towards Nacfa in March 1979, a fifth
offensive in July. The army Chief of Staff wrote a newspaper article
anticipating total victory, entiUed: "Di^ of remnants of seoessimiist bandits
luiking in bushes numbered.*^ Over 50,000 troops were deployed in the
attacks, together with large amounts ci armor. Most of the attacks were
destroyed weU short of their target. Between July 14 and 22, die army
lost an estunated 6,000 dead.^ Many died of thirst while trying to retreat.
Hie war was in a stalemate.
Indiscriminate bombing continued. Visitors to places behind the EPLF
front line told of an average of four or five sorties being flown each day.
Another offensive, launched towards Nacfa in December 1979, ended
in a disaster and rout for government forces. The EPLF was able to
counter-attack and push the army back as far as its headquarters at Afabet.
Along with massive and sustained bombing, the chief military tactic
used by the anny was the deployment of maîed m£uitry and armored
columns, driviog up the nairow rocky valleys towards the emfdacements
of the EPLF. Tens of thousands of conscript soldiers with minimal training
* Quoted in Htrnt of Africa, 4.1, (1981) p. 23.
^ Paul Brutsaert, Belgian Committee for Medical Assistance to Eritrea,

"Eyewitness RqNnt," 1979.
" Quoted in: Christopher Clapham, Transformation and Continuity in

Revolutionary Ethiopia, C^bridge, 1988, p. 208.
Âfrica Confidential, 20.17, August 22, 1979, p. 7.
115
were marched towards death, disablement or capture. Over 20,000 were
killed during 1978-9 alone. Many of their skeletons, bleached by the sun,
still litter the valleys of northern Eritrea, together with the hulks of tanks.
A prisoner of war commented: "Eritrea gobbles up entire divisions."^
The failure of the massive conventional offensives of 1978-9 led to
a change in military tactics in 1980 and 1981. Dawit Wolde Giorgis, who
was political commifliioner for Eritrea at flie lime, argues that policy
was abandoned in 19807 Otefloaioes note that from
of "sooidied earth"
1980 onwaids there were more attacks on noa-combataats, indndiqg
stepped-up aerial bombardment and a greater level of harassment in the
towns. What appears to have occurred is that fewer iaige-ecale offensives
were launched, with correspondingly less accompanying widespread
destruction of everything in the army's path. For two years, while the
government prepared its next offensive, the war was fought more as a
pacification campaign.
In December 1980 the government launched a relatively small and
ineffective attack, which petered out without military gains to either side.
1981 passed without a major military offensive.
The policy of setting up protected âirisoii-viDi^.aloqg mads, funiliar
from 1966-71, was revived, albeit mitudly oo a sbuU scale.* Peasant
Associations were also set up in south Eritrea, to provide doeer coDtioI
of the population. Curfews and restrictions on movement wm
reimposed.
Most villages had only one or two official entrances, and people attempting
to enter or leave through other routes were liable to be detained or shot.
Land mines were planted on military lines, to prevent pxinetration by
the EPLF and defection by soldiers, and around protected villages and other
areas used by civilians to constrain their movement.
Soldiers guarding villages and military patrols exacted a continuous
toU on dvflians. Hie Eritrean Relief AssodatioM (ERA) produced figures
for the civilian victuns of these regular patrols^ wldch roughly correspond
to independent estimates made by people in government held areas. ERA
daimed that between January and June 1980, the army detained 1,475 rural
people and executed 240, and soldieis raped at least 110 women. Nearly
500 cows and 80 tons of grain were confiscated from peasants, and 500,000
^ Quoted in: Pietro Petrucci, "The Kiemlui^ Africa Sidmctt", reproduced in:
Eritrea Infarmatumt November 1979.
^ Dawit Wolde Giorgis, Red Tears: Revolution, War and Famine in Ethiopia^
Trenton. N.J., 1989, p. 99.
•
Africa Confidential, 20,17, August 22, 1979, p. 8.
116
people were forcibly displaced by the joint policies of relocation and
harassment by patrols and aerial bombardment.
Bombing raids, which in 1978-9 had mostly presaged ground attacks,
now became a regular part of the war of intimidation and attrition. In the
first half of 1980, ERA estimated that bombing raids had resulted in 390
honses being destroyed and 463 cattle killed.
The fiist moadis of 1980 also saw significant devetopments within the
lebd&onts. IbeEPLF was able to launch a counter-attack on goveniment
positions in early 1980. A few months afterwards, the alliance between
tfie EPLF and the ELF which had held since 1975 began to break, and
quickly developed into an irrevocable split. There were some armed clashes
between the groups, for instance in August 1980, but large-scale civil war
was avoided in part because of military weakness of the ELF. The TPLF
assisted the EPLF in its attacks on ELF positions. Most of the ELF fighters
retreated into Sudan, where they were detained and disarmed by the Sudan
government. The last major group arrived in Karakon, eastern Sudan, in
1982. The ELF, akeady rent by schism, split still further.
Estimates for flie total numbo' of people, both combatants and dvilians,
killed between 1978 and 1980 agiee on a f|ôf between 70,000-80,000.
In 1978 there were 250^000 Eritrean refugees m
eastern Sudan (up from
100,000 in 1975); by Sqitember 1979 there were 390,000; and by March
1981,419,000. Tbe worst was yet to come.
The Red Star Campaign
The size of the army continued to rise every year. By 1982, the total
manpower stood at an estimated 245,000, and further mobilization was
proceeding apace.^ Material and logistical help was provided on an ever-
increasing scale by the USSR, and extensive Libyan support was also
provided.
After the conqMuative hill of 1980-1, 1982 was to be the worst year
of war mEritrea to date, m
y/hidtk die government made an all-out attempt
to crush the EPLF. It was also the year in which the government tried
its most systematic attempt to use less destructive oounter-insuigency
methods, including economic reconstruction.
In January 1982, Mengistu moved the national capital temporarily to
Asmara. By this time nearly two thirds of the army was stationed in Eritrea.
In a speech on January 25, Mengistu aimounced the Multifaceted
' NOVIB, "War and Famine

in Ethiopia and Eritrea: An Investigation into
the Arms Deliveries to the Struggling Parties in Eritrea and Tigray," Zeist, the
Netherlands, 1991, p. 12.
117
Revolutionary Campaign, and in response to the planned US "Bright Star"
exercises in the Middle East, dubbed the forthcoming offensive the "Red
Star." He described the EPLF as "anti-freedom, anti-unity, anti-people
and anti-peace bandit gangs" and "the pitiful dregs of history" and
confidently predicted their imminent demise. This campaign was the
end result of two yeais of planmng and preparations.
The offensive started with a campaign of satmatkm bombii^ NmAi
was bombed four or five times a day imd —
in a new development —
often at night as well. Phosphorous and duster bmnbs were used. Hie
EPLF alleged tiiat chemical weapons were also used, but this allegation
has never been proved.
The ground campaign opened with activity on seven different fronts,
including Tigray, and a thrust up the Sudanese border. The Sudan
government allowed Ethiopian tanks to cross Sudanese territory to attack
the EPLF in the rear. There was aerial bombardment on trade routes
between Eritrea and Tigray, to disrupt supplies and conmiunications between
the EPLF and TPLF.
Hie Red Star offensive involved the largest muaber of troops ever
deployed in Eritrea —
more tfian 120^000 were involved in the attacks
on the EPLF base areas. The sheer number of scddiers in the tenitoiy put
unexpected strain on the food resources, and the government was compelled
to institute an airlift of food to Asmara for the army.^^ The offensive
saw a return to the "scorched earth" policy of 1978-9, though on a larger
scale. The enormous level of sustained aerial bombardment and ground
attack devastated large areas of northern and western Eritrea.
The conscript soldiers in the Ethiopian ranks were used for massive
assaults on the EPLF positions around Nacfa, in the hope that sheer weight
of numbers would overrun the rebel lines. It did not. The EPLF were
outnumbered by eight to one but had the advantage of an doellent defensive
position. The advandng columns were repeatedly ambushed and then
machine-gunned as they stormed the EPLF-held BKmabdnsides. There
were perhaps 40,000 casualties among the government forces.
The Red Star campaign also involved other oounlaF-insnigency elements*
including forced relocation, attempts at economic reconstruction, and
attempts to obtain the return of refugees from Sudan. As initially conceived,
the campaign was to be "multifaceted," with primacy given to the "hearts
Clapham, 1988, p. 209.
" Africa Confidential, 25.5, March 3, 1982, p. 8.
118
and minds" component. As actually implemented, the military aspect
dominated.
Just before the military offensive was launched, the EPLF staged a
gnorilb nid on the militaiy aiiport in Asmaia, and destroyed a number
Qfaifplanes. Hie ITLF also made attacks near Meqele. Hiese emphasized
the government's need for more widespread counter-msurgency measures.
CoDtnd of the Popnlatkm
Throughout the year, tighter control was exerted on all civilians living
in government-controlled areas. Curfews were enforced from dusk or
slightly afterwards and movement was restricted. Those wishing to travel
needed to produce an ID card, an up-to-date rent book, tax clearance, proof
of future return, and (in the case of skilled people) a signed statement by
a guarantor who provided a bond of 25,000 Bin. Only ten could a travel
pennit be issned, though payment of bribes was also usually necessary.
A macabre joke common amoqg Eritieans was that in older to travel to
attend a funeral it was necessary to ap|^ for a permit a week before the
person died.
A number of means were employed in order to maintain surveillance
of the population. Apart from the regular activities of the security services,
such as phone-tapping and interception of mail, there were attempts to
encourage civilians to spy on each other, and to provoke signs of dissent,
so as to identify non-government supporters and enforce conformity. In
elections to kebele committees, all the residents of a neighborhood would
be called together. The government's list of nominations would be read
out, and then the assembled citizens would be adced for additional sugges-
tions. Those with the temerity to make a suggestion would be singled out
for surveillance and possible anest. Individual dtizoos were asked to help
organize frequent "political'' and "fimdraising" meetings* with obligatory
attendance and "voluntary contributions" by all. Those who participated
in the organization would have to report on the enthusiasm shown by the
co-organizers. People who failed to attend would be subject to reprisals.
Some "contributions," ostensibly for objectives such as reconstruction and
the literacy campaign, were deducted from wages at source, others were
donated at the supposed ''social'* functions. Non-payment would lead to
reprisals.
One consequence of the tighter restrktionB was that rural people on
both sides of the battle lines, îSiio had hiterto been able to cross the lines
with relative ease to obtain mariceted or relief food on tbe other side, could
now do so only with much greater difficulty.
119
The reconstruction element in the Red Star campaign included plans
to rebuild several schools, hospitals and factories. A special levy of ten
per cent was introduced on the salaries of all government employees. The
government claimed that $100 milliaa was spent on lecoastnictioD in
Eritrea, indtt^ng $3 million on flie demolition oCapaUic garden and its
re|dacement a concrete stadium and "revolntioo square."'^ The troe
amount ^nt is not known, bat by mid-1982, much of Eri1rea*s indnaHy,
out of action since 1977, was fonctioniqg agsin.^
ReftigMS
The Red Star campaign coincided with Ethiopia's (successful) attempt

to promote the return of refugees from Djibouti and the (unsuccessful)
attempt to encourage refugees in Somalia to return. An attempt was also
made to ensure the repatriation of Ibe 400,000 Eritrean refugees in Sudan.
The loots of Ibe aHeiwpts to obtain repatriatioii of Efilieaiis went back
to 1981. In April of that year, the EtUopiangoveinnieDt,1faioBgh the IJN,
prewnted to dbe fiist fntemational Qmference on Assistance to RefhgBes
in Africa (ICARA I) a project for an anticipated 100,000 Eritrean returnees
over three years. TTie UN and the government jointly daimed: "Over the
last 18 months, Ethiopia's contacts with the governments of Djibouti and
Sudan have provided a favourable context for the repatriation of refugees
from these two countries."^^ The government had not, however, made
contact with the refugees, who expressed no desire to return under the
prevailing circumstances. The government claimed that a pilot scheme
catering for 10,000 returnees bad already been set up in Keren, Eritrea.
Dr Abdel Rahman al Basbk, the Sudanese CnnmwsBMiPBr for Refugees,
denied that 10,000 refugees had left Sudan for Eritrea —
socfa an outflow
could not have gone unnoticed, and the staff in his offices on the bonier
had not seen any movement of refugees back to Eritrea. The Ethiopian
govmment and UNHCR did not comply with a request by the Sudanese
government for an independent mission to evaluate the numben of retoniees
and their condition.
" AfiruM Cmaemporwy Record^ 1981-2, p. B1S9.
" In the late 198Qs much of the industrial equipmept was difimanticd and taken
to Addis Ababa.
^ UN Coordinating Conunittee for Relief and Rehabilitation and RRC of

EAiopia, "Shoct-Temi RfiUefandRefadiiUtation Needs in Ethiopia,'* Manfa 1981.
120
In September 1981, the RRC made more substantial requests to UNHCR.
Claiming that refugees "had been forced to move [by the Sudan government]
with a view to lure international assistance," the RRC appeal document
asked for $116 million. This was based on the claims that 22,000 Eritreans
had returned so far, and a further 340»(XX) would be coming back within
four years. (58,000 letmnees from Tigray and Oonder were also
antictpated.)^
Tbe 1981 rqMrtiiation initiative fell short of the govenuDent^ hopes.
UNHCR donated only $13 millioa for the supposed 10,000 Eiitiean
returnees.^ Attempts to encourage the return of refugees continued over
the following years. In April 1982, the UNHCR proposed setting up a
sub-office for returnees in Asmara, several reception centers, and a
rehabilitation center at Ali Gidir, near Tessenei close to the Sudan border.
The UNHCR and the Ethiopian Embassy in Khartoum agreed that any
refugee in Sudan could register at the embassy and obtain free and safe
passage home. In 1982, 424 registered; in 1983, 142 did so.^^ Despite
the evident lack of demand for iqwtiiation, UNHCR attempted to set up
a tripartite commission between ilaelf and the Sudanese and Ethiqiian
governments in July 1983. The Sudan government declined, endoising
tfie view of its nutfor donor, the USA, that repatriation was not feasible
until political conditions in Ethiopia had changed. However, UNHCR lock
unilateral action and sent a mission to eastern Sudan to assess the prospects
for repatriation. The mission arrived in Kassala in January 1984 on the
same day that 500 Ethiopian soldiers arrived in the town demanding asylum
after their garrison at Tessenei had been captured by the EPLF. No more
was heard of "voluntary repatriation" after this embarrassing incident.
In 1982, the Ethiopian government stepped up pressure on the Sudan
government to cease giving asylum to Eritrean refugees. The Sudan
government had abandoned its offensive policy towards Ethiopia m 1978,
due to domestic political coosidentioaB, lecognitioo d the mflitaiy
ascendancy of the Deigue, and the foflue of Fkesideiil Jaafer Nlmeiri to
obtain support for Eritrean mdepeodence from te Oiganizatioo for African
Unity. A period of concUiation with Ethiopia followed. In 1979, Sudan
failed to protest an Ethiopian military incursion m pursuit of Ethiopian
"RRC, "The Returnee Problem m Ethiopia," Addis Abiba, September 1981.
This amounts to $130 per head. UNHCR give $15.9 milliott to the 441,000
Eritrean refugees in Sudan, or $36 per bead.
"Ahmad Karadawi, "Refugee Policy in the Sudan, 1967-1984," DPhil thesis,

Oxford, 1988, p. 331.
121
Democratic Union forces, and during the Red Star offensive, Ethiopian
tanks crossed Sudanese territory with the prior agreement of the government.
In July 1982, Colonel Mcngistu pushed further, threatening the Sudanese
Vice-President with unspecified reprisals unless Sudan withdrew support
for Eritrean refugees/* The Sudanese government refused, pointing out
that gave only access for the EPLF to Port Sudan $ad humamtarian
it
assistance. However, at that mcMnent the Sudan govern ent was itadf m
cfaaogiiig towards a more assertive poliqr of Buppoi^^
in E&opia. Hiougfi Sudan never provided smitary supplies, traimi^ or
bases, free access to Sudanese territory and intelligence sharing was
provided, a policy that was to continue until the overthrow of Prraident
Nimeiii m
April 198S.
The War in 1983-4
The Red Star offensive failed. By May 1982, it had failed to capture
Nacfa, and it was unofficially abandoned on June 20. The EPLF was even
able to counter-attadf and push government lines back. Having been
launched with huge publicity, the offensive ended in complete silenoe from
the government media.
1983 saw an offensive in March on the Halhal front, north of Keren.
Known as the "Stealth Offensive" because of the lack of publicity
surrounding it, government forces succeeded in overrunning EPLF lines,
but not in inflicting a significant defeat on the insurgents. Attacks continued
until August, and severely disrupted the planting of crops in Senhit and
Sahel districts.
In early 1983, the administration in Asmara made an estimate for the
total number of casualties that had been incurred in the war since 1975.
It estunated that SNlÔOO Ethiopian soldiers had been killed or wounded,
together with 9,000 guerrillas. These figures are very credible. The
estimated number of dvflian casualties was almost unbelievaUy Ugfi:
280,000. In total, over 250^000 deaths were attributed to tte war since
its outbreak in 1961 (presumably indudii^ those due to fauqger aod
displacement).^'
During 1983/4, the Ethiopian army underwent its largest growth to date,
surpassiqg 300,000 men. National military service, was announced in May
David A. KssmÊthiopui, Ae Umted Slates md Ike Soviei Umom^ London,

1986, p. 82.
"Dawit Wolde Giorgis, 1989, p. liy, Afirica Contemporary Record, 1983-4,

p. B133.
122
1983 and began to be implemented in January 1984, and the 60,000 new
servicemen were trained and ready for service six months later. Further
supplies of military equipment were provided by the USSR.
EPLF went on the offensive, making some significant
In early 1984, the
gains (the town of Tessenei was captured in January, causing the
I
endMnassmcnt to tbe UNHQl mi^^ Tlie govemment

tcsjpooM by another round of aerial bombardment, and by an offensive
launched on 27 October (see chapter 10).
Bombing "Everything that Moves"
Major Bezabih Petros, a pilot who was trained at both Williams Air
Force Base, Arizona, and Ligov Air Base in the USSR, was shot down
and captured by the EPLF in April 1984. In captivity, be had this to say
about the bombing:
We definitely know civilians will get hurt. But, knowing that the people
sympathize with the rebels, the Older is to bond) eveiyflih^ that moves.
In July 1979, Dan Connell witnessed the bombing of a group of pastoral

semi-nomads on migration. Two women and a boy were hurt in the attack,
together with five camels hurt, and one killed. Connell commented that
the family group could not possibly have been mistaken for a military target.
An important element in the government's bombing strategy was
instilling fear in the civilian populations. The bombing, and the measures
needed to cope with it, such as living by night, posed enormous practical
problems and were demoralizing for the civilian population. An Eritrean
refugee woman in Sudan explained why she had left her country: "I was
tired. I wanted just to walk outside in daylight without needing to look
into the sky and fiear for my life and my chikben.**' It was not even
necessary for the aircraft to drop bombs mOlder to inspire fear. They only
needed to screech overhead at unpredictable mtervals to remind people
of their deadly presence, and the need for constant vigilance.
These air raids not only caused direct physical damage to peqsle, animals
and material infrastructure, but damaged the very social, economic and
even psychological fabric of society. The constant fear of bombing in
EPLF-held and contested areas literally drove life underground. Everything,
whether civilian or military, had to be camouflaged from air attack. Schools
and hospitals were located in caves and under trees. Only essential
^ hiteiviewed by Alex de Waal, Wad Sherild, Sudan, Fefaruaxy 1989.
123
I
movements occurred by day — wheeled transport and much agricultural
all
activity had Markets had to be held at night. Animal

to take place at night.
herding, which requires vigilance in order to make sure animals do not
stray and must be carried out during daylight became dangerous, and
increasingly restricted to forested areas and twilight hours. Herders reported
that the Ethiopian waiplanes appeared to take particiilar delight in "hunting"
camels.
Even people not physicaUy wounded have been left witfi pennanent
psychological scars after bombardment. The survivors of air la^ luve
d^cribed being distraught with grief, delirious with fear for days or even
weeks, subject to sudden-onset panic attacks which leave them in a state
of acute anxiety, or prone to ever-present fears during the daytime and
nightmares during the darkness. Some children become frozen with tenor
at the mere sound of an airplane.
The war in Eritrea was fought on an unprecedented scale betweoi 1978

and 1984, with the 1982 campaign maricing the worst point Hie desbuctioo
and disruption caused by the war was instrumental in creatiiig the funine
which developed, thou^ a number of aspects of the situatioD prevented
the famine from becoming as severe as in neighboring Tigray.
Famine occurs when a society no longer becomes socially and
economically viable. In Eritrea in the 1980s, this occurred on account of:
(1) the direct destruction caused by war;
(2) the restrictions on economic activities caused by war and the other
cotmter-insurgency measures;
(3) the degradation of the natural environment, which was caused in

part by war and counter-insurgency measures; and
(4) rainfall failure.
This section will look at items (2) and (3); item (1) has essentially been
covered, and item (4) —
the repeated partial or complete failure of the
rains, starting with the winter rains of 1980/1 —
lies beyond the scope
of this report.
124
Ihe Impact cf War and Counter Însurgency Measures
One of the most disruptive aspects of the war was that it caused much
of the population of Eritrea to be displaced. By March 1982, there were
440,000 Eritreans in Sudan alone. In 1983/4, a further 120,000 fled to
Sudan, though many later returned. The 1985 offensives caused a fiuther
190,000 to be displaced inside Eritrea, and 30,000 to flee to Sudan.
The policy of relocation in protected villages caused severe disruption
to the livelihoods of farmers and herders.
Andu is a pastoralist-farmer from the small village of Adi- Werhi

Kifle
in Eritrea's Hamassien plateau. Andu and his two adult sons, Mehari
and Keleta, used to take advantage of tfie winter ndns (November-
February) by moving their livestock to the bahri (green bdt [down the
escarpment close to the Red Sea]), 140 km east of Adi-Wmi where
the rest of the family remained. Farming on the mainly state-owned
dominale land in the green belt, Andu and his sons would get a
reasonable harvest in February and March and return to Adi- Werhi by
April, bringing sacks of grain, butter, ghee [clarified butter] and salt
as rewards for their labour ...
At the onset of the May one son would return to the green belt
rains,
oxen left their with relatives. From May
to collect the pair of draught
until October the whole family was busy cultivating in their plateau
village. After the harvest and a few months* rest, the fnnung cycle
started again.
This was the case until 1985, when Andu Kifie and his famfly, along
with several hundred villagers from surrounding area8» were transferred
to a new security hamlet called Inwet, where movement is restricted
to a 10 km radius. Their whole agricultural system and pattern of life
broke down. Andu sums up the feelings of many peasants: "We are
like voiceless prisoners in these security hamlets. We have some oxen
but not the land to plough, we have the cattle but we cannot graze freely.
What choice do we have other than to starve?"^
In western Eritrea, many fanners were similarly forced to abandon

seasonal farming on the flood-ietreat of the Cash river. Many farmers
^ Zemnariam Fte, "The Legacy of War," m PANOS^ Crreenwar: Environment

and Conflict, London, 1991, pp. 140-1.
125
were thus able to harvest only one crop per year instead of two, and many
herders were forced to abandon their established practices of seasonal
migration for grazing. In addition, many areas were rendered unusable
by trenches, fortifications, free-fire zones, and land mines. A survey in
1987 found that land had been mined in almost ten per cent of the villages
covered, and that two huge aieas in Sahel and Akele Guzai had been
rendered completely off-lunits by mining.^
Overall Impact of Uie War on Food AvailabUUy
In 1987, a team from the University of Leeds visited Eritrea, femi both
sides,and calculated the food production of the territory, and the constraints
upon The team found that Eritrea was, even in normal times, a food
it."
deficit They estimated that in a "normal, non-war" year the
area.
production of staple foods would be enough to feed the population for
between seven and seven and a half months. In a "normal, war" year that
figure fell to 4.6-4.8 months. This implies that the war was costing
between 65,000 and 95,000 metric tonnes of lost food production per year -
- about half of the normal food deficit.
Impact on the Grain Trade
Normally, Eritrea's food deficit is made good by trade. The grain trade
was severely disrupted by the war. Up to 1975, much of highland Eritrea
and the towns were fed by grain imports from Sudan and the surplus-
producing areas of western Tigray. Over the following decade, there were
constant interruptions to one or other trade route.^
In 1975 the war interrupted the supply route from Sudan. In 1977,
export from Tigray was stopped by the insecurity there and the flight of
large Tigrayan landowners to Sudan. Fortunately, supplies from Sudan
re-started shortly afterwards. During 1977/8, flie key route to Sudan was
^ Lionel Qiffe, The Impact of the War and the Response to it in Dijfferent
Agricultural Systems in Eritrea," Development and Oiange, 20, p. 378.
^ Lars Bondestam, Lionel Cliffe and Philip White, "Eritrea: Food and
Agricultural Production Assessment Study, Final Report," Leeds, 1988.
^ The following information derives largely from interviews with grain traders
conducted in Kassala, Sudan, in March and April 1989 by Alex de Waal.
126
almost completely controlled by the ELF, which ensured physical
security, but the ELF issued very few permits for cross-border trade in
grain, restricting the number of merchants who could operate and the
amounts they could import.^ In 1978/9, the ease of trade briefly improved
with government control of the major routes. While the government never
imposed the same level of restrictions on trade as it did further south, it
did require all traders to move by escorted convoy. There was also, during
1960-2, increased security in western Tigray and taqpoits to Eritrea resumed.
Intennittent attempts by the TPLF to restrict the export of grain to Eritrea
were ineffectual, and at most a 10% tax was levfed. After Ae collapse
of the ELF in 1980, there was increased banditiy in western Eritrea, which
compelled merchants to organize their own armed convoys. From 1981
onwards there was greater harassment of traders suspected of trading with
Sudan. Greater restrictions on trade were imposed during the Red Star
campaign, and in 1983 the heavy fighting in western Tigray interrupted
supplies from there.
Markets in EPLF-held areas and transitional zones were accessible to
civilians from government-held areas only at considerable risk. Mohamed
Idris, a fanner in western Eritrea, explained:
When we return fmm here [an EPLF-held area] to oiir villages, we

don't know what happen to us, because there are spies who will
will
be asking where we have gone. We have taken a risk to come here.
In order not to be killed, we will have to lie and say we travelled to
Kassala [Sudan] to get food. The enemy looks after us like goats —
our whole living situation depends on tricking the enemy.
When we are away from our homes like this, we fear for our women.
The Dergue rapes and harasses our women. But what can we do?^
Women play a key role in the economic life of rural Eritrea. Petty
trading, short-distance migrant labor and oth« activities done by women
are to a household's income m
times of stress. As is dear from Uie
quotation above, fear for the safety of women greatly ponstnuned what
^ Only the town of Baieatu lemained hi government hands.
^ There is ^pectilation that the profits iMch accrued to a small cartel of traders
from Has practice were an in^Kictant source of finance for the £LF.
^ Quoted in: Barbara Hendrie, "Field Report — Eritrea," Enmgency Relief

Desk, Khartoum, April 1986, Appendix B.
127
they were willing or able to do, with resulting loss of income and food
to the family.
These disruptions had serious consequences for access to food in many
parts of Eritrea.The price of food rose as a result, and the poor suffered.
Ihe In^ct an the Natural Environment
Eritrea is a semi-arid area, with a fragile ecosystem. The fameis and

herders of the territory had evolved a range of strategies to be able to
ensure, so far as possible, a sustainable livelihood. The pressures of
sustained land use, exacerbated by rainfall failure, was already putting
pressure on Eritrea's natural resource base. The war contributed to this
process of degradation of the environment in a number of ways:
* Trees were cut to make trenches, gun emplacements and village

fortifications. The EPLF enacted a ban on the cutting of live trees, but
it is unclear to what extent tfiis was enforced. Hie los& of trees not
only directly aeated deforestation, but forced pastonlists to browse
their animate on other and possibly less suitable trees, causing furtiber
loss of tree cover.
* Large areas of forest in the district of Semhar were burned by the army.
The rationale for this was that the forests allowed guerrillas to approach
the strategic Asmara-Massawa and Afabet-Massawa roads without being
seen.
* Eucalyptus woods at Biet Giorgis around Asmara and some other towns
were cut down for the same reason.
* The mterruption of the charcoal and firewood trade from soutfawestem

parts of Entrea led to people obtaining wood ftiel from tfieir own
localities.
* The blocking of migration routes used by animal herders forced the

herders to keep their animals in one area throughout the whole year,
putting additional stress on the pasture and browse of these areas. The
same factor caused herders to switch to farming, which places more
stress on the land.
* Large areas of land were rendered unusable by land mines^ forcing

farmers to cultivate other areas instead, and forcing herders to move
128
their animals elsewhere. These other areas were usually less suitable
and more vulnerable to degradation.
* The general impoverishment caused by war forced larger numbers of

people to engage in marginal economic activities snch as seUing
firewood or the leaves dtdman palms (which are used for making mats),
which are destructive of the natural environment.
Mflitary engagements themselves also caused havoc to the environment.

For example, a fighter's description of the battle of Elabored, south of
Keren, in November 1978, indicates the level of destruction wrought m
the valley where the battle was fought:
[Elabored] was full of tanks, dead bodies and trucks from edge to edge.
All the dry grass was burned totally, and rows of trees were [knocked]
down from the tanks. The planes never stopped coming. They were
dropping different types of bomb including napalm [or phosphorous],
and at times the valley was so filled wi£ smoke you oonldn*t even
Every traveller to Eritrea has seen trees burned by phosphorous bombs

or with limbs blown off by cluster bombs or high ^(plosives.
Humanitarian Assistance
There was relatively humanitarian assistance to Eritrea during this

little
period. International aid to ERA

fell from $2.9 million in 1978 to $1.6
million in 1979, and only $1 million in 1980. Food relief was supplied
at a rate of under 6,000 tons per year. These amounts were pitifully small
compared to the level of need. Assistance on ttie govenmient side was
scarcely more generous, but Eritrea bad the advantages of containing both
of Ethiopia's ports, and being a sensitive political area and thus a priority
formuch government economic aid.
was distributed in such a manner as to ensure greater control
Relief
over the population. From 1982 onwards, it was made increasingly difficult
from government- to EPLF-held areas for relief.
for rural people to cross
In government relief centers, "applicants are accepted only if they come
as complete families, and then they receive rations [sufficient] for only
" Quoted in Horn of Africa, 4.1, 1981, p. 22.
129
such small periods that any hope of letuniiiig to their own village with
some food is illusory."^
Factors /Meliorating the Famine
Rural Eritrea suffered from deepening famine during 1982-4. However,

according to most indicators, Eritrea did not succumb to famine as severely
as Tigray or north Wollo. There was less distress migratioD and, apart
from the refugee camps in Sudan, there were none of the appalling rdief
shelters that were common further south. A
later investiption foond little
evidence for raised death rates durmg the fomine penod, tfaongli other
demographic signs of fanune such as a lower birth rate showed up
clearly.
A number of factors accounted for the lesser severity of the famine.
One factor was that although Eritrea is drier than the neighboring provinces
to the south, the relative shortfall in rain during 1982-4 was less. The
climatic adversity suffered by the rural people was therefore less severe.
A second factor was that the war was fought as a positional war —
with well-defined zones held by the contending armies —
rather than a
classic guerrilla war, so that outside the areas directly affected by fighting,
the population was subjected to relatively less haiaMnent A
thiid Utitait
the war actually brought some ecoaomic bmefils. The reconstruction
is that
element of the Red Star campaign significantly boosted urban employment.
The army in Eritrea was paid relatively weU and regularly, and Eritrean
towns received their quotas of consumer goods on schedule. Shopkeepers
in garrison towns such as Keren and Anseba reported doing a brisk trade
in watches, cassette recorders, coffee and clothes, mainly selling to
soldiers. Army officers benefitted from the flow of contraband goods
from Sudan and Saudi Arabia, and the parallel currency market that grew
up around the flow of expatriate Eritreans' remittances back home. While
growing rich, the officers tolerated this informal nunket. Vm
relative
prosperity for traders tridded down to certain sections of society^ including
the traders' relatives, domestic servants, house bnfldeis and otfieis.
^ Fritz Eisenloeffd, The Eritrea Durrah Odyssey, Utrecht, the Nelhedands,

1983, p. 38.
^ Alex de Waal, Topulation and Health of Eritreans m Wad Sherifei,"

ActionAid, London, 1989.
^*
Alex de Waal, interviews in Wad Sherifei, Sudan, March 1989.
130
deepened in 198S, the amounts of aid provided
Finally, as the fiunlne
to Eritrea became very The aid oo die government side was tied
laige.
to the continuing counter-insmgency strategy, bat it did have the positive
benefit of actually feeding some people.
Despite these factors, the famine of 1983-5 in Eritrea must not be
minimized. caused massive and avoidable suffering and impoverishment
It
among most sections of the rural population, and led to many deaths.
131
S. COUNTER-INSURGENCY AND FAMINE IN TIGRAY
AND ITS BORDERLANDS, 1980-84
In October 1984, the world was shocked by a film made by the BBC
and Visoews in Meqele (Tigray), and Korem (on the Tigray-Wollo border).
The pictures of a mass of destitute people, starving with a quiet dignity,
revealed a "Biblical famine" in the late 20th century. The famine had of
course been developing for several years with little attention from the
outside world — that was part of the horror of the story, which pricked
the conscience of the affluent west. This and the following chapter recount
the central story of how those images of mass starvation, and the wider
famine which they represented, came about.
Grinding poverty and an unpcedictable diinate played tfieir part in
ciealu^ the tragic pictures of October 1984. The social and agricultural
polides of the Ethiopian govemmoit were also important, and will be
analyzed in the next chapter. However, at die center of the famine —
Hgray and north Wollo —
the counter-insuigency strategy of the Ethiopian
army was the single most important reason why the drought of 1983-4
became not a "normal" period of hardship but a famine of a severity and
extent unparalleled for a century.
The counter-insurgency strategy followed in Tigray in the early 1980s
was different from that pursued in the southeast and in Eritrea. Starting
in August 1980, it involved a greater level of indiscriminate violence against
the civilian population, and there was no attempt to provide even the most
minimal level of compensatory assi^ânce to the stricken population. When
the level of brutality and destniction increased in early 1983, famine
developed directly.
The Ethiopian famine of 1983-5 was both "wide" — it affected a very
large area — and "deep" — there were places of exceptional severity.
The famine first southern Tigray and north Wollo, and spread
hit in
outward. At the nadir, up to a third of the country was gripped by famine.
The original center was the most severely hit, where the greatest depths
of famine were plumbed.
Drought and Famine, 1983-85: An Outline
The most remarkable fact about the funine of 1983-5 in Ethiopia was
by the time the drought struck, tfie famine was already well under
that,
way.
Every year, somewhere in northern Ethiopia, there is a harvest failure
due to poor rains and a food shortage, and people go hungry (see chapter
1). If there is a humanitarian agency workislg in the vicinity, calls for help
133
will be sounded, but this does not amount to widespread or unusual famine.
It isimportant, therefore, to distinguish the "normal" alarms for localized
distress from the "abnormal" alarms that accompany the development of
a major famine. We contend here that the droughts of 1980-3 were
unfenuukable; that localized surpluses existed; and ttat if nonnal prooesMS
of redistribution of food had been allowed to occur, there would haive been
no famine. 1984 was a drought year of unusual severity, it is trae but —
had the famme not already been in train, and had the artificial funine-
creating actions not continued, major famine could have been averted.
Starting in 1980, the "norm^" alarms began to sound, varying in severity
over the following three years. The Relief Society of Tigray (REST,
working with the TPLF), the government RRC, and voluntary agencies
sounded these alarms. In late 1982, the alarms became more urgent.
In February 1983, however, there was a change from a severe but
"normal" cry for help to the warning of a major famine. British relief
agencies made a major appeal on February 16. Hiis cry for help arose
because relief agencies woridng in the relief sheUeis ci Korem and Ibnat
(central Gonder) were suddenly receiving a large inflow of dnrtitute and
malnourished migrants. Attributed at tiie time to drought, the flow of
destitute migrants was in fact a direct result of the war (see below).
There is no evidence for harvest failures in northern Ethiopia over die
period from 1980 to early 1983 sufficient to cause severe famine.
No reliable figures are available for rural production in Tigray and the
adjoining areas. National figures for Ethiopia are available, however:
Table 1. Food Production in Ethiopia, 1977-84.^

Production: total
1977 99 95
1978 110 104
1979 122 113
1980 117 106
1981 115 102
1982 127 110
1983 118 99
1984 110 90
Note: 1974--76 = 100
Source: FAO Production Yearbooks. Also see: Bob Baulch, "Entitlements

and the Wollo Famine of 1982-1985," Disasters, 11.3, 1987, p. 196, where a seven
per cent higher production figure is given for 1983.
134
Not only do these figures fail to indicate any crisis until 1984, but they
show that 1982 and 1983 were, nationally, bumper crops —
two of the
best on record, and above or equal to the long-term average. There were
of course regional shortages, but the simplistic explanation of the famine
as a prolonged drought-induced food shortage does not hold up.^
The RRC produces estimates of the size of the crop and the availability
of food after each haivest. These are always pessmrisdc, as it is the job
of the RRC to identify deficits and appeal for aid. (Smveying the same
area, the Agricultuial Madcetnig Gcupoialloii (AM Q, with a brief to identify
snq)luses, usually produces a much more optimistic picture.)
After the 1980 harvest, food availabUity in Tigray, WoUo and Gonder
was "normal."^ 1981 was better. The main 1981 report identified surpluses
in the usual areas: Raya (eastern Tigray), Kobo (north Wollo), Borena (south
Wollo), Simien (north Gonder), central Gonder, and Gojjam.'* No
assessment was made for western Tigray, but surpluses existed there too.
For 1982, the reports for Wollo were encouraging —
food shortages were
localized, and most of the highlands was normal.^ Eastern Gonder was
suffering drought, but the main surplus-producing areas were less affected.*
In eaily 1982, Tigray was described as "encouraging/ and in late 1982
as "po(v,** but no surveys were done —
the mam measure used was ttie
Attempts to argue the reverse are based on assumptions that the data are
faulty. See for example: Gopu Kumar, "Ethiopian Famines 1973-1985: A Case
Study," in J. Dreze and A. Sen, The Political Economy of Hunger, Vol. 11, Oxford,
1990. Kumar is forced to argue back from "what we know about the progress
of the famine" to the ''defensible assumption" of a catastrophic fall in food output,
iwhich is not shown in fbt figures (p. 198).
^ RRC, "Food Supply Status and Forecast by Administiative Region," Addis
Ababa, November 1980.
^ RRC, "Food Supply System: Meher Synoptic Report, 1973/74 EC [Ethiopian
Calendar] (1981) Crop Season," Addis Ababa, March 1982.
' RRC, "Report on a Reconnaissance Trip in Wollo Admhiisliative Region,
(August 11 -September 4, 1982)," November 1982; RRQ

"Meher Synoptic RqxKt
1974^5 EC (1982) Oop Season," Marab 1983.
^ RRC, "A Report on a Reconnaissance Trip to Gonder

Administrative Rôn
(October 6 - November 6. 1982)" Addis Ababa, November 1982.
135
price of grain in Meqele town. Later in this chapter it will be shown how
government policy helped to cause the price rise over this period. A survey
done among Tigrayan refugees in Sudan in 1985 found that "their highest
yields in the last ten years occuned in IQSZ-SS."*
In Aprfl 1983, the RRC issued aievind fqKHt for Ihe 1982
Hiis was mudi move ahumiiig in tone —
despite fbt fact that ao oew
ii]foniiatk>n had been ooUectod abcmt 19 Giain
prices had shot up, and the RRC infened that major shortages existed.'
The real reasons for the change in tone were probably that a famine had
started, and the RRC needed to identify a drought to blame it upon.
The main 1983 season provides the first significant evidence for
widespread crop failure. But even at this stage, the failure was confined
to most of Tigray and some parts of north Wollo. In Gonder, the food
supply situation improved in 1983, with substantial surpluses in Gonder
Zuria, Chilga, Debre Tabor, and Simien. In most of Wollo the food supply
remained normal.^® The RRC had no data for Tigray, bot a relief agency
team visitu^ IPLF areas rq;x>rted suiplines
"
m
Sh&e and Raya, though
all other areas had nffered harvest fiuliires
Almost universal drought fiist occurred m the spring of 1984, affecting
the belg harvest. Belg crops produce only a small praportian of the food
produced in the north —about one quarter in the areas where the belg
rains fall, and none at all in 90% of Tigray. The belg failure was serious,
but should not have caused undue problems in the light of the average crop
performance over the previous few years, and the bumper national harvest.
^ RRQ "Pood Supply Statia and Foieciit by Admiaistntive Region'*, Mardi

1982, and Tood Supply Statoa and Fnecaat No. 1/ Deoeniber 1982.
* Cultural Survival, Politics and Oie EOaopian Famiae 1984-1985^ Cambridge,

Mass., 19S5, p. 192.
' RRC, "Food Supply Status and Forecast No. 2 (Oct-Dec [1982])," April
1983.
^RRQ Tood Supply Status and Forecast," March 1984 (baaed on daU from
late 1963);RRC, "»£her Synoptic RqNMt 1976/76 EC (1983) Crap Seaaon,"
January 1984. The snplus in Simien waa identified by REST.
John English, Jon Bennett, Bruce Dick and Caroline Fallon, "Tigray 1984:
An Investigation," Oxfam, January 1984, p. 62. The team reported that Raya had
reaped a surptna of 8-lQyOQO tona daapile a onc â
aon draosjht m
the area,
implying that the aurpluaea of pievioaa yaaia woe modi hnger.
136
But the RRCplayed up the belg failure, telling blatant untruths into the
bargain: "The highlands of Wollo, Bale and Shewa are the major belg
producing areas. Belg accounts for at least half of the annual production
in most parts of these areas. There are also areas in most of the remaining
regions wfaidi heavily depend on belg, partkularly in Tigray ..."^
RaiafBll data are veiy scarce —
for yean tli^ were ooooealed by the
gpveniment. No data are available for places in Tigray, but some are
available for Kobo, in north Wollo, which is close to the heart of the famine
zone. These data confirm that the drought only began in late 1983. The
following table shows the rainfall for ih&belg (February-May) andmeher
(July-October) seasons in Kobo.
Table 2. Rainfall in Kobo, North WoUo.^'
1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987
I
Belg 268 117 233 118 NA 168 270 19 214 174 331
Mdber460 352 497 499 646 362 150 51 243 450 233
All amounts in mOUmeten.
Further south, a similar picture can be seen in the fragmentary climatic

I
data that are available.Some rainfall stations in Wollo recorded a drought

in 1983, others recorded a good year. An example of the latter is Bati,
where the famine of 1984/5 was very severe —
but where rainfall in 1983
was the highest for more than 15 years.^
Satellitehnagery of vegetation also mdicates that spring (belg)
harvest of 1983 was likely to have been normal. Thus we see that the
drought started only after die famine was set in train. Ihe reason why
rainfoU data have been kept as state secrets now becomes clearer.
,
" RRC, The Belg Raul Failme and Its Effect on Food nroductioii, Speciai
Report," May 1984.
Derived from: Alemneh Dejene, Environment, Famine, and Politics in

I
Ethiopia: A View from the Village, Boulder, Co., 1990, p. 84.

^ Alemneh Dejene, 1990, p. 57.
I
'
137
I
L
Both rainfall data and satellite imagery confirm that the main 1983
season was satisfactory in Gondcr, implying the existence of the substantial
surpluses normally produced in that province.
1984 was a year of almost complete drought, lasting the whole year
and affecting a wide area. Production in Wollo was only 28% of 1983;
in Gender it was 86% (no figures are available for Tigray). It was the
results of this drought that observers saw when they visited the region in
late 1984 —
dry fields, withered crops, waterless wells. Hie fact is,
however, that a visitor can only see a siqgle year of drought, and that is
not enough to cause famine. The drought of 1984 was used as a scapegoat
for a famine that had begun much earlier.
In 1982 and 1983, the localized drought in Tigray was most severe on
the eastern escarpment. But the population hit hardest by the famine
originated in southern Tigray and northern Wollo, and was to be found
scattered in places such as Shire, Ibnat and Korem —
another indicator
that climate was not the fundamental cause of the disaster.
Another indicator of famine is grain prices. High grain prices indicate
a scarcity in the market. The following table shows the approximate
average grain prices during the famine period.
Table 3. Grain Prices in Noithern Ethiopia"
East Tigray NoiUi Wollo Noidi Gonder
Nov/Dcc 1981 100 50 40

Nov/Dec 1982 165 65 55
Nov/Dec 1983 225 90 45
Nov/Dcc 1984 300 160 70
Jun/M 1985 380 235 165
All prices in Bin per quintal (100 kg)
Prices rose to reach famine levels in Tigray in late 1982, but only rose
to comparable levels in Wollo between February and August 1984, and
in Gonder in mid-1985. Throughout 1983, prices in Wollo were stable -
- in some markets they actually fell. These prices are not consistent with
Derived from data in: Alex de Waal, Tigray: Grain Markets and Internal
Purchase," Oxfam, 1990, p. 44; Peter Cutler, The Development of the 1983-85
Famine in Northem Ethiopia," PhD thesis, London, 198; Baulch, 1987, p. 199.
138
the account that stresses repeated and widespread drought. They are
consistent with highly localized famine in Tigray, spreadii^ out into
neighboring areas in 1984 and 1985.
By 1984 the famine had deepened to such an extent in Tigray and north
WoUo that its effects were being felt far away. The high grain prices at
the epicenter of the famine were forcing up prices elsewhere, putting food
out of the leadi of the poor. Increasing animal sales were pushing livestock
prices down, and migrants were flooding the labor markets, llie famme
had acquired a momentum of its own and began to spread, helped by the
coercive and restrictive social and sgricultural policies of the government.
The drought and accompanying crop failures cannot explain the famine.
To understand why it occurred it is necessary to turn to an account of the
conduct of the war in Tigray.
The TPLF
The TPLF, from its inception, fought a classic guerrilla insurgency.

Itrefused to defend any territory, and instead relied on being able to move
among the population "like a fish through water." Espousing a mix of
Hgrayan nationalism —
for instance calUng itself the "second Weyant^
refuiing to the rebellion of 1943 —
and radical politics, the TPLF soon
came to have the tacit support of most of the rural population. The Red
Terror was waged with particular savagery in Tigray and drove much of
the urban population to support the rebels. The TPLF took advantage of
the government's military preoccupation with the Ogaden and Eritrea to
operate throughout rural Tigray during 1978/9. It set up a wide network
of councils, instigated land reform, and began some health care and
development projects.
In 1976, theTPLF took several European hostages, including a British
family of four and a journalist. It demanded a ransom of $1 million for
the British family, buteventnaUy released them after e^g^ months captivity
after pressure was broiigfat to bear by the Sudan government.
From 1980 onwards, the TPLF claimed to control 85 per cent of Tigray.
Because of the TPLFs guerrilla strategy, the government army was able
to launch offensives into almost every part of the province, and thus make
a counter-claim of having access everywhere. The two claims are both
true: TPLF fighters could move almost anywhere at will, with the support
of most of the people, and the army could, by force, reach most places,
but not hold them.
In 1980, the TPLF began to form militias throughout the countryside.
Hiis was partly in response demands from the villagers to have arms
to
to protect theuBdves from the casual and repealed violence of army patrols,
I
and partly in order to take the war to the government throughout the
province. In 1982, theTPLF launched its "Southern Operation" and opened
a new front in southern Tigray and northern Wollo. It followed the
formation of the Ethiopian People's Democratic Movement (EPDM) ayear

previously. Tamrat Layne was a prominent leader of the EPDM, which
started as a breakaway group of the EPRP. Always a close ally of the
TPLF, the EPDMwas militarUy active in WoUo and Oonder.
In 1982, the TPLF and EPDM started mountiiis major operations in
Wolio. Ayear]at€t,JlMca Confidential w^lhtA
are able to hit the whole length of the Dessie-Meqele road."" In March
1983, a joint TPLF-Afar Liberation Front (ALF) attack overran Bati, on
the eastern escarpment of southern Wollo. In April, TPLF-EPDM forces
occupied Lalibclla, the historic site of rock-hewn churches in central Wollo,
and kidnapped relief workers in Korem. Seqota, district headquarters of
Wag, was occupied. In September, joint TPLF-EPDM operations extended
as far south as Wichale, Jarre and Haik, close to Dessie. Eleven Swiss
relief workers were kidnapped in Jarre and later released. In August 1984,
TPLF-ALF attacks weie mounted near Milk, in eastern Wollo, and in
October Lalibella was captured again. On each occasion the army appeared
to be taken unawares by the attack, indicating that the rebels were able
to move with ease among the local population. In response, many elements
of the counter-insurgency strategy were applied thnmghoat northern and
central Wollo and Gonder, not merely in Tigray.
Until about 1986, the TPLF had virtually no heavy weapons and no
"base area." It had a core army of an estimated 7,000-15,000 men, and
a much larger number of militiamen. The EPDM
was a smaller force.
Counter-Population Warfare by the Government
The nature of the rebellion in Tigray led to a new variation on the aimys
counter-insurgency strategies. These strategies were instrnmental setting m
the famine in train. There were three mam aspects:
* Large scale military offensives, aimed at the sniplus-pioduciqg Shire

district.The Sixth Offensive of the Ethiopian army in Tigray was
launched in August 1980 and continued until March 1981. The Seventh
Offensive was fought from February to April 1983. (An eighth was
fought from February to May 1985.)
Africa Confidential, 24.19, September 21, 1983, p. 4.
140
* Aerial bombardment of markets. This started in 1980 and was intensified
in early 1982 to coincide with the Red Star campaign in Eritrea, and
remained at a high level thereafter.
* Tight controls on movement of migrants and traders* enforced in all

garrison towns, in eastern Tigray and northern Wollo. These controls
were introduced in late 1980 and widened over the following two years.
In 1982 restrictions were tightened in Wollo. In 1984 they were
enforced particularly strictly, with widespread detentions of suspected
TPLF sympathizers, and were extended to many parts of Gonder.
The logic behind the government's strategy was "draining the sea to
catch the fish." This amounted to counter-population warfare. Because
of its actions during 1976-9, the government had alienated all significant
sections of the populace, and could locate no secure base from which to
start a pacification strategy. As a result, thearmy engaged in counter-
population warfare. The increasiqg hostility of the population towards the
army combined with increasing TPLF militery successes, causing the army
to become more demoralized and more brutal.
The three elements of the military strategy combined to prevent the
normal redistribution of surpluses within northern Ethiopia. The offensives
effectively destroyed or made unavailable most of the surpluses in Shire
in 1980 and 1983. The bombing and the restrictions on movement
prevented the mobilization of the Raya surpluses from 1980 onwards,
becoming 1983 the restrictions prevented
particularly severe in 1984; in
much trade in Wollo and Gonder. The result was that peasants
in 1984 in
in the deficit areas of eastern Tigray and Tembien-Wag were unable to
provision themselves from the adjoiniqg areas, and bpgan to suffer famine.
The following sections will look in detafl at eadi of the three aq)ects.
CoiinterÎnsiirgency 1980-84: L Military Campaigns
In January and February 1980, a punitive expedition by the army burned

several villages in north-central Tigray. Credible reports indicate that 130
civilians were killed in Mai Kenetal. The village of Aweger was destroyed,
and when villagers from nearby Haile tried to help the inhabitants rebuild,
the army killed 51 of them and briefly imprisoned over 800. More than
300 civflians were killed in the entire operation. However, major military
action only occurred later in the year.
On August 22, 1980, the army launched its Sixth Offensive m central
Hgray. This madced a change m niilitary strategy, and flie beginnings of
wide^xead counter-population warfare. In the midst of the offensive.
141
Africa Conpdaitial oommented with remnkaUe presdeooe: "A ooaseqaeooe
of the fighting is likely to be widesnead fiumne.*^^ 4/Hiea Confidmdal
only got the time finme wrong, and underestimated the lestlience of the
Hgcayan peasants the famine took another two years to develop.
The offensive involved widespread military action over the following
seven months. 40,000 troops were involved, together with aircraft and
helicopter gunships^* — a small force by later standards, but enough to
economy. Tembien, the center of the
create major disruptions in the rural
famine area in the following years, was worst hit.
During the campaign, the army engaged in a number of activities that
directly affected ability of populatioa to food iladf. These
included:^'
* the destruction of grain stores: an estimated 6,000 tons of grain was

burned by incendiary bombs or destroyed by soldieis;
* the killing of cattle: REST estimated that 950 cattle were killed;
* the burning of crops and pastures: 142,000 hectares of farmland was

rendered useless in fighting in February-March 1981 alone by burning
or trampling;
* the enforced collection of taxes and contributioiis» often at a punitive

level, ostensibly indudiog "aneais;"
* the forcible displacement of farmers: about 80,000 fanners in central

Tigray were forced to leave their homes. 20,000 resettled themselves
in western Tigray and 5,000 became refugees in Sudan. Many could
not harvest in 1980, or had inadequate time to plough their fields for
the 1981 rainy season.
* destruction of villages: over 2,000 houses and five grinding mills were
destroyed.
^''Africa Confidential, 21.24, November 26, 1980, p. 7.
Mi 24 helicopter gunships were deployed for the first time in Africa in this
campaign.
Kirsty Wright, "Famine in Tigiay: AnEyewitness Acooont," London, REST

Support Committee, 1983, p. 7.
142
I
Abi Adi had been occupied by the TPLF since August 1978; on
September 11, 19803 was attadced by tbeaimy using hdto Many
houses were burned and much of the population was fiotoed to flee; four
people were killed. On September 29, a number of towns and villages
in central Tigray were attacked by airplanes and lielicoptei8» and, according
to one report, 27 people were killed.^®
After the campaign finished, the government had established a network
of garrisons throughout the province, many of them in important towns
for local trade. Abi Adi is one such town, critical for the trade between
the surplus-producing Simien district of Gonder and the deficit areas of
Meqele and Agame. For twelve months after the offensive there was a
garrison at Abi Adi, wiiich was aUe to enfoioe the lestrfctions on migration
and trade discussed m the next section. Some of these gsirisons were
withdrawn in 1981 and 1982, under pressure finm the TPLF. Regular
patrols contunied, and dvflians were subjected to harassment, robbery and
execution.
In a series of attacks in August and September 1981, over 400 people
were killed by soldiers and airplanes. For example, on August 26, the
village of Mezega was burned. 14 villagers were killed and 400 cattle
slaughtered.^^
The opening of the Red Star offensive in Eritrea in February 1982 saw
widespread bombing and an increase in army attacks in Tigray. Patrols
in southern Tigray became more frequent in response to the TPLF's
"Southern Operattoo.**
In one of many retaliatory attacks, on this occasion following a TPLF
ambush, soldiers burned a village near Adi Oudud, killing two women and
seven children who remained behind in their huts.^
The end of 1982 saw preparations for the Seventh Offensive, which
was to be the most brutal to date. A
month before the ground attack started,
helicopters and MiG fighter bombers started "softening up" the towns and
villages to be attacked. In December 1982, belicopter-gunships bombed
^Somali, Tigray and Oromo Resistance Monitor, (STORM), i.2, March 1981,
p. 6.
^ Information compiled by Baibaxa Hendrie, an IndqieDdeot consultant.
^ Afiica Coiuemporury Record, 1982-3^ p. B149.
143
Abi Adi town for half an hour. According to a member of the baito, ei;
houses were burned, "but only one person killed and a couple wounded."
The army offensive began in Janiiaiy> wifli iltadB in ocotnl and WMtem
Tigray Woik on llie huvest stopped at once, as people made jareparations
.
for concealing fc)od and other items from flic goMieiB, and
villages if need be. Migrant workers moved to other areas. Civilians were
killed by army raids near Axum, Enticho and Enda Selassie. On February
16, this escalated to a full-scale assault, involving 70,000 troops and all
the familiar abuses against civilians —
summary executions, burning of
villages, destruction of grain stores and the killing of cattle. The TPLF
attempted to defend its base in Shire using trench warfare, but was forced
to abandon the area. More than 100,000 farmers were forced to evacuate
their homes, and many of the estimated 375,000 migrants in western Tigray
were obliged to move to other areas to seek worlL
The dâstation was particnlarly severe in Shire —
a vital ampins
producing area —
wliere on Febmaiy 21 tiie army sncceeded in captunqg
the town of Sheraro for the first time since the late 197QB. Before attaddqg
the town, the army indiscriminately shelled it. Having occupied the town,
the army systematically looted it, taking over 150 tons of grain, taking oil
presses and grinding mills, together with burning other grain stores and
fodder grass collected to feed to animals.^ The health clinic was
ransacked. A medical team from Medecins Sans Frontieres saw "thousands
of hectares of land systematically burned, for example on the way from
Sheraro to Kafta." The team also saw the Tekezze Agricultural Center,
where soldiers had cut down the fruit trees and destroyed the irrigation
system.
REST had established several centers to ^ve assistance to funine
migrants conung from oentnd Tigray. Three of tliese —
Rndahngna, Az
Daro and Adi Nebreit —
were attacked.^
A
woman described the army's arrival in the village of Edaga Habret
on March 9 and the preparations the people aaade for survival:
We heard that the army was coming at 2 a.m. in the morning from
people who had run to our village from neighboring settlements where
^ Woreda Teka, interviewed by Geny MoCana and Sacab Vaygban, two viiiliia
to Tigray in 1988.
Wright. 1983, p. 8.
^ Gayle Smidi, Xountmg QnmUds: A Mbnitoring Report on Fnnine Relief

in TIgny." The Hague. 1983^ pi 19.
144
theanny had already reached. Theo around 6 ajn. we heard firing
between the kabrits [TPLF scouts]. As soon as we heard the news, we
tried to pxipm foodstuffs fat the future. You can't take injera
[Ethiopian pancakes] for a long period of time as it breaks and dries.
We roasted chick peas and cereals to make qolOy took the food and ran....
Two of us went up the mountain two hours away and we could look
down and see the Dergue [soldiers]. First the troops
lit one house using
a match, then they took burning grass from house to house. All the
houses were burned, houses belonging to 135 heads of families. Nobody
stayed behind. We
had tried to take important materials to the bushes
sononnding the village, bat these were discovered by the Deigue —
house matoiials, phrtes, jeny cans, soap^ salt, sugar, pepper, doth, sewing
machines. Three sewing machines were d^boyed and dl seven oU
presses were burned. My two beds were burned and the small garden
destroyed. A
lot of grass had been collected for feeding the animals
for the summer ... it was all burned. Seven people were killed and 20
wounded in the area. We stayed for three days in the bushes and on
the third day the troops left. There were so many of them that there
was a two-and-a-half hour line of them marching out of the village.
When we went back into our homes we found that all the grain in the
villagehad been burned. 1 lost three sacks of sorghum and 12 sacks
of sesame.^
This was the second occasion the village had beoi destroyed by the anny.
On July 2 and 3, 1983, the village was bombed again, wounduig two people
and killing some sheep and camels.
The major alarms for the famine were raised by relief agencies in
first
February 1983, when large numbers of mignmts started turning up in Korem
and Ibnat looking for food.
The farmers I got a chance to talk to in Korem had come from around
Seqota ... this stream of people looking for food had to go further afield
than usualwhen the rains failed there and the autumn harvest was small.
Some way over to Gonder, where there was still
traipsed west all the
agricultural work to be found in some places as late as November, before
hurrying back to WoUo when distribution started there.^
" Wri^t, 1983, p. 9.
" Judith AppLetoD, Save the Children Fund (UIQ field report, April 1983.
145
Others had travelled to western Tigray befoie feUuniqg to Koiem and
Ibnat in February. What is hnportant to note is tiiat the migration to tiieBe
areas was normal. It was not a preferred strategy —
the migrants would
rather have stayed at home, had the food bcca available but the —
movement west did not itself signal anything unusual. The famine was
because the migration failed. It failed because of the havoc
set in train
caused and the adjoining areas by the Seventh Offensive.
in Shire
As before, the offensive resulted in a wider spread of garrisons
throughout the province, and a large number of small atrocities followed.
For example, there are credible reports of the army killing 20 civilians while
burmi^ villages near Hagere Selam, Shiie, between June 17-26, 1983,
and IdUing two more at Alage.^
During and after 1983, lural people lemember the behavior of aoldiers
as being more brutal than before, and this is ooofirmed by members of
the armed forces. Lieutenant Yamani Hassan, a prisoner A
war held by
the TPLF, reported:^'
Civilians in the war zones have always been badly treated, but the
Red Star Campaign, in 1983.
brutality increased after the failure of the
I can think of four incidents in Tigray I have witnessed. One was in
Sinkatta where four men were questioned about the TPLF. They said
they did not know anything, and fliey were then shot. Another time
a 13 or 14 year old girl was raped. A
thud occasion was when soldiers
went to a group of houses near the church in Hansien. Three old people
came out, and the soldiers chose one and shot Um. Thcxe was also
a time when we were stationed at a village near Sanue, and the viUagers
came and brought us roasted maize and beer. They treated us very well,
probably hoping we would do the same to them. The order to leave
came in the middle of the night, and the soldiers burned the whole
village asleep in their beds as they left.
It is taken as read that these sorts of atrocities are all "part of the job."
Anyone who questions them, or talks about what is done, is picked up
by the Velfare" people." Hie soldieis are tnmed to act like machines
or anunals and not have any thoughts of their oiwn. These is no training
Âfrica Qmtenqforary Record, 1983-4, p. B136.
^ Interviewed at Tade Aaegar, Tigray, on December 1, 1988, by Saiah

Vaugban and Gerry MoCann.
^ The security aeivioe within the amiy.
146
in torture techniques or anything like that: soldieis are just given boxes
of matches and told to get on with it.
In July andAugust 1983, the army mounted operations in southern

TPLF offensives. In November, the TPLF was active
Tigray, in response to
in nortfiem Oond^, and its central command set up a tenîorary base in
the Simien mountains, wfaeie food was more leadfly avafl
Tigray. The aimy responded with a series of svmps Ifaroagjh the area.
These mvohred the bnmiQg of villages and monnds of haivesM
for threshing.^^ There was a larger army assault on Simien in January
IS^, followed by four other major attadss over the following eighteen
months.
Starting in 1980, the government utilized paid bandits, locally known
as banda or shimeq^ to engage in sabotage and terrorism. In November
1981, there was a number of sabotage attacks in Sudanese refugee camps.
Refugees who had returned briefly to Ethiopia reported being told by
soldiers that thev would be well-treated if they returned to Sudan to attack
refugee camps.^
Govntfr-rliMirg^ncy 1980-84: IL Bombfaig
From mid-1980 onwards, chronic day-in-day-out bombing began to

occur in Tigray. Most of by the air force were
the large-scale atrocities
perpetrated by attacks made in conjunction with ground offensives.
However, innumerable small attacks were made at all times. Airplanes
just needed to "buzz" a village in order to send the people scurrying for
shelter, and cease all activities for many hours.
The main target of the bombing in Tigray was the network of rural
markets in TPLF-controlled areas. Medebai, a market near Axum which
lies on the important Shire-Eritrea trade route, was bombed more than 100
times daring the 19808. Hausien, the most important maricet in northeast
Tigray, was booobed equally oAsb until it was completdy destapyed ui 1968.
Wdd, an unportant market which ludn Tembien, Raya and Wag, was also
frequently bombed.
The market at Chilla, near Axum, was frequently bombed. The worst
j
attacks occurred on March 3 and 5, 19S3: A TPLF fighter described the
scene:
Afrka Omteuôtary Record^ 1983^ p. B137.
^ Ahmad Karadawi, "RefMgee Policy in the Sudan, 1967-1984," DPhil thesis,

Oxford, 1988, p. 193.
147
L Copyrighted material
Yoli cannot believe what you saw —
it was not something for anyone
to see. The blood was flowing like rivers and sitting in poal% and tlMfe
were crushed bodies thrown everywhere, the blood of the people was
mixed with the blood of the animals that had been hit. You could see
a head there but you couldn't find the body, it was thrown some meters
away. The children were hysterical and screaming even after some hours
— the helicopters chased them and they couldn't get away. They cry
now even if they hear a plane. If they have seen a massacre when they
are only four or five years old, they will remember forever when they
shut their eyes that they saiw their mothen being kffled.^
A
medical team from M
edecins Sans FIroDtiflfes visited the town shortly
afterwards and interviewed survivors:
Four helicopters blocked the exits from the market and machine-gunned
the market place. MiGs then finished this "work." Even two weeks
later we
could still observe bomb splinters on the rocky ground and the
smell was unbearable. The ground was strewn with various broken
fragments, spilt cereals and corpses of donkeys. Everywhere there were
traces of blood —
on the ground and on rocks where people had tried
to escape. Here and there were the anlmown graves of more than a
hundred local people who had been massacred.
According to REST, 315 people were killed or wounded in this attack.

At least nine other markets were bombed during the Seventh Offensive
(between February and April 1983), causing at least 179 casualties.^
Phosphorous bombs were frequently used in attacks on markets and
villages, leaving horriblebums. Incendiary bombs were used to set fire
to fields and stores.
Asecondary target of the bombing was means of transport. Wheeled
transport became too dangerous to use. Caravans of donkeys, mules or
camels were fireqnently attacked. This forced traden to move at night,
and to move in small and less visible groups (while the thieat of bandits
and saboteurs m
some areas compelled them to do the oppoaitB, and move
in larger groups for self-pnitectian). A
peasant m
central Tfgny
commented:
^ Quoted in: Smith, 1983, pp. 100-1.
**
Smith, 1983, p. 100.
148
Of course the government knows that we can't survive from month to
month without a market. This is why the planes so often come and
circle on market day. We all scatter as soon as they come, but even
that causes problems: there are often thieves in the market place who
steal the goods that people drop when they run away.^
Hie inevitable lesolt of 4ie campaign against the madGets and tiadeis
was that maikets weie forced to be held at niglit —
thou^ the danger of
eariy-mcMiiing attacks on madnt towns where peo^
real. With no light other than candles and small gas lamps, the markets
could become chaotic —
people could not see properly what they were
buying, and vendors laid their goods out in the danger of them being
trampled upon. Social gatherings, an important reason for many people
attending market, were held less often. The TPLF regulated that all markets
should be held on Saturday nights, so that it was not possible for the air
force to bomb the market towns "in rotation," following the different market
days. A
negative result of this was that traders could not rotate between
the markets. Unable to travel by day, larger traders were forced into long
periods of klleness doring dayli^t, maldi^ slow progress to attend perhaps
just one small market per wieek. Hiey would not store laige amomits of
goods in any one jdaoe, bnt scatty them in different stores, so as to
mmimi?» the danger of losing everything to bombing or a ground attack.
Many traders were forced out of business. Markets thus contracted or were
closed down altogether. Combined with the restrictions on trade and
migration in government-controlled areas» the results were disastrous for
trade and exchange.
Bombing was also used against villages, churches, schools, and farmers
ploughing their land. Attacks appeared to be virtually random. In areas
of greatest TPLF control, such as Shire, the bombing even forced people
to cultivate at night.
Coiinter-Insurgraqr 1980-84: IIL Rcstrfetloiis on Movement
Whfle the military canîgns of 1980-4 were restricted to Hgray and

one small adjoining area H
Gonder, the restrictions on movement
encompassed a much larger area —
even further afield than the TPLF-
EPDM's most southerly military actions. In terms of limiting movement
and trade, the restrictions had a similar effect to the bombing campaign
in TPLF areas.
Interviewed by Sarah Vaughan
149
Restrictions on the Grain Trade
A particularly important counter-insurgency tactic was restriction on

the grain trade. This was implemented from a variety of motivations,
including the desire to restrict the movement of potential rebel sympathizers,
desire to stem the flow of food to the TPLF, suspicion of traders in general
(especially a fear that they might be smuggling arms) and a feeling that
private grain trade was incompatible with socialism.
Since 1974 there had been legal prohibitioiii oo certain co mm
eidal
activities (see chapter and local and andllaiy Ipgidation meant that local
administrators were able to harass petty traders at will and confiscate their
goods. From 1980 these restrictions were intensified in central Tigray,
and in 1982, they were extended to southern Tigray and northern Wollo.
Until then, Meqele obtained most its food from Raya and north Wollo;
this helps to explain the price rise in Meqele during that year. It became
more difficult for traders to obtain licenses, and they were subjected to
an increasing range of taxes. The use of wheeled transport for trade into
contested or TPLF-held areas was impossible —
everything had to be
"smuggled" by pack animals, at a cost of 3-10 times as much per item
carried.
Road blocks were set up at the entrances to all towns, to prevent
unauthorized trade. In addition, one of the functions of army checkpoints
on roads was to control the trade in grain. The soldiers also took advantage
of their position to extract bribes from traders, in effect taxiqg all movement
of grain.
Small traders entering towns were sometimes forced to sell their grain
at the government-controlled prices, which were very low, or were required
to pay tax arrears before being allowed to enter the market.
When possible, rural people prefer to take grain into towns in order to
have it ground into flour by mechanical mills, rather than having to do
it by hand. Regulating access to flour mflls becasM one way the
government restricted the movement of food. When occupying towns and
villages, the army also regularly destroyed flour mills and essential trading
equipment such as weighing scales.
The TPLF also tried to restrict the flow of grain to the towns, so as not
to deprive the countryside, but it attempted to facilitate the rural-niral trade
in grain.
Impact of the Restrictions: The Example of Tembien
These restrictions combined to prevent the movement of grain and were

therefore instrumental in creating famine.
150
In Tigray, these restrictions were invoked most severely in eastern
Tigray, especially the surplus-producing area of Raya, for the reason that
these were the areas where the government had most control. Proposals
by REST to send small traders to the area to buy the available surplus for
reliefpurposes came to nothing. In the adjacent areas of Kobo, Wag,
Tembien and Enderta, prices rose to previously-unrecorded highs.
In 1984, there were very tight lestnctions on trade in nordiem Gooder.
These were related to an upsurge in TPLF-EPDM activity norA of Oonder
and repeated army offensives aimed at dislodging flie TPLF positions in
the Simien mountains, as well as a number of ftiStm not associated with
the insurgency (see the follDwing chapter).
In late 1983, the farmers of Simien had been selling between 300 and
400 tons of grain in Tembien every week.^** Had this trade continued
throughout the dry season, at least 10,000 tons would have moved into
the drought zone by this one route alone. Instead, the trade was completely
blocked by the army.
What happened to the 8,000-10,000 tons surplus in Raya, or the 10,000-
15,000 surplus in Simien, is not known —
much of it probably rotted or
was fed to animals. known that in 1984 Tembfen, which lay right
It is
between these two surplus areas» wis probably the wont-hit funine nam
in the whole country.
Tbehu^gestsuiphisesmTigEay were found in Shire. Bombing of madcets
and army control of the trade routes into central Tigray efifectivdy prevented
any of this reaching the drought zone —
which was m torn one reason
why people from central Tigray migrated westward.
General Impact of Restrictions and Bombing of Markets
More generally, the restrictions increased vulnerability to famine

throughout northern Ethiopia.
In normal circumstances, the grain market consisted of an integrated
network of local markets. When the price rose in one area, traders would
m
boy food markets where it could be had more cheaply, and move it there
— both supplying food to the area and so briqgmg the price down, and
providing a market for farmers in surplus areas. This was no longer
occurring in northern Ethiopia in the 1980s. When the price rose in one
kx:ality, there was almost nothing to restrain it. Analyses of the level of
English et al., 1984, p. 72.
151
integration of the noarket reveal that there was almost no local trade io north
Wollo and Tigray.^
The most immediate impact of the government policy of restricting rural
access to the towns was to increase grain prices in the towns. In 1982,
the price of sorghum in TPLF-held Sheraro was 45 Birr per quintal (100kg);
in nearby govenunent-liekl Enda Sekosie tlie price was 130-160 Birr per
quintal. As the policy became entieiicfaed and the bombing campaign
became established, mral-ntral trade was disrupted too. In Meqele, the
price of grain in December 1982 was 181 Biir per quintal; in nearby areas
controlled by the TPLF there was either none available or it ranged from
140-200 Birr. In Shire it was 60 Birr and in north Wollo 40-90 Birr.
Tf unrestricted trade had been possible, the surpluses in the latter two areas
would have been taken to Meqele, at a transport cost of about 47 Birr and
23 Birr per quintal respectively.^^ The price in Meqele would have fallen
to 120 Birr and probably less. (If access from Sudan or Gonder had been
possible it would have fallen still further.) Thus the people of Meqele were
forced to pay 60 Birr per quintal, or an additional 50 per cent, as a premium
on account of this disruption. During 1982, ^riiea restrictkms were
introduced on the trade between WoUo, Raya and Meqde, the price in
Meqele rose by 67 per cent/^
As a substantial proportion of the adult male popolatkin of Tigray was
formerly involved in petty trade, indudiqg grain, ttie near-destruction of
the grain trade caused much rural unemployment.
Restrictions on Migration
Increasingly tight restrictions on movement were imposed from 1980

onwards. Restrictions started with controlling access to towns. They were
de Waal, 1990, pp. 56-7.
* James Firebrace and Gayle Smith, "The Hidden Revolution: An Analysis

of Social Change in Tigray ôrthem Ethiopia) Based on Eyewitness Accounts,"
London, War on Want, 1982, p. 37, fiKrtnote.
^ In the 1970s, 70% of the grain supply to Meqele origiBated from Raya and
Kobe, and 10% from Gonder. The Shire grain was eaqioded to Eritiea.
^ Calculated from: de Waal, 1990, pp. 44-9.
^ de Waal, 1990, p. 44; RRC, "Food Supply Status and Forecast No. 1,"
December 1982.
152
tightened in southem Hgny m 1982 and WoUo and Gonder during 1983-4.
While migration was never expressly forbidden, the welter of petty
legislation that existed acted as a license to pillage and haiass by any local
official or soldier.
The most direct manner in which these were implemented was through
a pass system — any individual needed a pass from the chairman of his
or her Peasant Association (PA) chairman in order to leave the vicinity
of the village. An individual caught without a pass was liable to arbitrary
detention or worse. People from TPLF-held areas could not obtain passes.
Where they existed, ?As were relucluit to issae passes. If someone was
caught in suspicious circumstances with a pass, the PA officials who
anihorisged tiie pass would be liable to punishment. Given the
vnpiedietability of local officialdom, especially propensity to refuse to
its
reoc^gnize the legitimacy of a document issued elsewfaete, this acted as

powerful disincentive to issuing passes.
In addition, PA officials would generally refuse to issue a travel pass
until the individual had paid past taxes and "voluntary contributions."
Similarly, a jxrson entering a town without evidence of being a dutiful
citizen in respect of tax payments was liable to be detained or harassed.
A number of government policies acted as powerful deterrents to rural
people visiting towns in Tigray. During 1980-3, rural visitors to towns
were subjected to routine hara^ment and robbery, on minor pretexts such
as coming from an area near where the TTLF was reported to be active,
failwe to have a PA membeiship card, or lade of proof of tax payments.
In January 1984, many routine checks were relaxed —
just as the
conscription canâign for the first round of national military service was
getting under way. Some visitors to towns were conscripted against their
will.
Also was a widespread campaign against
starting in January 1984, there
suspected TPLF supporters. Army
checkpoints were mandated to detain
any would-be migrants who were suspected of TPLF sympathies. Several
thousand people were detained, and more than 600 were kept in prison
or killed. Among those killed were 20 senior administrators in Tigray.''^
Waves of arrests later in the year swelled the number of detainees.
While most of the peasants detained hi 1984 were later either forcibly
reaettled, oonscripted to tlie army, or releaaed, some remainftd in prison
for more than 18 months. Together witfi many political detainees from
urban backgrounds, they were released on February 8, 1986, when the TPLF
stormed the central prison in Meqele and freed no fewer than 1,800 political
Africa Contemporary Record^ 1983-4, pp. B125-6.
153
detainees. An team of human rights workers later interviewed
international
some of thcm."*^ One 17
year old farmer was detained while visiting the
market in Axum; the security officials who were carrying out a check on
all marketgocrs' identity cards said he "looked the physical type to be a
spy." Another case was a child farmer aged 14, who was anested in August
1984 on the accusation of carrying paraffin to the TPLF —
which the boy
denied even having had in his possession.
In mid-1984 there was a sudlar though sauUer cnckdcywn in Gooder.
In September 1984, some of the Hgnyan dekgaiet to the Teadi
Anniveisary cdebsaticnis in Addis Ababa weie anested.
Sexual harassment was a strong detcnent to women nngnuits tliioughout
the period. When questioned by aid workers after the capture of most of
the Tigrayan towns by the TPLF in 1988, a number of women mentioned
sexual harassment as their greatest worry.
From November 1984 onwards^ fear of resettlement added yet another
disincentive for visiting towns.
A final deterrent on movement was the dissemination of land mines.
Though planting land mines became most common only after 1986, from
1980 onwards anti-persamiel landmines were planled go patfasi and ammd
army garrisons.
The result of these restrictions and deterrents was that normal pattems
of movement, trade, migratton and excfaaqge were stifled.
Hie Consequences for the TPLF

The famine profoundly influenced the fortunes of the TPLF. In the short
term the famine was a disaster for the front. Between 1980 and 1984, it
was unable to tighten its military grip on Tigray. In 1982-3, it expanded
into Wollo, but was gradually forced to restrict its operations. 1985 was
to be a year m which the araqr was abie to make significant territorial gains.
In late 1984, m re^xmse to tfie increasingly desperate situation, the TPLF
decided on tfie mass evacuation of people to refugee camps in Sudan (see
chapter 11). The loss of people from central Tigray and the diversion of
resources to controlling the mass exodus left the TPLF militarily vulnerable.
In the long term the TPLF was able to turn the Tigray people's
experience of famine into an asset — perhaps its greatest asset. A TPLF
leader, referring to the bombing campaign launched in September 1980,
said: "This is a new tactic to demoralize the people, but it will only make
Alex Lyon, Michael McColgan, Christian Rostoker and Didier Malapcl,

Human Rights in Tigray, Ethiopia," London, 1986.
'Torture and the Violation of
154
them hate the government more."^ By 1985 rural people in Tigray knew
that they would never be free from famine while the army remained in
the province. The experience of war-induced fomine was to be the greatest
source of volunteers for the TPLF.
The early 1980s also forced the TPLF to be intensely pragmatic.
Initially, it attempted socialist measures such as price control and the
creation of cooperatives in the areas of western Tigray which it controlled.
These were not a success. In particular, when the price controls were
introduced in mid 1983, tiadeis simply boycotted the maikets^ and after
a few weeks the TPLF backed down. A
policy of encouraging free
enterprise prevailed thereafter. Virtually no new cooperatives were created.
No attempts to control credit* or restrain moneylenders from charging
extortionate rates of interest, were made. The rationale was that the TPLF
had nothing better with which to replace the existing system. Above all,
migration and petty trade were encouraged. Migrants to western Tigray
were assisted by a network of checkpoints where relief was given, and were
helped in the reception areas on the system of "a family for a family."
The Consequences for Government Military Strategy
Did the government know what it was doing between 1980 and 1984?
The answer is: to a limited extent only. The government was determined
to restrict the food supply to the TPLF and to attack the economic base
of the population that supported it, but did not seem to realize the extreme
but inevitable consequences of these actions. This is consistent with the
over-rigid thinking that informed many of the government's social and
agricultural policies, that wholly overlooked the importance of mobility,
trade and local knowledge in rural people's subsistence. In the minds of
the members of the Dergue, famine was associated with drought, and
counter-insurgency with killing rebels.
The lack of realization is illustrated by the government's response to
the famine. In the early days, this was a straightforward portrayal of the
problem as one of drought, and appeal for mtemational assistance. As late
as 1983, Mengisttt felt able to draw attention to the famine. In his May
Day speech, Mengistu said:
Compatriots, there is a drought in some regions of our country. This

has brought famine among some of our people in the villages. This
situation tests our goal. We started off saying that we will at least satisfy
Dan Conneil in Ihe Guardian, London, October 17, 1980.
155
our food needs ... it is unacceptable that we cannot at least satisfy our
food needs. It is a mystery that we are starving when we have enough
land that can even produce surplus for other countries, and su^cient
manpower. We need to get out of this shameful situation.
It seems likely that having created one of the most severe and widespread
famines in modem times, the Ethiopian govenuaeat dSd not fully grasp
what it had done.
That state of ignorance was never total, and did not last. By the end
of 1984, there can be no doubt that the govemmeat was awaie Aat it had
been instrumental in creating the famine. It may not mlixed thehm
complexities or the depth of its culpability, but senior members of the
government knew that the war and the famine were inextricably intertwined.
In December 1984, Acting Foreign Minister Tibebu Bekele said to the US
Charge d' Affaires, "probably with more candor than he intended." that "food
is a major element in our strategy against the secessionists.
A belated government awareness of its role in the famine can also been
seen in tiie from te hpghming of 1985 oowaids,
militai^ stiategy adopted
îcfa aimed at utilizing food soppUes at a oounler-inBargency tod. Ibh
issoe will be taken op in cfaapleiB 10 and 11.
^ David A Kom, Ethiopia, the United States and the Soviet Union, London,
1986, p. 137.
156
9. "ECONOMIC WAR" ON THE PEASANTS AND
FAMONE
It is widely recognized that the agricultural and economic policies of
the Ethiopian government contributed to the creation of the famine. An
assessment of the merits and demerits of such policies mostly lies outside
the mandate of a human rights organization, with a few significant
exceptions such as the brutal manner in which many of the policies were
implemented, and the lack of any channel for debating issues of national
importance wad thereby changing policy.
Thb chapter wiU cover some of tbae areas of hun^
rights abuse, and
also the broader issues of agricdtnial and ecoBomic policy. Hie latter is
important for two reasons. One is that it is necessary to assess the entire
range of the man-made factors that created famine, so that the role of
human rights abuse can be placed properly in context. The second is that
the demonstration of the disastrous consequences of these policies should
mean that on any future occasions when they are knowingly followed, and
duly create famine, it will not be possible for their proponents to defend
themselves with the claim that they acted in good faith, but in ignorance
of the consequences of their actions.
The Logic of Food Supplies
Tbe Deigue power in 1974 during a wave of popular revulsion

seized
at the corraption of the rule of Haile Selassie. Some aspects of the
revolutionaiy government were radically new, but in certain respects it faced
similar problems to its predecessor and reqx>nded in similar ways.
One important stimulus to the revulsion against the ancien regime was
the discovery of the famine in WoUo by the British television journalist
Jonathan Dimbleby, whose film "The Unknown Famine" shocked the
western world into acting, and the Ethiopian government into admitting
the problem for the first time. Dimbleby's pictures of destitution and
extreme hunger are difficult to watch even today, for those who have
become accustomed to pictores of the 1980s famines, ®n Septeniber 11,
1974, tfie day before the Emperor was deposed, eioeipls from the film were
shown mteisplioed with footage from the weddiqg of the daughter of a
prominent government mmister, for which the cake had been spx^cially flown
from Italy. The following day, as the Emperor was driven from his palace
in the back seat of a Volkswagen beetle, the crowds shouted Ueba/ —
"thiefr
157
Hailc Selassie's symbols of Imperial authority were shattered for ever.
The Dcrgue at first tried to claim legitimacy through slogans referring to
"Ethiopia first" and "Ethiopian socialism." If these were not to be wholly
empty, radical measures were needed. One such measure that was adopted
was the creation of the Relief and Rehabilitation Commission (RRC), a
well-funded government department, assisted by expatriate advisors as
well as Ethiopian experts, to prevent fotnre fEonineSi distrilnite relief to
the victims of natural advostty, and rehabHitale thoBe kift destitute by the
famines of 1972-4.^ Another slogan was "Land to the tiUer," which led
to the land reform proclamation of March 1975, in which private ownership
of land was abolished, and with it all the feudal exactions that had so
oppressed the peasantry. It was optimistically hoped that the creation of
the RRC, together with other radical measures, would banish famine from
Ethiopia altogether.
While the ideology of the new government was radically different from
its predecessor, and some of its actions were truly revolutionary, the
economic constraints it faced were similar. It needed to feed the cities,

especially Addis Ababa, and the army.
Before the land nfom of 1975, most farmecs m
the south paid a large
proportion of their crop to landlords in the form of lent TUs payment
then constituted the bulk of the food that was marketed in the main towns.
After the land reform act was promulgated, this was no k)nger the case -
- much of the food was kept by the farmers for their own consumption.
Meanwhile, in the urban areas, wages stagnated and the price of food rose
fast. Standards of living for industrial workers dropped by between one
third and one half between 1974 and 1979. Economists argued that this
warranted drastic interventions by the government to purchase staple grains.
Placating the urban population was also an important political priority for
the government.
After 1977, the rapidly-growing army also needed to be supplied with
food. An army of 300,000 people consumes a minimum of 60,000 tons
of grain per year.
The government adopted an array of strategies to deal u ith its food
problem. Its ultimate aim was the complete transformation of the Ethiopian
peasant farmers into workers on state-run cooperatives. This was never
achieved, though two ambitious attempts were made in the later 1980s to
move rapidly in that direction — resettlement and villagization — which
are the subject of later chapters.
The RRC was in fact created by Haile Selassie, already bowii^ to the
revolutionary tide, as almost his last act as emperor.
158
During the period 1978-84, the main government strategies included:
* The Agricultural Marketing Corporation (AMC), which was set up in

1976, had its powers dramatically expanded.
* Heavy and evea punitive taxes were levied.
* Strict controls cm the private grain trade were enforced
* Restrictions on peasants' movement and laboring were enforced.
* State farms were developed, using forced labor.
* The RRC was mandated to transform destitute populations into a

potential collectivized workforce, and to obtain food from the
international community.
These policies, which amounted to an economic wir on the independent

peasantry waged by the state, were instnmiental in creatiqg vulnerability
to famme well beyond the area of connter-insmgency operations.
The Agricultund Marketing Corporation
The activities of the AMC

were described by a peasant in Wollo as
"robbing the poor to feed the rich."^
The government set up the AMCin 1976 and gave it increasing powers
over the following years. AMC
operations were gradually extended to
all the major crop-producing regions. By 1980, it was purchasing over
400,000 metric tonnes (MT) of food per annum, two thirds of it from
Feasant Associations (PAs). In 1982/3, 573,000 MT
was procured. Each
PA was given a quota of grain which it had to supply —
but was not
infoimed how large its quota would be until after the pUnting season.^
From the 1980/1 season onwards, the price for each type of grain was fixed
centrally and was the same in all parts of the country. In 1980, the prices
decided upon by the govemment were about 20-25% lower than those
' Dessal^ Rahmato, Tamine and Survival Strategies: A

Cue Study from
Northeast Bduopia,** Addis Abaha, IhstiUilB of Development Reaeaich (fDR% 1987»
p. 109.
' S. Franzel, Legesse Dachi, F. Colbum and Getahun Degu, "Grain Marketing
and Peasant Production in Ethiopia," Addis Ababa, IDR, 1989, pp. 2-3.
159
advised by economists,'* and stayed at the same level for eight years, despite
fast inflation in the prices on the open market. In 1984, the fixed price
was only about 20% of the free market price in Addis Ababa. The quotas
were also centrally determined, and often bore little relation to the size
of the harvest. In Wollo» the quota remained uncfaanged it 23,000 MT
in 1982 and 1983. Local officials in the Minisliy of Agrictittme objected
to the quota and the AMCs inflexible demands for pnmpt delivery to
coUectian points, bat the AMCwas mmiaved» aigiilngftatthae had been
an adequate haivest. (The harvest was of course adequate, bat te AMC
quota was set at a levd above the disposable surplus, and other government
policies were endangering the survival of the people.) In the drought year
of 1984 the quota fell to 6,000 MT (the government was still taking food
out of famine-stricken Wollo) but was reinstated at a high level in 1985/6.
In theory, market mechanisms could have provided the food that the
government needed. In practice, the government was not prepared to wait
for the time such measures would have needed to work, and its Maisdst
ideology was also hostile to the free maiket.
Delivery quotas to the AMCwere set for each PA. All fanneiSi
regardless of the size of their harvest, had to meet their quota. "Even the
poorest of the poor had to sell" complained one Wollo faimer bitterly.^
If they failed, the punishment could be confiscation of assets or
imprisonment. One
study estimated that as a result, half of the grain taken
by the AMC "surplus," but was taken from basic household
was not
reserves.^ The siime study found that as many as a quarter of the peasants
were forced to buy grain in order to meet their quotas. They had to buy
on the open market, often selling essential assets in order to raise the
money, and then sell at a considerable loss to the AMC —
which often
only made payments many montfis later. Many other peasinls had to
its
buy grain hiterm Older to have gnun to eat later in the seaioo. Some young
farmers even abandoned their land and went to look for work on state farms,
because the burden of delivering their quotas was so heavy that tiiey coold
not afford to save to set themselves up with an ok and a plow.
^ Alemayehu Lirenso, "Grain Nfarketiug and Pricing in Ethiopia: A

Study of
the hnpact of Grain Quota and Find Grain Prices on Gnin nDdooeng" Addis
Ababa, IDR» 1987, p. 38.
' Peter Cutler and John Mitchell, "An Evaluation of BandAid Funded
Agricultural Inputs Distribution Programmes in Wollo Administrative Region,
Ethiopia," London, 1987, pp. 43-4.
' Alemayehu Lirenso, 1987, pi 106.
160
Dr Dessalegn Rahmato calculated that a typical peasant in Ambasscl,
WoUo, harvested 14.7 quintals^ of grain per year, of which 5M quintals
were paid to the AMC, leaving 8.9 quintals —
about the bare minimum
for a family of five to survive on. Of the harvest of 6 quintals of pulses,
2.15 were taken by the AMC*
If the harvest were below average, the
family would suffer a shortfall —
on the basis of this exaction alone.
The grain procured by the AMC
was supposed to be resold throughout
the country. In Harerghe, the distribution quotas were 30% to the towns,
30% to PAs, 29% to pastoralists, 10% to government institutions, and
one per cent to jvivate agencies.' In fact, it was destined solely for the
laig^ towns —
in Haie^he, 100 per cent went to Diie Dawa and Haier
towns —
and the amiy. In 1984, the AMCs only and belated response
to the food shortage in WoIIo was to allocate 6,000 MT for distribution
in the towns.^° In the later 1980s, the army took an ever larger share
and by 1989, even Addis Ababa was not receiving its quota, which AMC
was entirely destined for the military.
A coercive government policy, implemented systematically and often
brutally, had the effect of impoverishing many thousands of peasant farmers.
Some were reduced to famine conditions, and others to a state in which
they could no longer sell food to poorer neighbors or migrants, or offer
them employment or other forms of assistance.
Taxes and Levies
The peasants of Ethiopia were subjected to a wide range of taxes and

levies. These steadily increased during the late 1970s and 1980s. All
farmers were required to pay agricultural income tax (20 Birr), a PA
membership fee (minimum 5 Birr), surtax (usually 25 Birr), plus levies
for road-building, school-building, literacy and other campaigns. Members
of the Women's Association paid a 3 Birr annual membership fee, and all
households contributed at least 3 Birr to the Youth Association. School
fees ranged from 2 Birr to 15 Birr. Starting in 1984, all were required
^ One quintal is 100 kilograms.
* Dessalegn Rahmato, 1987, p. 101.
' Ashenafi Mbges, "Food Shortages m Harerghe Region of ûopia," DFhii

Thesis, Ox&ad, 1988, p. 276.
^°
Peter Cutler, "The Development of the 1983-85 Famine in Noithem
Ethiopia," PhD thesis. London, 1988, p. 109.
161
to pay a levy for the woik of the RRQ nsually 20-25 Biir. Famine-
stricken peasants in WoUo could not M
to see the inmy
payment of this "famine tax" to the government."
of the enforced
From 1988 a
contribution "for the territorial integrity of the Motherland" was also levied.
In addition there were frequent demands for corvee labor (for instance for
state farms) and campaign labor (for road-building, reforestation, etc),
always levied without regard to the agricultural cycle. These labor demands
often averaged one day per week, with fines for non-attendance or ill-
discipline. Some taxes were levied by allocating quotas to districts, which
decided on the level of individual payments.
Jason Clay of Cultural Survival made an estimate of die total tax burden
on fanners in Harerghe, based on mterviews with refugees, and produced
the extraordinarily l^g;h figure of 546.92 Birr per annum —more than twice
the per capita gross national product of Ethiopia.^^ This figure is so high
that it needs to be treated with caution, but it is worth noting that about
40 per cent of the total consisted of irregular payments to the PA chairman,
cadres and local militia.
In areas of western Wollega recently occupied by the OLF, Dr Trevor
Trueman estimated that the tax burden had ranged from 70 Birr to 150
Birr yearly, plus charges for schooling and unofficial payments of health
services and other necessities.
Taxation was certainly punitive. Many investigators in rural Ethiopia
report peasants being forced to sell crops or aninuds or go into debt to meet
tax payments. In Wollo, after a small remission taxes in 1984/5 due
to the famine, tax payments increased in 1986. Agricultural income tax
was levied at a rate of 45 Birr, leading to suspicions that "aneais" were
being collected. "Arrears" of taxes were also collected by anny patrols
in Tigray and north WoUo.
Impact of the Policies: Grain Storage
Traditionally, Ethiopian farmers try to keep a large amount of grain

in storage, and will only empty their stores when they are certain of the
next harvest. This meant that the conq)lete failure of one year's harvest
would not lead to famine. The government policies described above made
this impossible by the 1980s. As a result, the harvest failure d
1984 led
" Dessalegn Rahmato, 1987, p. 101.
Jason Gay, "The Case of Haieigbe: The Testimony of Refogees in Somalia,"

and P. Niggli, The Spoils of Famine: EMapUm
in J. Clay, S. Steingrabber Fmune
Policy Peasant Agriculture, Cambridge, Mass., 1988, p. 188.
162
directly to famine in many parts of the country wliich were unaêcted by
war.
Restrictions on Trade
A number of the policies adopted in Tigray and its borderlands for

counter-insurgency were later expanded to areas under government control,
for reasons of sodal control and socio-eoonomic tranônnation. One of
these was restrictions on petty tiade, especially in grain. Outside insuigent
areas, the mtention of the policy was to direct as much as possible of the
marketed grain to the AMC.
Hie chapter on the Red Terror has documented the attack on large
merchants in the years 1975-8, and the accompanying legislation in the
Special Penal Code which acted as a deterrent to legitimate commerce.
A welter of petty legislation further restricting the grain trade, much of
it at the provincial level, followed after 1978. All trading licenses were
issued by the Grain Purchasing Task Force (an affiliate of the AMC). Grain
wholesaling became illegal in much of the country, for instance Gojjam
after 1982. The number of grain traders declined each year after 1980 —
1,100 had their licenses revoked in Gojjam in 1982/3,^ and 342 in Arsi
in 1986.^** Only in 1988 was the trade re-legalized; it was deregulated
in March 1990, although tran^Knrt remained strictly regulated until the fall
of the government.
No specific directives were issued relatioig to traders using pack animals,
but the complexities of licensing and the vague definitions of "hoarding"
and "profiteering" were sufficient to endorse any local official's attempt
to confiscate the goods of a petty trader. One Ethiopian economist noted
"it is unclear whether small scale grain trade is illegal or not".^
Road blocks were set up on all major and minor routes, in order to check
that no unauthorized trade was occurring. In some instances, the
"concession" was sold or granted to favored individuals within the local
PA, who would use the income from bribes, fines, ccMifiscations and levies
for himself —
providing a dear inoenHve to obstruct tlie trade as much
as possible. The tariff for a ten-ton ttack at a siqgle road blodc could be
as high as 5,000 Biir. Commonly, an individual was restricted to a limited
" Aiemayefao Lirenso, 1987, p. 73.
Franzel et al, 1989, p. 12.
" Alemayehu Lirenso, 1987, p. 52.
163
quantity of grain, such as one quintal. Tbe xestrictions were also extended
to people bearing gifts of grata. In 1964» some icndenla of Addis AMm
tried to send food to their relatives in the ftmhie ZQiie» on^ to have the
food confiscated at checkpoints.
Starting with the land reform proclamation of 1975, government
regulations required ''specialization." Each petty-trader was required to
make a choice between being a full-time farmer and a full-time trader,
and in the event of choosing the latter, to live in a town. As most petty-
traders engaged in trade during the dry season when they could not work
on their farms, thiswas a direct attack on their means of survival. This
restriction was never fully enforced, but provided another excuse for local
officials to exact bribes from traders, or to confiscate mules and grain at
will. On the occasions when it was enforced (such as Wollo, starting in
1984) the consequences were even worse.
The government regarded time spent at maikel as "wasted" or
"unproductive," and measures to lestrict markets were introduced depending
on tiie zeal and ideological correctness of local administcatois and cadres.
Some of the cadre's attempts to control tnde and marketing in a leaetttanent

site in WoUega induded:^^
* Restricting the time people were allowed to spend at a market.
* Imposing a synchronization of markets.
* Qosing small roadside markets.
* Defining market days as "work days," so that those who attended

markets lost work points in the collective work system of the setttement.
* Banning men from attending markets at certain times of year, so that

only women could attend.
* Banning all attendance at markets during the harvest season, for 2-3
weeks.
Such stringent measures were rare. More commonly, smaU markets

were closed, as being "unnecessary," or changed firam twice weekly to
weekly. Most markets were synchronized, being held oo Saturdays, whidi
Alula Pankhurst, "SettUqg for a New Worid: Fbopte and the State hi an
Ethiopian Resettlement ViUage," PhD Thesia, Mandiealei; 199Q, pi 332.
164
prevented traders and peasants attending more than one each week. In some
areas, a ban was introduced on peasants attending more than erne market
in tiieir locality.
Restrictioiis were partknlarly tight in Gonder during 1984. Hiis was
idaled to ftclois indudiqg counter-iusurgeucy operatians against TPLF-
EPDM, the hard-line administrators in the province and their hostility to
private trade, and several scandals involving local officials profiteeriog in
the grain trade. These restrictions were particularly damaging because the
1983 harvest in Gonder was good, and unrestricted trade would have
allowed much of the surplus to be redistributed commercially to the famine
areas of Tigray and north Wollo.
Before the revolution, there were an estimated 20,000-30,000 grain
dealers in the country. In 1984, the Ministry of Domestic Trade had issued
licenses to just 4,942, and by law they were required to sell a minimum
of half of their puicfaases toUie AMCWhereas over half the fanners in
die TPLF-Gontrolled village of Adiet engaged in tiade,^^ only two out
of 150 mone study in Haieighe did so, and a sunilarly small number
did so in Wollo.''
Markets are critical to peasants' economic activity in normal times, and
are absolutely essential to "survival strategies" during famine. By these
measures, the rural marketing system was seriously crippled, and the flow
of grain from surplus to deficit areas was effectively stemmed. The
consequences of this for Tigray (where the restrictions were combined with
the bombing of marketplaces) have been mentioned —
comparable, though
less severe, damage was done to rural people's survival chances throughout
the country.
Restrictions on Moneylendmg
A consequence of the restriction on trade was a sharp decline in

moneylending in many parts of the country. There was also a deliberate
assault on the practices of rural credit.
Debt is an essential part of rural life —
farmers rely on taking out debts
in order to obtain seed, food or money to hire a plough team. Before the
revolution, rural moneylending had been conunon throughout the country.
* Richard Baker, interviewed by Alex de Waal, November 1990.
" Ashenafi Moges, 1988, p. 244.
^ Dessalegn Rahnuito, 1987, p. 98.
165
In TPLF-held from
areas of Tigray, the majority of farmers took out loans
traders during the famine.^ While high rates of interest weie charged,
these loans were important in enabling them to survive.
However, in government-held areas, established sources of loans dried
up. Traders were driven from the market, either killed or driven abroad
in the early days of the revolution, or later forced out of business. Local
courts and administfatofs reused to leoognizB the validity of credit
agreements. The most important study of the fiunine in mnl WoUo noted
"Far more peasants wotdd have taken out loans than actually had ... if tfiey
liad more and better access to credit services."^ Asimilar huge post-
revolutionary contraction in credit was not^ in Haieighe»^ though in
some other (non-famine stricken) parts of the country* the supply of credit
was partly made up by other peasants.^
State Farms
After the revolution, the government nationalized all the commercial

farms in the country, and ran them as state farms. A
fixed wage rate of
93 cents per day (equivalent to US$0.44 or less) was paid —
well below
the market rate. Unable to recruit sufficient labor at this rate, the
government took to conscripting laborers either by force or with empty
promises. The results were not only abuses against the human rights 6t
the workers, but a dramatic drop in the availability of work.
The Setit-Humera mechanized farms close to the Sudan border employed
between 100,000 and 300,000 seasonal laborers annually in the early 1970s,
providing an essential source of income for those whose crops had failed.
16,000 laborers on the schemes fled to Sudan when the government began
forcible conscription into the army from among them. Rather than hiring
62 per cent of the sample interviewed by Gerezgerhe Bezabeh and Alex

de Waal, "Evaluation of the Impact of the Agricultural Rehabilitation Programme,
Tigray." London, AOORD, 1988, p. 19.
^ Dessalegn Rahmato, 1987, p. 251.
^ Ashenafi Moges, 1988, p. 322.
^ Dessalegn Rahmato, Agrarian Reform in Ethiopia, Trenton, NJ, 1985, p.

65.
166
J"
laborers, the government then began a program of forced labor, which was
documented by the Anti-Slavery Society.^
In July 1980, the government began to recruit laborers in the towns
with promises of payment of 49.50 Birr per month, plus food, medical
services and accommodation. When insufficient volunteers were found,
soldiers and kebele guards simply rounded up people from the street.
14,140 "volnnteas" were falno m Addis Ababa and abont 30,000 in other
towns. Thdr belongings (even their shoes) were confiscated and they were
crammed into truds, without even the most basic facilities or stops for
rest or sleeping; the trucks were so overcrowded that tiiey had to squat
for the journey of over three days. On arrival at the army camp ckise to
Humera, the soldiers rushed onto the trucks and dragged away many
women, who were raped that night, and many of whom were never seen
again by their relatives or friends.
At the Humera state farm itself, no facilities or accommodation were
available. Food for the workers was inadequate, and twelve hours of work
was enforced each day. No wages were paid. Minor disciplinary offenses
were dealt with by detention or beating; offenders were called "counter-
revohitionaries." Thewholecampwasguardedby armed members of the
"Fkoduction Task FofGe" who detained or shot dead those trying to escape.
In prison, the cells were groasly overcrowded and toitnre was routine.
Aocoiding to the farm supervisors, 1,626 people died from starvation,
disease, beatings and torture, or were shot trymg to escape. Several hundred
disappeared, including women abducted for sexual abuse by soldiers and
officials. Many others fled to Sudan.
This program not only involved gross abuses of the rights of the forcibly
recruited laborers, but removed an important source of income from poor
people in Tigray, rendering them more vulnerable to famine.
Similar practices, albeit on a smaller scale, were employed to raise labor
for the farms in the Awash valley. These farms had employed over 30,000
hOxMers, mostly from WoUo and Tigray, in the eaily 1970b. Bythe 1980s,
laborers were recruited for two to three weeks unpaid labor from central
Wollo, and hence no paid employment was available. Other state farms
recruited labor in similar wa>«» with similar results.
^ Anti-Slavery Society, "Forced Labour in Humera: Intervention on Behalf

UN
of the Anti-Slavery Society," presented to the Economic and Social Council,
Commission on Human Rights, Working Group on Slavery, Geneva, August 1981.
167
Other Restrictions on Wage Labor
Closely related to the restriction ofmovement and the recruitment of

labor for state farms was wage labor on smallholdings.
the restriction of
Since the land reform of 1975, wage labor was technically illegal. While
rarely implemented in the north, this ban provided yet another reason for
officials to harassand punish individuals. In soutlieni Eduopia, the tan
m
was enfoioed moie OQOsistently, paiticidaily die ciae of mi^^
In the early 1970s* moie than 50,000 migfanis came to Keflh evoy year
to pick coffee. Many came from eastern Gonder (an area which suffered
famine in 1984/5); some came from Wollo and Tigray. ^the eady 1980s,
the migration had come to a complete halt.^ A similar imatnnal «i%Mtinn
of laborers from Tigray to Illubabor also ceased.
Peasants were also required to do unpaid labor on the fields of PA
officials and militiamen.
Impact of the Policies: Employment
It can safely be assomed that hi a Doimal year, at least one mUlion

people mUgray and noitfi Wollo are reliant on off-fvm somoes of income,
primarily wage labor and petty trade. A
conservative estimate of the impact
of the government restrictions is that the available employment was
gradually cut by half between 1980 and late 1983. The result of 500,000
people — perhaps one in twelve of the population — being rendered
unemployed in a time of poor food availability was disastrous. These
people formed a large proportion of the total number of destitute migrants
seeking help from REST, the RRC or voluntary agencies.
Impact of the Micks: The Lhrcstock Economy
The restrictions on trade and migndiQB, together with the oonlnctioa

of credit, meant that most inral people had easoitiBlly only one opHon left
to obtain money for food: selling animals.
The famine of 1983-5 is remarkable in that it is the only recorded
famine north-east Africa in which animal sales outnumbered animal
in
deaths. In all other famines, most animal owners have preferred to keep
most animals in the anticipation of future recovery, obtaining their food
through means other than sellii^ them. Some animals die on account of
* Adrian P. Wood, "The Decline of Seasonal Labour Migration to the Coffee

Forests of South-West Ethiopia," Geography, 68, 1983» pp. 53-6.
168
hunger, thirst or disease, but that is a risk the owners are prepared to take -
-- they only sell when they are truly desperate. In 1983-5 in Tigray and
Wollo the reverse was true. Almost all the animals lost were sold in order
to buy food — 79% of oxen according to one survey.^ Losses due to
drought alone might have amounted to 50% at most — the much larger
number sold out of desperation represents another premium paid by the
poor for the military strategy of the govemmeot.
Aninqwclaiitoaiinqiieaoeisllieeiidiiri^ of Dorthem
Ethiopia as a lesnlt of the funine.
Animals —particularly plow tuxm — are essential to econoniic survival
in the Ethiopian highlands. Hie extraordinarily high levels of animal loss
in the famine — Uiia eifiess of the losses that would have been caused
by drought alone — represent a hard blow at the very basis of the peasant
economy. Six years later, animal numbers are still well below what they
were before the famine, and the lack of plow oxen represents the single
most severe constraint on rural production. The effects of the famine persist
in increased vulnerability to famine up to today.
The Relief and Rehabilitation Commission
TheRRCwasapankkndcalhutihitton. On the one hand, for the decade

1974 it enjoyed ooosiderBble institutional stiei^ and
after its creation in
lîthnacy, and a high decree of autonomy from other government
departments. The RRC was relatively efficient at collecting, analyzing
and disseminating information, and in coordinating relief programs. It
compared well to similar institutions in neighboring countries. The
Commissioner was able to travel abroad and meet senior diplomats and
politicians from foreign countries and the UN and appeal for aid on his
own behalf. At a time when all news of the famine was suppressed in the
domestic Ethiopian media (June-September 1984) the RRC was still able
to publish its bulletins on the developing famine.
The behavior of the RRC is a warning to those ^fvho would wish to see
the Ethiopian government as wholly monolithic and dedicated solely to
the shigle-minded deslractioa of Internal oppoaitloa. During the ytm
1980-4 th«e was a real paradox that while several arms of the Ethiopian
government were doing a great deal to create and perpetuate famine, one
arm was busy trying to relieve it — with a certain degree of success.
However, the RRC fulfilled three vital functions which fitted in
extremely well with the twin government aims of suppressiqg insurgencies
CuUer, 1988, p. 313.
169
and maintaining a food supply sufficient to ensure its iodepeDdeoce from
the demands of the ordinary rural people.
One of these roles was using the destitute population of Ethiopia as
the raw material for the creation of new villages, settlement schemes, state
farms and collectives. The villagization carried out in Bale and Sidamo
from 1979-84 is a case in point; resettlement will be considered later.
Hie RRCs second lole was propaganda. The RRC portnyed the ftmme
as a problem of drought and over-population, played down the eristence
of the war, and consistenHy clauned that all funiae victuoa were beiqg
reached by it and the voluntary agencies working alongside it. This meant
that the question of access to the millions of people in EPLF- and TPLF-
controlled areas was never given the publicity that it warranted. The
extremely needy populations in the war zones thus received far less than
they required, while generous relief could be distributed in less needy
government controlled areas in the center and south.
The third role played by the RRC was as a procurement agent for foreign
assistance. It had only limited success up to 1984, but after 1985 it was
exceptionally successfiil, and was repeatedly accused of underestimating
needs by the donors. While much of tiie aid given to tfie RRC undoubtedly
went to genuine victims of famine, much went to less wordiy lectpienlB.
The militias in Eritrea and Hgray are one example (see next chapter).
Another example is the manner in which relief agencies weie directed to
set up distribution programs in areas of the country which weie poor but
by no means suffering from famine — in some instances these were areas
in which large-scale procurement by the AMC was occurring. Finally,
diversion of food aid to the military and the marketplace certainly took
place (see chapter 10).
TPLF and EPLF Social and Economic Policies
The TPLF and EPLF both avowed radical socialist ecooomic doctrines,
but in fact followed far more pragmatic policies.
Both fronts bought grain on the open market. An attempt by the TPLF
in mid 1983 to enforce controlled grain prices in Sheraio maifcet failed
when the local traders boycotted the market, and the policy was abandoned.
Ironically, the TPLF came under criticism horn conservative govenmients
170
in Britain and the US for paying higli prices for grain for relief
distributions.^^
TPLF taxation policy consisted of levying a "voluntary contribution"
of 5 Birr.^* While there are no reports of coercion, a strong element of
social pressure was certainly present. Other taxes were levied on the export
of grain and animals, and the import of luxury goods. The EPLF had
similar policies, in addition to its more significant funds from expatriate
Eritreans' donations and fundraising, and a charge on their remittances to
relatives living in EPLF-controUed areas.
Both fionts moved firom opposing private enterprise to promoting it.
In tbd kte 1970s, the TPLF fought against oonseivalive political forces,
including the EDU, and tiius oootribnted to the exodus of traders from
Hgiay. It also set up cooperatives to compete with tradeis, and attempted
to restrict traders' activities. These measures failed. From 1983-4 onwards,
despite its increasingly hard-line Marxist ideology, the TPLF imposed no
hindrances on trade and regarded merchants as "strategic allies" in the
struggle. Petty trade was recognized as essential to peasant survival, and
encouraged, especially after 1987. There was, however, an assumption
that economic development would mean that it would ultimately cease to
be profitable, so that peasants would abandon it.^^ In the late 1980s the
EPLF revoked its earlier intentions to nationalize most of the Eritrean
economy, which had the immediate impact of leading to measures
facflitating private tradeand enterprise.
Early TPLFmeasures also served to restrict migrant labor. Though
the front never imposed any formal ban on w^ge labor, a number of policies
discouraged it. These included the attadis on Ae merchant-landowner
class of western Tigray, wiiose fnms were a major source of employment,
surveillance of the population to prevent infiltration of government agents
and saboteurs (particularly in 1980), and the TPLFs own land reform
program, which made membership in a baito (equivalent to a PA) a
precondition for having land rights. The first two of these policies changed
from 1983 onwards, so that between 1983 and 1985 the TPLF was actually
" In a letter to the head of an aid agency, dated July 11, 1989, the British
Minister for Overseas Development, Chris Patten, stated his opposition to "a system
which we believe is at the least giving merdiants excessive profits and may well
be helping to finance the TPLF."
* Oebro Tareke, "Fretimimuy Hisloiy of Peasant Resistance in Tigiai

pbiopiaX" Ji/rica (RomaX 39, (1964) pp. 213-14.
^ Meles Zenawi, interviewed by Alex de Waal, November 1988.
171
assisting migration, and itself piovidii^ wage laboring oppoitunities in
western Tigray, for example weaviqg badoels and tapping gum trees. Laige
landholdings in thii^-popnlated areas were also peimitied.
The Consequences for the People
The ultimate human cost of the famine is thenumber of people who

suffered a premature death. Most of ttiose who died were children under
the age of five.
Those with animals had to sell; those without had two options — to
stay and starve, or move to relief shelters. The phenomenally large
population living in the relief shelters m
1984/5 —
over 500,000 was —
the result of this.
Relief shelters and refiig^ camps are notoriously unhealthy places —
one epidemiologist has written that they "constitute one of the most
pathogenic environments imaginable."^ Epidemics of measles, typhus,
relapsing fever, dysentery and cholera swept through and decimated the
camp populations. Though debilitation due to undernourishment
undoubtedly played a part in the high levels of mortality that resulted,
overcrowding and unsanitary conditions were at least as important. It is
a rule of thumb that the death rates in a camp or shelter arc about five times
those prevailing in the community at large.^^ Some of this is due to the
fact that the camp mmates are already weaker than those who remain behind
in the villages, but much of it is caused directly by the denaded public
health environment. There were estimated to be iq) to 15Q|00D deaths m
camps during famine^
the —
half or more of these deaths could Inve
been averted had there been no need for relief camps.
The pictures of Korem and Meqele filmed by Mohamed Amin of
Visnews in October 1984 were thus the direct result of the counter-
insurgency strategy of the government.
^ Bruce Dick, "^Diseases of Refugees —

Caoaea, Effects and Control,"
Transaaions ofOte Royal Society ofliropical Medicine and Ifygiene^ 78, (1984)
p. 736.
Journalists commonly speak of the inhabitants of relief camps as "the lucky

ones" on account of the small amounts of relief they receive. Ihis is of course
not so.
^ John Seaman, "Famine Mortality in Ethiopia and Sudan," Liê, faltemational

Union for the Scientific Study of Population, 1987.
172
Copyrighted materii
In the yens 1980-3, lesliicled access to the Hgrayan towns meant that
mral people had restricted access to health services. In 1981 there was
an epidemic of malaiiat in 1982 there was an outbreak of meningitis, and
in 1984 there were serious epidemics of measles and other diseases. These
undoubtedly caused more deaths than they would have done had free access
to health facilities been available. In response to this problem, REST began
setting up health clinics, especially in 1981/2, but it was unable to cope
with the magnitude of the health problems.
How Many Died?
Ihe total number of people kiUed by the funine is not known. The
UN has gone on leooid saying that one millioo died, but this is no moie
than a guess. No systematic studies of mortality among the population
in general were conducted, and the government discouraged any attempts
to carry them out." The mortality data for the 1970s famines are better
than those foe the 1980s. The RRC never published its own figures for
deaths.
The death totals in camps can be estimated with some accuracy. 40,000
died in camps in Wollo between August 1984 and August 1985,^ 15,000
in refugee camps in Sudan, and a total of between 100,000-150,000 for
the whole famine zone for the whole period. (Deaths in the resettlement
program will be considered m chapter 12.)
Deaths amo^g the general population are open to greater problems of
estimation —
we know neither the size of the populatkm, the number
affected by £Eunine, the death rate in normal years, the length during which
the death rate was raised on account of famine, nor flie death rate during
the famine. Nevertheless, some approximations can be made.
The total country-wide affected population during 1984/5, according
to theRRC, was 6,098,000. In the north, there were 872,000 in Eritrea,
1,790,900 in Wollo,^^ 376,500 in Gonder and 200,000 in northern Shewa.
The real figure for Tigray was at least one million more than the official
estimate of 1,331,900; including these would give a total affected in the
A
study by the Ethiopian statistician Asmerom Kidane was subject to
censorship (see chapter 12) and independent aUempts to do demographic surveys
met with official obstruction.
^ John Mitchell. "Review of the Funine Relief Operation in WoUo

Administrative Region, Ethiopiay" mimeo, 1986, p. 20.
^ Later the figure for Wollo was revised upwards to over Z5 miUion.
173
north of about 33 rniUion. It can be asBumed that death sates rose among
the affected populatioiis of Tigmy and north WoUo a year before they began
to rise elsewhere. In 1983/4» 1.1 million were affected in north WoUo and
about two million in Tigiay. This gives a total of about 6.4 million
"affected person years."
The death rate in normal times in the north of Ethiopia is approximately
20 per thousand per year.
For death rates during the famine, the followii^ fragments of data are
available:
* An independent survey in 1987, which found that 30 per cent of sampled

households had lost an average of 1.6 funily aeoibeEB dnriqg 1994 and
1985.^ Assunung an average household of five meoaben, £is implies
a death rate of about 96 per thousand for the two years. This is
consistent with one year of normal mortality and one year of a raised
rate of 76 per thousand.
* An ad hoc Red Cross survey result, which indicates that 51 per cent
of households in central Wollo lost one or more family members during
the famine/'^ On varying assumptions, this implies a death rate of
102 or 163 per thousand for the whole famine.
* A survey among refugees arriving in Sudan, which indicated a death

rate of 70 per thousand m the preceding year (up to and including
migration).
* A large-scale survey m northern Shewa in late 1 985, which found death
rates of 96-108 per thousand over a short period for a large

population.^'
^ Dessalegn Rahmato, 1987, pp. 3, 131.
^ Cited in Dessalegn Rahmato, 1987, p. 131.
^ Cutler, 1988, p. 348.
^ Dr Mack Otten of Save the Children Federation (USl quoted by Seaman,

1987.
174
* AcompUationofdata by lUBST in eariy 1985 which iDdfcatedtlia^
people in Tigray were dying each day, implymg a death late equivalent
to 110 per tbofusand.'*^
* A survey in resettlement sites done by a professional demographer (the

only such professional survey) which found a peak death rate of about
115-123 per thousand.^'
Famine mortality usually shows a seasonal variation, with the peak

generally about one and a half times higher than the average for the year.
The peak of about 100 per thousand Implies a yearly rale of about 70 per
thousand, which is consonant with the findings of surveys coveriqg a loi^
period. The excess above normal which can be atbibnted to fumme is
therefore about 50 per thousand. With 6.4 million person years "affect^",
this imphes 320,000 deaths attributable to famine during 1983-5. Deaths
in camps over and above the general famine death rate need to be added.
100,000 people died in camps, about five times the number if the people
had remained in the countryside. Adding this extra 80,000, the total is
400,000.
More pessimistic estimates would expand the affected population and
raise the average famine death rate, each by 25 per cent, and give the higher
estimate for deaths in camps, resulting in an estimate of 590,000 famine
deaths. Ihese figures concur with the best-informed contemporary
estunates, which argue for a total of about 500,000 fiumne deaths for the
period 1982-6.^
How many of the deaths can be attributed to the counter-insurgency
methods of the Ethiopian government and other punitive measures
implemented to the cost of rural people? This question cannot be answêd
in anything but the crudest terms. The counter-insurgency strategy caused
the famine to strike one year earlier than would otherwise have been the
case, and forced people to migrate to relief shelters and refugee camps.
The economic war against the peasants caused the famine to spread to other
areas of the country. If the famine had struck only in 1984/5, and only
affected the "core" areas of Tigray and north Wollo (3.1 million affected
^ REST were assuming a population of Tigray and its borderlands amounting

to five million.
^ Asmerom IQdane, Demographic Oonsequences of the Ethiopian Famine,"

Demography, 26, (1989) pp. 515-22.
Cutler, 1988, p. 437.
175
people), and caused only one quarter of the number to migrate to camps,
the death tollwould have been 175,000 (on the optimistic assumptions)
and 273,000 (on the pessimistic assumptions). Thus between 225,000 and
317,000 deaths — rather more than half of those caused by the famine -
- can be blamed on the government's human rights violations.
In additioii, deafliB firam famine oocuned in £e south— in Harerghe,
Bale, Sidamo and in Wollaita district of Sliewa. These deaths afanost
certainly numbered in the tens of thousands. Oovonmeiit oonnlei^
insuigency strategies including militaiy offensives and tooed idocatioo^
were instrumental in creatiqg the famine in the tfaiee aoatheni pnn^Mes,
and its agricultuial policy was crucial in Wollaita.
^ Described in chapter 5.
176
10. WAR AND THE USE OF RELIEF AS A WEAPON IN
ERITREA, 1984-88
Id October 1984, the famiiie and war in Ethiopia took a dramatic turn,
with four sinmltaneons devdb]|Mnents. On October 23» the BBC screened
afihn of the starvation in Koran, unleashing a juggernaut of international
aid that completely transfonned the fEunine, which had up to that point
been developing without large amounts of external relief aid. Secondly,
on October 27, the army in Eritrea launched its largest offensive for two-
and-a-half years, which was to be followed up by an even larger series
of offensives during 1985. Thirdly, the Ethiopian government launched
its principal response to the famine, the resettlement program. Finally,
an ambitious villagization program was launched in Harerghe, as a counter-
insurgency measure against the OLF, which presaged the program in other
parts of the country. These four developments are the subject of this and
the following three chapters.
Tbe huge reliefopecation that swung into action meant that after October
1984, control of relief was a major component of the militaiy strategy of
both the government and the rebel firants. The systematte use (and denial)
of food relief for military ends was the most notable aspect of government
military strategy that also included extcaordinanly sustained and widespread
brutality against civilians.
In the southeast in 1979-82, and Eritrea in 1982, the government had
tried with mixed success to obtain humanitarian assistance from the
international community to use for the pacification element of its counter-
insurgency strategy. Between 1982 and 1984, these aid flows were drying
up, and the repatriation initiative was not meeting with much success. After
Odtober 19B4^ the massive mflflw of relief dlowied
when it wished to do so, to its prefened counter-insurgency strategy.
Hiis chiqiter focuses on the continued war in Eritrea, and the role of
aid in tiie government's counter-insurgency strategy. As elements of that
strategy were common to both Eritrea and Tigray, some details concerning
the aid programs in Tigray will also be included. It ends with the EPLF
victory at Afabet in March 1988. Building on less spectacular military
gains over the previous six months, this battle marked a turning point in
the war — from then onwards, both EPLF and TPLF had gained military
supremacy.
177
The Aid Rcspoow
The publicity suddenly given to the famine represented an earthquake

in the relief world. From donating only $11 million to Ethiopia in Financial
Year (FY) 1983 and $23 million in FY 1984, USAID increased its donations
to $350 million in FY 1985. The members of the European Community
increased their donations from $111 million in (calendar) 1983 to $213
million in 1984 and $325 million in 1985. Overall assistance rose from
$361 mUlion in 1983 to $417 miUioa in 1984 and $784 miUion in 1985.^
975,000 metric tonnes (MT) of food was deliveied to Hw flovemment side
between November 1984 and October 1985. About 80,000 went to MT
EPLF- and TPLF-hdd areas.
Dilemmas in Feeding like Nor&i
In October 1984 the aid donors faced an acute fiilemma; wliether to

channel relief through the government or rebel side.
In December, diplomats and relief agency staff in Addis Ababa estimated
that the government had access to only 22 per cent of the famine-stricken
population.^ The Eritrean Relief Association (ERA) and the Relief Society
of Tigray (RESl^ had aoocsB to most of the icoiaindcr, nsiqg "cnx»-boidei"
routes fiom Sudm, and tiavellii^ only at nigiit to awoM acrid
The Emeigency Relief Desk, a consortium of humanitarian agencies set
up in 1981, acted as an intennediary between the donors and ERA and
REST, avoiding many of the problems that those donors would have
encountered had they implemented the programs themselves.^ Nevertheless,
the great majority of the assistance was channelled through the RRC and
voluntary agencies working alongside it: according to most they
received over 90 per cent of the money and food.
The US shared with other donors a preference for working on the
government side, it was quicker, cheaper, and more public. However, unlike
every other major aid donor (save the International Conmiittee of tte Red
^
Figures from Organization for Economic Cooperation and Development,
"Geographical Distribution of Financial Flows to Developing Countries,** 1984-
1986.
^ Ftnil Vallely, Tamme: Russia and US on Golhision Ooune." The Times,

London, June 4, 1985.
^ Barbara Hcndrie, "Cioss-Bcndec Relief Opeiations m Eritrea and Tigiay,"

Disasters, 13, (1990), p. 355.
178
Gross [ICRC]X USAID was the only donor to contemplate giving substantial
support to the cross-border operation. Because of the politics of the relief
program, USAID did not in the end give the support that it promised, and
the cross-border operation remained grossly under-supplied during 1984
and 1985.
The US donated 5,000 MT
to ERA and REST in April 1984 for the
cross-border operation, a further 23,000 MT
in November, and another
"substantial" donation in December. In Sqjtember, USAID initiated
discussions to launch a much larger cross-border program, possibly of a
size to eclipse the program run firom the government side. According to
documents obtained bf the journalist Paul Vallely, several plans were
mooted; the most ambitious involved spending over $100 million and
building a road from Sudan into central Tigray; the least ambitious involved
donating 240 trucks to ERA and REST."* Only in late 1985 did assistance
in the form of 150 trucks finally materialize. Throughout most of 1985,
REST was still operating a battered fleet of 55 agii^ Fiat trucks, assisted
by a smaller fleet of ICRC vehicles.
Repeated demands by diplomats, humanitarian agencies and the rebel
fronts for a "food truce" to allow non-governmental agencies to supply
relief to the hungry in all parts of the country were consistently rejected
by the government. In December 1984, Acting Foreign Minister Hbebu
Bekele rejected an approach from the US Charge d'Albires, saying that
it amounted to a proposal "to make an anaqgement with criminals."^
Government ofificiiils consistently rejected the suggestion that there were
areas of the country which could not be reached by the RRC

In these circumstances USAID used the cross-border operation as a
bargaining tool with the government: unless greater access was provided
from the government side, the US would throw major resources into the
cross-border operation. However, this strategy was undermined by various
factors, including:
* The lack of a suitable private relief organization to take on the role of

implementing partner for USAID's relief food (CARE was approached
but refused in March);
* The unilateral donation of resources to the government side by all other

major donors, notably the UN;
* Vallely, The Thnes, London, June 4, 1985.
^Quoted in: David A. KonktEthkipUi, the IMited States and Ihe Soviet Unkm,
London, 1986, p. 135.
179
Copyrighted malarial
The AprO 7 popular uprising in Sudan, which oveitfaiew President Jaafar
Nimeiri (a close atty of the US and at the time very liostile to Mengistu)
and brought to power the Transitional Military Council, whkh hoped
for improved rdatioBS with the Ethiopian gpvejnmeiit;^
* The need for the relief program to be public, so that the television
viewers in the US could see that their government was acting. The
Ethiopian government was totally intransigent and refused to concede
any legitimacy at all to the cross-border program, so fliat it had to be
dandestiiie. It theiefoie could not be opn to te teteviaml viiiti of
US politicians, jonmalisls and media peraonalltiea.
Two other factors helped to swing the debate in favor of the government
side. One was that the army's military successes in the Eighth Offensive
in Tigray between February and April meant that many more areas became
accessible from the government side (see next chapter). The second was
that the US State Department realised that neither EPLF nor TPLF were
ideologically suited to playing the role of "contras" against the Communist
government in Addis Ababa. In favor of the cross-border operation was
the fear that a large refugee influx into Sudan would destabilize that country,
whidi icmained a western ally.'
Hius, due to essentially political considenitioiis, the crosi-bofder
operation fell out of favor with USAID — though a reduced level of
support for it continued. "Food for the Noiyi'' initiative was
Instead, the
proposed in March, whereby US \ oluntary agencies would work on the
government side. World Vision was selected for Tigray and Catholic Relief
Services (CRS) for Eritrea. The proposal and the decision to support it
were made before the full cooperation of the Ethiopian government was
obtained, this meant that throughout 1985 and afterwards the government
consistently had the upper hand in determining the conditions under which
aid was provided hi Eritrea and Tigray. Itame, not only did tfiecniaa-
border operation fafl to lecehre adequate support, but its value as a
bargammg counter with the government was never properly iodized.
Sudanese-Ethiopian relations dramatically improved during the rest of 1985,

but deteriorated during 1986 and 1987 as the Sudanese civil war escalated and
the Ediiopian government oontinoed to provide military support to die lebd Sudan
People's liberation Aony. Sudanese lelationB with the fionls improved as thoae
with the govenunent aouied.
^ Until January 1985 the Sudan government officially blamed food shortages
on the refugee influx, denying the existence of a domestic food problem.
180
Food Relief and Survival Strategies
The politicsof food assistance to Ethiopia over this period have received
much from journalists, relief workers and scholars.' Hiis chapter
attention
is ooDoemed with the pcoblon of funine more generally. Even in the worst
fanunes in AMca» food lelief provided by fannuuiitarian agencies or
governments plays a relatively minor ide in the survival of fie people.
For example, the enormous rdief effort to the famine-stricken Sudanese
region of Darfur in 1984/5 succeeded in providing no more than about
twelve per cent of the total food consumed by the people of that region
during the famine, and had little appreciable effect on their survival
chances.' Similarly in Ethiopia, famine relief was at best the last ten per
cent which assisted rural people in surviving. Moreover, in contrast to
other survival strategies such as gathering wild foods, trading, or taking
paid labor, the provision of relief food was unreliable and was often
accompanied by unpleasant side-effects such as the need to walk long
distances, absent oneself from the jEnm at aatical stages in the agri
cycle, or live in a disease-fidden relief shelter. Thus, while the provision
or mterruptkm of relief supplies was vitally important for the people of
northern Ethiopia after October 1984, other aspects of the counter-
insurgency strategy which adversely affected the survival strategies of the
affected population were, as during the previous years, even more important.
The fact that relief assistance is much less important than "normal"
economic activities is significant. does not make the disruption
While it
of relief supplies any less morally reprehensible, it draws attention to the
fact that other actions which create famine are even more deserving of moral
outrage.
The October 1984 Offensive
Within the same week that the BBC film of Korem was broadcast, the
Ethiopian government launched a major offensive in Eritrea. The following
year was to equal 1982 in terms of the suffering of civilian Eritreans on
account of the war; it was to surpass the horrors of that year because rural
people had to contend with the problems of famine as well.
' The best such book remains: Peter Gill, A Year in the Death cf Africa,
London, 1986.
'Alex de Waal, "Is Famine Relief trelevant to Rural People?" JDS Bulletin,
20.2, Sussex, April 1988, pp. 63-9.
181
By October 1984, the 60,000 new conscripts drafted earlier that year
had finished their military training and were ready for battle. The total
size of the regular army was 210,000, with 170,000 militia. Meanwhile
there had been a build up of armor —the army now had 750 main battle
tanks and 130 combat aircraft. These new forces were soon to be deployed:
for eleven of the next thirteen months, the army was actively engaged in
major cxffensives, its laost soatamed military action sinoe 1978.
On October 27, 1984, the Etliiopiaii army laondied a large offBiisive
which lasted untO January. Hiia inclnded the fnoiliar demeDta of
indiscriminate bombing and shelling of civilian targets as a prelude to
ground attack. The market at Molki, Seraye district, was bombed on
October 2, killing 42 marketgoers and wounding 90. In the EPLF base
area of Orota, a school was bombed. Villages close to the EPLF front lines
were subjected to indiscriminate shelling during a period of two weeks,
and there was further aerial bombardment towards the end of the offensive
in January.
After the fighting stopped there was little quiet in Eritrea. In April the
army attacked Nacfa, and there were numerous small engagements.
The Food for the North Inltiatirc, 1985
The huge inflow of aid in general, and the Food for ttie North Initiative
in particular, gave the Ethiopian government a new resource which it began
to use in the middle of 1985. The government preferred to have the aid
consigned to its own RRC, which could then utilize the aid as it pleased.
However, when some of the main aid donors, particularly the US, insisted
on using voluntary agencies, the government saw that it had advantages
in this arrangement too. The government knew that few relief agencies
would have the courage to speak out about human rights abuses or the role
of relief in the war, particularly if it threatened to drat down ttieir programs
m reprisal. The publicity which the agencies woold draw could thm be
used m the governments favor, and to flie disadvanti^ of flie rdiel fironts.
Paul Vidlely noted some of the ironies off tiie Food for the Nofth
Initiative:
The relief agencies were pleased. So was the Dergue, because the deal
would provide a programme of pacification in the rebel areas newly
under its control. Food could be distributed without the risk that it
might fall into the hands of the rebel army. Moreover, the presence
of western aid workers in the area would constrain the vigour of any
[rebel] counter-offensive. Having got the Soviet Union to finance the
182
operation, Colonel Mengistu had now got the US to finance its
consolidation with food handouts.^*
After an initial reluctance to let the Food for the North program go
ahead, the government allowed CRS into Areza and Barentu. The CRS
target was to open two more centers in Keren and Agocdit and disliilnite
food to 200,000 people. Over the fidlowii^mootliStgoveiiiiii^
for the program grew.
Hw 1985 Offinuhrcs
In July 1985, the EPLF took the important garrison town of Barentu.
The CRS feeding program, which had opened a few months beforehand,
was stopped. The government now had no presence in western Eritrea,
where the only functioning relief programs were those organized by ERA,
bringing in food cross-border from Sudan.
In August the government launched a huge offensive —
equalling the
Red Star campaign of 1982 in terms of numbers of troops deployed. In
two phases, it lasted tuita late October. In terms of losses by tiw EPLF,
it was probably
the costliest campaign —
estuaates for canalties among
the EPLFrange from 2»000 to 4^000 and higfier.^^ One aim ai the
campaign was to disrupt die supply routes from Sudan used by ERA.
While the offensive was waged, the army coounandeered all transport
in Eritrea. RRC distributions in Eritrea and Tigray came almost to a halt -
- falling from 14,122 MT in July to 2,069 MT in August.
Within the space of a few weeks, Barentu was recaptured by the army,
which went on to retake Tessenei, held by the EPLF for 18 months, and
a key town for access to Sudan. The whole cross-border route through
Kassala was closed on August 25, leaving ERA only the longer and more
difficult route from Port Sudan. A large irrigated farm nearby at Ali Ghidir
was also ciqjtured. ERA estunated that 20,000 MT of food aid was lost
totheamiy. 30,000 itfiMgcs fled across the bolder to Sudan, and 190^000
people were internally displaced in Eritrea. 22 villages were destroyed or
abandoned because of aerial bombardment or artillery shelling. There were
credible reports of reprisals taken against civilian residents <tf Barentu and
the surcoundhig villages, in which 37 people were killed.
10
Ihe Times, London, June 4, 198S.
" Roy Pateman, Eritrea: Even the Stones are Burning, Trenton, NJ, 1990,
p. 140; Africa Confidential, 26.30,October 30, 1985.
183
Aerial bombardment continued throughout this period. On September
20 and 22, the ERA camp for displaced people at Solomuna in Sahel was
bombed and 20 adults and nine children were killed. Tlie village of Badme
was also bombed, killing nine.
EPLF Strategy
Hie 1985 affenslves were a major bat tempomy seteck tete EPLF.
As in Hgiay, the experience of the government war stial^ and its role
in creating famine hardened popnlar support for the front. Between 1984
and 1987, £PLF stnqglfa sose from 12,000 to about 30^000^ and the
numbers of militia were increased.^^
The EPLF responded to the government advance of 1985 by
consolidating its control in northwest Eritrea, expanding its forces, and
protecting its relief routes. In late 1987, it began to go on the offensive,
disrupting government supply lines, and attacking convoys. In December
the EPLF overran the army's main defensive positions at Nacfa, a prelude
to the devastating attack on Afiabet in Maich.
Picificatioii in Erttm 19W-7 I: Tlw Amy

The military strategy adopted in 1986-7 was a familiar one: constant
military patrols and small-scale offensives, bombing "everything that
moves" in the EPLF-controlled areas, and a continued pcogiram of
villagization and pacification elsewhere.
The army in Eritrea exacted many bloody reprisals against the civilian
population. The following are some reported incidents:
* Nfaicfa 1986: Seaafe. Akde Guzd: two killed when a acddi^

gienade into a crowd.
* June 1986: lower Anseba and Dembezan aieafl, Hamanien: seveial

villages looted by aoidieis.
* August 6, 1986: Hamazu, Southern £ritiea: 140 lulled, 100 injured by

soldiers.
NOVIB, "War and Famine in Eritrea and Tigray: An Investigation into the
Arms Deliveries to the Struggling Parties in Entrea and Tigray," Zeist, the
184
* September 1986: Adibara, Barka: five peasants killed in letaliadon for
a land mine explosion which destroyed an anny trade.
* September 23-4, 1986: Akele Guzai: 24 peasants executed, 20 arrested.
* June 24, 1987: Adi Hadid, Akele Guzai: six killed.
* June 26, 1987: Haikota and Adi Shimel, Barka: 29 farmers were
ploughing their land when troops came and took to them to Adi Shimel,
vibm 16 were killed and Hailoota, where ten were killed.
* July 1987: Anseba, Senhit: 16 people killed and 106 cattle bunied by
soldiers.
Other forms of harassment continued. In one reported incident, a man

in Adikuta, Akele Guzai, was ordered to pay a 3,000 Birr fine after seven
of his goats stepped on land mines, which detonated. A
court ordered that
the fine be paid to compensate the army for "wasted mines."
The University of Leeds assessment team which visited Eritrea in late
1987 obtained figures for losses and destruction due to the war.^^ The
figures hidicate fiat since the begmmog of 1986:
* 22,500 hectares of land had been destroyed by military action or

rendered unisable by land mines.
* 3,500 tons of food had been confiscated by the army.
* 43,900 domestic animals had been stolen by the army (mostly sheep
and goats, but indudiug substantial numbers of pack animals, cattle
and camels).
1,500 civilians had been killed, 3,600 imprisoned, and 200 raped by
soldiers.
* 13 nuUioo Biir of money and property had been looted by tiie army.
" The figures came from ERA and were not therefore independent; some
incidents contained in them have ben ann-Gheckad, but no foil independent check
has been possible.
185
* 2^00 homes had been destroyed.^'*
Pacifying Eritrea H: The Air Force
Bombing in Eritrea continued constantly. Some attacks included:
* August 4, 1986: Dekidashin: three civilians killed, two wounded.
* September 15» 1986: Egela Hatsin, Akele Guzai: five killed, 20

wounded.
* September 20, 1986: Ha Tsaeda, Sahel: one IdUed.
* January 3 1 1987: Hawasheit relief center: four civilians killed, indudiqg

,
a four year old girl.
On at least one occasion, the MiGs crossed the Sudan border on their
bombing missions. In August 1987, an agricultural camp was attacked
and one woman was killed and five wounded.^
Pacifying Eritrea m: Relief
Relief food was a major strategic element in Eritrea from 1985 onwards.
The traditional relationship of regular and guerrilla armies to the population
— that they rely on the people to give them food —
was reversed. Both
the government and the EPLF
had more food resources at their disposal,
and used it to obtain the support of and control over the population. That,
however, is where the symmetry ended. The logic of the government's
position, as an unpopular presence trying to subdue an unwilling populace,
meant that food relief was used to restrict people, as the more acceptable
side of a violent and impoverishii^ omnter-insuxgency strategy. The logic
of the EPLFs position was that it already enjoyed wideqiiead popiuar
support, and it wanted to feed the peoê in the areas it oonboUed in ocder
to prevent them migratiqg to government-held areas or Id Sudan to look
for food.
" Figures reproduced in: Lionel Cliffe, "The Impact of War and the Response
to it in Different Agrarian Systems in Eritrea," Development and Change, 20, (1989)
p. 377.
^ Sudan News Agency, August 12, 1987.
186
After the success of the August 1985 offensive in Eritrea, the
government's attitude to the Food for the North Initiative warmed. "They
have done a complete volte face and are encouraging PVOs [private
voluntary agencies] to expand it still further" said Richard Eney of
USAID. CRS moved back into Barentu. The number of distribution
points was increased, including Keren. Critics of the program argued that
"when Ethiopian troops advance on a place like Barentu, and then a few
days later an American voluntary agency comes in to distribute US food
and medical supplies, yon become haid-piessed not to see this as an odd
kind of coordination."^^ CRS and USAD) ugned that the liumanitarian
necessity of pcovidiqg food ovenode any political-militaiy ooosideiations,
that the EPLF was happy to see Eritrean people fed, no matter who by,
and that a few months of emergency feeding would not win hearts and
minds after 25 years of bloody warfare.^*
The relief given in Eritrea was generous, in contrast to the years 1983-4.
However, outside the towns, it was tied to the continuing program of the
creation of protected villages. In order to receive food, rural people had
to bring their whole families, and register, often paying a fee of 5 Birr in
order to do so. The food was then given in frequent small amounts, making
it impossible for the family to return home with food.
Other means were used to oootcol the population in distribution centers:

most were protected villages, witii cuifews enforced and movement
restricted, with only one ix two permitted entrances, and ringed by anti-
peiscnmel land mines.
One way in which the aid was used to dcaw people into protected
villages was by threatening and harassing those who tried to obtain aid
from ERA distribution centers. Said Ali Mohamed was interviewed by
Barbara Hendrie at an ERA center in March 1986:^^
In my village we are completely encircled by the enemy. The Dergue

is in all directions; we drink from the same well, you can say. We are
afraid. All of the time they take our camels, our goats, our animals.
Quoted in: Jonathan B. Tucker, "The Politics of Famine: US Foreign Policy

in Ethiopia, 1982-1985," mimeo, 1985, p. 55.
Tudcer, 1985, p. 49.
Tudcer, 1985, pp. 49-50.
Barbara Hendrie, "Field Report — Eritrea," Emcfgency Relief Desk,

Khartoum, April 1986, Appendix B.
187
The food I get here, I will not take it home to my village, because if
Ido, and the Dergue finds out, they will kill us. So I will hide my food
in valleys or in the hills and will travel to the hiding place haai my
village to take some food at a time, daily or weekly.
The enemy has prevented us horn getti4g food before, so this is the
first time I have come here.
Securing Garrisons and Roads
Asupplementary functioD of the program, for the anny, was to keep

roads opca and give protectkm to military convoys. Od roads where the
EPLF was known to be active, the army would often send civilian buses
or relief convoys, and only if they got through without trouble, would they
send military vehicles. Some convoys were mixed relief and military, and
often the identity of different vehicles was unclear.
This policy inevitably led to tragedy. On July 13, 1988, a civilian bus
travelling between Asmara and Tigray struck a land mine and caught fire.
According to the government, 25 civilian passengers were killed
immediately and nine died later. Relief vehicles also suffered.
EPLFAttadks on Mitf Vehicles

In August-September 1987, two simultaneous developments meant that
relief vehicles became by EPLF ambushes. One was the
targets for attack
expansion of EPLF military activity further southward and eastward, so
that it threatened major roads. The second was the drought of the summer
of 1987, which meant that relief activities needed to be greatly stepped
up.
On October 23, 1987, the EPLF attacked a convoy of 34 trucks south
of Asmara. EPLF fighters drained the fuel tanks, poured the fuel over the
vehicles, and set them alight. One driver was also killed. 23 of the trucks
carried relief food, much of it supplied by BandAid, and the mddent
attracted worldwide condemnation. The EPLF **t««n**«i that three trucks
carried arms. This claim has never been substantiated. However, the relief
vehicles wm
travelling in a mixed convoy, which included conmicrcial
and government vehicles. The only marking identifying the relief trucks
were small stickers on the doors, which were invisible from a distance.
The attack on the convoy caused a flurry of activity in the humanitarian
community aimed to ensure that it did not happen again. Over the
following week, there was an attempt to negotiate a "safe passage"
agreement with the government. This was immediately rejected by the
188
government, wheiei^xm the EPLF said it would contiiiiie to attack convoys
to which it had not given prior deaiance.
Less ambitiously, there were consultations between EPLF, Emergency
Relief Desk and BaadAid in Khartoum.^ A
proposal was worked out
whereby relief trucks would be much more clearly marked (with large flags
indicating the beginning and end of the relief "bloc" in a convoy) and the
EPLF would be warned in advance of the movements of such convoys.
The EPLF for its part promised to give "instructions to the army units to
separate military from relief and to take all possible precautions to ensure
"^^
that relief materials and relief transportation are not harmed. The details
of this proposal were passed via the British Embassy in Addis Ababa to
Mr Michael Priestley, then head of the UN Emergency Office for Ethiopia
(UNEOE), who summarily rejected the plan.
Hieiealter, the safo^^ of the relief vehicles depended on the efficiency
of an infoimal process ol oommunicaiioo between the humanitarian agrades
in Asmara and Addis Ababa, Penny Jeoden, director of BandAid in London,
Emergency Relief Desk in Khartoum, and the EPLF.
Over the following months there were more attacks on relief vehicles,
but none on convoys of marked relief vehicles unaccompanied by military
vehicles. In November, there were two attacks on convoys, which later
transpired to have not been carrying relief. In December, a convoy of 13
vehicles was destroyed, and on January 15, 1988, another convoy, including
relief vehicles, was burned near Massawa. While the relief trucks were
unmariced and were accompanied by militaiy vehicles^ tfaeie is no evidence
that the attacking EPLF force tried to separate out the military and relief
vehicles as tiiey had earlier promised.
The attacks attracted mtemational publicity. Both the EPLF and the
government confined the argument to the issue of famine relief, which
ensured that the government would be the winner, at least in terms of
international public opinion: the EPLF undoubtedly carried the responsibility
for destroying the relief supplies. A more rounded assessment of the ethics
of the EPLF attacks on relief demands attention to the government's whole
pacification strategy. The relief operations were but a small part of an
overall government strategy. The ten per cent contribution to survival
provided by relief was more than ofifset by the government attack on the
Other 90 per cent of the existing economy, and the relief was itself an
integral part of the pacification program that was undermining rural people's
^ ICRC also participated but witfidrew pending its "Open Roads for Survival"
initiative (see ctmpter 11).
^ EPLF press release, December 2, 1987.
189
ability to provision themselves. Ensuring the "neutrality" of relief deliveries
would bave helped oidy a IHfle wlifle such an overall mflitiiy strategy
persisted. The EPLF attacks caused people to go hungry, but they helped
make the pacification strategy less viable.
In March 1988, the EPLF won its greatest ever military victory at Afabet
after a huge three-day battle in which it defeated a foroe of 15,000 men

and over 50 tanks. It was the turning point in the war, and has been
compared to the Viet Minh victory at Dien Bien Phu. However, such was
the international preoccupation with food relief —
a preoccupation partly
orchestrated by the government — that a headline in TTie Times of London
of March 31 ran: "Stepped up guerrilla raids threaten food deliveries."
Diversion of Aid to the Military
Frequent allegations were made during the early 1980s of huge-scale

diversion of food to the mOitary in Ethiopia. The evidence was largely
the testimony of refugees in Sudan, and visitors to EPLF- and TPLF-
controlled areas who saw relief food stockpiled in the stoies of captured
garrisons. In March 1983, these allegations reached a new level with a
report in the Sunday Times of London under the classic headline: "Starving
babies' food sold to buy Soviet arms." Repxâted visits by high-level
representatives of donor countries, including Canada, the UK
the European
Community and (a year later) the US, any substance to these
failed to find
allegations, besides odd occasions of "loans" of food to the army. In late
1985, the UN estimated the rate of diversion at about five per cent, which
is considered low under the circumstances. However, the diveision of this
amount of food would, from 1985 onwards, have been sufficient to feed
300,000-400,000 men — the entire armed fofces.
Government officials were frank about the practice of feeding soldiers.
In both Eritrea and Tigray (though apparently less so elsewhere) large
amounts of relief food were used to feed the locally-conscripted militia.
Dawit Wolde Giorgis, head of the RRC, reports a '^request** from Meqgistu
to divert food to the army in June 1984:
I understood the problem of the military. The conunon soldier was

the victim of government policies just as much as the peasant. Soldiers
were dying by the hundreds every week in the varicMis civU wars. I
didn'twant them to suffer more due to food shortages. I now suggested
that would be easier for the RRC to divert food to the peasant militias
it
rather than to the military establishment on the government payroll.

These militias recruited from the peasantry were not paid; they were
190
simply trained and ordered out to fight. I felt that feeding them was
l£ke feeding peasants.^
Relief allocations in Eritrea and Tigray were made by a Drought

Emergency Qumnittee, composed of a range of government officials, only
one of whom represented the RRC. Reportedly, allocations to the militia
were made against the wishes of the RRC representative in Eritrea.
Later, government Deputy Administrator for Eritrea, Yishak Tsegai,
told a reporter that the militia are "given priority because there is no fixed
salary or privileges.... 20,000 peasants are armed on the side of the
government."^ These militia came to be known as milisha sirnai or
"wheat militia" because of their method of payment.
In addition, about 15,000 militia in Tigray were fed. A
visitor to Tigray
discovered evidence of diis in Meqele in Ktocli 1989, after its capture by
fheTPLF:
Everyone I spoke to said that the RRC only gave food to their own
militias and their families, not to the poor. We found indications of
this in the looted RRC offices where file after file was titled "militia
of ... receiving assistance" with lists of beneficiaries.^
Extensive evidence of the use of food aid by army garrisons was

discovered by journalist Peter Worthington, who visited the garrison of
Afabet shortly after its capture by the EPLF.
I went to the Ethiopian army kitchen and store depots to see what had
been left in the hasty deparhire —and found, stacked against a wsdl,
a number of 50 kilogram sacks of flour, maiked "C.I.D.A. [Canadian
International Development Agency] Gift of Canada." Serving as curtains
to the shelves were other sacks that had once contained Canadian wheat
... I visited a dozen hole-in-wall shops, run mostly by Moslems and
stocked with items such as Kiwi shoe polish, soap, toilet paper, colored
^ Dawit Woide Giorgis, Red Tears: Wiur, Revolution and Famine in &hiopia,
Trenton. NJ, 1989, pp. 157-8.
^ Jane Perlez, New York Times, February 15, 1990. The real figure was closer
to 50,000.
^ Jennie Street, "Report of Field Trip to Non-Govemment Held Areas of

Hgniy, Ethiopia, March 1989," Khartoum, Emergency Relief Ded[, 1989, p. 34.
191
hail oil, sardines, banana chewing gum from Saudi Arabia — and bags
of Canadian wheat flour.
Every store, literally, had stacks of ftonr, Budoed CIJ)A. Questions

were answered evasively. They were keeping ^ to someone else ...
Yes, it was for sale in the markets of Asman ... no, it wasn't for sale -
- unless someone wanted to buy it ... the army helped them to get it,
sold to shopkeepers on the sly, everyone maldqg a small profit.
Worthington also discovered cooking and other foodstuffs donated

oil
as aid in the garrison. by the army and
In 1989, the sale of relief food
militia in Eritrea had reached such a scale that a brisk trade was being
conducted across the battle lines into Tigray, where it was contributing
a significant amount to the diet of people and keeping food prices low.
Despite its claims to the contrary, the UN in Etidopia was in fact well-
mformed about the large scale diversion of relief to the militia. In June
19S5, an international food monitor in Eritrea documented that militiamen
were regularly receiving 90 Idlogiams of wheat per month. Of this, they
kept about half for consumption, and sold the other half for money. The
diversion was concealed by entering six names on the ration list for every
militiaman. Often the family of the militiamen received regular rations
as well. Ordinary civilians on the ration list received between 10 and 20
kilograms, and when food was short, they received nothing: the militia
received priority in the allocation of supplies. The food monitor estimated
that appioxunately one third of the relief in Eritrea was actaally being
directed to the militia. Mr Knrt Jansson, the UN «wffHw^ fior the
emergency in Addis Ababa, chose not to piiblidze this well-documented
report, and denied press reports that referred to the diversion of food. Later,
Mr Jansson admitted that it had occurred but said that feeding militiamen
"could not be considered wrong" as the militiamen would have received
food anyway, had they remained as ordinary farmers.^^
It is quite possible that the total diversion of food aid to the army and
militia did not amount to more than five to ten per cent of the total.^^
^ Peter Worthington, The Sunday Sun, Toronto, April 24* IS^. Theie are
numerous other similar accounts dating back to the 1970s.
^ Kurt Jansson, Michael Harris and Angela Penrose, The Ethiopian Famine,
London, 1987, p. 56.
^ In 1985 the US General Accounting Office estunaled flienile of divexrion

at 4.3 per cent, but did not take into account the fiseduig of the militia.
192
This however obscures the fact that the percentages were much higher in
the critically-affected areas of the north; that even when the food reached
the hungry it played a strategic military role; and the assistance program
as a whole gave much needed foreign currency to the government.
Perhaps the most important way in which the relief effort contributed
to the govefnmeiit^ war efifoit and indeed sinvivalim
rate. All foreign assistance was excfaaqged into Ethiopian Bin at the rate
of 2.07 to the dollar, despite the fact thai the real madcet rate was two to
three times higher. The government thereby taxed all currency transactions
for relief by 100-150 per cent, in addition to port cfaaige8» impoft duties
and license fees. This came to be the major source of foreign exchange
for the government.
The relief effort also supported the fronts. This took several forms.
One was the feeding of militiamen, who were in other respects poor farmers,
and who received rations from their local baitos which distributed to the
poor on behalf of ERA and REST. Another was beneficiaries contributing
some relief supplies to fighters, without direct coercicHi, but undoubtedly
Witt some sodal pressure. The main strategic benefit the fronts obtained
from relief food was that the people were able to stay in their village and
were not obliged to migrate to Sudaa
Allegations of the straightforward re-consignment of relief to the fronts
— including fighters, support personnel, field hospitals, and prisoner of
war camps — have been made by the Ethiopian government, defectors
from the fronts, and Mr Jansson of UNEOE.^* They have never been
confirmed. The diversion of five per cent of the cross-border food would
have been sufficient to supply food to about half the EPLF's and TPLFs
fighters. With the exception of ICRC programs, internationally-donated
food travelled only in ERA and REST vehicles, and was distributed by
locally-appointed distribution conmiittees. While international agencies
frequently sent food momitocs to observe the tranqxxrt and distribution of
certain consigimaaents of food (mdeed the programs m rebd-held «
more intensively monitored than those on the government side)^ independent
comprehensive accountiqg of donations was never requested by the
donors.^^
ERA and REST exchanged foreign currency in Saudi Arabia at rates
close to the free market rate, though there was one incident in late 1989
^ Jansson, 1987, p. 52.
® The exception was the internal purchase of surplus foodgrains in Tigray,

which was funded by USAID in 1984, 1989 and 1990, and on ail occasions the
purchases were actually made by US mcniitont.
193
and early 1990 in which a significant cUveigeiice occurred, in the onkr
of 40 per cent. When the discrepancy was discovered by an American
monitor, ERA apologized and returned the additional funds to the donors.
International aid to the rebel side was much less than to the government,
and proportionately certainly contributed much less to the fronts' war effort.
Famine Continues
Almost all the fBdms that led to the creation of funme m

Eritrea m
1984 remamed in fdace after 1985. The only excqptioiBi were two yean
of better rainfall uid a much laiger supply of food od. However, the
pacification program and the war continued, and these continued to have
the effects described in chapter 7. The underlying causes of the &mine
were not addressed, and when there was drought combined with a smaller
relief program in 1989/90, severe famine returned.
A remarkable incident occurred in 1987 which illustrated the Ethiopian
government's In the early 1980s, the Eritrean Public Health
priorities.
Program (EPHP), a branch of the EPLF, developed a small, low-cost
civil
field microscope. In 1984, production of the microscope began under
license m
London. At a cost of idxmt one tcalh of the existing
commercially-avaiUble microacopes, and with a Uglit-weigpit fold-away
design, it had potential for use thnmglioiit the developiqg wodid. The Wodd
Health Oiganization (WHO) sponsored a series of tests, which the
microscope passed, whereupon distribution began in several countries such
as Brazil, Nicaragua and Indonesia. However in 1986 the Ethiopian
government began to object to the microscope, which bore the initials of
the EPHP, claiming that it was no more than part of a propaganda campaign
launched by the "secessionist elements engaged in the rabid dismemberment
of Ethiopia."'^ WHO and UNICEF were obliged to withdraw their
endorsement of the microscope.
Letter from H.E. Teferra Haile Selassie, Ethiopian Ambassador in London,

to Dr Neil Anderson, London ScbooL of Hygiene and Tropicai Medidoe» September
1. 1986.
194
11. STARVING HGRAY, 1984-88
Tigray during the last months of 1984 and the first half of 1985
represented the very nadir of the famine —
the most intense and widespread
suffering in the entire country.
As before, a primary reason for the severity of the famine was the
government's counter-insurgency strategy, including an extremely destruc-
tive army offensive. Unlike in other provinces, however, thete was no
agpificant relief program which could have ofibet this deliberately-indttced
suffering. Hie govenunent delibcfately witfibeld aid ftom the province
which it held, and tried to pievent aid from leachiqg TPLF-held areas.
The Evaciiation to Sudan
After the failure of the main 1984 harvest and the strangling effect of
the government's intensified counter-insurgency strategies, Tigray faced
unprecedented disaster. One way in which the TPLF responded to this
was a mass evacuation of people to Sudan.
From 1980, the TPLF had an established policy of assisting impoverished
migrants in western Tigray, through REST. In 1983, about 75,000 crossed
the bofdn to Sudan. In inid-*1984, the food sitnatiao became so diie that
the TPIJF warned the UN High OHiimissioner for Ref^^
a much larger number of refugees would be soon arriving in Sudan. The
predicted 300,000 did not hi the event arrive, but about 189,000 did.
UNHCR however had made contingency plans for only SQfiOO new refugees
(including Eritreans), and delayed implementing a program on the grounds
that the refugees were drought victims and therefore had no "well-founded
fear of persecution," that no appeal for assistance had been received by
the Sudan government, and that it could not deal with the TPLF as a "non-
recognized entity." Only when the Sudan government made belated requests
for aid, having futilely tried to close the border in November 1984, and
the aid bandwagon was akeady rolling, did UNHCR decide to treat the
inflow as a "special case"; it then provided generous assistance.
TPU? and REST assisted the westwanlmifâtioa, which took 4-5 weeks
oo foot Feeding oenleis were set up at key points along the route, and
food was brought across the border mm Sudan. From 5,000 new refugees
in September 1984, the inflow increased to 23,000 in October and 87,000
in December, tailing off during 1985. A disagreement between EPLF and
TPLF in March 1985 led to the former closing the key road from Tigray
to Sudan, which runs through Eritrea, forcing the refugees to use the longer
and more dangerous route through Gonder. Conditions in the camps in
195
Sudan were appalling, with some of the worst death rates ever recorded.
An estimated 10,00(^15,000 Tigrayan refugees died in camps in Sudan
in 1984/5.
About 150,000 intemally-dtaplaced Tigrayans were assisted by REST
in villages and shelters in western Tigray, and a further 500,000 migialed
in search of employment in western Tigniy and Gonder.
In March-April 1985, with assistance from TPLF and REST, the
refugees began to return to prepare their fields for plowing. The return
was resisted by most humanitarian agencies, including UNHCR (which
until a year previously had been anxious to promote repatriation together
with the Ethiopian government), which argued that the returnees faced
certain death from starvation. But the refugees knew that unless they
cultivated, they would remain perpetual paupers, and determined to return.
At one point they staged a hunger strike in support of their right to go Imck.
70,000 left by the end of June, and almost all of the remainder over the
following two years.
REST launched an ambitious (though underfunded) relief program to
assist the returnees, and to provide relief to the hungry inside Tigray.
Protecting the migration and relief routes became a major concern to the
TPLF, which consequently had to alter its mode of warfare from pure
guerrilla tactics to the consolidation of a "base area" in the west.
In retrospect, it is probably the case that the evacuation cost lives. The
very high death rate while on the road and in Sudan caused by epidemic
disease and exposure probably surpassed that which would have occurred
had the migrants renuuned at home in their villages.
Some ^servers have compared tfie evacuation to Sudan with the
resettlement program implemented by tiie government, equating it with
a crime against humanity. Hiis is inappropriate, for several reasons:
* there is not a single piece of evidence to suggest that the evacuation

was anything but voluntary; in fact there were more volunteers to
migrate than the REST "pipeline" could handle;
* it was temporary and was followed by a program of assisted return;
* in late 1984 the TPLF was led to believe that generous humanitarian
aid from the west would be forthcoming in the Sudanese idhgee camps,
and counted on that — the appalliqg conditions in the camps were thus
to a large extent the fault of the western donm and UNHCR;
* the TPLF never tried to implement such a program again.
196
The TPLF learned several lessons from the evacuation. One was the
military lesson that a guerrilla movement could not survive in an area which
was depopulated or disrupted by mass population movements, as the military
aellwcks of 1^ demonstrated. For Ifais and for aoniid tamnanitariaii
reasons, TTLF-REST policy after 1985 was geared to maintauiuig the
population in its home area. The second lesson was that the international
community was an unidiable ally. Laige-scale international aid to the
cross-border operation and the refugee camps, on which so much had
depended, never materialized. The TPLF turned inward; there was a
prolonged internal debate on the respective virtues of "pragmatism" and
textbook Leninism. The Leninists, who maintained that the TPLF should
rely principally on mobilizing the local population, won, and the Marxist-
Leninist League of Tigray, the vanguard party of the TPLF, was formed.
The Eighth Offensive in Tigray
The Eighth Ofifensive opened on February 17, 198S, and was waged
for three months. The campaign was fon^lit oo two fipoots: in Tembien
in central Tigray, the heart of the funine zone, and in western Tigray, where
harvest surpluses and employment were available, and where REST was
bringing in food across the border from Sudan, and evacuating refugees
to Sudan. While cutting off access to relief from REST was one of the
main objectives of the offensive, it also had the familiar effects of cutting
employment levels and trade flows, and disrupting agricultural activity.
The offensive also delayed the implementation of the Food for the North
Initiative.
The opening of the offensive coincided with the day before the tenth
anniversary of the founding of the TPLF, and large crowds had gathered
m various TFLF-held towns to celebrate. In an imcharacteristic lapse of
security and/or intelligence, the TPLF allawed the crowds to con^gate
m daylight in Abi Adl Adoctor woikii^ for a foreign relief agency
witnessed what happened next:
The celebrations were just getting under way in the marketplace when,
at 16:55 hours, there was a sudden roar and two MiGs flew low over
the square where we stood. People screamed and ran in all directions,
their faces frozen in terror. The MiGs circled and returned three times
to strafe themarketplace with machine guns and shells. The pilots then
spotted large numbers of people running away across open ground down
by the Tankwa Biver towards a eucalyptus grove, so the planes turned
their attention to this area, strafing it repeatedly. We crouched next
to awall notfar fin»n the maitet square as the MiOs roared m to attack
197
again and again. A
little girl ran screaming down the centre of the road
nearby. After 15 minutes the planes left, leaving behind them a scene
of appalling carnage. Initial estimates put the casualties at 25 dead and
more than 100 wounded, 36 seriously. The number dead would have
been far higher if not for the impressive efforts of first aid workers
whose prompt treataieiit saved many lives.
There was a second attack the following day, in wliicii the MiGs used
high eiqidosives to destroy the baildiDiga of the town, which had now been
almost completely evacuated, so casualties woe mnch ligfiter: seven killed
and 30 wounded. Several other Tigrayan towns were bombed during the
same day. For the next two weeks, regular bombing raids continued. Abi
Adi was bombed on two more occasions, as was the small town of Sambela,
and 20 people were killed.
Then, on March 1, the ground offensive began, with troops moving out
from Meqele and Korem. The Meqele column captured Abi Adi.^ Many
villages north of Abi Adi were attacked. In April, a survey of refugees
ui Sudan found fliat the majority came firom tfaii aiea, wime destniction
had been widespread.
Refugees, REST aid convoys and feeduig oenleis were aD targets of
the bombing. In December, a column of refi^gees walking to Sudan was
attacked near Shilalo in Shire, and 18 people were killed and 56 wounded.
Between March 27-30, nine transit and feeding centers used by REST for
giving assistance to migrants were bombed. 6,000 people had to be
evacuated from the center at Zelazelay, which involved moving ill people
and pregnant women from the clinic. The pharmacy and drug store was
later destroyed in the attacks.
Meanwhile, the ground forces moved into western Tigray to cut the
relief routes used by REST from Sudan —
also cutting the reads used by
refugees. The Henni Gorge link between central and weHem Tigray was
cut. On April 23, the mignuitsf transit camp at Edagi HMt WIS atlk^
and destroyed by ground forces;six days later Sberaro (an important relief
center as well as a TPLF stronghold) was captured, though held only for
ten days. REST was compelled to evacuate its field office nearby, which
was occupied by the army a week later. Cross-border relief operations
were suspended: a food convoy from the International Committee of the
Red Cross (ICRC) had to return to Sudan. A TPLF counter-offensive in
May recaptured these towns and villages, forcing the army to withdraw
Many of the following details of the offensive are based on informaticm

compiled by Barbara Hendrie.
198
to Enda Selassie, but the relief effort had been plunged into chaos for a
critical month.
There were further attacks on RESTs supply routes in June and
September, mounted from the garrison at Enda Selassie.
The offensive included maintaining the existing severe restrictions on
movemeiit and trade in eastern Tigray and nortiiem Wollo and Gonder.
The government also decided to cut off any access to food that the TPLF
and its sympathizers might have through feeding programs. Therefore,
rations at the camp at Ibnat (Gonder) were cut in March, and the camp
itself was brutally closed on April 29. Over 50,000 destitute peo|de weie
violently forced out by the army, which burned the makeshift houses, forced
patients out of hospital beds, and left the expellees with no food, water
or shelter. The incident created a diplomatic scandal, with the US charge
d'affaires in particular speaking out in strong terms.
In southern Tigray and north Wollo, the army was moving through the
heart of the famine zone. Its military progress was made easier by the
depopulation of much of the area: there were few villagers to feed, house
and provide intelligence for the TPLF fighters. Pushing west from Korem
in nud-April, an army column capimed Seqota, which had been held by
the TPLF-EPDM since February 1982. This column then joined with the
force in Abi Adi to cut off Simien, for another military assault from the
western side (this attack was another reason for the boming of Ibnat).
Simien is a major area of surplus grain and employment opportunities, and
the military activities severely disrupted labor migration and the grain trade.
In August, another offensive was launched on Simien* the fifth in 18
months.
One of the aims of the Eighth Offensive was to abort USAID plans
to donate substantial amounts of relief to REST, by showing that REST
could not safely deliver the food to Tigray. In this, as in its other short-
term objectives, the offensive succeeded. In early March, USAID
abandoned its ambitious friUms for cross-border relidf, and instead opted
for utilizing private US hinnanitarian i^gendes woik^
- the Food for the North Initiative. The private agency World Vision was
to open three feeding centers in government-held towns in Tigray. The
government was initially uncooperative, but by April belatedly agreed to
let World Vision open one feeding center, in Maichew.
Starving Tigray, 1985
In Eritrea, the Food for the North Initiative enabled the government
to undertake a pacification program, using relief as part of its military
Strategy. In Tigray, the strati^ wis different. For the most part, the
199
government showed no interest in pacification — the program for
population relocation was the resettlement program, which removed people
from the province altogether. Instead, it was interested in maintaining
Strategic garrisons and withholding food from the population which it
correctly saw as sympathetic to the TPLF.
The relief agencies in Tigray had an important military role. This was
not to pacify the conntryside, bnt to piolect military ganiMMii especiilly —
those most vuinenblelo TPLF attack —
and their ovedandsapply loatei.
Hins in April, when die aimy cqiliiied Shemo tfie govennnent adsed
World Vision to come and set up a feedii^g oenler there. The TPLF
recaptured the town before World Vision OOaU respond to the leqnest.
Similarly, when Seqota was captured the same month, the government
proposed moving some of the people in the camp at Korcm to Seqota and
inviting in foreign relief agencies. The agencies objected to the population
relocation but agreed to start programs in Seqota.
The presence of the agencies did not stop the government withholding
food from Tigray. This can be seen by comparing the relief deliveries to
the various RRC centeis in Ethiopia with the nniSber of famine-affected
people in the area served by each oenter, as shown in table L
TOjte 1. Ddiveridi bf tte RRQ April-AngMt 1965
% of affected people % of yaia delivered
WoUo 29 (25) 24.4

Nazareth 27 (23) 36.5
Eritrea 14 (12) 21^
Tigray 21 (33) 5^
Etoewhoe 9 (8) 11^
The figures err on the side of caution. First, they refer to the period
between April and August 1985, when the government had greatest access
to the different parts of Tigray, and before the disruption to supplies caused
by the August offensive in Eritrea fed through to the distribution centers.
Second, many of those in central and southern Ethiopia were in much less
need than those in Tigray and north Wollo. Third, RRC figures for
"affected people" are used. The RRC assumed that the total population
of Tigray was 2.41 million, witt 133 millioa "aKscM"; the Food and
Agricultnie Oiganiatian of the UN
(JPAO) piodnoed a slighUy higher figure
200
of 1.58 million 'âffected." REST argued that the population was ahnost
five million, with 3.8 million in need.
Mr Kurt Jansson, head of the UN Emergency Office for Ethiopia
(UNEOE), endorsed the govemment-FAO figures saying that "the scientific
evidence is indisputable."^ Mr Jansson was misinformed. As
At best,
shown evidence about the population in Tigray
in chapter one, the scientific
is highly disputable, and Ifae REST figure is almost oertainly closer to the
truth. Tbe RRC figure for tfiose in "need" was a work of imagination.
The FAO crop assessinent was tiased upon satellite imagery onl^ no ground
visits to fuial Tigray were undertaken. While the satellite imagery may
have given a reliable indication of the geographical extent of the drought,
the famine of course affected a much wider area. The FAO's inference
to numbeis of people affected was based upon the government's population
figures.
It can safely be assumed that the RRC omitted a minimum of one million
needy people in Tigray from its figures. This makes the neglect of Tigray
is even more striking —
these revised figures are shown in brackets in
table 1.
Hie show only RRC consignments; during Ibis period
figures
by voluntary agencies accounted for over half tte relief. Hiis
distributions
was deliveced in approximately the same proportionsi ttras not affecting
the percentages in the table above.'
For reasons discussed in the preceding chapter, the cross-border relief
effort was also starved of resources at this time. REST and ICRC
transported 32,000 and 8,400 metric tonnes (MI) respectively into Tigray
during all of 1985.
Thus Tigray, with one third of the total famine-stricken population,
received only about one twentieth of the food relief. Needy Eritreans, per
head, received about ten times as much.
b addition, oCBcial antipadiy to large-scale aid to Wollo was evident
to many relief wmkers, who reported on "a systematic attempt to deprive
[Wollo] of food."^ Mr Knrt Jansson of UNEOE
rqdied to the allegation
that Wollo was beu^ deprived, arguing that a low level of donations to
' Kurt Jansson, letter to Ms Gayle Smith, REST Support Committee,

Washington DC, August 29, 1985.
' Widi the exception of Eritrea, which vBoeived a muUcr prapoction from
private agencies.
* Paul Valiely, "Starving Wollo: An Empty Excuse," Jhe Tunes, London,

August 14, 1985.
201
the RRC and a relative shortage of voluntary agencies in the province
accounted for the relief shortfall.^ He failed to explain why the RRC did
not allow more voluntary agencies to work there, and did not increase its
own consignments to Wollo, taking food away from other provinces with
lesser need but larger programs.
In January 1985, the RRC was reaching 885,000 people in Wollo, and
the voluntary agencies about 400,000. Ova the ooaisp of itt year, the
voluntary agencies incieased their programs so that by Deoeniber tiiey woe
reaching just over ooemiUioa people. However, the people of WoUo gamed
little —the RRC took the oppoftunity to reduce its pn^gmms, so that it
was reaching only 153,000 by the year end —
making a gross total lower
than that reached twelve months earlier. The expanding voluntary agency
program was no excuse for the RRC cutbacks, as the distributions never
reached as many as half of the estimated 2.58 million people in Wollo "in
need."*
Sometimes the diversions were very crude. For instance, 7,500 MT
of grain earmarked by the donors for Wollo and consigned to the RRC
was sent "by mistake" to Nazareth in October 1985. It took four months
of lobbying from the weslon aid donocs before it readied its oonect
destination? CompariDg Wollo unfavorably with other parts of the oowitry,
one investigator noted that ratioas tliere "nuely eaaceeded 10 Idlognms [per
person per month]."*
Yet, as the table shows, famine victims in Wollo received on average
six times as much as those in Tigray. The UN claimed in August 1985
that 75 per cent of the (government-defined) needy people in Tigray were
being reached, a total of 1,126,468 people, with rations of between five
' Kurt Jansson, Michael Hanis and Angela Penrose, Jhe Ethiopian Famine,
London, 1987, pp. 70-3.
< MitcheU, 1986,

pp. 26-7. If the eariier "m need" total of 1.79 million had
been used, the performance would have been worse, not better. Many of the 1.79
million were extremely needy people in the war-affected northern parts of Wollo,
who were not reached by the distributions. Most of the 790,000 people later added
to the figure were less needy but more accessible, and received rations. Similar
"needy " popuktionB woe "diioovaed" in govenunent-held areas in 1985 and
afterwards as huge qoanlities of western relief became available.
^ Mitchell, 1986.
' Peter Cutler, "The Development of the 1983-85 Famine in Nocthem Ethiopia,"
PhD thesis, London, 1988, p. 414.
202
and 15 kilograms per month. Hiis was blatantly untrue. The RRC
distributed a mere 569 MT that month, and World Vision (the principal
agency in the Food for the North program) was well short of its target figure
of 180,000 recipients.
The UN claim was based on a four-day visit to army garrisons in Tigray
made in the company of senior government officials in July, which was
the month in which the largest government distributions to date had
occurred, by far. Allocations to Tigray and Eritrea then dropped by 85%
in August, while those to southern and central provinces rose by 76%.
Even the briefly-expanded distribntkms of July, however, reached forfewer
people than the daimed — that wmdd have required a program
eqnalliqg that in Wollo. Moat of the redpients were the 116»000 Tigrayans
in and around relief shelters in the govemmentîeld towns of the province,
resettlers, and the militia. In the spring and summer of 1985, it is unlikely
that as much as fifteen per cent of the needy in Tigray received relief fnnn
the govenmient side.
ControlUng Tigray 1985-7: 1 The Army

Following the Eighth Offensive, the Ethiopian army had a greater degree
of control in Tigray than at any time since 1977. The counter-insurgency
strategy already developed continued to be implemented with great
nithteness. This udoded military patrols, the activities of army irregvdars,
repeated attempts to block RESTs idief operation, the bondmn of miulcets,
forced resettlement, and the supply of only small amounts of food to the
general population.
Significant army attacks concentrated on attempts to control the roads
from Sudan, such as assaults made in April and July 1986.
Most of the military activity consisted of army patrols, which harassed,
detained and executed p>easants. Some soldiers captured by the TPLF told
visitors of their actions.^*
Mohamed, a former goldsmith admitted that his patrol had looted

...
a farmat Mai Kenetel [central Tigray] and then set fire to the building.
When the inhabitants tried to flee lii^ were mam
down by soldiers
who are trained to do the actual kiUfi^ Another wounded soldier,
Thomas, had experience of these killers. He said: "We were ordered
' Jansson, Harris and Penrose, 1987, p. 51.
Geny McCami, "Between Heaven and Hdl", Observer ScoOmd^ April 16,
1989.
203
to see everyone as the enemy. Ik01edtfanHigiiBiyeyes[Le.Iwa8mdy
to kill] many times. TWo peasants once strayed into my patrol — they
tried to run but were caught. The CO [commanding officer] interrogated
them and then made ns tie their hancte. Another soldier shot them both
between the eyes.
Many minor atrocities in Tigray were committed by government-

recruited banda, which translates as mercenaries, saboteurs, and terrorists.
A visitor to Tigray made the following notes after interviewing a peasant
farmer in Medebai sub-district (near Axum):
There were security problems in his "Abia [duster of villages], he said,

which borders Eritrea. Govenuaeat recruited armed personnel fiom
the locality who, since they are Imowledgeable of the surroundings,
guide government soldiers who are occasionally sent across the Mereb
River from Eritrea, especially to a district called Qohaine to disrupt
life. They are paid in food grain by the government. People refer to
them as "bandas." In September 1987 they killed some peasants.
We were informed that on 31 October 1987, the day of our arrival at

Wereda [sul^-district], they murdered a young peasant at his
the CSiilla
home near the mam town of ^nmi and peeled his driB off. It was
were perpetrated by thebandas
further explained that sunilar atrocities
from time to tune in "semi-liberated" areas, ijt, areas adjacent to
government garrison towns.
One of the most common undertaken by banda was the

activities
planting of land mines on routes from Sudan used by TPLF and REST.
In March 1988, the government launched its Ninth Offensive in Tigray,
with the intention of recapturing the garrisons that had been taken over
the previous five months by the TPLF. In moving out of the garrisons,
the army was militarily exposed and the TPLF was able to launch a
counter-offensive and udlict a series of defeats. After the batfle of Afebet,
the defeats quickly turned mto a rout, and m
late March and early April,
the TPLF captured most of the towns Tlgniy.m
Controlling Tigray U: The Ahr Force
The main targets of continued aerial bombing were markets, REST

convoys, and anything else that the pilots happened to notice. Three trucks
belonging to the ICRC were destroyed by bombing on February 3, 1986,
and one driver was killed. Adi Nebried market was bombed and strafed
204
in February 1987; 17 were killed and 55 wounded.^^ Sheiaio market was
bombed December 1987, killing ten. Phosphorous wis often used, for
in
instance as described by Tabey Kidane, aged 19: "I was guarding my cattle
near Edaga Habreit, when burning material came from the sky, burning
my cows." When examined ten
the trees and the grass and killing one of
months later by Dr Eric Charles, his wounds were still suppurating: "the
bums were deep and were a chemical type of bum ... they kept erupting
and wouldn't heal."
The Ninth Offensive in March 1988 was brief, but witnessed a number
of atrocities by die air fnce. These indnded:
* AbiAdi: "Within a few days of [the aimy's] departure, two helicopters

came and bombed the market square, IdUing and wounding 48 people.
Many of the people who died on that occasion were women, and it was
terrible to see their bodies lying in the square."^
* March 23: Nebelet village bombed, at least one old woman killed. This
was the 11th occasion that this village had been bombed, killing 17
people in total.
Controlling Tigray ni: Relief Agencies
The amounts of relief given on the govenunent side m Tigray were

small, and often given largely as an inducement for resettlement. RRC
relief was distributed exclusively in the towns^ and the continued existence
of the garrison towns came to depend on regular supplies from the RRC.
The main function of the aid agencies' operations in Tigray, as far as
the government was concerned, was to protect vulnerable garrisons from
attack by the TPLF —
the front would lose more in adverse international
publicity by attacking a garrison with a foreign relief agency present than
it could hope to gain militarily. In mid-1985 it began to allow more
international relief operations in and around Tigray, notably in Axum,
Adigrat, Abi Adi, Seqota and —
abortively —
Sheraro.
In December 1986, a TPLF-EPDM fnce (re)captured Seqota, closmg
an ICRC feeding program, and causing an outoy among relief woikers
and diplomats. In March 1988, an EPDM unit destroyed three relief trucks
with their grain, in eastern Gooder, causiqg another round of condemnations.
SuntU^ JUnes, London, March 4, 1987.
Interview with Wcweda Teka of Abi Adi Baito conducted by Sarah Vaugban
and Geiry McCann.
205
Because the trucks were painted green and unmariixd, the EPDM claimed
that the fightets believed them to be anny vehicles. (Four while UN traclGB
accompanying them were allowed to proceed.) In the same mcnotti the TPIi^
(re)captured Abi Adi. A western diplomat was outEHged: "essentially what
this means is that there is going to be mass starvation almost immediately,"
and the director of USAID added: "It's going to hurt us hard."^^ Two
days later, Abi Adi was bombed by the air force.
The most striking instance of the government's complete disregard for
the welfare of the hungry people of Tigray, and its sole concern with
military objectives, comes from Wukro, which was captured by the TPLF
shortly after taking Abi Adi. The TPLFs attack drew condemnation from
diplomats and reUef personnel. Under an ad hoc agreemenl; the TPLF
allowed the ICRC staff who remained in Wukro to distribute the remaining
food. An international food monitor described wlut happened next:
On April 8, 25-30,000 people gathered in Wukro in order to collect

food aid ... The ICRC distribution site, with a huge, clearly marked
red cross tent, is just on the edge of the town. The ICRC representative
of Adigrat, who had taken over responsibility for Wukro after it was
controlled by TPLF, arrived in the late morning ... with his mobile radio
he sent an open message to the ICRC headquarters in Addis Ababa,^"
giving details on the upcoming distribution ... At 2 p.m., MiGs appeared
and started bombing, very close to the orphan-center. Hie buflding
caught fire immediately, and the roof collapsed ... in thk bdlding, 52
dead bodies were counted, which were buried in two bomb craters.
While people from the distribution site and orphan center were fleeing,
the MiGs returned and bombed with cluster shells. In only one street,
I counted five big bomb craters.^
A total of about 100 people were killed in this bombing raid, and 14,000
people fled the town. Wukro was bombed again on April 13, and 31 people
were killed.
In northern Wollo, relief was more generous than in Tigray. However
it was government policies, such as
closely tied to the implementation of
resettlement, road-buikling, and the control of movement. On many
International Herald Tribune, March 10, 1988.
This was despite strong objections by the local TPLF commander.

" Maria Altstidl, "Report on a \ît to Tigray," German Emergency Doctors,
April 1988.
206
oocaskms, mial people would gather for a diitrilnitioii, only to find that
they did not meet the criteria forieceiving idief —
for what leamis, they
did not know. A
poem about lestrktioos on relief has been leooided:
What kind of grain was it which deceived us?

What kind of wheat was it which deceived us?
It returned to its country, and sent us to the forest^^
During this period, the government also continued to oppose the cross-
border operation from Sudan. Relief vehicles were bombed, and agencies
involved in tibe program were subject to hostile propaganda. Unfortimately,
the UN continued to decline to recognize die aoss-bozder program, even
though m
1988 it tnuspoited over 150,000 MT
of idief supplies.
Government hostfli^ did not even spare the ICRQ which was forced
to take unusual measures and operate clandestinely, with unmarked vehicles
travelling at night. Following the sunmier drought of 1987, the ICRC began
to promote a proposal for safe passage, whereby relief could be transported
across the battle lines from government-held towns to rebel-held villages.
This "Open Roads for Survival" initiative was launched on November 12,
1987. Unfortunately, the ICRC was either over-optimistic about the
prospects for success of the initiative, or it believed that sufficient political
pressure from the donors to make the proposal work would only be
forthcoming if the cross-border operation were seen to be unable to reach
the needy ui central Tigray. Hie result was tet flie IQtC withdrew iSrom
the cross-border operation, and dedmed to donate its fleet of 81 trucks
m Sudan to other agencies woildng cross-border, it also puUidy disputed
the claim made by REST and the Emergency Relief Desk^^ that cross-
border supplies could reach the highlands on a road newly-construcled
by theTPLF.
The EPLF and TPLF both agreed to the "open roads" proposal, though
they publicly accused the ICRC of being politically partisan to the
government. However, the government failed to agree to the ICRCs "open
roads" plan —
and indeed on April 6, 1988, it expelled the ICRC from
Eritrea and Tigray. The head of the RRC, Berhanu Jambere, justified this,
saying that the problem was "terrorist action supported by external forces,
and not an all-out external war. Therefore the ocganizatioo's neutral status
Quoted in: Alula Pankfamst, "Settling for a New World: People and the State
in an Ethiopian Resetacmem ViUage," PhD tiieiis, Manchester, 199Q, p. 121.
A ooDsortium of voluntary agencies cnnaignmg food to ERA and REST.
207
does not apply hoe."" Anotlier broadcast sbed fnitber light oo tins
dedsioo: Addis Ababa radio aocosed the ICRC of having an "anogant and
anti-people stand** and '^directly and indirectly supporting the bandits."^*
Two days later, the air foioe bombed an ICRC distribution at Wukro (see
above). Instead of its announced plan to feed one millioD people hi Hgny
by April, ICRC was feeding none.
A between suffering in Tigray
final ironic aspect of the relationship
and the central government is the role that the drought of 1987 played in
the government's strategy. The drought affected Tigray and its borderlands;
central Wollo and other government-controlled areas were scarcely affected.
On the basis of the drought in Tigray and the accompanying international
publicity, the goveimnent appetSsd lor 13 million MX
of lelief grain. This
was almost three times the request made in March 1984, when the sitnatian
was immeasurably more serious —
but this time the westem donon aocnaed
the RRC of ondeieatimating the size of the problem. By this time, famine
in Ethiopia was such a sensitive issue in the domestic politics of weatem
countries that the response was inmiediate and generous. Within seven
months, the target had been met —
a rate of donation 14 times as high
as in the seven months after the March 1984 appeal. The total amount
donated was well in excess of what the RRC estimated was needed, and
even more in excess of real needs. Most of the grain, of course, never
went to Tigray, but in effect served as an enormous subsidy for the
programs of titanic aodal enguieering that the govemmeat was
implement^g in the aouthem and oentral parts of tiie oonntry.
Insurgency Strategy of the TPLF

Following the famine and the 1985 offensives, both EPLF and TPLF
were in temporary military disarray. But by mid-1988 they were enjoying
great military success, and had begun their march to victory. The
fundamental reason for this is that the government strategy of using famine
as an element of war so deeply alienated the peasantry that they turned
in increasing numbers to the rebel fronts.
Introducing his authoritative discrorion of peasant suvival during die
famuie in Wollo, Dr Deasal^ Rahmalo of Adms AbabnUhivenily wtftea:
" Quoted in: BBQ Summary of World Broadcasts {SWB), ME 0124, April
13, X99S,
Radio Addis Ababa, April 13, 1988, quoted in; BBQ SIKB, ME 0126, April
15, 1988. This accusation was made ironically on the basis of Hie ICRCS tnttna
cross-border relief operation into rebei-held areas.
208
I
"for obvious reasons, I shall leave out of our account resistance and
rebeUicm as a form of survival strategy."^ However, aimed resistance
was an essential element to survival strategies.
Between 1984 and 1987, the number of TPLF fighters rose nearly three-
fold, and the front was turning away volunteers. When engaged in military
action, the TPLF could count on active support from the local population.
Even when attacking relief distribution centers, the rebels had local
support. Local peasants argued that the damage done by having an army
garrison in the vicinity frightening people and disrupting movement
outw^died any bendBt from the fb^ provided. "Hie food tiie Dergue
is a poison" remaiked one woman fismier.^

The TPLF militaiy stiategy was dramatically changed by the famine
and the evacuation to Sudan. Over the years 1985-7, it concentrated on
consolidating a "base area" in western Tigray, and controlling the access
routes from Sudan. In this respect, its strategy was more like that of the
EPLF.
Outside the base area, the main TPLF strategy was to attack isolated
garrisons at unpredictable times. As these garrisons were the key to the
government-sponsored relief effort in Tigray and its borderlands, the supply
of relief suffered too. In 1983, the TPLF relief workers at
had captured
Jane and Korem; in November 1984, it attacked Korem again. In 1985,
it staged raids on the Korem-Seqota road, though a relief monitor noted
"casualties from TPLF land mines or TPLF attacks were infrequent and
greater damage was caused by reckless driving."** On March 8, 1986,
the TPLF and EPDM attacked Alamata (north Wollo), and two World
Vision employees were killed and four wounded. According to witnesses,
the killings were not the "mistake" claimed by the TPLF, but were
deliberate. The following month, the TPLF destroyed two bridges over
the Tekezze river, disrupting transport in western Tigray. In December
1986, the TPLF-EPDM (re)captured Seqota, dosing an ICRC feeding
program.
In late 1987, the TPLF began to go on the offensive, snuffing out the
army's network of small garrisons. On October 2, it captured the strategic
Dessalegn Rahmato, "Famine Survival Strategies: A

Case Study of Northeast
Ethiopia," Addis Ababa, Institute of Development Research, 1987, p. 7.
" hiteiviewtd in Abi Adiby Alex de Waal, November 1988.
^ Mitdiell, 1986, p. 38.
^ Imematumal Herald Tribwie, May 13, 1986.
209
garrison of Rama, on the border with Eritrea, aad a flmaber of smaller anny
outposts followed. On February 22, 1988, it captured three relief woikBn»
who were later released. In March it captured Abi Adi. The govemment
launched a counter-oêosive, which ended in militaiy disaster.
210
12. R£S£TTL£M£NT
Three weeks after the media attention to the famine in October 1984,
the Ethiopian government officially launched what was to be the most
i
controversial aspect of its whole famine policy: resettlement. The plan
WIS to move alarge section of the populancm ham the noclli to the south.
The taiget was 1.5 milium people. In fint, about 600^000 peofde were
moved in three phases: November 1984-May 1SI8S, October 198S-Jamiary
1986» and November 1987-March 1988. The jiutification presented to
the west and to the people of Ethiopia was tfiat it was a famine relief
measure —
the north of the country was stricken by drought and
I
environmental collapse, and the only alternative was to move most of the
people elsewhere. Official justifications were embellished with such
I
manifest untruths as: "the fact is that much of Ethiopia, particularly the
northern provinces of Tigrai and Wello, are today an uninhabitable
wasteland" and "there have scarcely been any real rains in the drought-
prone areas since the 1972-4 catastrophe."^ In the domestic Ethiopian
media, the resettlement program was presented as the relief program —
the two were synonymous.
As weU as drought rdiet the program was described in glowing terms
as an opportunity to use the "viigm lands" of the south and west, as an
opportunity for socialist transformation and mechanization of agriculture
in the resettlement sites, and as the first challenge to the cadres of the
newly-set up Workers' Party of Ethiopia (WPE).
The TPLF and independent observers were quick to infer another motive:
counter-insurgency. Population relocation had been a central part of
counter-insurgency strategy in the southeast and in Eritrea, but had not
yet been tried in Tigray and north Wollo. A
policy of trying to remove
I
by force a large section of the population was consonant with both long-
standing military strategy in the country and the existing policy of "draining
the sea to catch the fish." A
Tigray an resettler, Hailu Kelela, was told by
his guards "Your whole woieda [sub-district] supports te TTLF, so we
will txeak anyone who lives here and we wOl not stop with the people,
but we will destroy the whole land unto the bst ttee."^ In addition, the
^ RRC, The Challenges of Drought: Ethiopia's Deoaie of Struggle in Rdief

and RdmbmOan, Addis Abiba, 1985, pp. 180 and 231.
'Quoted by Peter Niggli, "Ethiopia: Deportatioiis and Fofoed Labour Camps,"

Berlin, Berliner Missioiiweik, 1986, p. 9.
211
settler population provided a government stronghold in the resettlement
regions in the southwest, where the OLF insurgency was gaining ground.
Many Ethiopian government policies, including villagization and the
control of trade and migration, functioned both as counter-insurgency
strategies and as mechanisms for social and economic control of the
peasantry. Resettlement was the same. The details of the implementation
of the program varied fitom place to place; at ili wont, it was a bnitd ftxm
of OQaiiler-iiisiirgency, at ila best, a fiem attack oo Ae jadepeadeace of
ttie peasantiy.
Backgrannd to Rocttlement
Before the revolution there was a steady spontaneous outmigration from

the northern highlands to the south and west. Adrian Wood, an authority
on migration in Ethiopia identified 17 locations where resettlement was
occurring between 1950 and 1974, partly encouraged by the government,
and partly assisted by measures such as the eradication of malaria from
many lowland areas. Up to one million people are estimated to have
moved. Intematioiial agencies sadi as die Worid Bank agreed ftat tiie
nortfaem lughlaiidB were "overpopnlaled" and eaoouraged the govenuBCBt
to start programs for controlled resetUement.
The land refcnm of 1975 and accmmmnyigg policy changns stemmed
much of the movement, by making migration and the acquisition of land
more difficult. Rather than granting people freedom of movement, the
government sought to control the process of resettlement. In the ten years
between 1974 and 1984, 187,000 people were resettled under the auspices
of the Settlement Authority and RRC. They included the urban
unemployed, pastoralists and returning refugees as well as northern peasants.
The process was not a success. Although the settlements were planned
to achieve self-fdiaBoe withm tiuee years, by 1984 ttieie weie still 70,000
lesetdeis needing food assistance.^ The program was also very expensive.
Consequently, in 1983 the RRC stopped lesetttenent and started an internal
review of what bad gone wroQg. The review coedtided in September 1984
' Adrian Wood, "Spontaneous Agriciatuial Reaelll^^

in J. Clarke and
I.
London, 1982.
L A. KMiniki ^) Jbdtohhitfm
* Alula Pankhurst, "Settling

for a New Worid: People and the State in an
Ethiopian Resettlement Village," PhD Thesis, Manchester, 1990, pp. 39-40.
212
with the recommendation that a small-scale and slowed-down approach
be tried, using oxen and not tractors for cultivation.^
In the event, two months later, the exact opposite approach to
resettlement was implemented.
RMndtniciit of scttlcn: Non-Insnrgoift Areas
The Etfuopiaii government insisted to the western world that the

resettlement program was voluntary. This was a lie. In fact, each district
in the nnth had its own quota of resettlers, which it had to fill. If
volunteers could be had, that was good; if not, other means would be found
to recruit settlers. The existence of the quota system was itself
fundamentally incompatible with the notion that the recruitment was
voluntary.
In non-insurgent areas, various means were used to recruit settlers.
At the beginning, some famine victims were so desperate that they
volunteered. Resettlement camps and villages which had filled their quotas
for lesetdement were given priority in teims of aid litom the RRQ and
liieie were many instaooes of the witfabokUog of ay in Older ^
resettlement from certain areas. YImam, a settler explained:
It was a short term problem; how to "cross over" from March to April.
We could not last out till the main rains crop. If we could somehow
have survived through that period, we wouldn't have resettled.^
The food camps enticed many. One

available in the resettlement transit
settler reported: "If we could have slipped out after eating our fill we would
have done so" and another likened the camp to a rat trap set to ensnare
peopled All transit camps were heavily guarded to prevent people from
escaping.
Propaganda was also used to obtam volunteers* describing the easy and
comfortable life that setdersooold expect in their new homes. For example,
a video film of green pastures and forests was shown to the inmates of
Korem feeding shelter on December 31, 1984. A
similar film later shown
on Ethiopian television had shot the Wollo landscape through a filter that
^ P^mkhunt, 1990, p. 40.
^ Pankurst, 1990, pp. 95-6. March is the hungry month before April when
the harvest horn the short lains is gathered in.
^ Pankurst, 1990, p. 124.
213
made it appear red and barren, while the colois of the southera landscape
were distorted so as to appear blue -green.
However, an analysis of the origins of settlers clearly shows that
accessibility for government cadres and soldiers was the overriding factor
in determining whether people were resettled. 56 per cent of the rcscttlcrs
in Wollo in 1984/5 came from the easily-reached districts of Dessie Zuria
and Kalu, despite the shoct-tam natm of flie droufl^ tee. Bjr oooliait
Wag and Lasta, much wone-alfectod by fiunnie fiat alao man lemole,
pfDvided only 13 per oeot* At an Mividiial levd* coadnn wai wed.
Hiose in dispute with ttdr Ftasant Association (PA) chaiiimui» or in aireais
on PA dues or tax payments, were likely to be detained and leaettled. Some
traders were stopped at checkpoints and resettled, others were picked off
the streets. A woman from near Kombolcha reported how she and eleven
other families were taken: "We were called to a meeting and told 'your
land is on the mountain slope which is to be used for the forestry; you have
to go for resettlement!' We didn't even eat the maize we had grown on
our irrigated land."'
An agriculturalist studying conditions in Wollo delicately captured the
official approach:
Peiliaps the fine line between volimtaiy and coerced resettlement is

captured by the notion of "bego teseno", to which party officials and
political cadres carefully and skillfully reverted when voluntariness
failed. "Bego" means goodness or kindness, and "teseno" means
coercion. Hence, "bego teseno" literally means coercion for someone's
own good. That is to say, for those who do not know their own interest
coercion is a legitimate means of helping them realize it.^
The government guidelines for selecting settlers clearly state that a

wiUingness to go is only ooe of tfie oteia to be used. The other criteria
include being a member of the *uiban unemployed," living in an area
designated for conseivation or devdopmem, being in aoeacs on tax
payments, liviqg in a densely-populaled area, beiqg destitnle and having
* John MitcfadU "Review of the Funme Rdief Operadon m WoUo

Administiative Regum," 1986^ mimeo, p. 53.
^ Pankurst, 1990, p. 164.
^ Alemneh Dcjene, Environment, Famine and FolUics in Ethiopia: A View

from the Village, Boulder, Co., 1990, p. 99.
214
eaten reserves of seed, and being a pastoralist. A government survey found
that 38% of settlers belonged to these "non-voluntaiy'* categories.^^
Recruitment of Settlers: Insurgent Areas
In Hgray and noitfaeni WoUo, the means of recniiting setfleis was much
rncxe straightfofwaid: fiofoe was used.
Reliable reports of lesetdement at gunpoint were avaibd>]e Irom early
19^. bi the fbst week of Febmary, 17 track-loads of settiteis were focoed
fipom Meqele camp, and taken to the airport. On February 10, over 200
men were separated from their families and taken from the International
Conmiittee of the Red Cross (ICRC) feeding center of Wahrcb Sharti near
Meqele;^^ in the same week Afar herdsmen were rounded up by soldiers
in Adigrat/"^ On March over 100 were taken at gunpoint at Korem.
10,
The issue came to a head at the end of 1985. In October, a UN food
monitor was travelling together with two nurses from the French relief
agency Medecins Sans Frontieres (MSF) to Kelala in north WoUo. They
encountered a group of about 100 people beiqg escorted to a resetdement
transit camp by two soldiefs, wiio finely admitted tliat they were under
mstractions to shoot any wiio ran away. A
few weeks liter, die food
monitor reported to the UN
Emergency Office for Ethiopia (UNEOE) in
Addis Ababa on this and four other similar incidents. The head of UNEOE
fbfwaided tlie report to President Mengistu, and an official "investigation"
was mounted, which consisted of a guided tour of transit camps in the
environs of Dessie, the regional capital of WoUo. The guides were provided
by the RRC, and at least one Amharic-speaking foreign member of the
team was intimidated by his guide, being threatened with expulsion from
the country if he conducted his own investigations or publicized his
findings. Not surprisingly, the investigative mission reported that nothing
was amiss.
MSF, however, were less content, and in December went pnUic with
an account of tiie forcible resettlenient of 600 people m Korem on one of
duee such occaskmsm October and November. The i^gency also dauned
that the program was causing the deaths of 100^000 people. MSF were
unmediately expelled, and no more dissenting voices were heard among
" Pankurst, 1990, p. 161.
^ The Guardian, London, February 11, 1985.
The Times, London, February 9, 1985.
215
the relief agencies in Ethiopia. Shortly after MSF made their allegations,
an Oxfam Ashwood, confirmed that food had been withheld
nurse, Carol
from famine victims with the aim of forcing them to resettle.^'* The Oxfam
press officer responded to the implicit slur on the organization:
We continue to make representations to the Ethiopian authorities if we

have evidence tet te scale or speed of sesettleiiieiit or the metiiods
employed involve coerooB or are dinplive of haiveiting or feediqg
]MX)grammes. Based on oar ezperioBoes last jfcar, we have so fBifbiiad
this approach more efifecdve than liigb*iHofile pnbfic demmdations."
In late 1985, trucks belonging to the Save the Children Fund (SCF-UIQ
were forcibly commandeered to transport resettlers. SCF protested privately
to the RRC, but made no public statement, even when the trucks were taken
on a second occasion. Both agencies were anxious not to endapger their
ongoing relief programs.
At the end of 1985 the US-based human rights group Cultural Survival
released a report oo the resettlement program, based on interviews with
refugees in Swian who had escaped nom the resetttemeat CKD|^ TUs
report, based oo 250 interviews with refugees who liad escmed the
resettlement camps, repeated the aUâtioiis and provided many aoditioaal
details of human qghts violations.
The US government was outspokenly critical of the resettlement program,
and repeatedly criticized it for being forcible —
notably in a Presidential
Determination of September 1985. No US assistance was ever given to
the program. Apologists for the government accused the refusal of the
US and UK governments to support the program of being "a spiteful and
misdirected error.""
However, the UNconsistently played down the controversy. Over the
previous eight mondis the head of UNEOE, Mr Knt Jansson, had rep^^
asked the govenunent for exphmaticMs of reports of foicible resetdemeot.
^ "Oxfam Nurse says Ethiopia 'Blackmails' Refugees," African Business,

January 1986.
'K>xto OulUnes ill AppRMdi fa Eduopii.'^^^
Cultural Survival, Polilicf and lAe£UU(VlteF(^^

Mass., December 198S.
" Graham Hancock, Ethiopia: The Challenge Hunger, London, 1985, p.

110. Mr Hancock has since revised his position.
216
He never received a satisfactory explanation,"
later reported that "I but
neither did he publicly speak of any doubts he may have had. This silent
endorsement edged towards an open advocacy when the controversy re-
emerged. Mr Jansson related the problem specifically to the single incident
of 600 people being forced onto trucks at Korem. Speaking to journalists,
he "stressed that it was not certain that all 600 were moved against their
will," and on the basis of this and other similar arguments urged relief
agencies to support the resettlement program to avert widespread
suffering/^
In Januaiy 1986, the program was su^nded to neady two yeais,
tlioogii it is widely agreed tiiat tiie main reasons to this were the huge
expcDse and low returns, not international pressure. When resettlement
was restarted, similar instances of coercion recurred almost at once. On
January 1988, seven people were killed while "resisting" resettlement
3,
at Korem —
by trying to run away. On February 8, 20 were shot dead,
while 3,000 were forcibly resettled. The information reached the BBC,
based on eyewitness accounts provided by medical and nutritional staff
belonging to foreign voluntary agencies. The government instructed the
agencies to deny the existence of the incident, and a representative of SCF
was clearly discomfited when questioned about it by Uie British House of
Commons Foreign Affoiis Committee:
I diink we will never know. I Hunk it would be hard to find out what
really happened in Korem. ... The Ethiopian government has ofHcially
denied that the resettlement is forcible. I think their official eiq>]anation,
concerning the claims made by the BBC about Korem, is to refer to
it as a problem of disinformation brought about by foreign agencies
acting in conceit with the BBC

It obviously puts us into a very difficult
position ...^
In fact, there is a considerable amount of independent evidence that

corroborates the account originally given by the relief workers in Korem.
Hiis uiddent win be ciammcd In some detafl becanae fbe evidenoe has
Kurt Jansson, Michael Harris and Angela Penrose, The Ethiopian famine,
London, 1987, p. 67.
^ The QiumUan, London, November 7, 1985.
^ Minnies of Evidence taken before fûdm Atbaa Committee^ Febniaiy

17, 1988.
217
not been published before, and because it illustrates that nearly three years
after the initial accounts of forced resettlement, pressure brought to bear
on the government through '^private representations" had absolutely no
effect.
The personal diary of Eyob Goitam Naizghi, a visitor to Seqota in June
1988, gives an independent account of the background to those events:
By the "tella house" (local barX one man is attempting to explain his
experience of fiBpflyaepaialicM with tncBgnatiiiB art Rwasaome
time in December of 1987, when people in his village started a rumour
about food distribution in Korem by the "Commission". After a few
days of deep thinking, he said, a group of them decided to pack and
walk it to Korem with their families. After two days, approaching
Korem, they decided and the men went
to leave their families behind,
to Korem what they heard was real or a trap. After
to investigate if
having met the authorities of the "Commission", they were told to
register with their families and were given food-grain that will last them
for about two days. Thinking of it all, they were not able to sense any
trap for resetdement. Thus, they decided to bih^g tfaek funflies inside
Korem.
After two days, someone was mcyving quietly and telling people to sell
their pack animals. They cannot tell who he belongs to, and they
thought it was all a mad joke. Two days later, early morning, they were
rounded up by the army and loaded into brand new "Red Cross"^^
donated trucks, on their way to the unknown places. It was only then
they recognized it was a trap. Down-playing whatever guilt feelings
he may have felt deep inside, he told us, the uninvited audience,
boastfully how he sneaked out of the truck and managed to escape to
his village, leaving his wife with three diildm and a donkey behind.
What a loss, lie says to the donkey, because fliat was flte only tangible
property he ever owned. As to hb femily, he only hopes tliat they are
doing fine, for he has heard nothing since tliey sepanted.
In the middle of all this sadness, two Russian-made MiG fighter planes
unexpectedly roar the skies of Sekota town. Every single of us in the
town are in panic. Even those with shaky leg$ are atteoipting to run
away to the unknown ...
Rural Ethiopians commonly call any non-govemmental xeiief agency the

"Red Qoss."
218
In January 1989, a visitor to TPLF-controlled Tigray interviewed several
groups of escapees from the resettlement site of Pawe in Gojjam who were
returning home. These are some extracts from the testimonies:
1. We are returning to our homes in Wollo at a place called Sekota. We

escaped in the moonlight. We weie taken last year during February -
- about 11 mooHis ago. Hme was an amioancement from the
government that everyone from Sekota shouki come to get cards to
receive food ratfons. ... When we airived there we were endicled by
armed forces who beat and killed. Everyone, big and small, with or
without families, were being forced into the truck. Ihose who refused
were hnmediately shot. A lot of people died.
2. The government informed us that, because of the drought, they wanted

us to come to a center to receive food rations and supplies. As we
gathered at the place to receive the rations, we were surrounded by
soldiers on all sides. The military forced us to be loaded on a truck
which took us to the resettlement sites. Those who tried to claim
property or family memben left behind were beaten and shot dead.
We were loaded oo to 40 trodGi altogether the fint day, 30 tmcks the
second, and 60 trucks the third day. ... We were taken last year [1988]
in February. After this month it will be one year ago.
3. When they took us from Korem, a lot of people died, shot dead. Five
people were shot dead beside me by Kalashnikovs from the armed forces
who forced us to be loaded onto the truck. Some lost their legs. There
were tremendous beatings. We were surrounded in Korem as we went
to the marketplace. Most of us left our families and properties and
wives at home. We even left the donkey we had brought to market
for loading what we could purchase. Two people jumped off the truck
and died immediately.
4. We oxne from Gojjam Metekd resettlement area. Wewere forced by

the Dergue from Korem to go there last year. When we went to collect
ration cards off the Red Cross we were forced into armed trucks by
the military and sent to Metekel.... Many people were killed by gun
shots when we were forced away. Ten people died by jumping off the
truck. 30 more people were wounded. 1 personally saw 20 people die.
Six months later, another foreign visitor to Tigray met a group of 34

escapees from Village Settlement Area No. 102 of Metekel, Gojjam. Their
origiiial home was Andoik near Seqota, and he recounted their story:
219
They were starving due to drought, and in January 1988 were called
by the Relief and Rehabilitation Commission to go to the feeding centre
at Korem where they would be fed. They were part of a group of 1200
people from their village area who went into Korem on 8th February
'88, they were not issued with food but ordered instead to get into lorries
for a journey to resettlement. They refused, and in consequence some

20 people woe shot dead and some 30 womided, ate
heided Into the kxriei. In tliii fKOoeiB fimiilies wore sepaiaM; in this
group a man and wife are still with one son, but ihek two otber sons
have never been seen since diey left Korem.
The policy of forcibly taking people for resettlement was not only a
violation of their basic human rights, but also acted as a powerful deterrent
to rural people visitii^ towns, thus disrupting trade, migratioii and the
collection of relief.
Conditions in Transit and on Arrival
Despite the preferential taigetu^ of relief fbod to transit camps, diey

renuuned places where many died, eqwdally hi Hgrey. Figures for
Ambassel in Wollo indicate that 140 per thousand of those selected for
resettlement died in transit camps ahme" —
foftonately, a figure that
is unlikely to be representative. The use of unpressurized airplanes to fly
settlers south, at least during the first year, also led to many deaths.
Most of the resettlement sites were hastily chosen. The site of Metekel
in Gojjam was picked by President Mengistu during a helicopter tour, and
proved to be highly unsuitable.^ Most sites were chosen within three
weeks of the launch of the program, without the benefit of prior agronomic
or hydrological surveys. The setdeis were unused to locu conditions and
fell prey to nnftuniluur diseases; they were also mftHnilfaff witti the local
famdng conditions. Most of the "conventional" resetttemeat sites were
run as cooperative farms, using mechanized plowing, and with settlers
earning work points for their activities. Cadres of the WPE were able to
implement their dreams of socialist collectivization. By 1988, many sites,
instead of producing the expected surpluses to support the government.
" Cultural Survival, Ethiopia: More Ugfu om ReaetOmem, London, 1991,

p. 6.
^ This settlement project, perhaps the most disastrous in the country in terms
of loss of human life, was generously supported by the Italian government aid
program.
220
wm still far fitom self-sufifideDt, and became ledpieiitB of funine relief.
Hie gpvenunent was foroed 1o introduce major policy changes, such as

a return to the use of oxeo on individual smallholdings. In "integrated"
settlements, local people were compelled to integrate the settlers into their
existing villages, apportioning them land, and in these cases the settlers
formed as smallholders using traditional technology.
The appalling conditions endured by the settlers have been well-
documented elsewhere, and will not be further detailed here.^
The Counter-Insurgency Function of Resettlement Sites
The resettlement sites of Hareya and MeDoi Oda hi Bale have already
been mentioned as part of the connter-insurgency strategy adopted that m
area (chapter 5). Resettlement sites m
Woll^ from 1979 onwards were
also used m
a similar way. in 1983, a development expert noted:
Settlement schemes and state farms as centres of government presence

in areas may also have a role to play in controlling the rural population.
These may be used as listening posts and military bases should the need
arise. Where the settlers and farm workers are not locals, as is often
the case, they may help break up the ethnic homogeneity of an area
and provide a force loyal to the government. From this point of view,
these projects can be seen as modem day katamas [neftenya garrisons]
to watch over the local populations, whfle a military significance may
be assigned to iht roads buflt^
The concentration of resettlement sites in western Gojjam (where the

Ethiopian People's Revolutionary Party (EPRP) was active) and westem
Wollega and lUubabor (where the OLF was active) is no coincidence. Not
only were large "conventional" settlements established, but "integrated"
settlements were also set up, in which the settler population was mixed
in with the locals. The counter-insurgency components of this policy will
be further examined in chapter 18.
^ Survival International, Ethiopia's Bitter Medicine: Settling for Disaster

London, 1986; Survival International, 1991; Cultural Survival, Politics and the
Ethiopian famine 1984-1985, Cambridge, Mass., 1985, Peter NigglLu Ethiopia:
DeporUUkms aid Farted Labow Camps, Balm, 1986.
^ Adrian Wood, "Rural Development and National bitegiation in Etfaioina,'*

J^rican Affars, (1983) p. 53Z
221
Though counter-insurgency considerations were certainly important
in the planning and implementation of the program, once the government
had decided to proceed, the program generated a momentum of its own.
It has already been shown how the need to fill quotas resulted in more
resettlers being taken from the accessible areas of central and southern
Wollo, and fewer from the insurgent areas of north Wollo and Tigray.
Similarly, when fesettlemeiit sites were chosen, the ioitial setocdoa was
done extremely rapidly, and many of the sites decided upoo tnmed out
to be unsuitable for habitation. For the same reasons* the sites chosen —
at least during 1984/5 — were probably less flian ideal te counter-
insurgency purposes. The implementation of the program —
later officially
described as "hasty" — also led to local sapport for the OLF and £PRP.
£scape
Large numbers of settlers escaped or tried to escape. Even official

figures give estimates for population loss that range above 20%.^*^ One
estimate of the returnees to Wollo alone is 75,000; settlers from Tigray
were reported to have a much higher propensity to escape. Tbose who
returned home often represented just a fraction ctf iie attempted escapees:
the dangers of the trip included punishment (indoding summary execution)
by cadres, interception by the army, harassment and enslavement by
members of the Sudan People's Liberation Army (SPLA) and other Sudanese
flooded rivers, and disease
militias, banditry, the attentions of wild animals,
and hunger. In one reported incident in December 1984, 330 escaping
settlers were burned to death when government soldiers set brush fires
around their camp.^
A significant abuse against escapees was enslavement by soldiers of
SPLA. The SPLA, which has been fighting against the government of
Sudan since 1983, enjoyed close Imks with the Mengistu government (see
chapter 18).
SPLA soldieis frequently captured escapees, and subjected them to foiced
labor or concubmage. This amounts to enslavement. Most of the victuns
^ G. Sivini, "Famine and the Resettlement Program in Ethiopia," îrka

(Roma), 41^ (1986) p. 231.
^ Aiemneh Dejene, 1990, p. 98.
^ Gayle Smith, "Report on New Refugee Airivals to Blue Nile Province, Sudan,
January 13-14, 1985," Damazin, Sudan.
222
were Tigrayan from the fesetflemeat sites and heading
lesettleis escaping
tot the refugee camp ed Damazin in Sudan.
at
In Febcnary 1986, Sandia Steiqgiaber found 52 Hgrayan lesettleis who
had just anived in ed Damazin, after beiqg held by the SPLA. 30 were
adult women.^^ Tigrayans in the camp knew of almost 1,000 others still
held by the SPLA. The TPLF was unable to negotiate their release until
over one year later, and there were a number of subsequent stories of small
numbers of escapees being captured and held for varying lengths of time.
There was a remarkable absence of women among the southern Sudanese
refugee population: it appears that the SPLA soldiers decided to obtain
replacement women by force. One woman who spent two months in
captivity recounted her ordeal:
In October [1985] my husband and I escaped from that place [Qambela

resettlement camp] with a laige group —
over 1,000 people and —
we fled mtD the forest.
When we reached Sudan, we met people who at first gave us food.

Then they gave us money and led us to the next village. That is when
they took the children away. After that they took us, the women. I
don't know why they took the children —
for workers maybe, for slaves.
These men had uniforms. The people who gave us food did not wear
uniforms.
When was a battle. Our husbands tried to

they tried to take us, there
fight with tfiem, butwas stidcs against guns. Some husbands were
it
loUed, some wounded. The anny was victorious. The remainmg

husbands fled mto the forest. The army took us to their caim> by a
roundabout duection. We
stayed there six nights, then the ^ters
divided us amongthemselves, choosing the most tieautiful there were.
We were sent to the huts of each fighter and always guarded, even when
we went to the toilet. This is because they wanted to mate with us,
they wanted children by us. Other than this, we did not understand
them because we did not speak their language.
During the day we pounded maize for them. There were other wives,
women previously captured.... When we protested, they beat us. I lived
Sandra Stemgraber, m Oay cf a/., 1988, p. 89.

223
in a hut with three fighters. At night they exchanged me among
themselves. This went on for two months.
Like many otbeis, this woraan was pregnaat wilii a cfaOd whoae firifaer
wasamemberoflfaeSPIA. ShewasevealnanyidBaaed^llioiiilifllieooiikl
not lelocate her tausband, and intended to return to the lemainder of her
fBinilywho had beea left behind when she was fotcibiy leietded fram iier
home in Tigray.
The Anyanya 2 para-military force, which was supported by the Sudan
government and the Gaajak Nuer militia, which was aooed by the Ethiopian
government, were responsible for similar abuses.
How many Died?
Resettlement oeitaiiily killed people at a fnter rate than the fgamne.

The mortality rate was particularly high m
the early days of anival at
resettlement sites. Settleis in Keto, Wollega, lefbsed to talk about death:
We did not talk about it. Even the word was avoided. We used to go
round asking, "is there anyone who
has 'slipped away'?" Corpses were
carried off like sacks of maize; they were piled on a trailer and taken
to mass graves, Christians alongside Muslims. Children were placed
between the feet of adults. Grave diggers received extra rations of
food.^^
People apparently became numb to death. They no longer mourned.

Tbey slept, ate aad dnnk ooffiee next to ooipaes. They no longer had
the strength to dig proper ^nves. There were even occasions when
the graves were so shaDow that toes of corpses shick out ...^
In July 1985, Cultural Survival estimated that between 50,000 and

100,000 people may aheady have been dead on account of tiie lesettlement
^ Quoted in Sandra Steingraber in Clay et al., 1988, p. 97.
^ Ftakuist, 1990, p 235.
^ F^mkunt, 1990, p. 248.
224
program. This figure is open to much dispute, but the researchers laid
out their evidence for scrutiny. The UNEOE, anxious to give grand figures
for the deaths attributable to the famine, never produced an estimate for
the human cost of resettlement. The govemment-employed academic, Prof.
Richard Pankhurst, wrote of Cultural Survival's mortality estimate: "the
allegations made in such an unscholarly publication are so one-sided and
so extreme that they can only be accounted for in terms of the selective
use of data to support a preconceived political standpoint."^ Prof.
Pankurst cited not one single piece of independent evidence in his rebuttal,
but Mr Kurt Jansson of UNEOE was conviooed: "this [Cultural Survival]
survey has been convincindy debunked by the eminent Ethiopian scholar,
Dr Richard Pankhurst ..."^
Several investigations have been done into overall levels of mortality.
These can be used to obtain an estimate for the total number killed by the
program. In the following calculations, minimum estimates are consistently
used for deaths. If medium estimates were used, the figures might rise
by over 50 per cent; if maximum estimates were used, they would more
than double. These calculations suggest that the Cultural Survival estimate
was approximately accurate.
It is important to note that the population of resettlers was an abnormal
population —it contained very few diildrai and old people, and was
mosfly adults m
their prime of life (tiiese people were deUbeiatdy chosen).
As a result, the death rate would have been expected to be lower than the
20 per thousand per year that is the "normal" figure; it would probably
have been a maximum of about 17.5 per thousand.'*
RRC figures for recorded deaths during the first year of resettlement
indicate he^teoed death rates: 110 per thousand in Gojjam, 68 ia Ulubabor,
Cultural Survival, 1985, p. 99. The estimate was for total deaths; the
estimate wodced out bdow is for deatiis in excess cftbosc ftax would have been
expected to occur in the funine zone.
^ Richard Pankhurst, "The Ethiopian Famine: Cultural Survival's Report

Assessed," Anthropology Today, Z3, (June 1986) p. 5.
^ Jansson, 1987, p. 26.
^ Derived firom data in: Anwram Kidane, "Demographic Consequences of

the 1984-1985 Ethiopian Famine," Demography, 26, (1989) p. 518; and Central
Statistical Office, Ethiopia, "Report on the Results of the 1981 DcmogFaphic
Survey/ Addis Ababa, 1985, dau &om WoUa
225
42 in Keffa, 38 in WoUega and 34 in Gonder.^ In Jaiso and Keto,
Wollega, the rates were 93 and 51 reflectively.^ These figures do not
include deaths in transit camps, on the journey, or on arrival before the
settlements were fully established and registration of deaths began. Neither
do they include those who died while escaping.
The same RRC data indicate that in Pawe settlement, Gojjam, death
rates in the first four weeks of registration were equivalent to 332 per
thousand per year — almost 20 times normal, falling away over the
following weeks. In Keto, the recorded death rate over the first three
months was equivalent to 122 per thousand.
An investigation was also done mto death rates of newly aniving
resettlers and those who had aheady spent several months in the resettlement
sites — thereby inclnduig deaths mtransit. The sample mdnded people
from both famine-stricken areas (Tigray and north Wollo) and areas which
had escaped the famine (parts of Shewa).^' Only die results relating to
the newly-arriving settlers from Tigray and Wollo were published, due
to political pressure. These indicated a life-expectancy of around six years,
compared to the normal of over 40 for the area. This level was possibly
the lowest ever recorded in a scientific demographic survey, and for com-
parison was seven times worse than the mortality due to the 1972-3 famine
in Bangladesh. The crude death rate was 123 per thousand, which, allowing
for the over-representation of young adults in the population, is probably
equivalent to a level of 150-175 in a normal population.^
The author was obliged to blame the mortality rale on tfie famine.
However, closer examination of the data indicates that the deadi rate
recorded among the settlers ah-eady resident in the resettlement sites was
almost equal (115) and that the rates were similar for those from both
famine-zones and non-famine zones. The implication is that, instead of
blaming the death rates on the famine, it is more logical to blame them
on the resettlement program.
^ Sivini, 1986, p. 232.
^ Calculated from: Alula Pankhurst and Ezekiel Gebissa, "Report on a Study

Tour of Settlement Schemes in WoUega," Addis Ababa» 1987, pp. 34, 58.
Asmerom Kidane, 1989, pp. 515-22.
^ The figure is also an underestimate, because the settlers were asked about
deaths of family members. As many femcdlies weie qdit up, many deaths w«Nild
have occuned without the knowledge of other members of the family. Abo, whole
families could have died, leaving no-one to report on the deaths.
226
These findings indicate that death rates during the resettlement program
were — at a minimum estimate — in the order of 100-115 per thousand,
which is about six times normal for that population. In the famine-stricken
areas, death rates were raised by about three-and-a-half times. About
half the 1984-5 settlers came from such areas: this implies about 14,000
deaths over those attributable to the famine. One defender of the concept
of lesettlment drily noted that the program was "mvolviog human costs
higher than tiiose caused by the femine."^^
The ofhet half of the rested population was not suffering raised death
31,000 excess deaths oocniied among this
rates before resettlement: about
group.
Deaths during escape must also be included. At least 100,000 settlers
from Tigray and Wollo returned home. Interviews among refugees in Sudan
indicate a death rate of at least 20% among escapees. This figure may
be too high: assuming that only a minority escaped through Sudan, and
that the death rates among those travelling inside Ethiopia were much lower,
a minimum figure of 5,000 deaths during escape can be guessed at.
Thus, very roughly, a minimum of about 50,000 people were killed by
the resettlement program.
Resettlement and Famine Relief
In addition to the direct human cost of the resettlement program, it

involved enormous indirect human costs, by the diversion of resources.
Resettlement sites and transit camps received priority allocations of relief
food from the RRC. A
food monitor commented: "Because supplies of
RRC grain were insufficient, those beneficiaries registered for RRC distribu-
tion who had decided to stay in Wollo received no ration."''^ Voluntary
agencies opposed this policy, at least implicitly, which led to a climate
of suspicion between them and the RRC.
Between the conclusion of the first round of resettlement in May 1985,
and the resumption of the program in October, huge stocks of undistributed
grain buflt up around tie temporarily unused transit caAas, wfafle people
went faui^ry nearby. One agency rqxHt read:
^ Sivini, 1986, p. 235.
^ Sandra Steingnber, "Resettlement in 1985-1986: Ecological Excuses and

Enviroomcntai Conaequenoes,'' in Clay et oL, 1988, p. 47.
* Mitchell, 1986, p, 48.
227
L
There is a strong feeling of a serious scandal. The huge stocks are
unjustifiable ... The pattern that emerges is Huge quantities
quite clear.
of grain have gone to the resettlement sites or are being held in stock.
Wollo continues to suffer ... The general view is that the government
is not interested in Wollo. It is either appalling neglect or deliberate
mistreatment.^
Between June and September 1985, the RRC oonsinied an avenge of

8»630 metric tonnes (MT) of gmin to Wollo OpopolatKm: ftnee mimoo)
each month. The estimated need was 35,000 KfTper moBllL Addis Ababa
(population 1.4 million) leoeived 12,000 KCT per mouth over the same
period. As though this were generous, however, the amount allocated in
the following four months —
when stage two of the resettlement pi0gmm
was under way —
was a mere 1,535 per month.^MT
The suspicion that relief programs were hindered in Wollo in order to
facilitate resettlement is confirmed by a number of incidents in which the
RRC intervened to prevent voluntary agencies from distributing food while
recruitment of resettlers was proceeding nearby. "For example, party
officials foibade two EECMY [Ethiopian Evaqgdical Onudi-lidDme
Yesus] centres to distribute general rations when reaetlteiiient was carried
out.... Similariy, the Pbiladelphia Missioa at Knndi was ordered to postpcnie
distribution until resettlement in their woieda [sob-district] had finished.**^

The fear of resettlement prevented many
people from coming to
rural
towns and relief centers for food or other activities such as trade. When
600 people were forcible resettled in Korem in October 1985, 12,000-
14,000 others in Korem abandoned the relief shelter and fled to the nearby
hills. Fear of resettlement was a major reason for Tigrayaos failing to come
forward for relief.
The resettlement sites themselves were favored in relief distributions.
"A sixty-year old settler firom Kdala Woreda in WoUo pointed out, 'I have
received more relief food here in last three yen
tliao I bad to 80 in^
years in Wollo.*"^ Ironically, the setdenrnts needed to receive relief
food every year from 1985 to 1991, even in yean when the setften^ home
^ Quoted in: Paul Vallely, "Staiviqg WoUo: An Empty Excuse," The HmeM^
London, Aogust 14, 1985.
^ Mitchell, 1986, p. 56.
^ Mitchell, 1986, p. 51.
^ Quoted in: Alemneh Dejene, 1990, p. 102.
228
areas in the north were self-sufficient. The government spent at least $120
million overall on the program in its first four years.
Without the resettlement program, the relief program in Wollo, Tigray
and north Shewa could have been implemented much more effectively,
and an impotank hindrance to normal and esMOtial activities sadi as
migration and trade would have been removed An unknown and
unknowaUe number of people died unnecessarily as a result.
Resettlement and the EnTironmcnt
for the resettlement program was perhaps

The environmental justification
the most persuasive —
the land of the north was so degraded that it was
necessary to remove a significant proportion of the population in order for
forestry projects and other programs of land reclamation to be implemented.
However, there is good evidence that the program had the reverse effect.
The environmental impact of the program in the resettlement areas had
been disastrous. Large areas of forest have been cleared, often in an
mdiscriminate fuhion. Indigenous people have been displaced, and forced
to settle ebewliere. A
committee appointed by the Council of Ministers
reported m
1988 that
Unless concerted efforts were taken to arrest the accelerating rate of

deforestation and soil erosion [in the resettlement areas], there would
be a major imbalance in the ecosystem within eight years. The
magnitude of this imbalance and degradation would be similar to that
of the famine-affected areas of the northem highlands of Ethiopia.^
The resettlement program has also contributed to degradation in Wollo.

This is because the fear of resettlement is one factor that has contributed
tomany fumers lieuig foioed to behave m a purely short-term manner.
More generally, a farmer who invests labor and resources on his land, for
instance by buflding tenaces or planting trees, may not remain to see the
fruitsof his investment because be is arbitrarily plucked from his village
and sent hundreds of miles away. Woiie» his activities may even make
resettlement more likely, because they may arouse the envy or dislike of
the PA conunittee members, so that they select him for resettlement. Five
* Quoted in: Alemneh Dejene, 1990, pp. 105-7.
229
per cent of Alemneh Dejene's sample of fanners in Wollo cited fear of
resettlement as a reason why they failed to plant trees/'
More generally, while all land and natural resources remain state-owned
and subject to arbitrary disposal by an unchallengeable local authority,
conservation initiatives will be discouraged. This is illustrated by the case
of Abaselama in Wollo:
The landmaik of this peasant aaaociatifMi is a gully (caused by seveie

downhill floodiog), which begms at the top of Hie Ul and extends te
several kilomelen along the road ontfl it reaches the plain.... In its
twisted journey downward, came within 20 meters of the hoose
this gully
of one of the farmers interviewed for this study. Hie farmer took op
the formidable task of planting eucalyptus seedlings in 1979, adding
manure, building fences around the seedlings, and watering during the
dry season. The fruits of his labor became evident as the seedlings stood
taller than most of the huts around the area. His success became a
subject of conversation in the village. Either out of envy or a plot
organized by members or leaders of the peasant association, he was
accused in July 1987 of planting tiees on pathways bdonging to tiw
peasant assodatioo. The judidaiy cnmmittcc of the peasant assodatiop
revoked his ownership of the trees, instnictii^ him not to plant other
seedlings around the gully."
This story ended happily with the researcher intervening with the
regional Ministry of Agriculture to restore the enterprising farmer's right
to his trees. But there are innumerable examples of f)easants being
discouraged from conservation initiatives for similar reasons. One of the
most ironic is that those who are farming an area designated as a
conservation area are liable to arbitrary resettlement, and the self-organized
plantuig of tiees and protection of soil by fsmien is one fMlor ^4iic^
it more likely that the government will designate a place as a oonseivalian
area.
^ Alemneh Dejene, 1990, p. 40.
* Alenmeh Dejene, 1990, p. 43.
230
13. VILLAGIZATION, 1984-90
In late 1984, the Ethiopian govenuneot began a program of villagization»
wfaidi was intended to ngfmap the scattered homesteads, small hamlets
and tiadidbnal viUag^ of the entiie ooimtiyside into aooBqdetdy new
pattern (tf grid-plan villages, laid oot in accordance with ceiitrd directives.
The aim was ostensibly to promote social and economic development and
facilitate the delivery of services such as education and water supplies.
According to President Mei^gistu:
Collecting the farmers into villages will enable them to promote social
production in a short time. It will also change a farmer's life, his
thinking, and will therefore open a new chapter in the establishment
of a modem society in the rural areas and help bring about socialism/
In fact, like so many government policies, it had important counter-

insurgency elements too. The experience of villa^zatiao in Eritrea since
1966 and Bale during 1979-84 suggests that social transformation may
only have been a secondary objective. This is borne out by the fact that
the nationwide campaign of villagization was started in Harerghe in October
1984 primarily in order to combat the activities of the Oromo Liberation
Front (OLF). More generally, the program was conceived and executed
in a military manner. When a delegation from Ethiopia visited Tanzania
in 1978 to assess the results of that country's experiment with villagization,
the members concluded that Tanzania's failure could be attributed to a lack
of resolve — i.e. force.
Many of the aspects of the villagization program have been dealt with
elsewhere,^ hence this chapter will give no more than a summary of its
in^Murt The policy was abandoned in March 1990.
^
President Mengistu Haiie Mariam, Report to the Central Committee of the
Workers* Party of Ettiiopia, Afml 14, 1986.
*E.g. Survival International, For their Own Good ... Ethiopia's Villagization
Programme^ London, 1988; Jason Clay, Sandra Steingraber and Peter Niggli, The
Spoils of Famine: Ethiopian Famine Policy and Peasant Agriculture, Cambridge,
Mass., 1988; John M. Cohen and Nils-Ivar Isaksson, "Villagization in Ethiopia's
Arsi Region/ Journal of Modern African Studies, 25, (1987) pp. 435-64.
231
Connter-InsiirgeBQr aad ViMagiuittoa ia Hararghe
The government counter-insurgency measures of 1979-83 succeeded

OLE activities in much of the southeast la 1984, the Sidama
in restricting
Liberation Front was beaten decisively.
In the highlands of Harerghe, the OLF continued to be active in 1984.
The government response included heavy bombing of areas such as Maya
Qolo and Maya Guella, dose to Hater town, aad alarge campaign launched
in June, which lasted until Angpsl. The wont atrocity iqxnted wai
destnictioa of much of the district of Daro BiUiqa in July, wfaoe 1,000
bouses were reported Imiiied and 6,000 cattle confiscated.' Aoooidiqg
to the OLF, 102 peasants were killed, usually in front of the assembled
villagers, for resisting forced relocation. The offensive coincided with the
main planting season and directly contributed to harvest failure. Between
May and October over 50,000 Oromo refugees fled to Somalia, and by
December there were 30,000 displaced people in Dire Dawa. Immediately
after the offensive ended, the government started a large-scale viUagization
program.^
The viUagization in Harerghe was accomplished by force. Village
leaders and Mostem religious teacfaen weie detained fand sometiines
executed) whfle anny units instructed the villagDES to idocale to a new
site. Exemplary punishments were meted out to ohjectois, indndigg mass
public executiooSb Existing villages were burned, crops were often burned
too, and cattle were stolen or lolled. People were buried alive, and snivivois
raped, beaten or mutilated.^
In late November and December 1984, the air force bombarded several
areas of the Gobelle valley, southwest of Harer town. In January 1985
there was a military offensive through the same area, in which there were
several reported instances of killings of civilians. An unknown number
of men were shot and killed on January 29, at Gonda Abbadh, while
wollung asnuumallaboieisonagoveimnent project. Ttae was a second
influx of refugees into Somalia.
m
In the new villages, the govern ent (repiesented by cadres and
niflitianien)enfoioed a strict wod[ route. lii8omeplaoes»ancfaa8llabro
Âfhea Omiempormy Record^ 1984-^ pi B244

^ Jason W. Qay, nheCaRof HiiaigliKTcriimanksof Rd^^
in day ef a/., 1988.
^ Numerous instances are leported in Gay, 1988 and Survival International,

1988.
232
district, several clays' unpaid woric per week on state co^ faims was also
enforced.
The program included strict control of food supplies. This met two
government aims simultaneously: it could obtain higher quotas of crops
from the farmers, and it could deprive the insurgents of access to food.
It was combined with fiercely-imposed restrictions on trade.
A particularly irksome part of the whole process is the central control

over food supplies. Locals cannot eat what they want, and have
rqwrtedly been told not to give their children mfllc cows are state
property. Officials announced that the government would distribute
500 grams of rations daily to everybody. For fumiqg families which
have traditionally produced huge food sorpluses and mamtained a very
varied diet, such control is anathema, especially since local people
feared, justifiably, that the army and urbanites would cream off the
highest quality food, leaving the villagers with a meagre diet of sub-
standard grains...
Freedom of movement is greatly restricted in the villagized areas.

People are forbidden to travel through the countryside. They could be
shot if they do. The new villages aie connected to each other and to
cities by a netwoik of usually new roads^ built with fbned labour.
Troops can thus be deployed rapidly, the mflitia ocganizes security
locally, making sure its members are recruited from another area —
in contrast to the past, when the militia was locally-recruited. The
absence of social ties between the new militia and the locals probably
accounts for what some refugees have described as the ^stonatic rape
of women.'
In addition, the villagization disrupted food production in the year of

its implementation.
Village sites were selected with a view to defense, rather than access
to water, fhehwood, pastures, or fidds. Many finmen hid to walk much
greater distances to get to their fields, and some crops which require special
conditions and attention had to be dMmdoned. lliis was particularly the
case for villages ix^hkh were relocated from highland to lowland areas.
Villagers who previously followed a mix of farming and herding, moving
each year from the village to dry season pastures, were forced to become
Âfrica Confidential, 27.12, June 4, 1986, p. 7. The systematic rape of women

by PA officials, cadres and others was also reported in non-insuigeat areas.
233
wholly sedentary, and give up many of their animals. The cambiiuitiao
of impoverished agriculture, greater extraction of food by the govemment
and army, and the effective ban on local trade and migration, was
instrumental in tmniqg the droqght of 19&4 in the highlands of Hareighe
into a famine.
The down-hill relocation of villages also led to the alienation of grazing
land from pastoralists, including valleys used as drought-retreats. This
was one factor in the creation of famine in the lowlands in 1984/5 and again
in 1987/8, and a cause of inter-communal violence.
By mid-1985, over half of the Uglilaiids of Haieighe had been
villagized. The remainder was relocated m
a four month campaign l)etween
November 1985 and March 1986. In total, 2,115 new villages were
constructed, and more than two million people relocated.
The second phase of vilkigization in Harerghe appears to have been
conducted with less violence and more attention to local agricultural needs.
Some "model villages" were constructed at this stage, with facilities such
as an electricity supply provided. These were used as showcases.
The villagization program was effective, however, in restricting the
military activities of the OLF. The existing splits in the OLF also deepened,
with the Oromo Islamic Front gaining ground (see chapter 19).
Social Thinsformatlon and YIDaglzatlon
Following the successful conclusion of the first stage of villagization

program in Harerghe in June 1985, national viUagizatioB was declared as
a program the next month. By August 1988, the govemment reported that
over 12 million people — about half of the rural population in the areas
the government then controlled — had been villagized. While the
collectivization of agriculture was probably one of the ultimate goals of
the program, fewer than four per cent of the farmers in the country were
members of producers cooperatives by the time the policy was abandoned.
Cultural Survival draws a distinction between villagization implemented
as a counter-insurgency strategy in war areas^ and as a program te aodal
transformation in non-war areas. This dlstinctioo is inqxirtaal, bnt the
difference is one of degree. Tliere has been insurgency in the sontheni
provinces. The government was also waging a form of economic waifiue
against the peasantry, partly in order to exact higher kvels of pajyment,
and partly because it feared the sort of spontaneous rural uprisuig that
occurred in 1974/5, when there was no effective govemment control over
the rural areas at all.
The villagization program in most of the south was implemented with

relatively little violence. But there is no doubt that it was involuntary.
234
Cohen and kaksson, who recount the impleiiieatation of the pcogram in
Aisi in a manner sympathetic to the govemment, note:
Indirectly there was psychological force. The experience of Arsi's

inhabitants since Menelik's conquest [in the 1880s] has been that unless
the writ of central government is followed, then the army and/or the
police will enforce
it. Knowing this, Arsi's peasants dismantled their
moved them to new sites, and reconstructed them on assigned
houses,
oompoonds with few overt agns of lesistanoe^
The Aisi peasants' pcoxunity to Bale and Hueighe, and their biowledge
of the viUagization prognm there, imdoobtedly lemfoi^
m
to comply with official commands. Hie govem enfi piqiMiedness to
withhold food aid fram non-villagized areas abo put pressue oo viUageis
to comply with the program.
The implementation plan for each province was devised locally. While
this decentralization could have avoided many of the errors associated with
central planning, it also created a climate of competition for correctness
and zeal between the cadres of different provinces; in 1986 the most
successful officials were rewarded with prizes and promotions.
The creation of new villages involved a nnmber ctf measures which led
to unnecessary hardship and hunger. Government officials and cadres
surveyed sites and insisted on hoiise ooosliiiction m
the mi^^
crops. Labor was diverted firom essential household and agricnltoiai tasks.
Houses of reluctant farmers were arbitrarily demolished. Levies were
exacted from the peasants in order to finance the propam. The villagers
were moved before essential ancillary buildings had been constructed, such
as latrines, kitchens and stables. Inadequate space for housing animala
and long distances to pastures led to enforced sale of livestock.
In some
areas of south-central Ethiopia, the staple crop is the root of
the ensete (false banana) plant. This is a perennial plant which is
traditionally grown around the homestead. Several agricultural experts
reconmiended that villagization be deferred or canceled in eê/e-growing
areas, because the forced relocatioo would require the abandonment of
ezisthig ensele trees and it would be several years befone new trees would
achieve maturity. This advice was ignored, and villngjaartion proceeded
m
apace (though in some parts of southern Shewa, a compro ise was reached
whereby the existing dense pattern of settlement was merely rearranged).
Some of the hunger that afflicts these fertile areas in 1991 can be ascribed
^ Cohen and Isaksson, 1987, p. 452.
235
L Copyrighted material
to this policy. Similar problems affected villagization in coffee-giowii^
areas and places where the mild narcotic leaf chat is grown.
A number of villages had to be relocated after their sites were selected
by urban cadres in places without water, or where there was no drainage
so that flooding was prevalent.
ResistaiiM
TTie villagizatioii program was deqdy im|xipiikff Ifaoiiglioii

Most rural resistance to villagization and its accompanying programa
took the foim of unobtrusive sabotage. Peasants would do the minimum
possibleamount of corvee labor* on roads, schools and forestry schemes,
and produce shoddy work. There are accounts ^^^Hf^J^ planting saplii^
upside-down, certainly not from ignorance.
In some areas there was violent resistance. A notable example is
Gojjam, where there is a long history of resisting central attempts to meddle

in affairs connected with land (see chapter 3). Villagization in Mota district
was achieved in IS^ only wiA te assistaace of helicojpter gunships sent
from Addis Ababa. Ttee are other reports of villagers kuliqg or amrilaHng
cadres sent to enforee villagizrtion, for example, 25 people UDed near
Shashamane in Shewa, one in Opjjam, two in Sidamn, and an miknown
number in Gamu Gofa.'
In March 1990, President Mcngistu unexpectedly announced the
abandonment of the villagization program. Almost overnight, the existing
cooperative farms were physically stripped of their assets. Villagers —
for so long sullen and cowed — suddenly
displayed great energy and
initiative in redividing their farmland and returning to their original
homesteads, or claim to them in anticipation of a future
at least laying
move. In some were expelled torn tbe villages, or ied
areas» caidres
anticipating retributioo. In Mota, Gojjami, some cadres who tried to resist
the dismantling of the viUi^ were filled by te peasanlB.
Forced labor exacted as a tax demand.
' Survival hitemational, 1988, p. 26.
236
14. £RITR£A UNDER SIEGE; 1988-91
On March 17-19, 1988, the EPLF overwhelmed the Ethiopian army's

command at Afabet. Over 15,000 soldiers were killed, wounded,
northern
captured or di^)ersed, together with a vast quantity of aims and ammmiitiQii,
fndndiqg SO tanks. Three Soviet officeis were captured and another only
nanowly escaped. Building on the sigiificant if less qxctacular advances
by EPLF and ITLF over tibe previous six months, this marked a turning
point of the war. Three years and two months later, the EPLF occupied
Asmara and the EPRDF occupied Addis Ababa, bringing an end to ttieir
respective wars with the central government. These last three years of the
war in Eritrea saw no respite from mass abuses of human rights by the
Ethiopian army.
Afabet and Aftermath
m
The EPLF capture of Afabet left the Ethiopian army disarray, ^tfain
a week, the garrison atTessend was evacuated so as to bolster the defenses
of Keren. lUs evacuatiao was carried out in good order. Baientu was
also evacuated, on March 31, and the lelreating units burned several parts
of the town and looted many citizens of their possessions. The same day
the garrison at Anseba, north of Keren, retreated under EPLF fire, and on
the night of April 1-2, the EPLF overran the army trenches at Halhal.
In order to save Keren from apparent imminent capture, Agordat was
evacuated. The following six weeks saw fierce fighting around Keren,
in which the army managed to recover its positions at Halhal; otherwise
the new front lines did not change significantly.
The fighting saw many atrocities against civilians. One was witnessed
by Zahra Ibrahim, an Almeda woman from Halhal.
In ^nil the Deigue attacked Halhal and we were forced to run away.
We ran to Wadaq Sabra where there are some caves. There were many
people hidiog in the caves. The soldiers came to us to kill us. 1 begged
one of them not to kill me and my children — I offered him sugar —
and he left me and they killed only the others. They killed so many,
I couldn't count. It took one month to find the bodies and bury them.
For three days I was wandering in the hills and my little boy died from
hunger and thirst. There was a gas from the shells which made us cry
and made the children very upset — they cried all the time and became
very thirsty.
237
[InMay] we wm[still] in Wadaq Sabia. Tlw HLF ovenaa ana
and killed many ^liiqpian soldiers, and then left. We left with tlie
EPLF —
theie wne so many dead soldiers that we had to step on them,
and the streams wece flowing with blood. The £PLF took us to the
river Matafa/
Other credible reports of killingis in April-May 1988 include:
* April 5, Godeiti: 12 civilians shot dead and two wounded by soldiers

(ten of the victims were aged 60 years or older).
* April 15, Qazien: six civilians shot dead by soldien.
* April 20, Shebah: 18 civilians shot dead by soldien.
On May 12, in Sheib in Semhar district, a large scale and well-

documented massacre of civilians was peq)etrated by the Ethiopian aimy.
Idris Osman Enkersa was a survivor:
Itwas morning around 8 o'clock ... The enemy armored vehicles, 15

tanks, appeared on the Massawa side. They headed towards Sheib from
the coast side. The tanks kd [the aohlien] to SMb and sniroonded
thevillage. The soldien came mto every house and collected tlie people
by saying that "yon have a meeting today." They gathered children,
old women and men under a big tree.^
Amina Mohamed recounted what happened next:
We tried to run away but we were surrounded. The tanks moved in

on us, crushingpeople in their way. My entire family was killed, except
for my baby. For three days I lay in the midst of the bodies, pretending
to be dead. At one point my baby started to cry. A soldier aimed his
gun. I heard his companion say "don't waste a bullet. The baby will
die of hunger anyway." The soldien killed our animals and threw the
carcasses in the welb. They searched womenVi corpses for gold nose
' bteiviewed by Alex de Waal. Wad Sberifin, Sudan. Match 1989.
^ Interviewed by Jennie Street.
238
riqgs and earrings. After three days they kit the villiige and I walked
up here. Hiat's all.^
About 80 people were crushed to death by the tanks. A further 320

were killed by gunfire, from both the tanks and the foot soldiers. All were
civilians. Idris knew many of the dead:
Mohamed Shibeley Dery's wife with her son, his sister in law with her
two sons; All Gira Wad H8mid*s wife and his son, his soifs wife with
her three children, his dangliter with her chiMien;
Ibrahim Hamid Shihele^ Osman Hamid; Haadd Mohamed Gheiway
and his mother; Hamid Kunib witii his wife and his daughter; Hamid
Ahmed's wife with her two daughters; Suleiman Ali Oidir with his sister
and his sons; Hawa Osman Musa with her two sons....
On the same day and immediately following days, there were other
killings of civilians at 30 other villages in the vicinity, in which at least
100 others died. For example, three were shot dead by soldiers at Beet
Abreha the same day, and four at Fatna Arre at the end of the same week.
The air force also carried out a number of attacks on villages and other
civilian targets. These induded:
* March 31: Melebso: 15 IdUed, 25 wounded.
* April 2: Mensae Beit Shehaqu: five killed.
* April: repeated attacks on Afabet, casualties not known.
* April: Agordat: at least three civilians killed.
* April: Anseba: two civilians killed.
* May 7 and 8: Mensura: five civilians killed.
* May 3 and 13: Halhal: at least three civilians killed in this and two other
attacks.
* May 19: Afabet: ten civilians killed.
^ Quoted in a fibn by Alter-Cine inc. panicle

LaCourse and Yvan Fairy)
The Forbidden Land, September 1989.
239
These kUlings of dvflians seived no military purpose. Hwy
were
intended mealty to terrorize and {xuiiah tiie popnkdon.
The fighting and government reprisals displaced about 110,000 civilian^
including over 70,000 from the Sheib area. About 40,000 fled to Sudan.
Hiese refugees described how they hid to escape the Ethiopian army and
how it was not safe to inhabit a village during daylight hours. Instead,
people spent the day hiding in the hills, returning home only at dusk. They
could not wear bright clothing for fear of attracting the planes, and had
to hang their washing to dry in the shade of trees. They could not light
a fire to cook, because the smoke would give away their presence. Travel
to Sudan was possible only at night
One woman gave birth under a tree; the next night she had to oosthnie
her journey. She said: "I was Indgr, we had a camd. I know of women
who had to start walldog the same day that they had ^ven Uitfa."*
The fofced displacement of the population led to increased deaths, from
thirst, exposure and disease. There was a severe malaria epidemic in the
fall of 1988, and many of the victims were displaced people who had moved
from the highlands, where there is no malaria and so they had no acquired
immunity. A study of mortality among a population of refugees who
arrived in Sudan during 1988 found that death rates approximately doubled
during the period when the refugees were "on the road," and remained
ccmsiderably higher than normal m the refugee camp, chiefly oo account
of dianhoeal diseases and malaria. Most of tiioae who died were young
children. If the survey is assumed also to be representative te those who
were displaced inside Eritrea, it would imply that about 1,430 people died
on account of the displacement'
The State of Emergency
The defeat at Afabet led President Mengistu to make his first public
admission of the existence of the war for ten years. In a televised speech
on March 31, Mengistu said that the money spent on the war each year
could have built four major universities or ten large hospitals. He declared
that extra effort was needed to meet the threat: "fkom now on, everything
to the batdefnmt" Aweek later, after meetu^ wiA President Siad Bane
of Somalia, Meqgistn niade the surpdre announoenient of a peace i^greement
^ hiteiviewed by Ales de Waal, Wad Sherifei. Sudan, Mndi 1989.
^ Alex de Waal, Topulation and Health

of Erhreaos in Wad ShedH" Loodon,
ActionAid, 1989, pp. 77-87.
240
with Somalia (none had been signed after the 1977/8 war), allowing tiie
redeployment of troops from the Ogaden to Eritrea.
On April 6, the government expelled foreign aid agencies from Eritrea
and Tigray. This included the International Committee of the Red Cross
(ICRC). Only UNICEF was exempted.
This ban drew much adverse publicity from the international media,
which appeared to assume that the absence of foreign personnel was
equivalent to the complete cessation of the relief program. In fact, as the
Ethiopian government was quick to point out, more than 80 per cent of
the relief in govenmient areas was distribnted by Ethiopiaii ofganizatioiis
employing Ethiopian staff (chiefly the RRC and the cfauiches). The
intention of the ban was different: it was to remove witnesses for what
was going to happen next.
On May 14, the government declared a State of Emergency in Eritrea
and Tigray. As there was no pretence of civil administration in Tigray,
and in less than a year no government presence in the province at all save
a single garrison, it had little impact in that province. It is also questionable
whether the State of Emergency had any significant impact in Eritrea, as
the government already possessed an almost unlimited range of powers,
and the legal system was already subject to continuous and authorized
interference by the executive — the country was under a permanent virtual
state of emergency already.
The State of Emergency pcodamation gave an "Overall Admimstrator"
of Eritrea wide-ranging powers, accountable only to the President and the
State CouncO. Hie security forces (mchidmg army, police. People's Guards
and mflitia) were empowered to inspect any person or property and to detain
anyone. Ten kilometer strips along the Sudan border and tfie coast were
designated "prohibited areas," from which all people were required to move
— in effect making these "free fire zones." The Overall Administrator
was empowered to convene military tribunals and to appoint their officers.
These tribunals had jurisdiction over a range of crimes. The full list, as
enumerated in Section V, 19 (i) and (ii) is reproduced here, in order to
give a flavor of the military administration that followed.
(0 Military tribunals ... have jurisdictioo over crimes committed during

the State of Emergency in aocofdaaoe with the updated Special Ptenal
Code Prodamatioo No. 214/1974 [Ethiopian calendar, Le. 1981]:
(a) Crimes committed against the freedom of the country.
(b) Crimes committed against the constitutional order or against

organs of that order.
241
(c) Cdmes oommitted agauist Etliiopia's unity and the unity of the
people.
(d) Anti-ievolutioDaiy crimes.
(e) Crimes of armed terrorism and the waging of civil war.
(f) The crime of agitating and organizing.
(g) Crimes affecting the countiy's defense foioes.
(h) Crimes against puUic property and wealth.
(ii) In addition to the powers stipulated in sub-article (i) of this article,

the military tribunals have the right to judlge:
(a) Crimes committed in violation of this special proclamation.
(b) Any crimes which the Overall Administrator of the area decides
should be transferred to military tribunals from the ordinary
courts.
Death penalties required the personal qppioval of the Overall

Administrator.
Under the State of Emergency, the army continued to act with wanton
brutality towards the civilian population. Over the following twelve mooths,
some of the incidents m which civilians were killed included;
* June 2, 1988, Godeiti: nine shot dead.
* June 25, Mensae: four shot dead (including a six month-old baby), and
three wounded.
* September 16, Gdeb: four lolled by shellfire.
* October 22, Mai Harasat: eight shot dead, indndiqg three old people
and a five-year old girl, and eleven wounded.
* October 31, Degera: three shot dead.
* December 26, Tewro: eight shot dead.
242
* Febniary 1989, Semhar district: between 600 and 1,000 killed in Slieib
and sonounding villages.
* April 15, Sefea and nearby; 19 men thrown to their deaths over a cliff.
* April 17, Logo: 19 men stoned to death.
* June 3, Una Andom: seven shot dead (21 others were killed in nearby
villages on the same day).
Some of the air raids that took place indnded an attack on Lego on
June 10 (no fatalities reported), and a series of raids in Barka in September
and October. The bordler area was a particnlar target, and at on least three
occasions the MiGs crossed into Sudanese air space and attacked civilian
targets inside Sudan. In October, two Sudanese locust-spraying planes
were attacked, and in November there were two raids on Sudanese border
villages in Red Sea Province, in which a school was damaged.
The policy of forced relocations into protected villages continued. In
October 1988, about 5,000 people in four villages were forced to move
to the protected village of Elabored near Keren. They were given one day's
notice of the move, and the soldiers took the opportunity to loot much
property.
During the 18 months after July 1988, there was little huge-scale
military activity in Eritrea. In early 1990, that was to change.
Maasawa, 1990
On February8, 1990, the EPLF launched a surprise attack on Ethiopia's

second port,Massawa, which it captured after a three-day battle. During
the battle, the government forces retreated to an island some distance from
the main town, taking a number of civilians as hostages. After repeatedly
demanding that this force surrender, the EPLF attacked and defeated it.
About 200 civilians were estimated to have been killed in the fighting,
induding some of the civilian hostages.
After the battle the port remamed essentially intact, and lequued few
repairs before it could funcHoo again. Starting shortfy affeeiwards, ttie
Ethiopian government pursued a policy intraided to reduce Massawa to
rubble.
The first air raids occurred on February 16. For eight days there were
repeated attacks. The initial targets appeared to be the food stores, which
contained about 50,000 tons of US-donated wheat, for famine relief.
Incendiaries and napalm or phosphorous bombs were used, and about half
243
of this food relief was burned. Susan Watldii8» an official of Qxfam
Canada, who visited Massawa at the time, saw two warehouses and three
stockpiles of grain burning, and commented "it was clear that food aid was
the target of the bombardment."^ The grain was burned so systematically
that piles were still smoldering one month later.
After the initial round of bombing, March was much quieter. The key
issue in Massawa was whether the EPLF would be able to reopen the port.
Yemane Yohannes, a senior technician in the port, told a visiti^ journalist
"if a ship aiiives tomonaw, we can handle it."' He laid that loar dieiel-
powered cranes, three berths, four warefaoofles, and tfiree tugs were sdll
fiinctiooal, so that relief aliqpmenls could be unloaded. Reqpoadiqg to an
EPLF appeal for relief shipments, the German-based rdlef oigamzitiaa
Cap Anamur sent a ship loaded with relief towards Maaaawi.
On April 4, the Ethiopian air force began another series of sustained
attacks on Massawa. 30 people were killed and 54 seriously wounded.
In three raids over the following four days, another 41 people died.
The inhabitants of Massawa were compelled to spend the daylight hours
in air raid shelters —
in storm drains, under bridges, and in the cellars
of houses — or to evacuate the town altogether at daybreak, and spread
themselves over large areas under trees, in order not to provide a target
for the bombiqgs.
On April 13, one of the encampments of evacuees, at Foio just outside
the town was bombed. At least 25 civilians who were sheltering diere were
biunned. One victim of these raids described what oocured:
When we got out from Massawa we were under the trees. We are just
civil peoples.The aeroplanes have seen us [that] we are civil peoples.
They came at nine o'clock ... They bombed bombs and napalm bombs.
They have bombarded us for two hours.*
Hiese attacks also used duster bombs. Oustor bombs are a particularly
deadly munition, as they explode before strildM the ground and shower
a large number oi smaller bcnnbs, each one leUuu^ over a Uagjs area. They
are designed to IdU large numbeis of people. Hieir mflitary use is i^gainst
^ Quoted in: The Times, London, Much 3, 1990.
^ Quoted in: The Guardkm, London, March 16, 199a
" Quoted in: Alter-Cine ioc (Danielle Lacourse and Yvan ?a!tiy)AFi^ to
the Death, April 1990.
244
columns of infantry, but at this time Massawa was well behind the front
and empty of military personnel.
line, A
foreign visitor to Massawa wrote:
ERA officials familiar with GoE [government] military tactics are

shocked by a new kind of bomb which has only been used since tfie
Massawa takeover. Everyone in the town is taUdm about this new kind
of bomb, which destroys everything witfiin a 100 metre radius, and
which is parlucnlariy effective in shanty areas widi flimsy housing
structures.
These cluster bombs were ahnost certainly supplied by the Israeli

government.
On the evening of Sunday, April 22, a particularly devastating air raid
occurred. At 10:30 that morning, a raid had taken place. This attack
produced no casualties, because the population was either sheltering or
evacuated. The second attack occurred at 6:15 p.m., at dusk. This was
after the last hour when attacks normally occurred, as the bomber pilots
left themselves enough time to letum to base before nii^tfidL Hieiesidents
of Massawa had tiierefore left their shelters, and were in the streets on their
way to churchy market, or woik place. One woman described how she
was making kicha (unleavened bread) for the feast of Medhanie Alem (The
Savior's Day, a monthly Christiaa religious feast), with her family wh^
the bombs exploded. Another man was sitting with his family at home;
only he and his sister in law survived. One woman, sheltering with her
family in a drain under a road, described her situation:
The plane bombed us. The people are suffering ... entire families were
wiped out — not a single person left alive in the family. We did not
come here [to the drain] to have a good time. We
didn't come here
for iim. We're having a bad time. Mengistu has decided to bum us
like wood.'
Two cluster bombs cxfioded over a crowded street in die center of the
town. About 50 people were killed and 110 wounded, many of them very
severely.
Video recordings taken immediately after the bombing confirm that the
casualtieswere civilians. One video, shot after the raid, contains pictures
considered too horrific to be shown on public television: one shows the
body of a woman with her face entirely burned away, another shows a dead
Quoted in: A Fight to the Death.
245
child with a hole in his face. Other pictures, marginally less horrific,
include a close-up shot of a pile of bodies lying in the street, their flesh
punctured with fragments from cluster bombs, and pictures of survivors
with large areas of skin burned away; in some cases the raw flesh covering
the entire back.
Massawa is a town with clearly separated docks, commercial and
residential quarters, and the town itself lies some distance away from the
military installations. Tbe bombers flew low, it a height of a few hundred
feet. They deliberately targeted the residential areas of the town, and
attacked at an hour calculated to cause the maximum number of casualties
among civilians.
This series of raids also used demolition bombs, and did severe damage
to warehouses and other port installations, as well as destroying more than
100 houses. Parachutes were used to slow down the bombs' descent to
ensure that they detonated moment. By this time, several
at the correct
offensives by the Ethiopian army aimed at retaking Massawa had failed,
and it is probable that the bombing raids were now intended to destroy
the port entirely. A woman resident of Massawa, Fatna Ari, commented
"whenever Mengistu realizes that he is defeated he kills peojde with
aeroplanes."^
On May 1, the Ethiopian government threatened to bomb any ship that
docked in the port, forcing the ship chartered by Gap Anamur witfi its relief
cargo to be diverted to Port Sudan.
On June 3, the question of famine in Ethiopia was raised at the
Washington summit. Under pressure from the super-powers, the Ethiopian
government conceded that Massawa could be used for relief deliveries.
However, before this was officially announced on June 5, another air raid
took place. Raids in the Massawa area continued until June 10. An EPLF
spokesman responded; "the port facilities have been virtually destroyed
by air raids ... I don't think Massawa could operate as a port for many
months."^^ After early June, there were few air attacks on Massawa: one
occurred on Septeml)er 4 and another on October 24, in which one child
was killed. These raids were probably carried out in older to demonstrate
to the aid donors that any relief would be delivered to Massawa on the
government's terms, or not at all.
Quoted in: A Fight to the Death.
Quoted in: The Guardian, London, June 7, 1990.
246
The capture of Massawa by the EPLF also led to the government
unleashing air attacks on other towns and villages in £ritiea. Some of the
attacks included:
* April 3: Afabet: 16 killed, 24 wounded.
* April 4: Aftbet: 51 killed, 125 aerixMisly wounded.
* Apta 17: Afabet: no fatalities.
* May 21: Afabet: two killed.
* September 13: Gedged: two wounded.
* September 17 and 21; Matalili; no casualties, but several fishing boats

destroyed.
* September 22: Haicota: no casualties.
* October 3: Koatit no casualties.
* October 11: Adi Ma*alim: Ifaiee wounded.
* October 14: Tikombia: thiee wounded.
* October 14; Mai Shiro: three killed, one wounded.
* October 16: Godeiti: two children killed.
The Siege of Asmara
The of Massawa led to the government being confined to Asnuun

fall
and the snnounding area. Jko stale of siege lasted until Ae defeat and
surrender of tfie garrison on May 25, 1991. The siege witnessed the
development of famine conditions in the enclave, which will be discussed
in chapter 16. It also saw many abuses ugainst civilians conunitted by
the army and administration.
Flacing Civilians in Danger
One abuse against civilians was a systematic attempt to retain them in

and around the battle zones. This first occurred shortly after the fall of
247
Massawa, when the army made a series of attempts to recapture the port
by attacking from Ghinda, which lies half way between Asmara and
Massawa. All the attempts failed. On March 11-13, the army prevented
the civilian inhabitants from evacuating Ghinda, so that their continued
presence would provide a human shield for the army and deter EPLF
artillery barrages. When it became impossible to live in Ghinda, many
residents had to move to caves and other makesWft ihelten aeuby.
Elsewhere in the enclave, flw army also prevealied dvOiam from leaving
go to Aaniara, or leinoved them 00^
villages near the froot line to
distance. For people displaced from Massawa who tried to
fmrtancfft tiie
travel to Asmara were ccMifined to Nefasit, jnal behind the front line at
Ghinda.
Killings
The army continued to kill civilians on frequent occasions, though no

massacres on the scale of Sheib occurred —almost certainly because the
army was no longer able to penetrate into EPLF-controUed territory.
The worst single jnddeBtoocaiied on June 9»lSm. Agroupafioldleii
left their mflitary base and eniefedtedty. They shot dead 31 yoolhs and
injured at least 15 mofe. The killings took place in several parts of the
city as the youths were returning home from watching a televised World
Cup soccer match at Kidane Mehret in the city center. They were shot
in the street and in the doorways of houses* just before the oufew hour.
Other incidents include;
* August 9: Hagaz, near Keren: two mothers killed.
* Late August: Decamhare: about 20 peasant farmers were summarily

shot by die anny dni^g a mflitary engagement with te EPLF.
* September 12: Asmara: soldiers shot dead 16 dvflians.
* September 27: Keren: two civilians were killed when soldiers opened
fire (they claimed they were shootii^ in the air» celnhratiqg the Ethiopian
New Year).
* January 1991: Tsazega: a girl of 12 was shot by soldiers while they

were allegedly training.
* January 18: Anseba: a giri of 16 was kflled by soUieiB with a knife

whfle lesistutg attempts to lape her.
248
* Febmaiy 7: Adi Gaima: two killed by soldiers.
* February: Girgir, near Keren: two elderly women were killed by soldiers^
one shot and one strangled.
* March 28: Adi Quala: one man was killed by soldiers while returning
from market.
* April 3: Sheikha Wadi Bisserat: one was killed by soldiers.
Deteniions and Restrictions
Another abuse against civilians was detention under the emergency

powers given to the overall administialor. Civilians were arrested and
detained on the slightest suspicion of sympathizing with the EPLF. Some
were detained simply so that officials could obtain bribes for their
release.^^ Some detainees were killed in prison. For example, on April
9, ten were executed in Mariam Ghimbi prison in retaliation for the EPLF
assassination of the prison governor. One of those killed was Tsehay
Gebremedhin, an employee of SEDAO Electric company.
Niimeions lestrklKms weie |daoed 00 dvflians in Asman
towns of the enclave. A cmfew was miposed from 9 p.nL until dawn —
later it was brought forward to sundown. During the hours cf the curfew,
peofde were liable to be shot on sight On several occasions, dtywide
searches were made, as on December 21, 1990, and again on February 21,
1991. All people were stopped in the street and required to produce their
identity cards. Numerous people were detained on both occasions,
especially women who were not carrying their cards, and people who
originated from Tigray.
The army prevented people from leaving Asmara. People who tried
to leave on foot were subject to summary arrest and detention. If they
could, people tried to leave by air. Ethiopian Airlines flights came
irregularly, and stopped altogether after an airplane was hit by a shell while
standing on the tarmac on April 26, 1991. Ihe cost of a ticket plus all
the requisite payments and bribes to obtain travel documents was
prohibitively expensive for most
The administration began to Ofganize the civilian population into civil
patrols. In rural areas»villagess were required to take on security functions^
For more details see: "Ethiopia: Human Rights Crisis as Centnl Power
Crumbles," News from Africa Watch, April 30, 1991.
249
protecting ioads» bridges and hOlsidesby mountiiîcmiid-tl^^ guards.
They were required to report suspicions activity. Hiose who objected were
detained and physically ai>used. In Asmara city, civfl patrols were organized
in February 1991. Each kebele was required to provide a guard for its
neighborhood for four-and-a-half hours each night, and was held
responsible for any suspicious activity during those hours. Qvil patrollers
were given no weapons, torches nor uniforms.
Shelling
The residents of thetown were also at risk from shelling by the EPLF.
Starting in March 1990, the EPLF began regular shelUng aimed at the
airport. The artillery was located at Bizen and Ala, some 30 Idlometers
away to the northeast.
The target of the shelling was the airport, and the apparent intention
was to damage the military installations and aircraft tbere, and put the
runway out of action. The shells damaged military transport planes, MiG
fighter-bombers, a plane belonging to the RRC, airport installations, and
an ammunition dump. Numerous soldiers were killed while leaving through
the airport in April and May 1991. The airport was forced to close on
several occasions on account of the shelling.
The shelling caused civilian casualties at the airport. In early January,
1991, three women were killed while waiting for an Ethiopian Airlines
plane to Addis Ababa. On March 1, an airplane used for tfie relief airlift
was hit whfle on the ground and one civflian was Idlled. An Ethiopian
Airlines fuel tanker airplane was hit by a shell on March 22. On AprU
26, an Ethiopian Airlines passenger plane was hit iyy a shell and three
passengers were killed.
The shelling also caused some civilian deaths in the town. In the first
few days of the attacks, shells landed over a wide area, causing a number
of civilian casualties, both from the blast directly and from fragments hurled
over a large distance. After that, all the shells landed in the airport and
the neighboring two quarters of the town, Godaif and Sembel. The EPLF
warned the residents of these areas to evacuate their homes, but not all
did so, m
part because of tiie fear that their houses would be lequisitioned
by the army. Between March and June 1990 about 60 civilians were lêd
in the town by the shelling and 100 houses were badly danuged.
In late August, the shelling caused its higliest number of civflian
casualties. On August 20, one child was killed in a tank and artillery
bombardment of Akria, Arbate Asmara, Inderase and Gabriel Church
quarters of the city. On August 25, 17 civilians were killed. In the two
incidents, 20 were also injured. Hie government reported the incidents^
2S0
acknowledging foe the first time that the city was besieged by the EPLF."
The EPLF, however, blamed the incidents on the government.
Circumstantial evidence suggests that the government was indeed likely
tobe responsible — certainly, residents of Asmara believed so. This was
based on several considerations. One is that the source of the shelling was
Balazar, to the north of the city; EPLF shelling came from the northeast.
Second, the shells from the EPLF artillery took precisely sixteen seconds
from firing to landing; these shells took less. Thirdly, the shells were cxf
a diffnent Idnd, and made a different sound.
Assassinations by EPLF
Throughout the last years of the war, up until the weeks before the fall
of Asmara, the EPLF continued to assassinate civilians accused of
collaborating with the security forces. On average, the EPLF announced
approximately one such assassination every two weeks in the last months
of 1990 and the start of 1991. Victims included members of the Workers'
Party of Ethiopia, informers for the security forces, and the governor of
the Mariam Ghimbi prison, who was accused of torturing detainees.
According to the EPLF, the offenders were tried in absentia^ and were
warned twice to desist fitom their activities. If they refused to heed the
warnings, an assassination squad was dispatched. Tliis practice was
criticized by Africa Watch.^^
Eritrean Opposition to the EPLF

Throughout the 1980s the EPLF was the militarily dominant Eritrean
military front. However, it faced competition from no fewer than nine rival
fronts or political organizations. The Eritrean Liberation Front-Popular
Liberation Forces (ELF-PLF), headed by the veteran nationalist Osman
Saleh Sabbe, split from the EPLF in 1975, and itself split in 1979, with
the formation of the ELF-PLF-Revolutionary Council, headed by Osman
Ajib. After Ajib's assassination, leadership was taken over by Abdel Gadir
Jeilani, who is a Baatiiist who enjoyed patronage firom Inq. The ELF also
split in 1977, with the fofmation of the Eritrean Democratic Movement
(EDM) headed by Hiruy Tedla Bairu. In 1981 the ELF split into three
factions, led by Abdalla Idiis, Habte Tesfiamariam and Gioigis Te^^
î4FPAugust29, 1990.
News from Africa Watch, "Ethiopia: Human Rights Orisis as Central Power
CSrambies," ApiU 30, 1991.
251
Pterscmtl rivalries wm «t least as important ai Ideological dilutes in
cieatiiig the fsctiooalisni. By 1991 the HJ' oiganizations coiwirted of:
* ELF-Revolutionaiy Gomicil: headed by Ahmad Nasir» secular in

orientation.
* ELF-Umted Organization: headed by Omer Buraj, Islamic in orientation.
* ELF: headed by Abdalla Idris, Islamic in orientation, with close ties

to Sudanese security (commonly known as the "ELF-Abdalla").
* ELF-PLF: headed by Abdd Oadir Jdlani, Baattlst
* Eritrean Islamic Liberation Front (^Mu}ahideen^\headed by Ibrahim

All, and supported by the
Islamic fundamentalist in orientation,
mtemational Moslem Biothen and the Sudanese Natiooal Islamic Front
From late 1986 until about 1989, the ELF-Abdalla was involved in
negotiations with the Mengistu government. According to the EPLF and
rival ELF groups, the government provided military supplies. The ELF
groups received support and sanctuary from the Sudan government, which
also contimied to support the EPLF.
In addition, there are Afiur groups m
finvor off a united Afar teiritory
and two non-oombatant Eritrean oiganizatiaos» both secular m
orientation
and headed by Christians:
* Sagm ("Return"): beaded by Tewolde Gebre Sellaasie.
* Democratic Front for the Liberation of Eritrea; headed by Gebre Berhan

Zere.
Militarily, the most important of these organizations has been the ELF-
Abdalla. Operating from bases around Kassala in Sudan, the ELF-Abdalla
was mvolved m guenilla attacks on the EPLF in southern Barka. The main
activity was the planting of land mines (both anti-vehicle and anti-
personnel) on roads and tracks used by the ^LF
and ERA. Someofffliese
mines were even planted inside Sudan, and in December 1969, Dr Lais
Bondestam, a Swedish academic who had long experience of studyii^
famine issues in Eritrea and northern Ethiopia, was killed when his car
struck a land mine inside Sudanese territory. In August 1989, there was
a battle between the Mujahideen group and the EPLF close to the Sudan
border, which involved an intervention by the Sudanese army to support
252
the Mujahideen. Some militaiy engagements have also occuired since the
EPLF victoiy in May 1991.
The Fall of Asmara
In February 1991, the temporary lull in the war was hroken with
smmltaneoiu offeonves by the EPRDF into Oonder and Gojjam, and the
EPLF mto the Danaldl towards Assi^b. Ka mid-May flio EPLF aboUnmdied
an aasanlt on Decamhaie, and caiUmed die town on May 24. Aimy
idnfoicemenla sent from Asmara were surrounded and captured, leaving
the road open towaids Asmaia. Hie gairison at Aamaia snivendered the
following day.
Throughout the siege of Asmara the residents of the town had feared
a repeat of the massive bloodshed that had accompanied the ELF-EPLF
attack on the city in January 1975. Ethiopian military commanders had
also warned that Asmara would only be captured as a ruin. The air force
had shown its willingness to bomb major towns such as Massawa and
Meqele. These fears deterred an EPLF assault until the air force bases
at BaUr Dar and Debre Zeit had been Gaptnred — the latter on May 24.
By fliis time, fear of EPLF reprisals deterred aimy atrocities. To the great
rdief of allEritreans, the city was captured intact, without either widespread
destruction or loss of dvflian life.
Most of the other garrisons in Eritrea followed immediately, with a
significant exception. A
large contingent of the Keren and Asmara garrisons
decided to fight its way out to Sudan. This included about 75,000 soldiers,
some family members, and some high-ranking members of the
administration and security forces. They left Keren on the Agordat road,
and fought with EPLF units at Barentu and on the road to Sudan. There
are no accounts of them attacking civilians. More than one thousand
soldiers died; mostly in action, but some on account of thirst. About 14,000
soldleis arrived in Sudan seekmg asylum over the following few days.
In total, the EPLF captured 82,000 prisoneis of war and 44,000 dependents.
253
15. ARMED DECISION: THE NORTH, 1988-91
In Tigray, the year 1988 was the most savage in the entire history of
the war. Atrocities were committed on an unprecedented scale by both
army and air force. Many of the government actions were designed to
reduce the population to a state of famine, such as the deliberate killing
of azen, buniLDg of grain stores, and bombing of REST food convoys.
Ibis came on the beds of a drought m
the sommer of 1987. However,
the number of war- and drought-displaced people never appioacfaed the
scale of 1983-5. The reason for the failure of drought and war to result
in famine was largely because the government was restricted to the towns
and main roads in a way that had not been the case previously, and military
action in the countryside was shortlived. The restrictions on movement
and trade that had been so devastating four years earlier were no longer
so effectively enforced because of the reduced government presence.
After three years of cool relations, the TPLF and EPLF began to
coordinate their military activities again in April 1988.
In January 1989, the Ethiopian People's Revolutionary Democratic Front
(EPRDF) was formed by the ITLF and EPDM. Two new ofgimizations
were added: the Ethiopian Democratic OCGceis* Revohitionary Movement
(EDORM), whicA consisted larg^ of cqjtnred officers fi^
army, and the Oromo People's Democratic Organization (OPDO), which
was formed from among the Oromo of the WoUo escarpment and Oromo
prisoners of war. The formation of the OPDO
reflected and deepened a
split between the TPLF-EPRDF and the OLF.
In February 1989, the TPLF-EPRDF occupied all of Tigray save one
small garrison. Six months later it struck southward, right into Shewa.
During 1990, the EPRDF concentrated on consolidating its gains, and in
early 1991 launched three offensives in quick succession which finally
destroyed the army and government of President Mengistu.
May 1988: The Army hi Disarray
Following the EPLF victory at the battle of Afabet, the TPLF quickly
succeeded in overrunning many ganisons in Tigray, includiqg Enda Selassie,
Axum, Adwa, Adigrat, Wukro, and Maichew. Government troops also
withdrew from other areas.
In Enda Selassie, the retreating troops destroyed the town's electricity
generator, which had been built by public subscription of 500,000 Birr in
1985. At the town's health center, the staff were ordered to load all the
equipment and medicines on to trucks, which were then set on fire.
255
Retreating government troops committed a mraiber of atrocities ^gidnst
the civilian populatioa* inchidiqg the foicible evacuation and burning of
Faida village, in Raya district, on May 14.
The TPLF advance was followed by a number of punitive air raids.
The raid on Wukro which coincided with food distribution by the
International Committee (^f the Red Cross (ICRC) has been described in
chapter 11. Wukro was bombed on two other occasions. Other raids were
carried out on Axum, Hausien, and other towns and villages. Korem was
bombed on May 26, when people had gathered for a food distribution; there
were 24 casualties reported.
The Bombing Campaign of June 1988
On June1, the government counter-offensive was lanncbed The ftst

stagewas a bombing campaign of unpiecedenled severity. Some off the
bombing raids included:
* June 8 and 10: Seqota: four people killed or wounded, 71 houses

destroyed, and the church of Endagabriel badly Hynwgy^
* June 10; Amdo: five people killed, including a mother and child.
* June 12: Dcjcna: REST food convoy bombed twice.
* June 14 and 15: Sanue: in two raids, 17 people killed and one tfalid
of the town destroyed.
* June 16: Ruba Kaza, Tsegede: no human casualties, Imt 24 domestic

animals killed.
* June 18: Samre: casualties not reported.
* June 19: Enda Selassie: casualties not reported.
* June 19: Dande, Raya district: 29 killed, 74 houses destroyed.
* June 20: Abi Adi: four killed.
* June 20: Sheraro: two attacks, casualties not known.
* June 21: Enda Selassie and suxroundmg areas: casualties not known.
256
* June 22: Hausien was destroyed; an estimated 1^00 maikietgoets weie
killed (see below).
* June 22: Samre and seven surrounding villages were badly damaged.
* June 22: Abi Adi attacked:
"IVo MiOs drded over tiie town and IdUed a pair of oxen that were
ploughing a field just outside the town. Hie fumer escaped, but
one woman was IdUed and four others wouided."^
* Jmie 25: Entidio: 21 IdUed, seven bnildiqgs destroyed.
* June 25: Mai Kenetal: three killed or wounded.
* June 27: Atsbi: four killed, eight wounded, 109 houses burned, some
animals also killed.
* July 1: Adi Ramaz: casualties not known.
* July 2: Mai Humer: casualties not known.
* July 3: Adi Daro: casualties not known.
* July 3: Slieraro: casualties not known.
* July 3; Edaga Habriet: casualties not known.
* July 8: Sheshebite: casualties not known.
Bombing attacks continued, albeit less regularly, in the following months.

No air raids causing more than ten fatalities have been confirmed for the
numtlis July-December 1988, but many caused smaller numbers of deaths.
An unknown number were killed when 96 houses were destroyed in Adi
Hageray on August 19; were killed in Shenuro on December 12.
^ Woreda Teka,
fanner, trader and member of Abi Adi baUo^ intemewed by
Sarah Vaugban and Gerry McCaim» November 16, 1988.
257
The Destruction of Hauslen
The air raid on Hausicn on Wednesday, June 22, 1988, was the most
savage on record in Ethiopia during the three decades covered by this report.
Hausien was a market town in north-eastern Tigray. The market attracted
people from all parts of the province, and from as far away as Eritrea and
Gondar. On a normal market day the town was packed with several
thousand people, coming to trade in animals, grain, salt, coffee, and other
commodities. Though Hausien was attacked eight times in mid-1988, the
residents did not consider themselves to be at serious risk from air attack,
because the area was not controlled by the TPLF, and most of the market-
goers came from areas controlled by fhe government.^ Unlike the practice
m TPLF-controlled areas, the weekly Wednesday maiket therefore
continued to be held during daylight hours.
The following account is reconstructed from the interviews conducted
by two visitors to Tigray, Sarah Vaughan and Gerry McCann.^ Though
the interviews were conducted in November 1988, the memory of the
atrocity remained fresh in people's minds.
The bombing started in the late morning and continued until nightfall,
following a carefully coordinated plan.
Blata Aragabi, a 57-year old farmer recounted the day:
Itwas Wednesday, and I was in the market square. At about 10 o'clock

in the morning two helicopters came low overhead and circled for about
an hour. An hour later they came back with two MiOs. They circled
for a while and then bombed the market area, which was packed with
people, and animals waiting to be sold. Apart from the market square
itself, there was compound.
also a big animal market in the old school
The MiGs concentrated on the markets: no-one could have stayed alive
in those areas. Meanwhile the two helicopters circled round trapping
people as they tried to escape, cutting them down like leaves.'*
^ The fact that eight attacks warranted consideration as below average risk
indicates the intensity of the bombing campaign. Only a small minraity of the
attacks are mentioned in this report.
' They Vaughan, "Extracts from a Report on a Visit

are reproduced in Sarah
to Tigray, October 18 - December War on Want, and Geny
13, 1988," London,
McCann, "Between Heaven and Hell," Observer Scotland, ^nil 16, 1989.
^ Hausien is in a valley and there are only two

roads out of the town, SO h
is easy to seal off with just two helicopters.
258
[The bombing] started at 11 in themoming and went on until about
4 pjn. another
[at this point man said it was later —
5.30 p.m. and —
Blata agreed]^ almost until it got dark. Each time the MiOs and the
helicopters had finished bombing they went away, and more kept
coming, two by two. I don't know how many times new planes came;
because of all the dust it was dark and people were crying all around
me. It seemed to be about every half hour or so that they would leave,
and there would be a few minutes interval. People would come out
of where they were trying to shelter and pick up the bodies thinking
it was all over. Then the planes would return.
Hie bombeis used hig^ exptosives and duster booiba; the helicopters
used machine guns and rockets. Suivivon were paiticnlarly disturbed by
the "burning liquid" yMdk feU from the airplanes» piesnmably napalm or
phoq)hofous. "A," a priest, aged 41, described flie scene:
It was so dark, the smoke hung over the town as if it were night. People
were crying, confused, and hysterical. There was something that fell
from the sky, like rubber, but it burned your flesh.... There was a lot
of blood in the market place. So many animals were killed by being
burned or poisoned. I don't know what the poison was but it was
something that burned them. Those of us who were left wouldn't even
eat the carcasses, or have them near our houses.
Blata again:
[In the market] most of the peofde and cattle were being burned by
something that seemed like rubber. It bomed as it dropped off the sl^,
and didn't cut like metal does.
A cluster bomb remained unexploded in the schoolyard after the raid

and was photographed by Gerry McCann. It was unfortunately impossible
to ascertain its origin.
Most people sheltered in houses, but the bombers turned on these too,
using high explosives. When the day finished, there was scarcely a building
left standing.
Iquar Gd>re Giorgis, a woman beer-seller:
' Other infomiants said that the bombuig finished at 6 p.m.
2S9
I was selling sewa home. Market day is always a
(local beer) in oqr
good day were about 20-30 people in the
for selling beer, and there
house. I had my 13 year old daughter N^;jiti with me. When the
bombing started we thought the safest thing would be to stay inside.
They only seemed to be bombing the immediate area of the market.
Soon, though, they started on the houses, and mine was hit. The roof
on my house was not the usual tin, but wooden with heavy beams.
The whole roof caved in, and the walls came down too. TTiere was
one woman who escaped being buried. She was crying and digging
at the rubble, and she managed to dig out a man who had been buried
up to his neck. Bveotnally tiie two cxf iwni dog me oot, tfaongh I was
buried fitomDooo until 3 o*doGk. We were the only tinee who snivived
firom my house. Negisti died aloqg with evefyone else. Some of them
we managed to 4g out that evening were still caoidons, but they died
soon after. Tliey were all farmers or tnklers, some tarn Hansien, some
from the villages around. I knew most of them qnite well; 1 can
remember about half of them by name.
Haile Geresadie, a poor farmer aged 18:
[The bombing] started in the moming, and I was at the animal market
with my parents and grandparents. The animal mirioetWMlUl of people
and donkeys, but we ran as quickly as we could into the nearest nooe.
Ibe house was bombed, and out of about 20 people there were ooly
three of us [who] survived. We were on the srae nntfaest from where
the bomb landed. My parents and grandparents were all killed. [The
others] were just fnmers and traders who nn in from the animal market.
Some of them are still buried there.
I am always thinking about my parents, and often I just can't stop.

Sometimes it is very difficult to get to sleep because it is always in my
mind.
Zimam Hamenur, a woman spice trader:
All six of my family were in our house when it started* and we stayed
there all day. After several hours s bomb came tawgjh the window
and hit my daughter Fatima. Everyone else was rahnrt but her right
hand was cut oft. Even then we stayed in the house —
we were too
frightened to go out, and we just sat and cried round my daughter.
260
After nightfall when bombers had left, people contemplated the
the
aftermath. The true number bombing will never be known.
killed in the
At first the TPLF claimed that 360 were killed, and 500 buildings destroyed,
including nine stores, 15 shops, and a mosque. Later estimates by the TPLF
were hiêr: 600 fatalities, then 750, finally an "official" figure of 1,300.
Africa Watch believes that these figures are aU underestiiiiates. Most of
the people attending the maiket in Hausien came frm
was no list of the people present on June 22, and many of the dead and
injured were quickly removed to their home villages. Some bodies remained
buried in the rubble of buildings months later. According to the testimonies
of the people of Hausien, the number of &talities was as high as 1,800
or even 2,000.
Blata again, in response to the question of how many were killed:
You can't count grains of sand. Even now, six months later,^ we are
still finding bodies. The last one was on Friday week: we found the
head of a man and buried it. I heard from the [TPLF] fighters that
thousands were killed, and a Deiê radio broadcast said that they had
killed 3,000 bandits at Hausien.^ We buried between 100 and 200 m
the churchyard, but there were also lots of bits of bodies heads and —
limbs —
and people took many of the corpses back to be buried in
their own villages. ... People came here from different parts of the
country, from Eritrea, from Wollo, for the market day.
Priests are responsible for burying the Christian dead, and so have a
better idea than most people about the extent of the carnage. From priests
"A" and "B":
As soon as the MiGs went away everybody started to work together

to dig the bodies out of the rubbleand bury the dead. We buried as
many as we could immediately tiiat night. Ihe task of digging through
the stones went on for wet a month... [Asked: how many bodies do
you think are still buried?] We could estimate the number that were
found, but have no idea about those who were not. They came from
Shoraro and from all over the region. In every tabia [viUage cluster]
' The interview was conducted on November 20, which was in fiict just under
five months after the bombing.
' No government broadcast concerning Hausien was published by the BBC

Monitoring Service, so Africa Watch has been unable to check this.
261
or village you go to in this area you will hear that 40, 60 or 80 people
were killed. Those hom farther afield are harder to count.
An elderly man estimated that in the ruins of a hotel and a row of shops
on one side of the market place, there were 250 bodies still remaining.
Priest A
gave his estimate:
I think about 1^00 were killed in total, when yon indnde tlioie Âio
were taken back to dicar villagBs. Hiere were piobabiy aaoliier 1,000
wounded, and some of them may have died later. We buried 150 jut
in this churchyard, but many were too badly cot vp for burial, or are
buried just where they died. There were many priests killed, from
Eritrea, Tembien, Adwa, even from Wollo. Sometimes as many as
10,000 people gather here for market day. No one can count the
destruction of cattle and grain and money and all the property that was
buried.
Not aU bodies could be buried, or even identified Haik agîn:
bury [my fiunfly raembeis] because they were bamed away

I coukin*t
There was only one body that could be brought out [of ttie
to ash.
bombed house, in which about 17 people died], and even that woman
was very badly burned.
Kesi Gebrc Hiwet, a farmer and priest aged 54, came to Hausien the
He spent the night digging people out of the ruins,
night after the attack.
and estimated the casualties at 2,000 dead and 800 wounded.
We found so many corpses, we went on digging until morning. Whilst

we were digging we found dismembered hands and feet, but oouldnt
find the bodies iiey came from. In die momiqg we started to buy the
dead and brought the wounded away. Four of those whom we brought
back to this iabia died after two days.
Medical facilities are poor and many of the wounded died later. Coping
with the injuries was a major problem. Zimam Hamenur needed to find
treatment for her 15 year old daughter Fatima, whose hand had been cut
off:
The next morning we and some neighbors carried Fatima to Nebelet;

I had heard previously from some fi^^ters that there was a clinic there.
The TPLF gave her mfusions and bimdaged her, and kept her diere for
262
a month. They wanted to take her to Tsai where there was better
treatment, but on the way we heard that the enemy was coming so we
decided to go to Wukro instead. We were quite frightened of going
to a government town but to help my daughter's hand we had to take
the risk.
I had to come back to Hausiea to look after everyone else, and Fatima
stayed in Wukro. We
have a message to say die is better now, and
we are expecting her back at any time. A
while ago someone took my
yomigest daughter Neehma to see her sister in Wukro. She was very
upset and crieid whenever she thought about her sister. She used to
wake up in the night cryiqg.
There arc reports that some injured victims were denied admission to
government hospitals.
Many of the survivors were deeply traumatized by the bombing and
will probably suffer from psychiatric disturbances such as post-traumatic
stress disorder for the remainder of tiieir lives. Tsehai Geredche, a woman
i^ged about 30, lost her husband m
the attack and spent many hours partly
buried in the rubble. She was six months pregnant at the time.
After I was dug out I was delirious for some time —

I have been told
it was for about three weeks. All I can remember is that I had a pain
in my legs and that I was very frightened I had lost the baby. I think
I was feverish. It wasn't for quite a few weeks I could take in what
had happened, and think about what to do. The six other children were
all safe, and also the baby; as you see, I gave birth two months ago.
My husband was buried at the church.

Hie evidence of eye witnesses suggests that about 1,800 innocent men,
women and children were delibeiately killed in Hausien by aerial
bombardment. Many others were severely mjured and manned, or
psychologically traumatized. Apart firam the scale of the carnage, what
makes Hausien a particularly brutal atrocity is the systematic nature of the
attack. It required careful planning to arrange for a succession of MiGs
and helicopters to be present at Hausien, far away from the nearest airfield,
in coordinated shifts throughout an entire day.
The motive for bombing Hausien can only have been terrorism against
the people of Tigray, in part revenge for the military successes of the TPLF
over the previous months, and in part "softening them up" for the
government offensive. Hausien was probably selected as a target because,
not being in a rebel-controlled area, the market still met during daylight,
263
and there were no TPLF fighters in the area with anti-aircraft tftilkiy to
make an attack dangerous. It had no militaiy Mgnifioancft.
The Summer 1988 Comiter-Offensive
The ground offensive started simultaneously with the bombing campaign,

with te tioqw inoving in to towBB and villages a li^
After some fieice fighting in north WoUo on tfaeopoung days of the attack*
the ITLF stuck to its strateg^r of not holduig tenitoiy, and allowed the
government troops to leoocnpy most of die towns on the main roads, and
to pass through others. The army quickly moved from Woldiya to Korsm
through the Amba Alage pass. On June 4, soldiers attacked Harako village
nearby: five civilians were killed and three wounded, 77 houses were
burned.* Another force moved south from Meqele to retake Maichew.
On June 4-5, the soldiers burned six villages near Meqele (Adi Gera,
Gobozena, Grarot, Rabea, Issala and Bahri), and on June 7, a further two
villages were destroyed (Mai Wewe and Adi Guguad). Casualties in these
atrocities are not known. After re-occupying the main towns of southern
Tigray and northern WoDo, the anny then moved mto central Tigray,
occupying Abi Adi.
On June 24, a large number of in&ntry —

three brigades came —
through the town. There was a large battle at Hagerai Selam, and
Dergr was going to Adwa retaking the towns. No-one was killed,
but they stole or destroyed a lot of property. A
lot of goats were taken,
and when they found oxen they would just cut off one leg, or cut out
the liver without even killing the animals. They collected people's
[farm] tools together and destroyed them. 1 lost 400 Birr and some
of my furniture. From Abi Adi Dergi went on to Mai Kenetal.^°
Government forces pushed north, taking Wokro, Adigrat, Adwa and

Axum and then attacked western Tigray, taking Sdekkka and Eoda Selassie
in early July. Another army odunm moved nordi from Gonder. However,
* The following aoooont owes mncfa to infonnatifln cnmpfled by Bsibaia

Hendrie.
' Tigrayaos often refer to the Deigoe in the first person singiilar.
^ Woreda Teka, farmer, trader and member of Abi Adi baiio, interviewed
by Sarah Vaughan and Gerry McCann, November 16, 1988.
264
the TPLF regroupedand finally eog^ged the anny in Shiie and north Gonder
in July,and defeated it.
On a number of occasions, government troops killed civilians. The
following incidents have been reliably reported:
* June 28 and 29: Adwa: 50 people killed, including 19 bayonetted and

thrown over a diff
* June 29: Hageiai Selam and soiioiinding village: 341 peasant
* July 5:Netsege:30£amieisbiiiiiedintbeirbousesby8Qklieis»60hoi^
destroyed.
* July 5: Hagerai Selam: an unknown number killed.
* July 6: Mai Mekden: ten killed, three wounded, 25 houses burned.
* July 20: near Maichew: a young girl thrown to her death over a cliff.
* Inly 31: Adi Nebiid: 15 people killed, one wounded.
* August 2: Kdiih Boini: 13 pople killed, two wounded, three houses

burned, seven tonnes of grain destroyed.
* August 9: Bahra and Senkata: many people beaten, two recently married
couples taken away, four women raped.
* August 9: Adigrat: seven women raped.
^ August 16: Mai Mado: five people killed, one wounded, 22 houses
burned, grain steles emptied and the grain mixed with soil, many
MwtiMih killed.
* August 29: Adi Hagfiiai: 23 kiOed, 193 injured; the dead ii^^
children deliberately burned, and many of the iiguied were cut and
mutilated with knives.
By the end of August, large-scale military action had ceased, but army
patrols continued to be routinely brutal. In October, at Tselessi Bit and
Selekleka, 20 people were killed in five separate incidents, including people
locked in their houses and then burned alive.
265
The bombing and ground offensive caused an estimated 60,000 people
to be displaced from their homes by the end of July.
In the ground offensive, the army used mass columns of conscripts to
attack TPLF positions. On at least one occasion, this amounted to mass
slaughter. On July 7, an army column advanced north from its base at
Dansha in north Gonder, straight into an ambush laid by the TPLF. The
conscripis in Ibe vanguard: they weie cauglit in a beaivily-niined valley
with TPLF fighters in the hills on both sides. Which ever way they tiuned
they were cut down by gunfire or blown up by land mines. The TFLF
claimed that nearly 3,000 were lolled, wounded or captured. TPLF fighters
later spoke of their distaste at the carnage. They said that in later
engagements TPLF tactics changed, and concentrated on destroying the
command unit in a military force.
The months from September to December 1988 were relatively quiet
in Tigray, north Gonder and north Wollo, though intermittent bombing
continued. For example, on September 8, Nebelet was bombed and several
houses burned.
The Govemmcnt Evacnatloii of Tigray, Febniary-MiBrch 1989
In January 1989, the TPLF began to take the offensive, at first in north
Gonder, and then in western Tigray. In a series of battles between February
15 and 20, a joint TPLF-EPLF force captured Selekleka and then Enda
Selassie. The fronts claimed that 26,000 soldiers were put out of action,
and it was certainly the government's worst defeat since Afabet. The army
evacuated Humera (on the Sudan border) and Adigrat, and on February
27-28, the provincial capital, Meqele, was abandoned, leaving the
government with only an outpost at Maichew. In effect, all of Tigray was
under TPLF control.
The TPLF was stunned by the unexpected evacuation of Meqele and
waited for three days before entering the town.
In each of the three towns of Enda Selassie, Adigrat and Meqele, the
army and government officials caused widespread destruction before Ifac^
left. In Meqele, on February 26, two army tanks shelled the electricity
generatiiig station, destroying completely five huge generators, each capable
of producing one megawatt of power. Bedding and instruments from the
hospitals were systematically looted by soldiers. Residents of the town
looted many furnishings.
At 11 a.m. on March 21, airplanes bombed the generator at Enda
Selassie, inflicting some damage. One woman was killed. Other raids
included:
266
* March 26: Humera: casualties not known.
* March 26: Adwa: casualties not known.
* March 27: Adwa: casualties not known; a rare example of a night attack.
* March 30: Axum: three killed.
Plnnlthre Fitrob in tlw Tchdfii Vall^
The success of the TPLF (now part of the newly-fonned EPRDF) in

taking control of all of Hgiay was matched bf less spectacular but equally
significant progress in its penetration into moie soutberly areas of WoUo.
This began to ignite disputes that had lain dormant, because until it became
clear that the government was losing control, subjugated people had not
dared to challenge the government's authority. The army responded to these
local disputes and threats in a punitive manner. An example of this comes
from the Tcheffa Valley in southern Wollo and northern Shewa.
The Tcheffa valley is a grazing area for Oromo and Afar pastoralists
and drought refuge for all groups, including Amhara farmers from the
highlands. In 1986, duriiig villa^zati(m, An£ani higbl^
in the valley and pfedmont. Hie new villagers complained of mosquitoes
and that Ifaey wete being setded in a giazi^g area, that was not appropriate
for farming, and that this would upset the pastoralists. Hie government
paid no attention.
During 1987/88 there was drought in the lowlands, and the Afais
penetrated to the valley, leading to some armed clashes. The situation
deteriorated during 1989, due to continued drought in the eastern lowlands
and the presence of the EPRDF around Dessie and the perception that the
government was losing control.
The Amhara farmers wanted to return to the highlands, the Oromo
natives became increasingly militant, but a series of negotiations came to
nothing. In March 1989, a party of aimed Oromo horsemen came to Fuisi
sub-district. Hiey dashed with the Amhara fsimers (who had a militia),
and eight were killed (on both sides). A rumor spread throughout northern
Shewa that "the Oromo are rising in rebellion." The army was sent to
pacify the area, arriving a few days later. The army mission became a
punitive expedition. The soldiers went and hunted and killed as many
Oromo as they could find. Only Oromo were killed.
Under the guise of preventing EPRDF activity and keeping the peace,
the army occupied Oromo villages. They controlled all Oromo movements
and demanded food from the villagers. There were numerous small
267
incidents of killings andwonndings of Oromo dvilians in nuuketplaces,
at wellsand on roads. Hiis continued until October 1989. Over200Oioino
are estimated to have been killed.
The EPRDF Southern Offensive, Late 1989
In late August 1989, as the government prepared an offensive into

southern Tigray, the EPRDF struck first. By September 8, EPRDF forces
had captured Maichew and Korem and were advancing along the main
highway in Wollo. In October, Woldiya was captured and Dessie was
almost taken, and in November, the advance guard of the EPRDF penetrated
into northern Shewa. In late December, EPRDF forces captured the town
of Debie Tabor in southern Oonder, to t)e driven out by the ainqr a month
later.
Hiere were, as always, numerous bombiiig attadcs on sites throu§|ioot
northern Wollo and Tigray. The foUowmg have been idiably reported:
* September 9: Chercher, Tigray: the marketplace was bombed and sHafied;

148 people were killed and about 100 wounded.
* September 9: Gobye, Wollo: one killed.
* September 10: Gobye, Wollo: 21 killed, 100 wounded (market day).
* September 12: Gerarsa, Tigray: four wounded.
* September 19: Raya, Wollo: no fatalities reported.
* September 20: Kuhnelsk, Tigray: three killed.
* September 22: Axum, Tigray: two killed, eight wounded.
* September 25: Tekezze Bridge, Tigray: three killed.
* October 15-November 2: Kobo, Wollo: four attacks. Casualties not

known, but the clinic was strafed by helicopter gunships.
* October 27 and 29: Meqde, Tigray: 31 killed m the first attack.

* November 5: Adwa, Tigray: no fatalities reported, two tracks destroyed.
* November 15: Zalembesa, Tigray: 14 wounded.
268
* Novonber 21: Sheraro, Hgray: 31 killed, 60 wounded.
* December 27: Adwa, Tigiay: four wounded.
* January 1, 1990: Adi Nebried, Tigray: two wounded in a church.
The attacks onMeqele town deserve special mention. They were

significant because they showed that the government was prepared to attack
a major Ethiopian city, a provincial headquarters and former imperial capital
(1871-89). lliecHywwmioietiiaDalniiidrediiiileubeliiiidtef^
and had no military importance. It was abo an unusual attack because
news of it filtered back to Addis Ababa and caused widespread, thongli
muted, public outrage.
The Meqele attack caused many residents to leave the town. The
hospital was evacuated during daylight hours, with all patients returning
at nightfall for medical attention. Even at night, no electric light was used
for fear of attracting the attention of overflying high-altitude Antonov
planes, which were occasionally used for bombing at night.
For the first time, there was also widespread bombing in Gonder and
central Wollo. Some of the raids included:
* November 13: Kara Mishig, Shewa: one killed, 25 houses destroyed.
* Noveniber 15: Tenta, Wollo: no finalities reported.
* November 16: DogoUo, WoUo-Shewa boiden no finalities lepoited.
* November 16 and 17: Woldiya, Wollo: no fiatalities reported.
* January 4, 1990: Tenta, Wollo: two killed.
* January 10; Wurgessa, Wollo: one woman killed, five houses destroyed.
* January 13: Nefas Maucha, Gonder: 23 killed.
* January 23: Deha, Wollo: seven kiOed.
* January 25: near Debre Tabar» Gonder. 15 killed wfafle sheltering in

a stoim drain under a load.
* January 28: Isitayoh, Wollo: 40 killed (most of them in the church of

Kidane Mehrat).
269
* January 29: Wegel Tena, WoUo: two killed.
The army was mostly during these months. There was

in retreat
widespread looting towns, including Kobe and Dessie, but fewer
in several
reported incidents of attacks on civilians. One incident occurred on
September 17, at Gobye in north Wollo, when at least one civilian was
killed in an army rocket attack, which destroyed four houses. The same
day at Zaremma, nearby, two civilians were killed by soldiers.
The EPRDF treatment of the civilian population during this offensive
appears to have been remarkably good. EPRDF fighters and cadres were
reported as entering villages and lelliqg the inhabitants that they would
not be harmed and their possessions would be respected. They took
inventories of commercial and private stoies in tfie towns they occupied
to prevent looting. No incidents of violence against civilians have been
reliably reported.
Relative Quiet: February 1990-February 1991
In the twelve months from February 1990, there were no major

offensives by the EPRDF. There was more-or-less continual skirmishing
in north Shewa and southern Gender, but the only major battle was a failed
assault by the aimy at Alem Ketena in June. It was, overall, a lemaricably
quiet year in terms of abuses against civilians by the army, lliis is probably
because the army was now Qg^ting in mainly Ambara areas, and the officer
class of the army, which is dominated by the Amhara, was less wiUiog
to sanction abuses against other Amhara than against Tigrayans, Oromos,
Somalis and others. The pattern of abuses is more akin the result of a
breakdown in morale and discipline than to deliberately-planned mass
killings. However, a number of incidents warrant mention.
In March 1990, the army first evacuated and then re-entered Bahir Dar
after a battle with the EPRDF. On re-entering the town, soldiers killed
an estimated 50 civilians. According to some accounts, the soldiers were
drunk and ill-disciplined.
Between March 15 and April 7, in Desste town, soldieis killed 16
civilians, including two children. CHi at least some cf the occasions, the
soldiers were off-duty and had been drinking.
Another incident occurred in June. Membere of the army garrison at
Mclkawarer in the Awash valley of northern Shewa became involved in
a dispute with the local Afar inhabitants. The origin of the dilute is
270
unknown but is rumored to be related to chat^^ cbewiqg. Over 20 Afar
civilianswere killed in the fight that resulted.
If these and other similar incidents could be described as occurring
without official sanction, the same cannot be said of the continuiiig air raids
against civilian targets. Some of the raids included:
* March 28, 1990: Wire Ilu, Wollo: grain stockpile bonied.
* May 2: Rama, Wollo-Shewa tx»der: two lolled.
* May 2: Alem Ketena, Sliewa: fonr killed.
* May: Meifaabete, Sbewa: casualties not known.
* May 5 and 9: Kolesh and Ambat, Sbewa: six killed.
* May 11: Kinche, Shewa: 21 killed.
* June 10: Ticha, Shewa: 23 killed.
* June 23: Adi Abun, near Adwa, Tigray: one wounded.
* August 1: SQnjero, Wollo: no casualties reported.
* October 5 and 14: Meki, Wollo: five killed, six wounded.
* October 19: Mehal Meda, Shewa: four killed, five wounded.
* October 22: Temsa, Wollo: a family of eight killed, ten others wounded.
* November 7: Woldiya, Wollo: one killed, one wounded, and relief

offices burned.
* November 7: Kul Mesk, Wollo: seven killed, eight wounded.
* November 27 and 29: REST stDfts near Sudan border bombed and
strafed: 3,000-4,000 metric tonnes of grain burned.
* December 27: Woldiya, Wollo: nine killed, close to a relief store.
Chat is a mildly narcotic leaf widely grown and chewed in Ethiopia.
271
* Februaiy 13, 1991: Ddne TUwr, Gooder two killed, tea womded.
Operation Tcodrof
OnFebruary 23, 1991, the year of relative quiet came to an end, when
the EPRDF announced "Operation Teodros," aimed at destroying the army
in Gender and Gojjam. The announcement of the launch of the offensive,
its aims and timetable, showed an increased confidence by the EPRDF.
The EPRDF clandestine radio also instructed the citizens of the towns in
Gonder and Gojjam to guard the civilian infrastructure of their towns to
prevent looting, such as had oooned kk Meqele before its capban.
One fadPT that a«Mted Hbc EPRBF pffemivc was the gowii^g ilknition
of tiie local people from te govennuBBt. Tbli wai lebtod to tie heavy
conscription campaigns of the previous year, and the disarming of the load
militia in western Gojjam following a revolt in Maccli-April 1990^
coinciding with the abandonment of the villagization program.
Within a fortnight Operation Teodros achieved its aims. The offensive
was so swift that there was little chance for the army to undertake reprisals
against civilians, though two incidents deserve mention.
One incident was the systematic execution of prisoners in Gonder town
during the three days before its capture. Most of those killed were
Tigrayans detained in the town's prison for suspected sympathies with the
EPRDF, and tiie executions ifpear to be an act of ptt-etapdye vengeance.
Jenny Hammond, a Briliih writer who visited the town the day after its
capture by the EPRDF, spoke to townspeople \v1k> repotted that about 120
detainees had been killed, and tiie executions of 100 or so more had tieea
scheduled for the day of the EPRDF takeover. Later, Ma Hammond met
Dawit Berhane, a Tigrayan merchant who had spent three years in prison
on charges relating to alleged irregularities in obtaining a truck license.
Dawit related how 19 of his cellmates (17 of them Tigrayans) were taken
out and executed the day before the town fell. Dawit himself was scheduled
for execution, but the official authorization mistakenly had his father's name
made out as "Berhe", so he was sent back to his cell for another day while
this administrative error was rectified. Due to be executed at 6:00 p.m.,
he was released by tfie EPRDF at 4:00 pan. Dawit believed that 300
detamees had been eiecnted m the final days, and said that dnii^g the
previous three years over 3,400 people had been eaecated in the pmoo,
90 per cent of them Hgiayans.
Asecond incident was the long-distance shelling of Dejen town in
Gojiam on April 16, after its capture, in which six people were killed.
272
The EPRDF advance brought it into conflict with the Ethiopian People's
Revolutionary Party (EPRP), which had an anned presence in western
Gojjam (see chapter 18).
In the aftermath of the offensive, the EPRP and the government made
a number of allegations about killings, detentions and looting by EPRDF
forces. One incident was the killing of three senior government offidal&
incliiding a security duef and a milituy ooamumder, by peasants."
Reports indicate that the officials and their armed escort opened fire on
the peasantSi and were killed in the ensniqgbitde. Other alleged incidents
hichide the use of kthal force against anti-EPRDF demonstrators and the
detention of taaay political opponeats. The focts smiouading these cases
have not yet come to light.
Following the capture of the towns in Gonder and Gojjam, traders from
Tigray and Eritrea immediately entered them to buy grain for transport
north — grain was cheap in these areas and expensive in Tigray and Eritrea.
According to one visitor, "deals were struck before the corpses were buried."
The EPRDF forces also sealed government grain stores. These actions
caused panic among some local people, who feared that their grein was
beiog confiscated.
In late Match, as the government attempted to counter-attack into
Gojjam, the EPRDF launched Operation Dula Billisnma Welldta (Oiomo
for "Equality and Freedom Campaign") into Wollega. This captured
Nekempte, headquarters of Wollega, on April 1, and then advanced
southward and eastward, towards Addis Ababa. Following the battle for
Nekempte, retreating government troops looted several parts of the town.
By this time, the army was on the verge of complete rout, and was unable
to regroup for any significant counter-offensive. By the same token, it
was unable to engage in systematic violence against civilians.
Shortly afterwards, the EPRDF occupied Fincha'a, which is the site of
the hydro-electric power station which serves Addis Ababa. Ihe power
wasnotcutoff, though theEPRDF contacted tlieAddii Ababa municipality
by telephone to ask for senior cqgnieen to come to cany out urgent
maintenance tasks.
On April 28-30» the air force bombarded Fincha'a town and hydro-
electric station, and one civilian was killed in the town and one worker
wounded at the power station.
^ See: "Ethiopia: Human Rights Crisis as Central Power Cnunbles," News

from Africa Watch, April 30, 1991.
273
The Final Days of the Mengistu Reghne
In what can only have been an act of deliberate spite against the
Tigrayan people, on May 8, the air force carried out a raid on Sheraro in
Tigray. Sheraro had had no military significance for more than four years,
but retained its symbolic significance as the first town occupied by the
ITLF, and its "home". Aoooidiqg to leporti* 15 civiliins were ki^
90 wounded.
One week hrter, the EFRDF hnmched "Opeiatkm WalleUm"" oe the
Wollo front. Dessie and Kombokha were captued the fduowuig day.
In the battle for Kombolcha, an ammunition daiap was blown np^ causing
extensive damage to the town and an unknown number of civilian casualties.
The EPRDF claims that the dump was deliberately ignited by retreating
soldiers, but this has not been confirmed. By May 20, the government
lines throughout southern Wollo and into northern Shewa had been overrun.
This coincided with a government defeat at Ambo, west of Addis Ababa,
and the city was effectively undefended on two sides. President Mengistu
fled the country the next morning.
President Mengistu had always boasted tfiat he would fight to the last
Publicly, he compared himself to the Emperor Teodros, who oommitted
suicide rather than surrender to his eneasies. Meogisln^ reputation for
intransigence and courage was the last asset the government had; loyal
soldiers respected him and were prepared to continue <%h^M«fl When
Mengistu fled, the keystone that had held together the remaining elements
of the government and army was gone. The army —
450,000 strong just
months before —
disintegrated. Tens of thousands of soldiers abandoned
their posts and flocked into Addis Ababa, selling their weapons or using
them to intimidate people into giving them food and drink. Looting became
common. Other soldiers took off their uniforms and went home. Some
senior officers in the army and air force fled abroad, mostly to Djibouti.
Only afew elite units inside Addu Ababa auintaiiied loyalty to the acting
head of state. General Tesfaye Gebre iQdan, bat a mutiny broke ont on
May 27 and there was fighting around the Presidential Palace.
The final week of the war consisted in a slow EPRDF advance on Addis
Ababa itself. EPRDF forces surrounded the city, cqituriqg tile cnidal air
force base at Debre Zeit after a small battle.
During May, western diplomats and the UN repeatedly urged the EPRDF
to refrain from attackipg Addis Ababa before the US-convened peace talks
^ It was named after the Oromo student leader mentioned in chapter 4.
274
I
opened in London. They expressed fears for the safety of the civilian
population should there be fightmg in the city itself.
The final assault on Addis Ababa took place on the morning of May
28. Almost all of the army had melted away, and resistance to the EPRDF
advance was light. There were pockets of street fighting throughout the
city, but the only sustained exchange of fire occurred at the Presidential
Palace, where an ammunition dump also exploded. According to the ICRC,
there were about 200 deaths, both combatants and civilians. Those who
died were either caught in the crossfire or killed by the explosion; tfaoe
is no indication that either side targeted civilians.
tianspiied that a second explosion had also occnned at an
It later
ammunition dump at Shogole the same morning. Eye witnesses said that
local residents began looting the arms depot, whereupon a fighter from
the EPRDF fired a rocket-propelled grenade, which caused a huge
explosion. An estimated 500 people died. When a German pastor spoke
to a camera crew from a news network, claiming that the explosion had
been caused deliberately, his words were cut by an EPRDF censor.
However, as one journalist commented, "even if a rocket was fired, no one
could have imagined the appalling consequences."*^
Before dawn on June 4, another explosion occurred at an ammunition
dump in the Nefii^ Silk area of the city. Approximately one hundred people
were killed and 130 wounded, including several firefighters and members
of the EPRDF who were trying to assist victims. "Hieie was extensive
damage to property. The EPRDF claimed Hat it was the work of saboteurs
loyal to the previous government, and said that they had detamed several
suspects, one of whom was a former army ofHcer caught while trying to
launch a rocket-propelled grenade at a fuel truck. Hiis account was
confirmed by at least one western diplomat."
On entering Addis Ababa, the EPRDF prohibited all forms of public
demonstration. However, protests against the EPRDF occupation soon took
place. EPRDF fighters responded by firing, at first over the heads of the
demonstrators, and then into the crowds. Eight were killed on May 29
and two more on May 30, and a total of 388 were injured, according to
Red Gross estimates. Tbe EPRDF churned that members of die crowds
were armed, and pointed to an mcident m
which two EPRDF figliters were
" Richard Dowden, The Independent, London, June 2, 1991. At previous

explonoas at nmnunition dump^, such as AnnaER aiip^
in May 1991, tfaeie had been a series of smaller explosions and fires, allowing
most people to escaqpe.
" Jennifer Pamielee, Waskmgum Post, June 6^ 1991.
275
killed by an assassin at the university campus. Some protestors were armed
with weapons, including hand grenades, and the crowds had pelted the
fighters (who had neither riot shields nor training in crowd control) with
stones, and on at least one occasion opened fire. One journalist commented:
"Even street kids have automatic weapons ... It's as if the millions of dollars
of Soviet military aid have all arrived in the capital at once."^^ After these
killings, and after the explosion of June 4 had shown tet members of the
pieviDiis rpgiiiie weie still active in amwd oppcMition, the protnti
disappeaied.
Jomnadists abo reported the nunmaiy ciecBtioa of tiro membew of
m
the security service of the former government hy me beiB of an BPLF
unit which had participated in the assault on the city.
The occupation of Addis Ababa cost between 600 and 800 civilian lives,
most of them in the explosions at the ammunition dumps. Much of the
violence was the work of retreating government soldiers, and it is likely
that the EPRDF occupation of the dty prevented further lawlessness and
loss of life.
Jenifer Panneiee, WashingUm Post, May 31, 1991.
276
16. TH£ POUnCS OF R£LI£F 1989-91
In the summer of 1989, there was widespread drought in Eritrea, Tigray
and parts of Wollo, leading to fears of a repeat of the famine of 1983-5.
Those fears were intensified when the summer rains of 1990 were also
poor, and there was a near-complete harvest failure throughout Eritrea.
Fortonately, to date, these fsm have not been lealized. Ttae ks been
considerable haidsh^ throughout Eritrea, Hgray and nortbem WoUo, and
pockets of suffering amountiqg to severe fimuie in one or two places in
Eritrea. But there has been neither tbt mass migration to relief shelters
and refuge camps that were characteristic of the 1983-5 funine, nor mass
starvation.
This chapter looks at the causes of the famines in Tigray/north Wollo
and Eritrea respectively, and also at food shortages elsewli^ in the country.
Scarcity in Tigray and North Wollo
In both 1989 and 1990, the harvests in Tigray and north Wollo were
very poor. Rainfall was as low as m 1984, and production was furthw
hampered by lack of oxen — a legacy from the 1983-S fiunine, the
presence of land mines» and the inadequate maikeAing system, with many
rural markets still held at night for fear of bombii^. Neverthel^
according to an independent crop assessment mission, in both years there
were surpluses m
Shire and Raya, of alwut 50,000 metric tomes (Ml) and
100,000 MT respectively. Surpluses in north Gonder were not assMed,
but were certainly substantial.
Throughout the period 1989-91, all of Tigray and north Wollo was
controlled by the EPRDF, which meant that the famine relief could not
be distributed by the government RRC or voluntary agencies working
alongside it. Food relief could only be brought in with the consent of the
government, or by wmkiiK omi-4ioider from Sudan.
Because the war had amrect hnpact on aH funine relief operations for
the first time, the liiik between war and funine b^gan to be identified
the western media. In die case of Tigray and north Wollo, this is ironic:
from 1989 onwards, this area was very largely at peace. The qwcific effects
of the war were confined to sporadic air raids, Portages of consumer goods,
and obstacles to the delivery of relief. These burdens were much less
onerous to rural people than the military offensives and government
restrictions and exactions of the early 1980s.
277
In the areas where fighting occurred, it now consisted largely of
conventional battles between the opposing armies of the govemmeot and
EPRDF. Several factors exacerbated food shortages:
* In the immediate vicinity of the front line there was disruption to

villagers' lives by the fighting itself.
* Hie anny garrisons imposed a ooosidenble boiden on local waonuou,

usually requisitiaiiiqg food, and ofiten ^^^snmnAin^ ftui local women oobbo
and pcepare it for them (see duqpter 17 for oinmplcs fnam northem
Shews).
* In the Oromo areas of the Wollo-Shewa escarpment and Tcheffa Valley,

army patrols and checkpoints imposed tight restrictions on movement
and trade.
* Surveillance of migrants and trades was generally stepped up, and more
local militia weie mobilized to peifoim these lontine functions.
However, with a few Isolaled eioeptionSk the level of huassment and

restriction never approached that imposed in TIgray and its boidalands
m the early and mid-1980B. This was probably becannse army was
operating in mainly Amhara areas, and persisted in seeing the conflict in
ethnic terms —as against Tigrayans. The army was also unable to
penetrate into EPRDF-controlled areas to inflict damage, and from February
1990 onwards, there was little fighting on the central Wollo front on account
of the international relief operation being mounted there. The constant
skirmishing occurred in southern Gonder and northem Shewa, both areas
m whidh food supplies were bettor. Consequentfy, irrespective of levels
of relief assistance, the fomine of 1989-91 was always going to be much
less severe than that of 1983-5.
EPRDF PoUcies
From EPRDF consistently implemented a policy of trade
1989, the
liberalization,and the intra-regional movement of grain was not hindered.
Migrant labor was also possible, though in Tigray the TPLF tended to
discourage it and preferred people to remain in their villages to engage
in programs of environmental protection such as afforestation and terracing.
The high degree of internal security and ease of mobility within Tigray
and the aborning areas led to die functiooiiig of ibt economy in a way
that approximated "nonnal" for the first time for 15 yean. Together with
278
the absence of the government counter-insurgency strategies, this was
undoubtedly the main factor in preventing the severe droughts leading to
severe famine.
Relief Programs
A relatively efficient lelief program also contributed to the lack of

funine. Hie piognm consisted of three ekmetilK "imemal purchase,"
cross-border food deliveries, and the Jomt Relief Ptogram (JRP) of the
Ethiopian churches. The government tried to obstruct all tfaiee, but bad
neither the detennination nor means to succeed.
Internal peace made possible the large scale purchase of local surpluses
for redistribution as famine relief. Tliis program, known as "internal
purchase," proved to be the quickest and cheapest way of providing relief.
Although the actual prices paid for the grain were relatively high, this was
offset by cheaper transport costs.
Throughout 1989 and most of 1990, the cross-border relief program
fiom Suchin was the single largest oontribntor of relief food to Tigray.
Hie tnn^KMrt was slow and eipensive, because of the long distances, the
rough roads, and the fact that convoys could move only at night because
of aerial bombardment. There were numerous attacks on relief trucks.
The JRP arose out of the recognitioa in late 1989 that the heavily
drought-stricken areas of the country were under the control of the EPRDF.
The plan was for the Ethiopian churches to organize transport and
distribution of relief to Tigray. Initially, the relief was to be provided
through government -held areas of Eritrea, but after the capture of Massawa
this changed to the "southern line" through WoUo.
Initially, there was much skepticism about whether the JRP would
actually work. Those doubts seemed to be confirmed when the program
became mired in a set of disputes:
* The government and churches claimed that the roads needed repair;
REST and the EPRDF said that their vehicles had no difficulty passing
* The government JRP vehicles should not leave the main

insisted that the
north-south road; REST argued that the most needy areas were away
from the road, and that it would be unnecessarily disruptive for people
to migrate to the roadside to receive relief.
* The government and churches msisted on doing their own registration

of beneficiaries, and required all famfly meaabers to be pfcsoit at the
279
registration; REST replied that it already had lists of needy people, and
requiring all family members to come for registration required a three
to four day walk for many, and would create chaos. (After such chaos
did indeed ensue, and several people were injured, the requirement that
allfamily members be present was dropped.)
* The JRP wanted to start diitiibiitkm ionoflhW^

this was the aiea haidest to leacfa by the cnMa-hoider loate; REST
wanted distriMkms to start shrndtaneoiisly in Tlgray.
The first tracks moved north from Bessie and crossed into EPRDF-held
territory on March 20, 1990, only one week behind schedule. However,
during the following six weeks, progress was slow only four per cent —
of the target amount was distributed, all in Wollo. In May, one third of
the target was met, including distributions in Tigray, and from then until
the following March, distributions averaged over 80% of target. The
program became a success, and matched the cross-border operation.
Despite the government's recognition that rebel-held areas actually
existed and needed relief, and the net that the size of the JRP opeiatioB
was calcuhited on the assninption that REST would provide cnMfr4xMder
relief to much of the population, the government continued to liond> cnaa-
border relief routes. On January 29, 1990, a REST food convoy was
bombed in western Tigray; three trucks were burned, one local herder killed,
and two REST employees wounded. On November 27 and 29, 1990, REST
food stores near the Sudan border were bombed and more than 3,000 tons
of relief food destroyed.
A food monitor noted the resulting ironies:
It was interesting to compare this [the JRP's] very impressive fleet of

white Mocedes trucks with the REST fleet of trucks. Hie REST fleet
is also made up mainly of Meioedes tncks, hut due to the Ethfopjan
govenunent's propensity for bombing relief oonvoys oomii^
Sudanese border, these have all been painted acamouflage gieea colour.
In Tigray the JRP fleet is only allowed to tnvd during the day. The
REST trucks, on the other hand, can only move during the hours of
darkness ... during the daytime they have to remain hidden under trees
or buried in trenches and covered with canvases.
The success of the JRP owed little to the government, which continued
to harass the program at frequent intervals. It did not allow food monitors
to travel with theJRP convoys —
all the monitoring was done by aid
agency staff who had come in on the oosa-boider route. Therequirement
280
that all beneficiaries travel to the roadside was feared by many people as
the prelude to a government offensive up the main road. The government
even bombed some of the towns where JRP distributions were taking place.
On November 7, Woldiya in north Wollo was attacked. One woman was
killed and one girl injured, and the offices of the Ethiopian Relief
Organization (the counteq>ait to REST woridng alongside the EPRDF in
Wollo and Gkmder) were bnined. On December 27, Woldiya was bombed
again. Both these attads disrupted vdief distributions.
However, on the whole the JRP had the additional benefit (tf bringing
tranquillity to tiie people in its vicinity. Neither side lanndied significant
military action on the Dessie front for almost one year after the program
started. The bombing attacks along the JRP route, though prominent
because well-documented, were much less frequent than elsewhere. This
allowed people to travel and work in the day-time and markets to meet
during the daylight. Asemblance of normality returned to the towns along
the main road.
Problems with the JRP intensified in February 1991, when the EPRDF
knmched Opciation Teodfos. Though the fighting wis confined to Ponder
and Oojjani» and so did not affect tfie enviioos of M JRP fonte, govenunent
interference intensified. On Maicfa 12/13> the govemment launched an
attack on the EPRDF--held town of Wicfaide, the first rebel garrison on
the JRP route. This held vp a food convoy. On March 18/19, the EPRDF
counter-attacked on the west side of Dessie. Though this did not endanger
the JRP route, the government decided to halt the JRP at once. (In fact,
the timing of the decision suggests that it was made before the EPRDF
attack was launched.) On March 20, the EPRDF issued an ultimatum that
the program should restart within one week. Three days later the
government complied.
In early April, the government detained seven drivers working for the
JRP. This may have been related to an attempt to requisition their vehicles
for m ilitary use. This immediately led to other driven, who were in
EPRDF-held areas, refusing to return to Dessie for fear of arrest. Fearing
the halting of the program, the EPRDF also made die "release" of trucks
conditional on the arrival of new relief convoys — it was attempting to
hold some vehicles as '^collateral'* to ensure the continuation of the program,

but this served to slow down the rate of delivery.
These disruptions meant that by May the JRP was delivering relief to
only one quarter of its intended beneficiaries.
On May 16, it was the turn of the EPRDF to halt the program, as it
launched its Operation Wallelign to capture Dessie and Kombolcha. The
EPRDF daimed that the government had already halted the deliveries before
the attack was launched, but the sequence of events is not clear. The final
281
days of Mengistu and the followiqg week of Geneial Tesfiiye Gebie Kidan's
rule saw no further deliveries.
Though it takes up most of this account, the story of the relief programs
of 1989-91 is in fact relatively marginal to the story of how rural people
succeeded in surviving the drought of 1989-91. The main components
of that survival were the absence of ground war and the absence of
restrictions on trade and movement, which enabled the eoonomy to fanction
in an integrated manner. The absence of outright fimiine, despite mote
severe drought than in 1983-5, serves as witness to the ftct tfaat drought
need not create famine, and that the reason why the appallmg fiuime
developed in 1983-5 was not because of the weather.
Famine In Eritrea
In 1989, most of Eritrea was controlled by the EPLF, including, from

February 1990 onwards, the port of Massawa. This created a mirror-image
of the situation in Tigray: relief for the government -held enclave around
Asmara could only be brought in either with the agreement of the EPLF,
or by using an expensive airlift.
Famine in Eritrea during 1989-91 contrasted with TSgiay. The siege
of Asmara led to famine conditions developing In the dty of Asmara and
the surrounding enclave. Throughout the 1980s, Eritrea had always been
more dependent on food aid than Tigray and norlAi WoUo. The substantial
cutback in relief aid that coincided with the siege was therefore more
serious. In addition, restrictions on the commercial movement of food and
requisitioning by the army garrison played an important role in creating
famine in the enclave. Government tactics of enforcing a food blockade
arc familiar from the first siege of Asmara, in 1975. Finally, western Eritrea
is economically integrated into eastern Sudan, and the unprecedented
shortages and famine in eastern Sudan from mid- 1990 onwards a^ravated
the problems caused by drought and war. Therefore the fnnine m
Eritrea
during 1989-91 has proved to be more severe than duiiqg 1983-5.
Famine in the Asmara Enclave
Eritrea is, even in normal times, a food deficit area. For the city of
Asmara and the surrounding area, there is an even greater relative food
deficit. The 1.1 million civilians in Asmara and the surrounding area which
was the government-controlled enclave would normally consume about
15,000 tons of food per month. Usually, the great majority of that is
imported, either by traders operating in western Tigray and Gender, or by
282
government marketing organizations. Since the mid~1980s, regular delivery
of food relief has also been important.
The disruption of all the supplies of food simultaneously in February
1990 brought Asmara very quickly to the brink of famine.
Before the fall of Massawa, grain was cheap in Asmara. Shortly after
the siege began, the price of grain in Asmara rose more than ten times to
over 700 Birr per quintal. This was fat mm than the great majority of
the residents could afford, and made famine inevitable. Nonnally, the hire
of profits would have brought grain traders to Asmara, paying bribes to
cms tihe battle lines, hi the first eight months of the siege» this hardly
happened. The army pievented large quantities of commercial grain
reaching the city. There were instances in which grain was confiscated
by soldiers when people tried to bring it in. Residents who were caught
with grain traded from EPLF-controlled areas were regularly detained and
punished. There were even instances in which people travelling by air from
Addis Ababa to Asmara, and bringing food with them for their relatives,
had this food unloaded from the airplane at Addis airport.
The policy was partly dictated by the increased bribes that soldiers could
charge because of tfie grain scarcity, and partly by a deltberale plan to make
the civilian population of the town softer.
From February until October, the army's ban on firee movement of
commercial food into the endave, together with a ban on free movement
of people out of the enclave, was the single most important reason for the
hunger affcctiqg the civilian population.
In October, in recognition of the severity of the food situation, the
administration lifted the ban on free movement of food. The normal checks
on traders continued, and bribery remained rife, but punishments were no
longer meted out to those found in possession of food from outside the
enclave. General Tesfaye Gebre Kidan, the Overall Administrator, justified
this change in policy. He told a meeting of Asmara residents that
government relief supplies had in the past ended up feeding the rebels;
now it was the tnm of the rebels to feed die government. vSs pragmatic
policy did not end the fauqger, and nor did it reduce tihe price of grain by
very mnch, but it prevented the famine in the dty from developiog into
mass starvation, as had appeared inevitable.
One factor that contributed to the famine in Asmara was uicreasing
unemployment. The siege led to many enterprises being cut off from their
suppliers or their markets, so they were forced to lay off workers. Some
government -owned industries were also closed down, dismantled and
relocated in Addis Ababa.
Requisitioning by the Army
The army in Asmara and the surrounding areas often requisitioned food
from the residents. This was probably the second most important cause
of the scarcity. While some of the requisitioning was looting by
undisciplined soldiers, much of it was G^tainly based oa directives from
the military command. The wheat mflitias ceased to be paid their latioo
and instead turned to looting.
The rcquisitioniitg of food and other commodite and lemovd of peojde
was based on powers given to the mflitary authorities under State of
Emergency legislation.
The impact of the army^ requisitionii^ is conveyed in a tetter from
a civilian in Asmara:
March 15th 1991

Dear [brother],
Asmara has become a living hell and I can't see how we are going
to survive for long.
The food consignment from Massawa is unreliable and inadequate.^

Worse still is what little we get from the UN and the churches is stolen
by the town boys [i.e. government soldiers] at night. In the Edaga
Hamus area many families including mine have been broken into by
hungry soldiers. This happened to us twice during February when three
armed soldiers broke into our house and took away our two week ration
of flour, sugar and oil. Five of my friends in other parts of Asmara
told me of similar incidents. This is happening all the time.
My brother, we are facing a slow, tenjUe, undigpiiOed deatfal

Asmara is now dying Sadly, Ker [Keren] is also in a stale worse than
Asmara ...
^>art from the instances mentioned in this tetter, some incidents of

requisitioning included:
* The confiscation of half of the food stocks of an orphanage in

Decamhare;
* The shipments were in fact arriving

regularly but the government did not
inform the citizens of Asmara of that fact.
284
* Hie coiifiscatioD of large numbers of cattle and other livestock in at
least four villages in Akek Guzai district and the aiea sonoiinding
Decamhaie;
* The people of Senhit were required to feed the garrison there;
* In the area of Chendek (near Keren) there was an order prohibiting

people from picking beles, a cactus fruit eaten as a famine food in times
of shortage, so that it was resetved for the soldiers;
* ktlieAdiTedezBnaiea,iioithof Asiiiaia,ttier6sideDlSQf tenviU

were moved by military order and all their property and food stocks
were then teqpMdooBd, A
similar instance ooconed at Adi Beyane.
Fortunately, the supply of food and money to the garrison (which

numbered about 120,000 men) never broke down entirely, so that the army
did not have to depend completely on requisitioned food.
The military also requisitioned other scarce items, such as fuel and
medicine. There were three military hospitals in the city, with 6,000 beds.
Because these were too overcrowded, the army also requisitioned half of
the 820 beds in the civilian general hospital, and with them, half of the
time and equipment of the hoq)ita] staff. Two dviliaD dodois and several
muses and auxiliaries were also required to serve in the sulitary hospitals.
Much of the supply of dragis to the hospital was often bought by the
military; it was paid for, but the hospital could not obtain replaconent
supplies.
Food Relief through Massawa
When the EPLF captured Massawa on 8-11 February, 1990, the port
was closed to relief shipments. A month later, the EPLF announced that
the port had been repaired and appealed to the UNsupply relief. The
to
Ethiopian government shunned this offer and repeatedly bombed the town,
killing over 100 people, destroying many bufldings, and bumiqg about
25,000 tons of food aid. A
ship chartered 1^ the Geittian relief organization
Cap Anamur was sent towards Massawa loaded with food, but on May
1 the government tfueatened to destroy it with aerial bombardment. Despite
diplomatic pressure from the western aid donors, the government did not
yield and the ship was diverted to Port Sudan, from where its cargo was
transported across the border by ERA.
At the Washington Summit in early June 1990, Presidents Bush and
Gorbachev discussed the issue of Massawa and called upon Ethiopia to
285
allow the port to reopen. The Ethiopian govemmeBt complied aad ceased
the bombing raids two days later.
This was followed, in July, by an attempt to open Massawa. A ship
chartered by the World Food Programme (WFP) with four port technicians
sailed from Djibouti towards Massawa on July 14. Their mission was to
assess and repair the port facilities so that large shipments could be handled.
Several days of negotiation Moiwed, concemiqg « m
cooditioiM ndcriiîicfa
the technidaBs would enter itc port. No ayeement was readied. On July
21, the ship abandoned its mission and returned to DJibontL
Tht foilure of the WFP mission appears to have been the result of too
many adois playiqg different roles in the whole process^ each with a
different agenda, and inadequate communication between them. The
Ethiopian government allowed the mission to go ahead at a moment when
it suited its diplomatic and military needs. Thus it coincided with a visit
to Washington by Foreign Minister Tesfaye Dinka, with the intention of
obtaining US blessing for Israeli arms deliveries to Ethiopia (the resumption
of the emigration of the Ethiopian Jews was announced at the same time).
The Ethiopian government also tried to impose conditions on the mission
which it knew the EPLF was unwilling to accept, such as the ship remaining
in radio communicadon with Addis Ababa. TlieWFPwasanzknistosend
its tiyJinid«n» to die port as quick^ as poasibk. The BPLF wm
communicating at dififeient times with the WF? in Rome and USAID in
Washington; the latter not only wanted the technicians to land in the port,
but the ship to dock there as well, in order to make the political point that
this was possible. The EPLF also demanded full information on the
consignees of any relief grain that arrived in the port after it was opened -
- it did not want to lose total control of the operation to the UN, nor see
the grain consigned to the RRC in Asmara. The mission was aborted when
an EPLF demand that the WFP technicians be allowed to land in a small
EPLF boat was not passed on to WFP, but was rejected by USAID. WFP -
-which might weU nave aooqiled this demand —
bdieveddwtftemisrion
had been rejected outright, and turned tlie ship loond.
Claims made in the press at the time, for mstanoe tfiat the EPLF had
turned back two ships canyiqg gnun, were onfounded —
the additional
ships did not exist.^
Despite the agreement in principle to open the port, the government
bombed it again on two occasions in September and October, to demonstrate
that it still reserved the final decision.
* New York Times, July 22 and 25, 1990.
286
After proloqged negotiation, and the near-complete absence of further
media attention, on January 18, 1991, Massawa finally opened for relief
shipments. Over the following four months a ship chartered by WFP, the
Far Suez, made six round trips from Djibouti and brought over 60,000 MT
of grain. This was divided into halves, with one half allocated to ERA
for distribution in areas, and the other half to the RRC and
EPLF-held
the churches in the government -held enclave.
Despite skirmishing on the Asmara-Massawa road (initiated by the
government), there were no interruptions to the supply to Asmara.
Hie difificnlties started when the grain anived in Amara. 6,000 MT
of the fiist shipment was distributed by dmiches, but the larger quantity
consigned to tfie RRQ over 20,000 MT, remained in stock for several
months. The excuse given by the administration for the lade of distribution
was lack of fuel, but this was extremely lame because fuel supplies
continued to arrive regularly until March 22, and much of the food was
due to be distributed in Asmara city anyway, where the horse had become
the preferred mode of transport. The distribution only started belatedly
in May. It is likely that the administration was keeping the food as a
reserve to distribute to the military in case the siege was tightened further.
Airlift of Relief
From May until January, the only relief supplies that arrived in Asmara
came by air. Like the supply through Massawa, the airiift was subject to
and manipulation.
political controversy
Hie Ethiopian govemment proposed a relief aidift to Asmara in February
1990. In March, the UN came up with a plan acceptable to the Ethiopian
govemment for an airlift from Assab and Djibouti. The donors, however,
were unhappy with the proposal that the food should be distributed by the
RRC, fearing diversion to the military. Instead, they insisted that the food
be consigned to the Ethiopian Catholic Secretariat (ECS). The airlift began
on May 3, using two chartered Hercules planes flying from Assab and
making four trips each per day. This allowed the delivery of a maximum
of 4300 MT
per month, considerably less than the 8,000 MT
that ECS
estimated that it needed to provide hidf latioiis for the 1.1 millioii people
in need. In addition, due to shelling of the aiiport, bad weather, and the
need for maintenance work, the UN
airlift was able to deliver less than
the maximum figure: by the end of February it had delivered 36,000 tons,
or 75% of maximum capacity. A
plan to airlift food using Soviet transport
planes was proposed in June but never materialized.
287
The Ethiopian gpvenuneiit had 26 dvfl and 16 mflitary aiioaft availabfe
for tran^rting food. Not one of Hhese aiicraft was ever employed for
this tasJc.
The EPLF warned against the airlift, and later renewed its "warning
to those quarters which are involved in military intervention by
still
transporting military materiel to the Dergue government pretending it is

relief and humanitarian aid.""^ It warned that the airlift would continue
at its own risk. This referred primarily to the delivery of supplies by

Ethiopian Airlines planes and the proposed airlift using Soviet aircraft.
(Soviet-supplied Antonov transports were also airlifting military equipment
to Asmara.) The EPLF also said that an aidift was unnecessary because
food could be delivmd to Asmara thnmgh Massawa or ovoinid fiom
Sudan, using the icutes used by ERA.
The suggestion that the UN
relief airlift was transporting military
supplies was untrue. However, the civilian and mflitaiy airports in Asmara
are not wholly separate, but share the same runway and other facilities.
The airport was also used for the airlift of military supplies (usually ten
flights each day) and was the base for bombing missions by MiG fighter-
bombers. It could thus be considered a legitimate military target.
In the event, an implicit understanding was reached whereby the EPLF
did not shell the airport while the relief flights were arriving or on the
ground. The EPLF was always aware of the timing of the arrival of the
relief flights. This inevitably gave immnnity to attack to any militaiy
aircraft that happened to land at the same time. However, the EPLF dd
continue to shell the airport at other times, with the intention of patting
it out of action.
On March 1, an airplane used by the UN-ECS airlift was struck by

a shell and one employee of ECS was killed. This brought the airlift to
an end. It is not clear whether this incident was an accidental violation
of the tacit agreement, or whether, after the reopening of Massawa, any
such agreement had lapsed. The EPLF certainly argued that the airlift was
no longer necessary after January because of the opening of Massawa.
However, the food supplied through Massawa remained inadequate for the
needs of the Asmara enclave, just as it was not enough to feed the needy
people in EPLF-controUed areas.
The shelling of the relief airplane and the haltiqg of the ahrlift again
demonstrates the complexity of flie ethical issues involved in determining
when starvation is bemg used as a weqxMi. If the actual incident of shelliqg
^ Voice of the Broad Masses of Eritrea, BBC Sunmuiry cf World l^vadcosts

(Sm), ME 0806, July 3, 1990.
288
was an aoddeirt, it was part of a moie geneinl atto^
the aiipoft* which would have stopped the airlift anyway. If any tacit
agreement not to fire when relief airplanes were in the vicinity had lapsed,
it had done so in the context of two developments: the suf^ly of food from
Massawa (which was not for the most part being distributed) and the
imminent military collapse of the government, which was hastened by the
intensified bombardment of the airport.
Relief in EPLF-Held Eritrea
Both 1989 and 1990 were years of severe drought hi EPLF-heM areas
ofEritrea. OMobmed with the eoononiic dedhie of the two mam soikg^
of enqiloyment and maiketed goods —
Asmara and eastern Sudan this —
spelled famine.
ERA continued its own distributions in the areas of Eritrea controlled
by the EPLF. In 1990, over 100,000 MT of grain was transported across
the border from Sudan and donated to people affected by drought and war:
a far larger amount than that distributed in the enclave. This relief was
allocated to needy people according to lists drawn up by village committees.
A spokesman for ERA conmiented that some of this food later found its
way into Asmara, as recipients gave help to their needy relatives: "this food
is always transferred from one hand to anottier at the end of the day.'* ERA
offered to set iq> distiilmtion centers m
EPLF areas, to which the residents
of Asmara coiiJd come and collect a ration. Implementmg this proposal
would have required a massive extra donation of resources to ISBLA^ but
in any case the govonment did not respond.
The Ethiopian government remained resolutely opposed to the
humanitarian activities of ERA: "any attempt to supply food aid across
the border is contrary to the sovereignty of the country ... and has absolutely
no acceptance by the Ethiopian government."^ ERA food convoys were
still subject to aerial bombardment. On January
5, 1990, at Tserona, an
ERA food convoy was bombed. On September near Tessenei, one truck
3,
carrying relief was burned in an air attack. On May 10, 1991, at Tekombia,
Barica,two trucks carrying relief were damaged in an air raid.
Shortages hi the Soath
In 1990/91, there were also shortages in a number of parts of southern

Ethiopia, such as Hareigbe,GamuGofe and parts of southern Shewa. These
^ Voice of Ethiopia, Addis Ababa; see BBC SWB, ME 0704, March 5, 1990.
289
occurred despite the fKt that the abandonmciit of Mandst eoooomic policies
in March 1990 and relatively good rainfiBU had contributed to a national
bumper crop. These localized shortages can be attributed to a number of
factors.
The most important factor is the legacy of the previous decade of
unremitting attacks on the economic base of the peasantry. The legacy
of villagization and other disastrous policies left many rural people
extremely vulnerable.
A second factor is that when the "change in direction" was announced
inMarch 1990, the Agricultural Marketing Corporation's procurement for
1989/90 was already half complete, so ttiat many areas were damaged by
the old policies. A related factor is the incomplete liberalizatioii; for
example there was no deregulation of motorized transport, whidi meant
that farmers who lived more than a day or two's walk from the nearest urban
market did not benefit from the regulated market in grain and other
foodstuffs. In addition, the "change in direction" meant that provincial
and district administrators had more local autonomy in decision making,
and some enforced hard-line restrictions on trade and other activities.
The area in which human rights abuse most directly impinged upon
food production was conscription. The conscription campaigns of 1989-91
were unprecedented in their size and scope (see dbaptoi 17). Not only were
large numbers of young men taken from tfieir homes and fanns to serve
in the army, but the fear of conscription forced them to be in a state of
readiness to flee to the hills at any sign of soldiers or government ofiBdals.
Though it has not been investigated fully, it seems likely that some of the
areas of greatest shortage in 1990/91 were areas in which the conscription
campaign was conducted most extensively.
A final factor contributing to food shortages was the temporary
breakdown in law and order in some parts of the country following the
collapse of the Mengistu government. Retreating soldiers sold or abandoned
their guns. In areas such as Wollaita (southern Shcwa) and Harerghe, the
opportunity was taken for looting, which contributed to an interruption of
relief programs. For example, alwut 800 MT of relief food was looted
from a relief agency store in Shashamane, southern Shewa, in tiie days
after the government fell. The food crises a^ctii^ Sudanese refugees in
Gambela and refugees and returnees in Harerghe will be examined in
chapters 18 and 19.
290
17. THE RAGE OF NUMBERS: MENGISTU*S SOLDIERS
Oatside the areas of anned conflict, the mam way in wfaicli the wars
had a direct impact on the lives of ordinary Ethiopians was through
conscription to the aimy. Throughout the 1980s, in the towns and villages
of Ethiopia, the talk was not of defendii^ the nation, but of who had been
taken for the army.
After the revolution, the main strategy of the Ethiopian army was sheer
force of numbers, in both men and material. Mengistu was obsessed with
what the 19th century military strategist von Clausewitz called "the rage
of numbers." He built the largest military machine in sub-Saharan Africa.
Numbering about 50,000 at the time of the revolution, the army included
about 450,000 regular soldiers and militiamen in early 1991. The majority
of tfwse soldieis were conscripts. Many of tiie metfiods of cooscri|ition
violated the basic human rights of the conscripts,^ and their treatment whfle
in the army also involved many abuses.
Violations included the conscription of boys under the age of 18, in
contravention of Ethiopian law, and under 15, in contravention of the
internationally-recognized rights of the child. Conscripts were also
commonly taken an arbitrary and violent manner, without warning or
in
the chance of communicating with their families. Once conscripted, the
recruits were then subject to ill-treatment.
Conscription, 1976-S2
Mass
conscription to the Etliiopian armed forces began in 1976, for
tfie Mareh" on Eritrea. In April 1977, the "Call of the
"Peasants'
Motherland" was issued to nise recnuts te the "peasants^ miUtia,
on the north and Eritrea. In August this militia was diverted to face the
Somali aimy in the Ogaden. In what was to become the normal procedure,
each Peasants* Association (PA) or Urban Dwellers' Association (kebele)
was given a quota of recruits which it had to provide. After the offensives
against Somalia and the Eritrean fronts, most of this peasants' militia force
was not demobilized, and became in effect part of the regular army.
Over the following years, a variety of measures were used to obtain
conscripts. Most of these measures were never formalized, but were
implemented by administrative fiat. Possibly the most conmion was the
For a more detailed analysis of human rights abuses during conscription,

see News from Africa Watch, June
1, 1990, "Ethiopia: Conscription, Abuses of
Human Rights Durii^ RecndtmcDt to the Aimed moes."
291
rounding up of young men in villages and marketplaces in the south, usually
during military operations or forced relocations. Conscription to the army
was often cited as a reason for flight by refugees from Oromo areas in both
southeast and southwest Ethiopia. Others were picked up in one and twos,
mostly in Tigray and north Wollo, usually when they were away from home.
Journalist Dan Connell spoke to some conscript soldiers had been
captured by the EPLF:
One middle-aged tuwet torn Tigray^ Endeila diittict said that a

representative of the Derg had demanded five "volunleeiB" fiom each
village in his region. The appointed head of his Peasant Associatioii
had selected him to go. Several others said they had been told they
were going to a political rally and would be brought home afterward.
One said that peasants who had been resisting the Derg's heavy taxation
were told that they would be pardoned if they turned themselves in.
Those who did were put in trucks and sent north. A 42-year old peasant
from Woldiya in Wollo said that he was walking towards a coffee house
in town when he was forced into a police wagon and later transfened
to a tnick which earned him to the fiiODt Aaolher explaiaed Aat he
had been haWog his pants mended by a local tdlor wheo he was
shanighaied by an army polrol, and stood up to tfiow that all he had
to wear was a burlap sadk aroand his waist, became he had not beeo
allowed to wait for his panls.^
National Service
government announced plans for the organization of the

In 1981, the
"entireworking people into a national military service and civil defense."*
This was formalized by the National Military Service Proclamation of May
1983,^ which provided for the conscription of all men aged between 18
and 30. On reaching the age of 16, all young meo wne to i^gisler with
their PA or kebeU for "pre-faidiictian tnining" prior to natioiial servfce
at age 18.^ Then Ifaey were to ondeitakB six months mOiiaiy tahuBg
' Dm CoDDelU "Rcpieiiiai as a Way of Life," Horn of Africa, 3.2, (1980)

p. 15.
^ MeogiBta HaUe Mariam, May Day addien. May 1, 1981.
' Decree no. 236 of 1983, May 4, 1983.
' Tbc pre-induction training app^s never to have been implemented.
292
followed by two years' service in the armed forces, remaining on reserve
until age 50. The Proclamation also specified the duty to remain in service,
even after the end of the normal period, in times of mobilization and war.
The conscripts served either in the regular army or in the civil defense
units, generally known as "people's militia." The people's militia were
enlisted to aecve in thdr home areas and did not generally have combat
dntieSto except fot those in Eritiea. Eight new tramiqg camps wece built
for the natioiial servicemen, with Soviet and Libyan help.
R^pdar louids of oooscriptkm ooconed after 1983. OfHcial figures
for the numbers of conscripts were never published, but reliable estimates
have been made. The first campaign was carried out between January and
April 1984. Its target was 60,000 men but it is likely that only 50,000
were actually conscripted. After these recruits completed their six months'
training, a second batch of the same size was recruited in January 1985.
The third batch, starting in December 1985, had a target of 120,000.
Subsequent campaigns usually had targets of 60,000-80,000 recruits. The
fourth campaign was recruited between November 1986 and January 1987;
the fifth between November 1987 and January 1988.
Fdlowuig the EPLF victory at Afobet hi March 1988 and later rebel
successes^ cooscription hitensified. A
sixth round was implemented
immediatdy, under the dogan "eveiytfiiqg to die waifiront." A
huge part
of this campaign was the re-moUlizatioo of men who had served in the
first and second national military seivioe intakes of 1984-5, and wlio
remained on reserve. In practice the re-conscription net was thrown wider,
and ex-servicemen from other intakes were also taken. Servicemen due
for demobilization also had their length of service extended indefinitely.
Financial contributions for the war were also solicited from the general
public —one month's pay was deducted from the salaries of public sector
employees, and new levies were raised from the peasants.
Following the TPLF capture of Tigray in February 1989, there was yet
another round. In the by-now fomiliar code for the launching of a
conscription campaign, Ethiopkn radio annomioed that on March 26 each
of die 284 Melaf m Addis Ababa held a meetmg "00 the possibility of
translating into deeds the slogan *L6t Eveiyone be Vigilant to Safeguard
the Homeland'... [and] the residents reiterated that they would contribute
their share - from preparing provisions to strengtheniqg the revolutionary
army, falling at the front and other spheres."^
The first batch of national servicemen was demobilized on schedule
in November 1986. The second batch was also demobilized. Both these
• BBC, Summary of World Broadcasts (SWB), ME 0420, March 29, 1989.
293
intakes were called upon to remobilize in April 1988. The third batch,
due for demobilization in July 1988, was never demobilized, which was
also the fate of subsequent intakes.
Following the TPLF advance in Wollo province in September 1989,
another round of conscription — at least the eighth since 1983 — was
begun. This was intensified in November following further military disasters.
In a speech on October 31, President Mengistu said:
The main strategy for defendiQg our coiintry is the proper strengtfaeniog
of our revolutionary army with manpower and material so that it can
meet its obligations. There should be a vast popular participation which
must be well-coordinated and mobilized for the decisive victory ... In
defending ourselves against the danger foisted on us and overcomii^
it, manpower is the decisive factor.
This speech heralded one of the largest and most brutal rounds of
conscription. A national campaign center was set up on November 10,
headed by Vice-President Fisaha Desta. Two days later each of the kebeles
in Addis Ababa held a meeting Id nominate a committee of people who
would be mandated to find the required quota of oonscripls. Conscription
carried on through November and December. Ethiopian radio carried
regular leports of "many tfaousoids of volunteeis" amviog at tiaining camps,
singing patriotic songs and expressing their eagerness to go to the front
for the "decisive victory" against the "anti-unity, anti-revolutionary, anti-
people bandit gangs." There was a lull in the New Year, but after the EPLF
capture of the port of Massawa in February 1990, conscription restarted,
especially in the rural areas. All other social and economic programs were
made to take second priority to the strengthening of the armed forces. Over
100,000 conscripts were taken in the year to June 1990.
Worse was to come. In his 1990 May Day address. President Mengistu
called upon the people to make sacrifices for the government, saying "I
do not think the people should expect miracles firom the limited number
of patriots, revolutionary army, and few militias in the vangDan)."* In
June, the National Shengo (assembly) called for "non-stop recruitment"
to the armed forces to be undertaken urgently. This included the
mobilization of retired army and police personnel to serve in the armed
forces, and the encouragement of civilians in the war zones to engage in
' BBC, SWB, ME 0602, November 1, 1989.
• BBC, SWB, ME 0754, May 3, 1990.
294
were urged
guerrilla warfare against the insurgents. In addition, all citizens
to make war effort. Above all, it meant the
financial contributions to the
constant threat of press-gaagioig, with no "safe" periods between discrete
campaigns.
The stepped-up mobilization included veterans, taken to mean anyone
who had prior experience in the army or other uniformed services. In 1991
it was expanded to indadesMcmdar^rac^ All
Ethiqiia's institutions of higher educatioo weie closed and die students were
made subject to conscription. Students at Addis Ababa univeisity were
required to register for military service, under the implicit threat of having
their education terminated if tiiey refused. With the university closed, many
students (particularly those without relatives in Addis Ababa) were also
left without means of support, and had little alternative but to join the army.
Apparently no other punitive measures were used, and some students were
sufficiently fired by the patriotic call to arms made by the government to
have readily volunteered to fight. The great majority — over 80 per cent -
- are estimated to have registered. Of these, about one quarter actually
went for military training. The others, realizing the harsh conditions and
dangers they would face at the front, and under pressure from parents and
friends, mostly went into hiding. Outside Addis Ababa, most students were
sim|dy rounded up and taken off to trainiqg camps.
Addressiqg the nation on April 19, 1991, President Mengistn called
fot an even greater mobilization to defeat the insurgents, and called for
an army equal to that of Iraq one million men. Subsequently, the
Le.
Shengo called more than ever before" and authorized
for "mobilization
the recruitment of all able-bodied adult males, using all means available.
Fortunately, the government had neither the time nor means to implement
this ambition, and within six weeks the army had disintegrated.
Conscription of Under-age Children
There were many instances in wbkh children younger even than the
de facto minnnum age of 15 years weie ooosoipled into the aimy.
Joumdist Tom Lansner visited EPLF-conbolled Eritrea in May-June 19^
and found 50 boys aged under 16 in a prison^-of-war camp <n 1,500 totel.
One 14-year old, Thebether Sawra, described how he had been taken by
three militiamen while playing football in a neighborhood alley the previous
January. "I told them 1 was 14 but they didn't say anything," he said.
295
Another 14-year old had also been snatched at a football game, and a thiid
had been taken while attending a village meeting in Bale province.^
The government consistently denied that it was conscripting under-age
children, and failed to respond to protests made by Save the Children.
The army commonly defended the conscription of under-age children
on the grounds that these boys served as "aides" in the military camps,
as messengers and the like, and did oot have combat duties. Even if true -
- and there is plenty of evidcooe that Ihey did Indeed engage Id oonbn -
*- this ^vonld have been no jullfiGadan. Sndiio-ealled "iSdes" hid been
abducted and lost thek liberty, and were subject to many of thesane ifgm
and dangers of life at Ae wufipont as combat soldieis.
Methods of Conscription
was compulsory in principle, but the government lacked

National service
the means implement this. The comprehensive conscription of all young
to
men aged 18 would imply about 350,000 recruits annually; which was
beyond the capacity of even the Ethiopian army to absorb. Instead,
conscription was selective, using a variety of methods of obtain the required
nuniber of young men aad bo^
m
One of flie com oneit nettiods of oooscriplioa was Huaoffk liie PAs
and the uiban keMes. Each PA and kdfde wis set a quota of peoptei
which they were to fill by whatever means they chose. This system was
notoriously open to abuse by individuals. PA or kebele officials rarely
conscripted their relatives or friends, and used the conscript quota as a way
of settling grudges, obtaining sexual favours from the wives and sisters
of those they chose to detain, or —most commonly —
soliciting bribes
from conscripts' families. Paying these bribes was a heavy burden on the
poor. As well as PA and kebele officials, others could demand payment.
At the thne of conscription, people sold as nmchM they could — food

and ammals; they did not hm nnich, as It was a finnine-prone, fbod-
deficit area.... Hie people used the money pay tebribes to get Ibeir
to
sons released. When men and boys were taken as conscripta, they were
first of all put in camps, in the area where they weie taken, before being
given medical examinations, there were two opportunities for families

to get their sons retened. Ibe first was to pay oiBdals so that te son
' The (^server, London, June 5, 1988.
296
failed the medical examination. The other way was to bribe the military
guards who were on at night, to let their sons out.^°
The PA or kebele could also nominate other government servants, such

as health workers or agricultural extension workers, to be responsible for
collecting the quota. The nominated individuals had no choice but to take
00 this unpopular job. Hie people's mflitia wm
oomnionly giveo the task
of lecniitii^ and mflitiameii themselves weie likely to be conscripted if
they fuQed to fill the quota. PAs and ibebefes also lesoited to pidong up
stiaQgecs and vagiants from the streets in order to fill their quotas. On
one occasion they conscripted a Nigerian visitor to Ethiopia named Ibrahim
Garba.^^
The quota system became deeply unpopular with kebele officials and
party cadres. As a result, other methods of acquiring recruits became
conmion. One such method was based upon workplaces and schools. At
its simplest, factories and offices were allocated quotas in a similar manner
to kebeles. This was common in 1988. During 1989-91, methods became

progressively more arbitrary. Many workers in the public sector were
sunply detahied and thereby conscripted. One example was woriceis m
the construction andioad-builduig industries: inNovember and December
1989, almost aO male woriGcis in these industries were either taken for
military service by military police who stationed themselves at their
workplaces, or hid themselves and did not turn up for work. For a time
it was common to see only women workers on building sites. Workers
in the private sector were more secure; Ethiopian law requires that the
government compensate private-sector employers when their employees
are taken for military service.
The conscription of schoolboys followed a similar pattern: policemen
loitered near school entrances and detained pupils entering or leaving. Two
ninth-grade pupils at the Menelik II Senior Secondary School in Addis
Ababa disappeared Ghcmnstancesm September 1989, without the
hi these
knowledge of their families and friends. They are believed to have been
fofcibly conscripted. During the foUowii^ months many pupils only went
to school when they could be accompanied there and bade by their parents.
Others wnpky stayed away, but any pupil who was absent for 21
^ Refugee from Wollaita, Shews, inteiviewed m Sudan, October 1989.
He spoke no Ethiopian languages and could communicate with nobody.

He was later captured by the EPLF, who tried to negotiate his release through
the Nigerian Embassy in KhartouoL "Shocker from Eôpia,** Eritrea Infonnaiian,
ZIO, November 1980.
297
consecutive days without a doctor's note was automatically expelled.
Instances of forcible conscription also occurred at Addis Ababa university.
A foreign medical team visiting TPLF-controlled Tigray in 1988 spoke
to two university students who had been conscripted and subsequently
captured by the TPLF, and conscription on the campus occurred in late
1989, prior to the main attempt to mobilize students in 1991.
Prisoners
Prisonen were very vulnefaUe to ocxncriptioa. AMol Watch has

interviewed Getachew, who was conscripted in Much 1990 ate being
detained while trying to flee the country. Tedgai, a oomcript from south
Gonder captured by the TPLF, told a visitor to Tigray in late 1988 that
"some soldiers put me in prison for ten days. They told me I could get
out if I joined the army, so I agreed. "^^ Soldiers, cadres, and kebele
were notorious for detaining people without charge or on trumped-
officials
up charges, so that many of those conscripted in this manner had not
committed any offense.
Anybody suspected of an offense, however minor, was at risk from
conscription. Tliis wisapedaUy tnieif heweieciiig^c^
area, so that his lektives and firicDcb weie not there to
A visitor EPLF-contiolled Eritrea
to m
1984 met a prisoner of war who
originated from Tigray. He had gone to visit his grandmother in a
neighboring village, but had no travel permit from his PA. In his
grandmother's village he was detained and conscripted. At the time he
was eleven years old.
Press Ganging
An extreme version of forcible conscription is the press gang. This

is oftenknown in Amharic as afesoy which translated as "sweeping up",
and might be termed the "vacnimi cleaner" approach to leonitment A
group of armed policemen or party cadres wonM roam tiie streets and
marketplaces, piddng up any individuals or nmnduig up any gronpa Ihey
come across. Alternatively they would surround an area and fofoe every
man and boy to sit down or stand against a wall, using the threat of opening
fire; all those considered eligible would then be forced on to a truck and
driven away. Young men and boys were conscripted while playing football
in alleyways, going to school or market, or attendii^ religious festivals
^ Gerry McCann, Ihe Guardian, London* February 20, 1989.
298
or football matches. Teenage boys who worked in the informal sector
selling cigarettes, matches, and lottoy tickets were a particular target. Many
of these boys were under age.
Press ganging mostly occurred in rural areas, but was common in Addis
Ababa during November and December 1989, April 1990, and from July
1990 until the fall of the government. The following incident of afesa,
which occuned on April 11» 199(V was described Giorgis, a busmessman:
I left my oCGce at 430 madoBt

in the afternoon, just as the [Meicato]
was bediming to close. There was a commotion in ttie second-hand
clothes section [of the market]. Some civilians were pushing the boys
who sell clothes from Dire Dawa, forcing them into one place. There
were also five or six policemen there. Suddenly these policemen got
out their guns and shouted at the boys to stop [stand still]. Then about
ten of the civilians - probably they were security men or [party] cadres
- got out guns too. They formed a circle about 20 meters across,
enclosing these boys, and shouted at them to sit down. There was a
lorry to take the boys away. I couldn't see how many were taken: I
had a boy of 14 from the office with me and I was frigUened for him
I was even frightened for myself too - so we disappeared from there
as soon as we saw the guns.
People were forced to resort to different ruses to escape afesa. As

people riding in cars were usually safe, employers, friends and relatives
with cars would pick up men and boys from school, university or place
of work when they heard that there was a danger of afesa. If there was
an afesa in a neighborhood, local women would patrol near the area and
warn men and boys to stay away, or give them gabis (shawls) to hide under
to disguise themselves as women. If all failed, people would search for
a hiding place in a nearby house. One student escaped a press gang by
hiding in the back room of a local bar; asked why he did not use the
adjacent phone booths to telephone for his father to pick him u]^, the student
replied that "I would have been conscripted wfafle standing in the phone
booth."
There were numerous instances in which people trying to resist or esci^
press-ganging were summarily killed by the conscriptors.
Food Aid as an £nticement to £nlist
A variant method of conscription used in drought-stricken rural areas

involved the use of food aid. Since the large-scale provision of relief food
to Ethiopia in 1984, it was common for tfie Relief and Rehabflitation
299
Commission (RRC) to withhold relief food from villages which failed to
meet quotas for conscription.
their A
more dramatic abuse involved enticing
rural people to come to towns to receive a distribution of relief food —
and then conscripting the young men.
One of numerous examples of this occurred in Senafe, Eritrea, in January
1990. Senafe was then the most southerly government-controlled outpost
in Eritiea» and was smioimded by nnal vmagea controlled by the EPLF.
FoUowiqg the fofluie of the nuns in this aiea hi 1989, faitenwdaaal aid
donors provided relief food to the RRQto distiibiite to peqde hi Ibe
drought-affected areas. In early January the RRC began such a distribntkm
at Senafe. On the first day, only women, children and old people came
forward from the villages to collect the rations. They were given their food
and allowed to return to the villages in safety. Assuming that it was safe,
young men came for their rations on the following day. At least 600 were
promptly seized by the army for the local people's militia. "A United
Nations monitor stood by helplessly ... as soldiers rounded up teenage boys
for military service" according to Jane Perlez, correspondent for the New
York Times}^ There are reports that some of those seized were transported
to Asmara and released, Imt visitors to Senafe reported seeing a large new
contingent of militiamen drilling just outside te town. Afinct Wateh
believes that many of the men seized on this occasion were foioed into
the people's militia.
Similarly, punitive measures were often taken against villages which
failed to provide conscripts, or against the families of conscripts who had
escaped. These included detention, beating, and the confiscation of assets
such as livestock.
Conscription of Women
There was no systematic attempt to conscript women into the armed
forces, though members of the govonment leiened witfi admiiation to
historical military campaigns in wiiich women participated. PAs, kebdes^
and press gangs did not usually take any women tauKdp^
However, there was de facto omscription of women. Women who lived
in a town with a large army garrison were at risk from the attentions of
the soldiers. Frequently they were harassed and raped. Many women in
garrison towns had no means of livelihood other than becoming the
^ New York Times, February 15, 1990.
300
ooncubines of aimy ofiBceis or piostitiites for the cammon soldieis.^
oil the Tigcayan town of Enda Selassie repocted:
Yomar, a womaii from I I
[the soldiers] did what they liked. They took the girls by force, even
married women. If you refused they would take out their pistol. They
would arrest your brother, and when you went to visit him in prison
they would ask you to sleep with them. In order to get your brother
released or to stop his murder you had to choose whether or not to give
your body.^^
The lepeated rape and other abuse of women is a serious violation of

linmanxjs^ Wlurt made these abases into a foon of conscription is
when the soldiers in question weie transferred, they sometimes insisted
on taking "their" women
with them. These women were not only denied
their libnty, and and economically, but were subject
a[|iloited sexually
to many of the same dai^gns as the soldiers, inchidipg shelling,
bombardment, and capture.
In 1983, there were several hundred women attached to the army garrison
at Tessenei, Eritrea. These women had been recruited from 1978 onwards,
ostensibly as "aides" to the soldiers, to cook, sweep, and wash clothes for
them. In fact their main purpose was as sexual servants. They were not
volunteers. While some had come with army units as they were posted
to the town, others had been specially flown in by military airplanes. Most
came from Hgray and Shewa. When the gsnison was caplmed by the
EPLF in January 1984, the women and their giowiqg band of cfaHdien were
left behind, and their quarts of the town was bombed by government
airplanes. In 1984, at the small front-line army outpost of Mersa Teklai
on the Red Sea coast, there were five involuntary female "aides" serving
the soldiers. In 1989, Jennie Street, a relief monitor visiting Meqele, the
capital of Tigraywhich had been recently abandoned by the government,
reported: "A mantold me that the army had forced both his daughters to
marry cadres, against their and his will, and that they had been taken to
Addis Ababa when the Dergue pulled out. He said many girls had been
forcibly taken in this way."
Ihese abuses continued until the fall of the Mengistu government. One
example comes from Senbele,mnQithem Shews. When an army battalion
Large numbers of single women in the towns of Tigray and north WoUo
needed letief aid Mowing their capture by TPLF-EPRDF.
Quoted in: Jenny Hammond, Sweeter ûm Honey: Testbmn^ cfTlgrayan

Women, Oxford, 1989, p. 150.
301
(of 600-800 men) was stationed at Sehbete, in early 1990, the local people
weie foiced to provide food for it. Protests that the local villages
themselves were short of food, on account of drought, went unheeded.
When the battalion left the area, the soldiers rounded up women and forced
them to accompany them — ostensibly as cooks and cleaners. Elsewhere
in northern Shewa, local women were forced to come and live at army
camps to cook food and serve the soldiers, abandoning their families for
long periods.
Sufferings of Conscripts' Families
On occasion, conscripts were not allowed to commnnicafe witli tlieir

families, who therefore had no way of learning of their fate. The acale
of conscription was such that more than ooe in four eligible young men
and boys were conscripted into the army and people's mUitia. Almost every
family was affected. Hundreds of thousands of families in Ethiopia had
sons who simply disappeared into the armed forces, and they had no way
of knowing if they were still serving, killed, captured, or maimed. The
Ethiopian government refused to recognize the existence of the tens of
thousands of prisoners captured by the rebel fronts.
The psychological effects of thii prolonged aepaiatiao could be
devastating for the fiunilies. Aieaeaicher studying woaen hi Addb Ababa
in 1988 found that 94 out of a sample of 113 women wm
suffering fram
what they described as "oppression of the soul" in — something
equivalent to chronic depression — and that 90 of these attributed the cause
to the fact that their husbands, brothers or sons were serving in the arnqft
often forcibly conscripted, and usually they had heard no news from
them.^'' This depressed psychological state in turn led to neglect of their
young children, who as a consequence suffered more £rom illnesses such
as diarrhoea.
Fears of conscription plagued those not directly affected. One Ethiopian
woman refugee illustrated some of these worries:
A finend of nune ... deliveied a baby boy at that time {1969]. When
she was congratulated she said "but IH only have fain for about 12 or
13 years, and then hell go to Meqgishi; if 1 had a ^d, I would have
her a bit longer."
Asticr M. Almedom, "Aspects of the Health and Growth of the Suckling

and Weanling Quid in Ethiopia,' DPhii thesis, Oxfiud, 1991. pp. 175-6.
302
Asociologist woikiQg in Manz, oortfaem Shewa, veooided aioqg which
expressed the same sadniBss:
The mother of a boy,

Tie your stomach with rope,^^
It will be a vulture
And not a relative, who will bury him.
Another song is an ironic comment on government propaganda:
Asthe ants swann

Tbebirds fly.
Woe is the child of Manz
He fought for his country.^'
Treatment of Soldiers
Soldiers were provided with poor housing, food, and medical care, and
they were subject to arbitrary and often brutal treatment. Training of
conscripts was increasingly basic. Africa Watch obtained the following
testimony from Getacbew, a schoolboy who was forcibly conscripted in
Marefa 1990:
We arrived at the trainiqg camp in Debre Zeit on March 22. Hiere

they shaved our heads were shaved and bmned our dotfaes. Hiere was
a medical examination, and two failed. They gave us a green [army]
uniform, a blanket, a bedsheet, a plate and a drinking cup. We were
sent to sleep in a large store belonging to the Defense Construction
Authority which had been turned into a sleeping place for the camp.
We slept on plastic grain bags filled with the leaves of eucalyptus trees
and Christmas trees. About 2,000 slept in our store, and there were
three other stores used for sleeping, which were much bigger. An officer
told us that there were 20,000 in the camp. During the first week,
thousands arrived every day; then they stopped arriving.
There were people from all over the country; all the same, all
conscripted. We were divided into units of 48. Each group was to be
Womm bind their stomachs afier childbirth to help lose the cxum weigjbt.
^ Helen Pankhurst, "Women, the Peasantry and the State in Ethiopia: A Study
£nim Menz," PhD Thesis^ Edinburgh, 1990, p. 77.
303
commanded by a lieutenant, but we hardly saw any officers in the camp,
only trainers. Our group contained boys from Moyale [a district on
the Kenyan border]. Many did not speak Amharic and they needed
a translator. The youngest was 14. We drilled every morning from
7 to 12, and then afternoon from 2 to 5. We did not use guns
in the
or sticks, we you were ill they gave you medicine, but
just drilled. If
if you lagged they would beat you with sticks. We
ate injera [Ethiopian
bread] made fiom maize and beans: nothing else, the llood was bad.
Hiey told as that we weie goiag to be tnuMfened to a seoood oamp

where we would karn how to use a gun, but there was a riioatage of
transport so we were staying for now in the first training camp in Debre
Zeit. They said that the previous batch had been 43,000, and that they
had stayed 15 days and seat straight to the north [the war front].
The camp was not a proper military camp: it was a place belonging
to industry. There was no wall or fence around it, only guards. After
one week there I tried to escape with four others, but we were seen and
captured. We were lucky; others were shot dnd when trying to run
away. They beat us witt sticks. There was one officer who best
especially hud. 1 had wonods here [on the left fofeann] and here [below
the left eye]. But they were not serious hi beathig as: diey beat us near
to death but none of us is dead. We were put in a odl made from
corrugated iron. It measured 3X4 [meters] and there were 60 people
in there: we could not lie down to sleep, we could only
sit. Some of
the people there had tried to escape — one had awound where he had
been shot — and
others had done other things wrong. They gave us
no [medical] treatment there, but they did later. We were kept there
24 hours, and then sent back to training.
The next Saturday night [April 7/8] I decided to escape this time —
on my own. At midnight it started to lain haid, and Aie guards went
back under their shelters. I aept out oo my atomadi. Ths time I got
away — as you see.
After a few weeks of such training, massed columns of conscripted men

and boys were thrown against the guns of some of the most hardened
guerrilla fighters in the world. Each month, hundreds or thousands were
killed, wounded, or captured, often without even firing a shot or seeing
their supposed enemy. Many did not even know how to fight. Three
teenage boys conscripted into the army and captured by the EPLF at Afabet
said:
304
When the battle began, we didn't know what to do. We asked the
officer. He said: "you have a gun, shoot Do like those in front of
you."^
A 15 -year old boy was conscripted by soldiers from his village in

August 1989, while herding sheep. After four weeks "training** he was
seat to flie firant at Woldiya (WoUo), wfaeie he was captpied by the EPRDF
hi his first engagement In November, he was mterviewed by a visiting
relief worker:
One day in I was ordered to quickly jump on a truck and

the morning
we left the town. We
drove for a short time and stopped in an area
where there was a lot of shooting going on. Together with the other
soldiers from my truck I walked a short distance and then we arrived
in a place where I saw lots of troops fighting, running around and laying
on the ground. I did not know what to do and asked one of the people
who came with me. He told me to go ahead and shoot at people who
were wearing a different uniform from mine.
EdbraeowDt oTDIsclpliM
Conditions withm the Ethiqpum amy were at best poor and at worst
a livuig mghtmare.
Mohamed, a former goldsmith, was conscripted 14 months ago and

received four months' training during a lull in the war. He was already
a veteran of four battles before his capture. "But it's not the fighting
I remember," he said, "just the fear. A
man in my old unit tried to
desert but he was caught and the officer told one of my comrades to
shoot him. He had to — we all would. Otherwise, we would have
been kiUed."'
Sergeant Booetsioa Kidan Mariam Fecado, a deserter in Eritrea, gave

the following account ofhow the Ethiopian army treated ordinary soldiers
foUowii^ the fsilore of the Red Star o^BDsive:
Quoted in a film made by Alter-cine Inc. Qluuele LaOoane and Yvan Faiiy)
Jhe Foi^kUem Laid, September 1989.
^ Geny McCann. ''Between Heaven and HeU% Observer Scotland, April 16,
1989.
305
As army authorities decided
the troops retreated to their base areas the
to make an example of those who were
be blamed for the failure
to
of the offensive. Thirty two people were picked out from the 23rd
Division and shot. The authorities also declared that it was an offence
for anyone to mention the strength of the EPLF at any time or to
criticize the army in any way.
The Deigue's cadres at the base camps also inade en^^

of other individuals who were publicly cxeoited for "crimes agaiBSt
the aimy." Those executed induded:
* Sergeant Tcsfaye Ayena who was shot because he was accused of

leaving the battlefront because he was sick.
* Captain Hailu, who was accused of giving up a water-well to the

enemy and leaving his rifle behind when he retreated.
* Militiaman Eshetu Kebede, accused of running away from tiie firaot

near Alghena.
Tes&ye, who was accused of discussiqg the EPLF with oter soldiGSB
and speaking favourably about them.
* Kemal Abdu, a private, accused of retreatii^ from the battlefroot
Soldiers were regularly shot, accused of wounding themselves in order

to be hospitalized, retreating in the face of the enemy, or simply
"gossiping".
The general conditioos in the army were also bad. Soldiecs' pay was
regularly withheld and they were toM it was sent to tiieir parents or
wives, when people never received a birr from the
in fact, those
authorities. Those who
reported sick were often refused treatment on
the grounds that they were already suffering from heart troubles, eye
defects and other illnesses before they joined the army and it was not,
therefore, the army's responsibility to give them medical treatment.
Even from soldiers' families were either held back or were opened
letters
and kept from the intended recipients for long periods of time, leaving
the families without a reply and uncertain as to whether their sons or
husbands were still alive.
306
All these problems, together with (he constant fear of being shot for
some real or imagined offence, created a very insecure atmosphere in
the army and many people became so desperate that they decided to
desert at the first opportunity. A
number of people even committed
suicide rather than face the continuous strain of fighting and reprisals
from the army.^^
Sgt. Bocretsion himself saw three men — Corporal Teshome, Private

Misgame Fantaw and a militiaman whose name he did not know — shoot
themselves.
Senior officeis were regularly executed after the faflure of offensives
or for expressing the opmion that the war was unwimiable. Iheie were
executions in Eritrea m
June 1982 and February 1988, and m
Addis Ababa
m May 1990 (see below).
Within the army there was a tight security netwoik:
There are officers throughout the army who are called "welfare officers",
but who are in fact secret service personnel. These ones are relatively
safe because you know who they are, but there are other secret service
people who operate clandestinely, and their presence creates a lack of
trust amongst soldiers. These people are extremely powerful and they
set up a sort of infonnation netwoik. Any punisbment or action that
is taken doesn't just arise by itself, bat is the result of this spy network.
On the field disciplme is sometimes brutally oiforoed. Theie are stories
of infantry gomg mto battle with lines of machine guns behind them.^
During the 1980s there were numerous stories, mostly unsubstantiated,

of firefights between army units and of special units deployed to gun down
soldiers who tried to retreat. Flight-Lt Habte Luel, a helicopter pilot who
defected to Sudan in August 1987, claimed that he had done so after
disobeying an order to fire on retreating troops. One substantiated incident
occurred at Bahir Dar, Gojjam in March 1990. The army garrison blew
up one span of the bridge across the Blue Nile, in order to prevent EPRDF
forces crossing it. The bridge was blown up while retreating government
soldieis were still oo it, IdlUng an mdmown number.
Another rnddent occurred in eady 1991, when the army m
Eritrea tried
to deploy mflitia forces as front line troops at Gfamda, north d
Asmara.
^ Somali, Tigray and Oramo Resistance MonUar (STORM), 3.2, June 1983.
^ Lieutenant Yamani Hassan, a prisoner of war held by the TPLF, interviewed

by Geny McCaon and Saiah Vanghan at Tade A2ieg«r,Tigny, December 1, 1988.
307
Hiis led to firiction, as tbe fnilitiamen felt ttit llie oonliact governing their
deployment had been violatDd. Tb€ results were a apala of deeeitiaiis from
the militia in Eritrea and the refosal of some btttuions to be tramfened
from their home areas. The army respcmded by confiscating all the property
of deserting militiamen and burning their houses, detaining others, and
threatening summary execution for any caught while attempting to desert.
The families of militiamen who deserted were subject to reprisals, including
detention. While looting, burning and detention certainly took place, Africa
Watch is unable to confirm any incidents of summary execution.
A serious incident occurred in early February 1991, when a militia unit
that originated from Qohayn in Seraye withdrew tmm ft» Ohinda front
towards Asmara. Hie mut was met by a fbroe fitom the rmilar army at
a place between Mai Haber and Adi Hawedia, and a fiiallglit enmed.
Acooiding to reports* casualties ran into the hundreds. Hie divisional offioor
of Seraye, Ghezay Sebhatu, was killed in the fighting. The militia were
defeated, and over 300 were detained in a railway tunnel, without even
the most basic facilities, and lacking fresh air. They were later transferred
to the military training camp at Adi Nefas, near Asmara, where an unknown
number remained in detention until the fall of Asmara in May.
The Attempted Coup of April 1989
On May 16, 1969 a group of senior officers attempted to stage a coup

d'etat m Addis Ababa, wfafle Pkesideiit Meqgiitn was oat of the ooontiy
on a visit to East Gennany. Hie staled aims of tlie plotteis indnded a
negotiated end to the wars and political and economic liberalization.
Preisident Mengistu's inteUigenoe had learned in advance of the planned
coup, and his security forces struck first, precipitating an attempt by the
plotters to seize power before they were fully prepared." Two generals
were killed in a shoot out with security forces sent to arrest them by the
loyal Minister for Internal Affairs, Tesfaye Wolde Selassie. The mutineers
took control of the army and air force headquarters and the defense ministry;
all were besieged by loyalist troops. There was fighting at all three
locations before the coup plotters surrendered. The Minister of Etalme,
Maj-Gea Haile Gioii^ HMe
Mariam, was killed by tliB pk^^
for telling them to surrender. The security foms also made a sw^p
through the city, arresting several hundred subjected sympathizen wim
the coup. In toted 44 officers were reported t^led.
Âfrica Qmfidemial, 30,11. May 2&, 1989.
308
The commanders of the gairison in Asmaia mutinied simultaneously
but were also overpowered.
After the coup had been crushed on May 18, 176 army officers were
detained. The former Commander in Chief of the Air Force, Maj-Gen
Fanta Belay, disappeared and is believed to have been executed. In
December, 13 generals and one navy conmiander were brought to trial
belbrethemililuy divisiODQf IfaeSiipi^^ The fictlliat they were
tried latfaer dian summarily executed appeared to indicate that Mei^gista
was prqared to grant them demeacy, wfakfa would have been weD-reoaved
in the aimy, the general popnlation and the international community.
However, in a surprise announcement on May 21, 1990, the government
radio stated that 12 of the defendants had been found guilty and executed
the previous Saturday night in the basement of the Presidential Palace.
The generals were given no chance to appeal against the sentences, which
were carried out immediately.
The news of the executions and the speed of their implementation came
as a shock to many Ethiopians, and contributed to student unrest at the
university of Addis Ababa over the following week. In the longer term,
the executions also contributed to the demoralization of the aimed foices.
PriaoDcn of War
Many soldieis were captured by the rebel fronts and (in 1977/8) the
Somali army. There are reliable reports that the WSLF and Somali army
regularly killed prisoners of war. Treatment of prisoners by the OLF has
not been properly documented. The largest number of prisoners was taken
by the EPLF and TPLF-EPRDF, and these were well treated. Captured
soldiers had their weapons and their boots confiscated and were then taken
to prisoner-of-war camps.
At different times in the war, the EPLF and TPLF-EPRDF held
dioiisaads or tens of thousands of prisoneis. In kte 1989, die EPRDF
claimed to have 37,000. Their liviijg conditioiis were basic, widi meager
but adequate food and accommodation —
but ni this respect diey lived
little differently from the local population or indeed die members of the
rebel fronts. Blankets, clothes, soap and cigaretles were supplied when
available, but rarely footwear. Medical care was provided: in late 1988,
there were 4,000 wounded prisoners receivii^ some treatment by the TPLF,
309
including 1,200 caied for in field hospitals. There are no reports of
physical abuse or execution. The prisoners were sometimes able to
correspond with their families. Prisoners were used for manual labor on
road construction and other infrastructural projects, but the work was not
excessive and discipline was not enforced in a humiliating manner. Some
were given training in literacy and nursing. Other social and recreational
activities — notably soccer matches —
were organized by the prisoners
of war themselves.
In its first congress in January 1977, the EPLF oommilted itadf to
respecting the Geneva ConventiQns witb respect to flie ngjbts of pnsoneis
of war.^ The TTLF made a similar promise. However, the International
Committee of the Red Cross (ICRC) was unable properly to fulfill its
mandate with regard to the prisoners. Details about the disagreements
between the fronts and the ICRC are not public. However, it appears that
the EPLF and ICRC were unable to agree on the circumstances in which
the ICRC could interview prisoners. Relations between the TPLF and the
ICRC were warmer, and some discreet assistance was reportedly provided
to prisoners in Tigray, but soured in late 1987 when the ICRC withdrew
from tiic cross-border relief operation.
Prisoner of war camps were attacked by air foioe bombers on several
occasions. As the locations of the main camps were well-known to the
government, this must have been deliberate. Orota camp in Eritrea was
attacked several times. war were killed and
In Tigray, ten prisoners of
20 injured in an air raid on June 28, 1989, and in Novembor 1989 air force
planes bombed a wood outside Adwa the morning after a contingent of
3,000 prisoners of war had left it.
The Ethiopian government consistently refused to recognize the existence
of prisoners of war held by the EPLF and TPLF. This led to dangers when
prisoners tried to communicate with their families and when they tried to
return home after having been released. This was an important factor
impeding ICRC efforts to carry out its mandate; it could not assist prisoners
to correspond with their ftonilies, nor initiate or monitor prisoner releases

and exchanges.
Health Action Group, "Report on a Recent Visit by Medical Team to Westem

Tigray, London, 1988; Health Action Group and REST
" UK
Support Comminee,
"Report on the Emergency Health Gaie Project, July 1968-Janntty 1969,** London,
1989.
^ The Geneva Conventions apply to international armed conflicts, but ako

set forth the humanitarian principles that are relevant to internal conflicts.
310
The EPLF released prisoners of war on the occasions when it was unable
to provide for their sustenance or safety, or as goodwill gestures. Releases
occurred at regular intervals from the late 1970s onwards. Many prisoners
were however kept for extended periods. The TPLF did not keep rank-
and-file prisoners for longer than a year at most. After a six month period
in which the prisoners were introduced to the aims of the TPLF, prisoners
of war were given the choice of trying to fetnm home, remaining in the
TPLF-held area, going to Sudan as a refugee, or joinmg tiie TPLF (or,
after 1989, tiie EPRD^. Ttae are no documented cases of stgpificant
variations in this practice.
The Somali army captured about 10,000 prisoners of war between July
and October 1977. There were no central directives concerning their
treatment, which depended entirely on the inclination of the commanding
officer who captured them: some were well-treated and sent to prisoner
of war camps, others were abused and even executed. In October 1977,
the Somali Ministry of Defense reportedly issued an instruction to all
commanders that there were to be no more prisoners. Thereafter, many
were routinely shot on cqMure, ^fle offioen who refined to do Ais passed
their prisoners to senior officers who would deal with them aocordmg to
their preference. Hiousancb wereoertainly summarily executed; fortimately
the order was given after the Somali army had made its mahi mflitary gwns.
The WSLF treated prisoners of war in the same manner.
The government reguded all ex-prisoners of war as deserterSk who were
liable to be imprisoned, executed, or rc-conscripted. There were "re-
education" schools for released prisoners of war in Gonder and Meqele,
in which they were detained for varying periods of time, and subjected
to physical abuse and torture. Some were killed.^^ Many were re-
conscripted into the army. In later battles some of these ex-prisoners were
recaptured. Some soldiers captured by the fronts had been conscripted,
captured, and released as many as three times.
The government treatment of captured rebel fighters and Somali soldiers
was poor. Many were subjected to torture and prolonged imprisonment,
and some were lolled, lliey were denied tihe amenities and rights granted
to prisoners of war held by the fronts.
^ An unooofinned claim by the EPLF is tfwt 75 prisoners of war wAio had

been militiamen were killed while passing through towns after their release in early
1979. Another claim is that five were killed in Maichew, Tigray on December
13, 1988, after having been released by the TPLF.
311
The Forces of the £PLF
The size of the the EPLF remained a closely-guarded secret
army of
throughout the war. During the early 1970s, the manpower of the ELF
and EPLF certainly outnumbered the Ethiopian army in Eritrea, but after
1978 the position was reversed, with the build-up of the army and the
collapse of the ELF. Common estimates of the size of the EPLF were
12,000-18,000 in 1982, rising by 1989 to 40,000-50,000, plus 30,000
mflitia.^ Both the EPLF and tiie Ethiopian govemmeiit however had
leasons to imdeiestmiate the nometkal stieqgft of te ftCMifs fotoes^ so
Iflcely to he somewhat hipier.
the real figoies aie
The majority of the fighters in the EPLF were ondoubtedly vohmteen.
Many young men fled from the conscription operated by the Ethiopian
government and instead joined the EPLF.
Throughout the 1980s, the EPLF operated a draft to fill the remainder
of its ranks. The principle and the implementation varied
from place to
place, but essentially consisted of a quota of conscripts levied on all
it
Eritrean communities inside Eritrea, and occasionally was extended to

refugee camps in Sudan. According to the testimony of refugees in Sudan,
each community was left to decide how to fill its quota, but on occasions
the EPLF would itself choose whom to take if no conscripta woe ddiveied.
The draft was imposed together with the provirioii of aecvkes such as
education. Women were encouraged to join the front as well. There are
no confirmed reports of the conscription of under-age children.
Anumber of refugees fled to Sudan to avoid being omscripted, or
having their sons or daughters conscripted. Many Eritreans from the
western district of Barka are conservative Moslems, and often a family's
reason for flight was not opposition to the draft itself, but fear of the
secularizing and modernizing influence that membership in the EPLF would
exert on the conscripted son or (especially) daughter. (Conversely, escaping
from such a fitunOy enviromnent was one reason why some young women
volunteered to join the EPLF.)
One significant inddent of local resistance to EPLF conaciiptkmoocuned
at Asela in the Danakfl district of eastern Eritrea. The people of this area
are Afar, and Moslems. The EPLF had occupied the areamare-or-
all are
less continually since 1978, and had supported a militia drawn from the
local Afar. In September 1988, the EPLF tried to conscript a contingent
of the people to its regular forces. The people of Asela objected to the
" NOVIB, "War and Famine in Ethiopia and Eritrea: An Investigation into
the Arms Deliveries to the Struggling Parties in Eritrea and Tigray," Zeist, the
312
leqaiiement that the conscripts move to Sahd distikt for Indiiii^ lather
than remdniog in the DanaldL This followed some eadier incidents in
which the local people had objected to the EPLF requiring women to attend
political education classes. The conscription dispute ledtoananned clash
between the Afar militia and the EPLF. According to reports, an EPLF
unit opened fire and caused about 20 casualties among the Afar men
resisting conscription. The Afar themselves had ready access to arms, and
a battle ensued. Afar community leaders "officially" claimed that there
were 300 Afar fatalities,^ but in fact the total number of casualties is
unlikely to have exceeded 250 dead and wounded. Further clashes occurred
in Febroaiy and May 1989.
The Forces of the ITLF-EPRDF
The fighting stiei^ of the T7LF and latterly the EPRDF was an even
more closely-guarded secret than that of the EPLF. The most common
estimates put their strength at 5,000-7,500 between 1980 and 1985, rising
to 30,000 by 1989 and 70,000 by 1991.^' No estimate for militiamen
has been made. However, these figures are likely to be underestimates.
In 1980, the TPLF began to arm village militias throughout Tigray. Noting
that by the mid-1980s every village had a part-time militia which included
a substantial number of the adult men, and recognizing that the population
of Tigray is closer to four million than two million, this implies a very
coosidenble leseive sticugth of wdl over 100^000.
The TPLP-EPRDF never had a problem with lecnnts. If anything, the
problem was the leveise — there was too much pc^xilar demimd to jom
Its ranks, or at least receive anns from it to form a militia. There is a deep
attachment to armaments in the Ethiopian highlands, which has long been
heavily militarized. One of the reprisals that was most resented by the
population of eastern Tigray after the suppression of the Weyane revolt
of 1943 was the confiscation of firearms from the people. Traditional
culture extols warlike values, and the possession of a rifle is seen as a mark
of prestige. The TPLF built upon this entrenched tradition, composing
and popularizing its own songs which vaunted the importaDce of joining
tiie aimed struggle. To lie a TTLF fighter was to adiieve a heightened
sodal status. Young men were thus under considenAile sodal and
* As lepoited Africa ConfiOaM^ Aqgust 25, 1989.
^ NOm 1991, p. 12.
313
a
psychological presmie to join. Many vohinleaed, and llie TPLF was
tfaeiefoie able to scieen wouM-be lecniits and adect only Ihe nusat snilable.
Similar attitudes greeted the EPRDF during its advance southward.
On occupying southern Wollo in late 1989, the H^RDF was CQofronted
with large demonstrations of peasant farmers demanding to be armed. On
moving in to Gojjam in February-March 1991, the EPRDF was immediately
met with a demand that the local militia (which had been partly disarmed
by the government in mid- 1990 on account of participation in local revolts
in April 1990) be given back their weapons — the EPRDF complied. Many
government soldiers and militiamen who were captured by the EPRDF cither
immediately volunteered to fight against their erstwhile colleagues, or
decided to do so after a brief spell of captivity. There is no evidence that
coercion was used to make prisoners of war join llie fraot.
The TPLF-EPRDF official requirements for figfatois indnde a lower
age limit of 18 years. Some fighters interviewed ^
jonmalists in Addis
Ababa in May- June 1991 certainly looked younger than that age, and some
admitted to being in their mid-teens.
Once a fighter was in the TPLF-EPRDF, discipline was strictly enforced,
and there was no method of leaving except through injury. There are
reliable accounts of members of the TPLF wanting either to leave active
service, or to leave the organization altogether, and being prevented. The
EPRDF has indicated that this policy will change now that the war has
been won.
More generafly, the TPLF-EPRDF has further entrenched a popular
culture that centers on firearms and fighters. Hie TPLF-EPRDF ideolagy
stresses that a health wofker, relief woiker or local administrator is also
a "fighter** in the '^people's struggle." It also stresses that war in general
is an evil, but that the war against the Dergue was a necessary evil and
therefore good. It is questionable, however, the extent to which these higher
principles have been understood by the population at large, or even whether,
outside the TPLF heartlands, a serious attempt was made to inculcate them.
A fundamental principle of the TPLF is that "the people, oiSLanized,
politicized and armed, cannot be ruled against their will."* The
widespread ownership of firearms, many of which are modern automatic
weapons, does facilitate popular resistance to central government, but it
also creates potential problems for the enforcement of taw aad order and
the implementation of democratically-arrived at decisions.
^ Meles Zenawi, mteiviewed by Alex de Waal. Nbvonber 1988.
314
*
Other Forces
The forces of the OLF and most other rebel fronts were relatively small
in comparison to the EPLF and TPLF-EPRDF. There is little information
available about their practices of recruitment or their treatment of fighters.
The fact that most of the WSLF fighters in early 1977 were in fact members
of the Somali anny has been mentioned in chapter 4; the forcible
recniitment of Ethiopian lefngees to the WSLF, Somali anny and possibly
the Somali-backed Qcomo m»t win be discMBcd in chi^pler 19.
315
18. WARS WITHIN WARS: THE WESTERN AND
SOUTHWESTERN LOWLANDS
Ethiopian provinces spread out from the central highlands like spokes
of a wheel. Most provinces consist of a highland area, usually inhabited
by Amhara or Oromo, and a lowland hinterland, inhabited by marginalized
people who are oft^ semi-nomadic pastoiali^ Hiis is paiticolariy the
case for the west and die southwest.
West and southwest Ethiopia is the most eoonomically productive and
ethnically complex part of me oomitfy. Gqjjam province is one of the
Amhara heartlands, but contains a huge peripheral area to the west,
inhabited mainly by Agau and Gumuz people. Ethiopia's main export,
coffee, is indigenous to the southwest,^ which is mostly fertile and well-
watered. In the nineteenth century, the Oromo states of the Gibe region
(straddling modem day Keffa, Wollega, western Shewa and eastern
Illubabor) were the most prosperous part of the country, and were the center
of the regional trade in coffee, slaves, gold and ivory. Apart from coffee,
these commodities originated in the surrounding lowlands, which are
inhabited by a variety of people, including Gumuz, Berta, Koma, Mao,
OanzB, Amiak, Nver, Nyangatom, Chai, Passeaatcfa, Kwegu, Muisi, An,
Hamar, and otfaen. Many of tese cHudc teims overkqp^ or are used m
dififeient ways by different groups, and many ethnic ffoupii have two or
more names. In the west of Wollega and Gojjam, these peoples are refened
to as ''Shanldlla" by the hightonders, a derogatory term that they themselves
reject. These groups are incorporated into the state to varying degrees —
some may be considered to be subjugated, otbecs are marginal but have
maintained a high degree of independence.
The existence of the international frontiers with Sudan and Kenya is
a central fact of this area. Many ethnic groups straddle the border. The
civil war in southern Sudan and the repression in northern Kenya have often
meant that life on the Ethiopian side of the border, where the government
has at times shown flexibility m local administrative arrangements, has
been preferable to life in the neiglilioriB(g conntiy*
Hiis chapter will ouflme the wan and femines that have affected the
lowlands of the west and southwest, province by province, from north to
south.
^ The word
"coffee is even said
" to derive from die name of Kefb pnivinoe,
where the trees grow wild.
317
Western GqUan: The EPRP RdrUtoi
Hie westem regioB of Gojjam is one of the Amhan heaiflaiids, with
a traditional independence firom Shâ (see chapter 3). The western
lowlands of the province, however, are inhabited mainly by peoples who
have more in common with their neighbors in lowland Wollega than the
highland Amhara. These people include Gumuz, Agau, Shinasha and others.
The Gumuz (also known as Begga), who numbered about 53,000 in 1970
will figure most prominently in this account.
Westem Gojjam has long been incorporated into the extended domains
of the highland states, which have raided for slaves in the area. In the 19th
century, Ras Kassa of Quara became a renowned shifta leader in the area,
rising to become the Enipenir Teodros.
Followii^ its fiist defeat in Addis Ababa dniiqg te Red Tenor, and
its second defeat m Tigray at the hands of the TTIf m 1978, the EtUoman
People's Revohtionary Party (EPRP) set up a base for guerriUa operatioos
in westem Gonder. In 1983, it expanded its opeiatiaDS into northwest
Gojjam. The leaders of the EPRP forces were mostly educated Amhara,
but the rank and file were drawn from the local people of the area. Growth
was slow but steady. In 1984, the EPRP held its second congress in Quara.
From 1985 onwards, the EPRP became more active, particularly in
northwest Gojjam. This was partly in response to the resettlement program,
which created deep resentment among the local people.
Hie establishment of the huge Metdcel setOement oonplez m late 1984
led to thedi^lacement of the local Onmnz. Hie Oumnz of this area rely
on shifting cultivation and gathering wild foods from the forest The
gpvemment declared that any arable land cnnently mcdtivated and forests
were "unused," and designated them resettlement areas. According to
Dessalegn Rahmato, the established land use system was thereby dismpted:
new land could not be cleared for shifting cultivation, and as "the main
difference between hunger and a full stomach may depend on forest
resources," local famine followed.^ Some Gumuz tumed to armed
resistance; in 1985 and 1986 there were incidents in which settlers were
fired on, and some joined the EPRP.
According to the testimonies of resettlers who escaped from Metekel
and letumed to Tiây, the EPRP harassed setflements and side food
and other commodities, but gave food assistance to escapees and helped
them return home.
' Dessalegn Rahmato, "Settlement and Resettlement m Mettekd, Westcni

Ethiopia," j^ca (RemaX 43, (1988) pp. 14-34.
318
As elsewhere, the government lecniited amflitia foioe firom among the
resettlers and used it not only to police the lesettlement camps, but the
local people as well.
In December 1985, the government launched its sixth attack on EPRP
positions in westem Gojjam and south-west Gonder, deploying about 4,500
troops. By mid-January, the attack had been repulsed, but retreating
Birawe (December 21) and Arema
soldiers burned several villages, including
(January 1986), and summarily executed at least 20 civilians. Two
8,
further attacks were launched in each of the following two dry seasons.
In 1988-9, activity intensified.
On June 21, 1989, the EPRP attacked a cooslniction project within the
Metekel area and kidnapped three Italian wxuken. Several vehicles and
a water point were destroyed m
the attack, and (according to escapees)
food from a store was taken. The kidnapping incident gained much
international publicity and government reprisals were quick to follow.
There is a credible account that on December 20, 1989, government
troops entered Ambela market, Ankesha Banja sub-district, and opened
fire, killing 14 and wounding 20 marketgoers. In August 1990, the EPRP
made military gains, leading to another round of reprisals. These included:
* August 28: two villages in Ankasha sub-district were shelled, and three
civilians killed.
* October 1 and 2: villages near Daqgiki were shelled and ten dvflians
were killed.
* December: soldiers killed a number of peasants in Laye Zigem, and

reportedly then cut the genitals off the nude corpses and displayed the
mutilated bodies as a detment to support for the insuigents.
Conflict between EPRP and EPRDF

During 1990, the military advance of the EPRDF brought it into contact
with the EPRP. The wounds from the battles of 1978 had not been healed -
- if anything, they had intensified due to the fact that the EPDM, a
constituent of EPRDF, was originally a breakaway group from EPRP. Ibere
were a number of mutual accusations of aggresston. When the EPRDF
occupied highland Gojjam in February-March 1991, diese intensified. Hie
EPRP accused the EPRDF of "declaring war" against it, of detaining EPRP
supporters in the towns it occupied, of shooting unarmed demonstrators,
and of taking away infrastructure from the resettlement sites. The EPRDF
on its side accused the EPRP of aggression, of ambushes and of minmg
319
loads. QnApifllS, ISK^l, tiieiewBBalHrtllebetweeaEPRPaMlEPRDF
foUciwiiig occupation of Dangila.
tiie latter's
Following the fall of the Mengistu govemment in May, figMiqg betweoD
the EPRP and EPRDF contimied in western Gojjam. The Sudan
govemment, which had hitherto given support to the EPRP, withdrew that
assistance and closed the border. Four EPRP leaders in Sudan were arrested
and handed over to the EPRDF. Intense fighting continued into Jiily„ with
the EPRP admitting serious losses.
The conflict with the EPRP presented the EPRDF with its first challenge
concerning its conduct of warfare in a situation in which it possessed
overwhelming superiority in manpower and fluterid. At tfie tune of writing
it is too eaily to tdl how the EPRDF Ibnes hive acqnitted themadves hi
terms of treatment of the civilian popidation bdievtd to be sympatfaetie

to EPRP, and treatment of EPRP combatants taken priaooer. There hive,
however, been no reports of such abuses.
Highland Woltega: LHtherans and OnNBO Nithwaliats
In the 1880s, the western Oromo states and Keffa were incorporated
into the Shewan empire of the Emperor Menelik. Some were conquered
by Menelik's armies, under the command of an Oromo general, Ras Gobana
Dacche, and were subjected to the alienation of land and the imposition
of neftegna Amhan setflers. Others, notably Leka-Nefcempte (eastern
Wollega) submitted voluntarily, and retained id^gree of inteoial antoacmy.
Ihe peripheral areas, most of them already subject to Gnnno domination,
were conquered in die 1890s.
The ethnic heterogeneity of the area and the variety of manners in which
the peoples submitted to Abyssinian rule meant that local administration
was uniquely intricate and frequently anomalous. As elsewhere in the
empire, many conquered people did not submit readily to their new over-
lords. There were frequent if localized rebellions. The Gibe states had
themselves subjugated neighboring peoples, and there were occasional
violent clashes between the Oromo and the peripheral groups.
During the Italian occupation, the Oromo leaders of Wollega and
niubabor mituted a movenrat for hidependenoe fran Ethiopia, bned on
the premise diat they had voluntarily jomed hi the 1880s and had not
diereby forfeited any sovereign rights, and m
1936 pethionBd the Britidi
government to secede and become a British protectorate. The attempt was
unsuccessful, but Oromo nationalism remained at least as potent in this
area as in eastern Ethiopia. Many Oromo from WoHî
were active in
the Mecha-Tukma Association oi the 1960s.
320
An important element in the growth of Oromo dissidence in the
southwest was the Evangelical Church of Ethiopia, known as the Mekane
Yesus. This was founded by Lutheran missionaries from Sweden, who
were also active in Eritrea. The Imperial government forbade the
missionaries to operate in the Amhara highlands, but permitted them to
evangelize among the Moslems and foUoweis of traditional religions in
teKmUL Wolkga provinoe was ^pdim Md[iiie Yesos ocmoeii^^
and by the 1970i it had a laigB imaaber of folkswen and, equally
imporUaHy, had provided edncartional fadlities on a scale that outstnpped
aU other provinoes save Sfaewa and Eritrea.
An indigenous proteitant churdi, the Bethel Bvaqgelical Church, was
also influential in the growth of Oromo political oomdoimeaB. This gained
a stroQg foUowii^ in the Dembi Dolo area.
Repression and Insurrection 1975-85
The Oromo Liberation Front (OLF) was slow to begin military operations
in western Ethiopia; it only began small-scale activities in 1981. The
conditions wne right, to resistance to grow. The govenunent
hofwever,
had already instigated a of cnckdowns in tfie area, munly auned
series
atthe mflncace of the educated elite assori atrd with the protesiant chmdhes.
The militaiy commander of WoUega fiom 1976 onwards was Sergeant
Negussie Fanta, who soon acquired a reputation for ruthlessness. In early
1977, 15 Oromo students were executed for campaigning for the right to
an education in Oromo (ironically, the government-sponsored literacy
campaign was soon to cede them that right). The period of the Red Terror
saw a crackdown on educated Oromo, especially members of the Mekane
Yesus. A prominent pastor in the Mekane Yesus, Gudina Tumsa, was
imprisoned in June 1979, and later "disappeared." His wife, Tsehai Tolessa,
was imprisoned with over 400 other Oromo women in February 1980, and
later released. Between May and Deoeniber 1981, 300 Mekane Yesus
churches were closed, and 600 pastors and odier church wodoers anoted
and church property confiscated.
In an apparent attempt to entom a cultural change and a break from
the church, people were forced at gunpoint to attend literacy classes and
public meetings.^ The government sent students to WoUega with
instructions "to make 80 farmers literate;" instruction in the ideals of the
revolution was also given.
^Somali, Tigray and Oromo Resistance Monitor (^STORAf), L4, October 1981.
321
Lowland Wolkga: Multiple Marginalization
The OLF had its natural constituency in the Oromo-inhabited highlands

of Wollega. However, when it started military operations, it did so from
a base in Sudan, and therefore operated in the lowlands. Most of the

lowland people are not Oromo, but Berta, Komo, Gumiiz. and othen. These
peoples weie already sabering at the hands d
govw rnHnent militaiy and c
resettlement policies, following a history of deletion at the hands of

raiders and conquerofs from all diiections.
Life on the border consisted of shrewd calciilatkms as to where short-
term security could best be had. For example, a substantial part of the
Komo crossed the border several times. They originated on the Ethiopian
side of the border, but crossed into Sudan earlier this century to escape
Oromo raids, when Sudan became safe from Arab slavers.'* In the 1960s,
many Komo crossed back into Ethiopia in order to escape the forced labor
demands of Sudanese military outposts. In 1966, a Sudanese army unit,
annoyed at the departure of its servile labor force, crossed the border,
burned several of ê villages newly built by the KioBio» killed animals and
took several hundred people back to Sudan. Hie vilhiges weie also rakled
by Nuer groups associated with the Anyanya insunection in southeui Sudan.
Itie Komo who remained protested to the Ethiopian authorities, who gave
them arms and set up a police post, together with flag poles aiid flags so
that they could advertise whose protection they came under.
The Ethiopian government also brought in highland settlers to secure
the border. Started under Haile Selassie, this was intensified under the
Dergue.
In 1979, resettlement camps were created
at Asosa, close to the Sudan
border, with about 25,000 resettlers. These involved alienation of land
from local residents, and many of the settlers wm
armed. Locals weie
also forced to work without pay on the resetdement projects. This was
also one of the few areas m
which agncultuial oollectivizalioii wasenfoioed.
Forcible conscription to the army was implemented on a large scale. In
early 1981, nearly 10,000 refugees, mainly Oromo, fled to Sudan to escape
these abuses. This coincided with famine in western Wollega. Despite
good climatic conditions, government policies had induced a severe localized
food shortage, affecting an estimated 30,000-40,000 people in western
Wollega. The OLF estimates that one thousand died.
^ This .section is derived from: Wendy James, "From

Aboriginal to Frontier
Society in Wcstem Ethiopia," in D. Donham and W. James
"Working Papers
(eds.)
on Society and History in Imperial Ethiopia: The Southern Periphery from the
1880s to 1974," Cambridge, African Studies Centre, 1980, pp. 49-50.
322
In the late 1970s, the Beni Shangul Liberation Front was active in the
lowlands, engaging in gueirilla activity on a small scale. During the 1980s
it had no efifective presence in the field.^
Early OLF Activity and Army Reprisals

The first OLF cadres arrived in Wollega in 1981 from eastern Ethiopia,
and began to recruit.
Military activity by the Ethiopian army on the "western front" began
with an offensive in January-February 1982. Counter-insurgency activities
quickly intensified, and a year later a major military camp was constructed
in the Didcssa valley. Both routine patrols and larger campaigns were
conducted with the ^discriminate violence against dvflians £at is so
fEoniliar from elsewhere in the country. During 1983-5, the OLF benefitted
from the offensive stance towards Ethiopia taken by tfie Sudan government.
In August 1984, according to credible reports, army lepmals in Begi
lolled over 200 civilians and destroyed numerous homes.^ Another
army campaign in western Wollega during November-December 1984 led
to the burning of villages and the killing of civilians. The largest offensive
took place in June 1985 and was centered on Asosa. This followed an
intensification of OLF activity, and a joint statement by the OLF and TPLF
that they planned to coordinate their attacks and ultimately open a joint
military front. Reports indicate that the army used scorched earth policies.
The OLF accused it of burning villages and other revenge atrocities.^ This
offensive coincided with the planting season and contributed to local famine
conditions.
In the period 1985-8, the counter-insurgency operations in Wollega
were closely related to the implementation of the lesetHement and
villagization programs, to which attention must now turn.
* hi1989 the EPRDF trained some Beni Shangul Liberation Front fighters,
who re-entered the area in early 1991. They clashed with the OLF and were
defeated near Shirkcle, northwest Wollega, in early March, whereafter they came
to an agreement with the OLF.
''Africa Contemporary Record, 1985-6^ p. B2!Hi.
323
Impact of Resettlement
The first major resettlement program in the weft was implemented io

Asosa, starting in the late 19708. Hie aettlen weie bfought frooi WoUo
and Gojjam; many were given military tnuniqg. Tbe load people were
compelled to give land to tlie settlers, and to aqpply free labor te the
construction of the reaetdement sites. The settler militia were rôrfsdl^
used to exact taxes from the locals and obtain conscripts for the aiay.
The much larger resettlement program of 1984-8 had comparable
consequences for the local population, on a larger scale. In "integrated
settlements," the settlers were mixed in with local people, who were obliged
to share their land and other resources with the newcomers. The larger
"conventional" settlements involved the displacement of indigenous people
and the disruption of existing systems for land use, reducing many to a
State of huQger and destitntion.
Investigations by Qdtoral Survival indicate tiiat Hie impact of die
resettlement program of 1964-6 on die hidigraotts population of westem
Wollega was disastrous.* According to die testiaiaay of nfugees in Sudan,
local people lost land to the settlers, and were forced to imdertake large
amounts of unpaid and coerced work constructing the re5;ettlcmcnt sites
and accompanying infrastructure. The loss of forests and forest resources
to the settlements was also disliked. These reports have to be set against
more positive descriptions from central and southern Wollega, which report
much less tension between settlers and locals.'
Based upon the existence of the settlement militia and other
considerations, the OLF declared that "the settlement program is a legitimate
military taiget.""^ It attadDedsetdemenlsoo several oocasiooB, for enmple
die setdement of Jarso on April 28, 1988, wfaea two Mh idief woikMB
were captured.^^
* Cultural Survival, IVîîc^ and the Ethiopian Famine 1984-I985tCuaiMgl6,
Mass., 1985; Sandia Stdngiaber, "RaaetdcaBent in 1986->198r, in J. W. Clay»

S. Steu^ber and P. I^ggli, The Spoils of Famine: EMopkm Pmdm FoUq^ and
Peasant Agriadlure^ Cmbddgpt Maia., 1988.
' Alula Pankhurst, "Settling for a New World: People and the an
in
Ethiopian Resettlement Village," PhD thesis, Manchester, 1990.
^ OLF, "Sutement to Agencies Working in Oromo Land," May 4, 1988.
" The two were releaaed unhmned in Sudan a mondi later.
324
ViUagizaiion
Villagization began in Wollega in late 1985, and was implemented in
the adjoining provinces starting the following year. The program was linked
to the construction of roads, and relocation near army garrisons. The
program in western Wollega was implcmentBd wifli tiioroughness and
coerdon, though the levd of videooe did not nurtcfa Ifaat of Hareigbe.^
All the villagers' possessions were g^jsteied, and many were confiscated,
indttcUog idow oxen. PMple were detained, tortured, nped and executed;
houses and grain stores were burned.
Western Wollega was unnnal in that villagization was also accompanied
by enforced collectivization; on the collective farms the produce was entirely
t£ên by government officials, and the villagers were instead given a ration.
Outside the insurgent zones of western Wollega and Illubabor (see
below), the villagization campaign in southwest Ethiopia was more akin
to that in Arsi —
implemented with an implicit threat of violence, but with
little actual force used.
The Role of the SPLA
Ihe Sudan People's Ubecation Aimy (SPLA) was set up in 1983

following a mutiny by southetn Sudanese soldiers in the town of Bor. From
the b^gimnqg it was led by Colonel John Garang. Political marginalization
and economic exploitation of the south and increasing human i^ts abuses
by the Sudan government were all factors contributing to the mutmy and
the subsequent rapid spread of the revolt throughout much of southern
Sudan. The SPLA turned
to the Ethiopian government as a natural ally
and Col. Mengistu for his part saw Col. Garang's movement
in its struggle,
as a useful counterweight to Khartoum's continuing support for the EPLF,
TPLF, OLF and other smaller fronts.^^ Close links were quickly
establishedb^ween the SPLA and the Ethiopian government at the highest
level The Ethiopian government provided the SPLA with militaiy
equ^xnent, bases and a radio station.
Until the M
of tlie Mengistu goveinment in May 1991, there was close
military and security coordination between the SPLA and the Ethiopian
army. Ihe Ethiopian army assisted the SPLA in attacks on Kurmuk in
Sandra Steingraber, "^Villagization in a War Zone: Rcfqgee Reports from

Western WoUega," in CUy era/., 1988, pp. 200-213.
" In the late 1970s and early 1980s, mutineers in southern Sudan had sought
aiKi obtained refuge and sometimes support from the Ethiopian government.
325
Blue Nile Province (across the border from Asosa) in November 1987 and
November 1989, and in Eastern Equatoria Province in early 1988. The
Ethiopian army also allowed the SPLA freedom to operate in several areas
of western WoUega and Illubabor. From 1986-90, the SPLA had a military
base at Duul, just inside Ethiopia opposite the Sudanese army garrison at
Kurmuk. Local administrators and senior civilian politicians reportedly
objected to the SPLA pieaeaoe, bot were oveaiiled by die mflitaiy
commaiid. For its part, the SPLA fought against the OLF. This oocuned
right up until May 1991, when SPLA contingents fonÂ aloiigside the
Ethiopian army at Denibi Dcdo, near Gambela.^'*
The SPLA presence in WoDega Ethiopia led to a number of abuses.
The enslavement of escaping resettlers has been discussed in chapter 12.
Cattle raiding in Keffa and Gamu Gofa will be discussed below. It
attacked Oromo refugees and displaced people in August 1987 and

November 1988, killing several civilians on each occasion.*^
On November 9, 1989, the SPLA attacked a refugee camp at Yabus
in Blue Nile Province, Sudan, and burned it to the ground. Fortunately,
the 10,000 refugees had evacuated the camp when they learned of the
likelihood of the attack, but were foiced to spend many days in the wild
without supplies. Yabus camp lay m
an area contestod between the SPLA
and the Sudanese anny, and relief items had been regulady bfought to the
camp with OLF and Sudanese army military escorts. It is therefore
understandable that the SPLA n^gbt have suspected that it also performed
a military or intelligence function.
Perhaps the most common abuse was the requisitioning of supplies and
stealing of cattle from the local populations, with accompanying violence
against civilians. Simon MoUison, a visitor to Berta areas of western
Wollega, controlled by the OLF, in March 1990, described some of the
damage caused by SPLA units:
There are two examples of the Sudanese army giving comparable military
One was support for the Ethiopian Democratic
assistance to rebels in Ethiopia.
Union offensive in Gender in 1977, and the other was the OLF offensive in 1990,
described below.
Sudi raiding was also undertaken by local paitofal gmupi armed by the
Sudan government to fight aguoit the SPLA.
Unconfirmed claims by the OLF are that 37 were killed in the first attack
and 19 in the second.
326
The damage caused by the SPLA was greatest near the Sudanese border
and here the ruins of many villages are the only signs of the area having
been inhabited. Some of these are now being rebuilt but others are only
shown to have existed by the often-singed groups of mango trees in
the bush. In villages destroyed more recently, ruined houses still have
the charred renmants of human habitation - broken pots, lamps, etc.
Deeper inside [Ethiopia] (50-6(Naii) village were not destroyed but

had been regularly looted by the SPIA. In one village people told how
over 1,000 SPLA troops had regularly set up a camp in the village.
They would demand food and money from the villagers and strike them.
Twelve people had been shot. Even the clothes they were wearing
would be taken. Eventually this treatment had impoverished them to
the extent that many of them fled to the bush, where they were mainly
living on wild foods.... In another village I was told how the SPLA
had stolen many of their animals and had burned their grain stores.
They had been scared to cultivate in the immediate vicinity of the
village.
The OLF offensives of 1990 and 1991 effectively drove the SPIA from
the northern and southern parts of western WoUega respectively.
Groaps armed by the Sudan Govemownt
The Sudan government employed a very different strategy to its
when faced with

Ethiopian counterpart insurrection. Instead of using a
large and well-equipped conventional army, the Sudan government chose
to give arms, support and training which
to locally-based militia groups,
would then attack the forces of the SPIA with a greater or lesser degree
of coordination with the army. This policy started in 1983 and persists
up to the present. Some of these niilitias, such as the Mumhaliin of
southern Darfur and KordofEui regions^ have been responsible for some
of the grossest abuses of human rights witnessed in modem Africa,
including laige-scale massacres of civilians, slavery, destruction of villages,
and deliberate starvation.^^
Close to the border with Ethiopia, the Sudan government armed several
"friendly" groups. These included the Anyanya 2 paramilitary force, mainly
drawn from members of the Nuer ethnic group, and Toposa and Murle
" See: Africa Watch report. Denying "The Honor ofLiving: " Sudan: A Human
Rights Disaster, March 1990.
327
militias. into Ethiopia as well as attacking the SPLA
These groups raided
and its sympathizers. One
cattle-herding people, the Chai, who live doie
to the Sudan border, lost almost all their cattle by 1988. Many people were
killed, and the remnants of the group was forced to letieat from Ibeir
existing grazing land to the Maji motmtains.
Military Activity m Wollega 1989-91
In Noveniber 1989, an SFIA attack across the border from EthiofMa

succeeded in capturiqg tlie town of Knimiik. TUs wis flie second time
that Kutmuk had beeo taken, and as on the ivevions oocaskm two yean
before, the Sudan government launched a major campaign to recapture it.
In the face of superior fioioes> the SPLA withdrew. There then followed
a period of close military cooperation between the Sudanese army and the
OLF. In the first days of the new year, the OLF launched a major offensive
in western Wollega, with assistance from the EPLF and possibly the
Sudanese army, which certainly launched a simultaneous attack on SPLA
positions inside Sudan. Within a few weeks, the Ethiopian army and the
SPLA had been driven out of their major positions in northwest Wollega,
indttding the strategic town of Asosa. Though a fierce counter-attack
followed, and Asosa was recaptured, the OLF gained the ascendancy hi
the area.
The OLF Capture of Asosa
The OLF capture of Asosa on January 5-10 witnessed a number of

abuses against civilians. Hie first incidents were against the refugees in
Tsore camp.
Tsore camp contained about 42,000 Sudanese refugees. The majority
had been displaced by fighting in the area in 1987—8, though some were
very recent arrivals. About 60 per cent were speakers of the Uduk language.
Accorduig to reports, a warning was delivefed to the SPLA (by either
the OLF of EPLF) that an attack was hmnuient and that the lefi^gees shonU
be removed firom the area for their safety. However, SPIA foiled to
effectively pass such a warning on to the residents of Tsore, who learned
of the impending attack only ên the gunfire came within earshot. Most
then collected their possessions, abandoned the camp, and headed southwest
towards Yabus Kubri, just across the border in Sudan. According to
testimonies later obtained from the refugees, they passed safely among
people they identified as "Eritreans," but while travelling through a mountain
ravine they were shot at by the Du walla people, who were armed and
instructed by the OLF. An old man and several young girls, some of them
328
canying babies on Ifaeir backs» were killed by being shot and falling into
a lavine.
About 120 refugees who did not evacuate the camp were killed in the
OLF attack, some of them deliberately.
At Yabus Kubri, the displaced refugees and a contingent of SPLA
fighters were bombed by the Sudanese air force. They fled to nearby hill
villages. There was then a ground attack on Yabus Kubri by the Sudanese
army, which used artillery to shell the hill villages where people were
hiding. The people were forced to run southward again, leaving sick people
behmd on tilie road. Tlie refugees moved through a snccewion of places.
On the way, they suffered fam^ and finlfaer aerial bombaidment; and heaid
of threats of more ground attacks by the Sudanese army. Finally the
refugees were directed by the SPLA to cross back into Ethiqna, at Pagak,
where they arrived in batches between March and June. A representative
from UNHCR visited the refugees in their makeshift camp in Pagak and
provided some food rations. The camp at Pagak (which is inside Ethiopia)
was then bombed, presumably by the Sudan air force. For reasons of cost,
UNHCR decided against a new camp at Pagak and instead transferred the
refugees to Itang camp between April and July.
The OLF denied that the attack on Tsore camp took place. It did not
attempt to justify the attack on the grounds that the camp was also used
as amilitary base, though thm Is strong evidence that the SPLA ntiyzed
an the refugee camps m Ethiopia for military purposes, and the camp
occupies a strategic location ck»e to the Koniiuk-Asosa road, iiîm tiie
OLF could not tmve safely allowed the SPLA to remain. There is also
a striking similarity between the attack on Tsore and the destruction of
Yabus camp by the SPLA just two months earlier, so a motive of simple
reprisal cannot be ruled out.
The OLF also attacked trucks briqging suf^lies to refugee camps on
at least one occasion.
The OLF overran Asosa town and the resettlement camps in the area.
Both of these may be counted as military targets — the town had an army
garrison and the resettlement camps (collectively known as Gojjam Sefer)
contained about 6,000 militiamen who had been used to secure the area
and extract taxes, forced labor and conscripts. Both also included
substantial dvflian and non-combatant popohmons. In the attadcs, there
were civilian casualties, and there was at least one incident in which
Amharic-speaking civilians were deiibeiatdy killed; accoiding to different
versions, shot or burned to death.
The Ethiopian government made great propaganda out of the atrocity.
Supposed eye-witnesses to the incident were interviewed, who claimed
that non-Oromo people were rounded up and instructed to gp for a meeting
329
in a school, where they were nwchine-gunned. One man claiming to be
an eyewitness said that the attackers spoke Tigrinya (the main langm^
of Eritrea and Tigray). Full details of the incidorta have never fully cone
to light.''
While evacuating, the Ethiopian army burned and looted at least three
villages: Shigogoo, near Asosa; Shaoto, near Bambesi; and Koiôo, near
Kobar.
The Army and Air Force Counter -Attack
The army re^xmded with a bombing campaign. Aooofding to OLF

claims, mosdy sttbstantiated
by mdependent somoea, Ihefolkiwing airiaida
occorred:
* January 7 and 8: Asosa: nine killed, 15 wouided, indudiiig women

and children.
* January 10: Asosa: ten killed, five wounded.
* January 15: Bambasi, east of Asosa. (For this and the following raids,
no casualty figures are available).
* January 23: Mandi, east of Asosa.
* January 23: Dalatti, east of Asosa.
* January 26: Bambasi.
* January 27: Hopha, north of Asosa.
* February 7: Hurungu: many houses burned.
* February 8: Arge: many cattle killed.
* February 8: Buldugilin, north of Asosa: many houses bonied.
This was followed by a ground assault,which took the fonn of a military

action to recover the (now-deseited) town of Asosa, and a aeries of punitive
^*
In one respect at least, this prapiganda backfired: tfw OLF lepoited a new
influx of recruits from govanment-ooatroUed areis^ who vohimccwd in aider
to "kill Amharas.**
330
expeditions in the rural areas.Simon Mollison, visiting western Wollega
in March 1990 reported some of the actions of the army and their effects:
Closer to Mandi ... some raids on Berta villages had been carried out
by government forces. Such raids appear to have been a recent
phenomenon. They have commonly taken the form of government
troops entering a village and stealing animals, food, money and
possessions from the houses of the people, who had fled at the first
sight of the aimed force. Those late to leave had been shot at and some
had been killed. This happened at the village of Feidos» for example,
a little more than a month ago [mid-Febniaiy]... They say that at least
one man was killed. Houses were bfoken into and looted but no burning
took place.
The village of Sirba, inhabited by the Sese people, was attacked by a

government force in January. At least two of the local people were
killed and their bodies were tied to trees, but others had time to escape
because the attackers were seen while still some distance off. The
village was looted and much of it was destroyed. The people have not
returned to rebofld the village as they feel it would remam a target.
Sirba was the centre of some missionary project and had an airstrip.
It can thus be seen as having been quite an unportant centre for the
area. Because of the burning and the method of dealing with the
corpses, this raid seems to have been motivated by a desire to punish
the villagers.
Harangama is a small village by the Blue Nile about six

in the hills
hours'walk from Oda. It is a Gumuz and was attacked by
village
government forces in September 1989. Six people, who were slow
escaping, are said to have been killed. The village was burned and most
people lost most of their possessions.
I talked with some Qromo people in the lowland bamboo forest about
two hours from Ferdos. They told me something of the reason that
they had fled their homes to diis previously unsettled area.
Government forces had regularly and systematically "raided" their

villages.They would take the villagers' money and cattle and also their
children (who would be taken to fight in the army). One old man I
talked with told me his story. He had arrived to settle in this lowland
area two months before [i.e. in January] having finally decided to leave
the collective farm, Baamichee, where he had lived for the last eight
331
years. Previously he had lived in a village called Gumbi where he had
been "a rich man". He had owned many cattle and a lot of coffee
bushes, he said. His farm in Gumbi was "given to others" (possibly
Tigrayans [i.e. resettlers]). Life had got worse and worse in Baamichee.
At first it had not been bad but now the government had taken nearly
all he produced (maize, luba, peppers, tef). They had taken some of
his children and some of his cattle. In the end he was so anxious to
escape get away that be left 20 cows behind, taking only two and a
litfle money. He escaped with ninB otber funito Iw w
one who managed to get iway with anything. lieisnowlivhKmthis
ww
the oo^
lowland area (they seem to call it Buche) where people would never
have dreamed of living once. It is okay in the dry season but not in
Iheiains. The malaria will be terrible and the aoflk a heavy black clay.
Other Oromos I talked to in Buche had similar stories to tell. Oromo

refugees from Asosa town who I had talked to in Bikorri in Sudan also
told stories of terrible taxation, straight theft and always of their children
being stolen and sent to the war. The number of Oromos who have
"voluntarily displaced" themselves to the Buche area where they live
m great poverty and in an area Aat will obviously be a swamp in the
wet season is further testimony to the conditions tey decided to leave.
In the first three months of 1991, about 4,000 refugees from this aiea
crossed the border into Sudan. They included a disprqxvtioiiate munher
of young men and teenage boys, fcariog coosaiptioiL
Burning Villages
Throughout 1990 and into early 1991, the army was active in destroying
villages, spreading southward and eastward out of the border area as the
OLF gradually gained ground. The testimony of an army deseiler obtained
by Dr Trevor Trueman, a physician woiking with a fanDumituian agency,
and circumstantid evidence suggests that the army utilized special 'ta
squads". Members of tiie squads would enter a village after warning shots
had scared the inhabitants awiy. collect farm tools and other implements
and place them inside the houses, and then systematicaliy bum all the
houses. Crops were also burned.
The foUowii^ incidents have been confirmed:
* September 1990: Al Amir area, several villages: 600 houses burned.
332
* September: Bcla Bangaa village, TuiUu district, bumed and attacked
from the air, about 700 houses burned.
* October: Gibao, near Kobar: about 300 houses bumed and looted by
soldiers.
* November 8: Shirkale village, near Asosa, bumed in an attadc by

helicopter guoships, killing one civilian and woondiqg five others.
* November 21: Fongo, near Kobar: over 90 houses bumed by soldiers.
* November 26: Ego Kummk village attacked by helioopler gunships.
* November 28: Ego Gomooo village attadsed by helicopter gunships.
It will be clear from the details given above that the government military
activity from 1989 onwards was causing great impoverishment. Certainly,
visitors to the area report that local people said that they had become much
pooler. Dr Tmeman estiinated that in Al Amir the aveiage area fumed
by one family had fdlen from 1-2 hectares before the ndUtaiy actions to
about half a hectare afterwards. The cumulative effect of enfioioed
villagization, collectivization, destroction of houses and farms by the army,
and forcible di^dacement were added to by the prevention of trade. Donkey
trade almost came to a halt in the late 1980s on account of the bmtality
and demands for money of army patrols and checkpoints. It only restarted
after the OLF occupation, but the retreating government forces took with
them cars and trucks, severely dismpting bulk trading.
all Food prices were
climbing fast in early 1991.
There can be no doubt that if the government's counter-insurgency
policies had continued for another year or so, and spread to a larger area,
fandne would have ensued m
westm Wollega* de^te the fact that it is
an exceptionally fertile and well-watered area.
Conclusion
Most of the population of the lowlands of Wollega has gone through,

inan accelerated form, a version of the destruction, impoverishment and
displacement that occurred in Tigray and Eritrea over a longer period.
There were certain added complications such as resettlement, and the
manner in which the Sudanese civil war directly spilled over into Ethiopia.
333
While the OLF is in ocmtiDl tit the greit mafority of the area at the time
of writing, certain lowland areas adjacent to the Sudan border reoEuin
beyond any form of civil authority. Whether peace and stability are
established in these parts depends crucially on developments in the Sudanese
civil war. Re-establishing civil administration and the rule of law, and
resolving the many local ethnic disputes that are certain to arise in this
area, will demand considerable skiU and patience horn the incomii^
administration.
The Anuak of Ulubabor
One group of peripheral people who suffoied particiilarly fitom the

government's policies woe the Annak, ^fbo are a NDotic people who inhabit
an area straddling the Sudan boida. Many Anuak bore the brunt of
violently-enforced resettlement and villagizatiaii, and then were
overwhelmed by an inflow of Sudanese refugees, accompanied by the
SPLA. However, as in the case of most peripheral people, the government
followed a strategy of "divide and rule." Certain groups of Anuak certainly
benefitted from government patronage and policies. The government armed
an Anuak militia, which however had uncertain loyalties. The Anuak in
Sudan also suffered from the neglect of their area fay successive Sudanese
governments, and the outbreak of the Sudanese cml war hi 1983, and thus
were more sympathetic to the Ethiopian govemmeot.
The Ethiopian Anuak nunibered an estimated 56^000 m
1970.
Historically and culturally they have greater ties wifli tfadr neighbora in
Sudan than they do widi the Ethiopian highlands. Their well-watered area
has, however, been coveted by successive Ethiopian governments. In 1979,
many Anuak were evicted en masse when the government set up irrigation
schemes on the Baro river. Amhara settlers were brought from the north
to farm the schemes. This coincided with an intense conscription campaign
for the army.^' Several hundred Anuak were killed by the army, and in
response the Gambela Liberation Front (GLF)^ was set up in 1980. It
had links with the OLF and operated through Sudan. There were clashes
between the army and the GLF m
1982, with government lepiiBals against
the civflian population.
In 1979, 4,000 Anuak and Nuer fled to Sudan, daunmg their hmd had
been confiscated by the government. Shortly afierwardb, 3,000 ci their
^'
Survival hiteroational, 1991, pp. 3-4.
^ Renamed the Gambela People's Liberation Movement in 1S)8S.
334
southern neighbors, the Begol, also fled to Sudan. Without any
investigation of conditions on the Ethiopian side of the border, the UNHCR
floated a proposal to repatriate these refugees in 1983.^^
The resettlement program in Gambela in 1984-6 involved another round
of land confiscation. It also meant that the Anuak population was matched
by 70,000 new settlers.
In June 1986, villagization began among the Anuak population of
niubabor. The program was implemeiited in a particularly severe manner.
The Anuak populatioo was oompeUed to relocate in villagBS integrated with
tiie recently-set up resettlement sites. The new villages have been described
as more akin to forced labor camps.^
In April-May 1987, the le-named GLF staged seveial attacks on the
resettlement sites that were now "integrated" with the new Anuak villages.
A ch'nic in Abol settlement was attacked, apparently with the intention of
killing government cadres. Government reprisals included the killing of
a number of Anuak, and the enforcement of a strict curfew and related
restrictions in Gambela town and nearby areas, by a militia drawn from
among the resettlers.
The Anuak were also victims of attacks by the SPLA, with numerous
credible but unconfirmed reports of IdUings of civflians. After 1988 the
SPLA gained full control of the Sudan border, and GLF mflitary activity
ceased.
In May- June 1991, when the OLF and EPRDF occupied Gambela, the
GLF was given a large role in the administration of the area. The Anuak
militia was partially disarmed. The GLF is reported to have engaged in
attacks on resettlemeitf sites, ktoting villages and killing tens of civîans.
The SPLA and Sudanese Refugees
Starting in 1983, Sudanese refugees began to flee the war in southern

Sudan. Many headed for Ethiopia, where the SPLA and Ethiopian
government gave them protection, and intematicmal ag^cies provided
assistance.
Two camps were set up in western IQubabor and one in the adjoining
area of Keffa. By 1990, Itang camp had 270,000 r^fiteved refugees, though
in reality it probably held about 150,000, due to some having double-
» Ahmad Kaiadawi, "Refiêe Ftolicy m the Sudan, 1967-84,** DFhtt thesis,

Qiford, 1988, pp. 112, 332.
^ Sandra Steingraber, "Integrated Settlement in Gambeiia: Araied UfnisiDgs

and Government Reprisals," in Clay et al, 1988, pp. 236-7.
335
registered and others having left the camp but remained on the register.
Fugnido had 85,000 registered, though actually probably 50,000-60,000.
Dima, in Keffa, had 35,000. The population of these camps was about
75% young men and boys. This reached an extreme at Dima camp in 1988.
where the population was 97.8%, 86% of them aged between 15 and 45.
Part of the explanation for this is that the women and young children had
fled the war and famine in a different direction (to northern Sudan), and
part is that the refugee camps also operated as military bases for the SPLA.
Security was always camps. Foreign viaUm was
tight at the refugee
given only guided torn, and usnally not allowed to atay ovemight The
SPIA presence was strong, and mifomied SFLA members were often
piesent.According to a former ConmiiSBioaerte Relief and Rehabilitation,
internationally-donated food aid was diverted to the aoldiers of the
SPLA.^ In 1991, it was commonly estimated that 20 per cent of the food
destined for the camps was diverted to the SPLA, but visitors to the
neighboring areas of Sudan report that much of this "diverted" food was
in fact being eaten by civilian relatives of the refugees inside Sudan.
However, the diversion of ten per cent of the food would have been
sufficient to feed half of the combatant members of the SPLA.
According to an arrangement reached with the Ethiopian government,
the SPLA was given a free hand in mnch of Dlubabor proviiice, in letmn
for keeping the GLF in check.
In the lowlands of KefEs province, cattle raiding by SPLA units was
common. SPLA soUiers would demand cattle from the local populatkm.
The Refugees Return to Sudan, May 1991
In a few days in May and June 1991, almost the entire population of
Sudanese refugees in Ethiopia returned to Sudan, precipitating a
humanitarian crisis and a major relief operation led by the UN. Allegations
were made — some by the SPLA —
that the refugees had been forcibly
driven out by the EPRDF and/br OLF, or diat aSndaneae amqr attack on
the camps was planned. None of these claims woe trae. The SPLA later
said that they had led the refugees out, fearmg for die safety of refogees
during a Ineakdown of law and order.
° UNHCR Technical Support Service, "Ethiopia: Health and Nutrition

Assessment of Somfaem Sudanese Refugee Gnups m
KeAi, Ulnbabor and Wokgi
Awiajas, 8-22 March 1988," Geneva, 1968, Annese F.
^Dawit Wolde G\oxp^, Red Tears: WBr,F4muneandRevoluikM mEOiwpia,

Trenton, NJ, 1989, p. 158.
336
In late March 1991, EPRDF forces crossed from Gojjam into Wollega.
In the first ten days of April they captured the provincial capital, Nekempte,
and other key garrisons. Defeated government soldiers looted several parts
of Nekempte town, including stealing property and vehicles belonging to
the UN relief operation for Sudanese refugees.
In occupying Wollega, the EPRDF cut the route between Addis Ababa
and the Sudanese refugee camps. Oa April 5, an EPRDF rqxesentative
assuied the UNHCR in Geneva ttiat the Flnxit was wiUiqg to cooperate
with cootiniied relief programs to the camps^ allowing fine passage of relief
commodities. Hie UNHCR never took up the Q£6»r» presumably because
the negotiations necessary to obtain an agreement from both sides would
have taken too long, the situation was too unstable, the position of the OLE
was unclear, and the UN was fearful for its staff following the looting
incidents in Nekempte and the OLE attack on Tsore in 1990. This meant
that there were no more food deliveries to the camps. However, grain
remained in store, and some distributions occurred.
More EPRDE said it was willing to see the refugees
generally, the
remain would not accept the armed presence of the SPLA.
in Ethiopia, but it
In A|^ and May, four separate attenqits were made by mediators to bhng
representatives of the SPIA and EPRDF together to discuss flie plight of
the refogees. On each occasion tiie suggestion to meet was rejected by
the SPLA "why should we talk to tiiem?" was the SPLA attitude.^
Meanwhile there was a series of meetiqgs between the EPRDF and the
OLF, and the two organizations agreed to coordinate their military strategy
in the southwest. This led to a cooidinated advance southward towaids
Illubabor.
When President Mengistu fled the country on May 22, the imminent
collapse of the government was clearly evident. in the camps
The refugees
were tense, apprehensive of a repeat of the Asosa incident. The relief
agencies feared for the safety of their staff, and began to withdraw from
Itang, the most norAerly camp, on May 24-25. UNHCR evacuated its
staff at the same time, taking the keys of the food stores and other vital
equipmoit. Thehospitij at Itmg was left without doctow or admim'strators.
The Uust staff members left on the momhig of May 26.
On the evening of May 26, OLE forces approached the Ethiopian army
ganrison north of Itang, and EPRDF forces approached Gambela town.
The advancing forces shelled both garrisons, and the garrisons replied.
However, the govenmient soldiers quickly left, setting fire to their houses
^ "Sudan and Ethiopia; No Rebel Unity," Middle East InsermtUmal, June 14,
1991.
337
and an ammunition dump, which caused a series of explosicms. Several
bridges across the Baro river were destroyed, almost certainly by
government soldiers. Shells fired by the advancing forces landed close
to the camps. The shelling and explosions alarmed the refugees —
especially those who had previously fled from Tsore.
Earlier in the year, the SPLA had laid contingency plans for an
evacuation of the refugee population." The refugees had been warned
beforehand that they might have to leave, river transport was arranged for
community leaders and adndnistraton^ and the migratioo to Sudan was
conducted in a lenuukably oiderly manner.
Throughout the day of May 26 and the fSoUowipg nigj|it» the camp of
Itang was evacuated. Most people headed for Nasir; smalkr numbers went
to Akobo and to the other re^gee camps, from where they returned to
Sudan. Some people went back to Itang the following day to collect
possessions, but members of the government Anuak militia were present,
engaging and this deterred them. There is some evidence that
in looting,
this looting had been planned in advance, for example Anuak militiamen
had prevented the camp administrators opening the food stores the day
before. The Anuak militia and unaffiliated bandits also preyed on small
groups of refugees as they trekked towards Sudan. A
Gaajak Nuer militia
(also aimed by the Ethiopian government) crossed into Sodanese territory
(partly because of a conflict with tfie Anuak militia) and preyed upoo tlie
refugees there. Dead bodies of those killed by these militias and bandits
floated down the Sobat River the followiqg week.
The camps at Fugnido and Dima were evacuated over the following
weeks. The refugees from these camps neither saw nor heard any sign
of the OLF or EPRDF forces. In Fugnido they were reportedly warned
to leave by armed local people, and migrated to Pochala in Sudan in
accordance with SPLA instructions. In Dima, the SPLA closed the camp,
looting and destroying vehicles and other property, and ordered the refugees
to leave to Pakok mSudan. The SPLA forces then made a stand inside
Ethiopia agamst tiie EPRDF forces untU tihey were forced to leave m
early
July.
The Sudan government was aware in advance of tlie likdy lelnm of
the refugees, and closely monitored the return movements, by listeniqg
in to radio traffic and sending airplanes to overfly the area. The Sudanese
aur force bombed Nash: on May 14 (killing 49 people and woundiqg 50,
^ Proposals had been circolatiitt for some time for a gradual return of the
refugees to their homes. Western ODnois had disagreed over whether it would
be better for the refugees to remain or return. Some refugees had also made
independent plans to retum after the harvest of late 1991.
338
and forcing the evacuation of the town) and on May 22, killing one.
Columns of returning refugees were also bombed, at Jokau (oa the way
to Nasir) on May 30 and Akobo on May 31. The bombing was carried
out from a great height and was highly inaccurate and caused at most one
fatality.
A UN-led relief program, including an airdrop of food to Nasir, was
implemented almost immediately. Like other such programs in southern
Sudan, it has been subject to delays and restrictions by the Sudan
govenunent. To date, ^fle a laige relief-dependent population exists in
Nasir, POchalla, Akobo, Pakok and in the snnoiiiiding areas, the prevuus
good nutritional state tit the refugees, the local resources of the area, and
the relief ppogram has pcevented die eAtieuies of fnnine.
Conclusion
The continued support of the SPLA for the Mengistu government until
its final days was a debacle for the organization, and particularly for its
leader Col. John Garang who was personally identified with the policy.
The SPLA lost military supplies and bases, its radio station, and a haven
for its civilian sympathizers. These factors contributed to an attempted
coup by the SPLA military commanders in Upper Nile province in August
1991. Hie outcome of the split in tfie SPLA remains imcertaia at the time
ofwriting. Relatkms between the EPRDF, OLF and the two wiqga of the
SPLA will be an impoctant detenmnant tit the peace and stability of the
border region in the foreseeable future.
Lowland Gamn Gofa: Carrien of New Gmis

The lowlands of Gamu Gofa, adjacent to the frontiers of Sudan and
Kenya, is the remotest periphery of Ethiopia. The peoples of this area have
never been fully controlled by the highland states —
they are peripheral
but not subjugated. They are mostly cattle-herders, and have a history
of inter-conmiunal violence. However, in the 1980s, this violence changed
maiicedly, with the supply of modem automatic weapons to some groups.
This weaponry upset tlie pievioiisly existuig state of approximate balance
between different groups, and led to anprBoedenled numbeis of civilian
deaths. The increased level of violence was also caused by^ and in turn
caused, direct military intervention by the Kenyan aimy, and may yet
provoke a similar response firom the Ethiopian government.
339
Conflicts uptoOie 1980s
The river Omo drains into Lake Turkana in Kenya. In its lower reaches,
itpasses through territory inhabited by pastoral groups, such as the
Dassenatch (also known as Marele), Nyangatom (also known as Dongiro
and Bume), Mursi, and Hamar. Their immediate neigbbois in Sudan
are the Toposa and in Kenya are the Turkana.
These peoples have traditionally conducted armed conflict between
themselves. Some of this conflict consisted of cattle raiding, and some
of disputes over territory. Social anthropologists have observed the rules
foUowed in this local waifBre, wfalcfa indnde attempting to maintain
leciprodty in attacks, andfbmaliziflg idations and btnmdaries after imods
of hostility. Dr David TVuton, who has been studying the Muni K>r over
two decades, describes a typical raid carried out by the Hamar, which
occurred on December 25, 1969, in the Eima Valley:
In the early hours of the morning a rifle shot was heard by people living
nearby but it was assumed that the stock of this [cattle] camp were being
worried by hyenas. Later it was discovered that the camp had been
raided and three people killed —
the herd owner, who had been shot,
and his two sons, aged about seven and thirteen, who were lying where
they had been deeping wifli their IfaiOBis cot All the cattle had been
taken and their tracks led m
the diiectioa of te Mago Valte^^ Hie
tracks of the laideis mdicated that there were no mm
than ikwr of
them.^
can be seen that the fighting involved "civilian" loss of life. The
It
hostilities also contributed to recurrent food shortages, not just because

loss of cattle or farmland meant loss of food, but because fear of raids led
herders to take measures such as keeping their animals in large, well-
protected groups, thus not utilizing grazing resources fully, and caused
fanners not to cultivate outlying fields.
^ "Before the introduction of firearms, this particular group called themselvM

Nyam-Etom CElephant-Eaters'), which stressed their hunting abilities, but after
the acquisition of guns, they rephrased this slightly to Nyang-Atom ([carriers of]
'new guns') which stresses their bellicose qualities instead." Jan-Ake Aivaisson,
Omo Valley, EOdopia, Uppnla, 1989, pS7 (quodqg SeigD Tonty).
^ David Turton, "Warfare, Vulnerability and Survival: a Case from

Southwestern JBthiopia," Cambridge AiUhrapology, 13.2, (198S-9) pi 71.
340
David Turton also describes a series of wars between the Mursi and
their immediate neighbors in the highlands, the Bodi, over territory. Wars
occurred in the early 1950s and between 1971 and 1975, and consisted
in occasional raids and ambushes, with long quiet periods m between, until
a formal peace agreement concluded the conflict and re-drew the territorial
boundary.
These conflicts and the measures taken to preserve security directly
contribnted to fiemiliie in the area in 1971-3, wlien they ooincided with
drought, llie drought and fnnine itself kd to increawdpcesBiiie on nat^
lesources and led to moie oonflict with die BodL Disease, fanqger and
homicide all accounted for high levels of mortality during tfiose years.^
Between 1968 and 1971, a war was also fought between the Bodi and
their eastern neighbors, the Dime. The Bodi enjoyed the advantage of
superior access to firearms, and were able to undertake raids with relative
impunity. About 700 Dime men, women and children were killed and a
further 1,000 forced to leave the area, a considerable loss to a population
totalling no more than 11,000. The war was brought to an end by Mursi
attacks on the Bodi and a government punitive expedition which confiscated
cattle and fireaims.^
Another local war was fought between the Dassenatcfa and the
Ny angatom.'^ In the late 1960b, the Dassenatch, nnder fxessnie from the
rismg wateis of Lake lîdfana, which was flooding their fiundand, b^gan
to press on Nyaqgalom territory. The Kenyan pofice enforoed a peace in
1966 between Dassenatch, Nyangatom and Toposa, which involved homing
sevend villages and trying to niake the Ilemi Trianigle a "noHi^
* David Turton, "Response to Drought: The Mursi of Ethiopia," in J. P. Gariick

and R. W. J. Keays (eds.) Human Ecology in the Tropics, London, 1977, p. 180.
" Dave Todd, "War and Peace between the Bodi and Dime cf Soutfawestem
Ethiopia," in Katsuyoshi Fukui and Davkl Turton (eds.) Warftre amamg East
AfirUan Herders, Osaka, 1979.
The following is based upon a detailed account of this conflict, in: Serge
Tomay, "Armed Conflicts in the Lower Omo Valley, 1970-1976: An Analysis
from within Nyangatom Society," hi Bdaii and Tuton (eds.), 1979.
^ The Ilemi triangle is an area of Sodaneae teii iUa y, acâcent to Ethiopia,

that hasbeen administeied by Kenya, under agreement, since colonial days. The
international frontiers in this area have been drawn without lefnenoe to the
boundaries and migration patterns local ethnic groups.
341
This peace began to bieak down in 1968-71, with killings by all groups^
including three Kenyan policemen killed by Dassenatch in July 1970.
In 1972, this developed into a serious Dassenatch-Nyangatom conflict,
with each side raiding the other and killing between six and ten people
in four separate incidents between March and early June, followed by a
major Dassenatch attack on three settlements south of Kibish on June 20,
in which 204 Nyangatom men, women and children were killed
at least
as they slept or awoke. The Nyangatom were driven from their fields befoie
tliey could harvest, and lost many cattle; faniiBe lesnlled.
In Januaiy 1973, a joint Hamar-Kaia war party attadoed the Nyangatom,
killing between 80 and 100. After the Ethiopiaa police Med
to respond
to Nyangatom appeals to intervene, the Nyangatom setaliated and killed
104 Kara at the village of Kurdam the following month. On June 21, the
Hamar-Kara alliance attacked the Nyangatom at Aepa on the Omo River,
killing about 60. The Nyangatom did not retaliate, as they were preparing
(jointly with the Toposa) a The raid was only
raid against the Dassenatch.
a partial success: the intended victims managed
escape and only five
to
were killed, 3,000 animals were taken, but 20 of the raiders died of thirst
on the way home. The Dassenatch counter-attack in December at Kibish
left20 dead.
Father clashes continued mto 1974, with at least 41 fMalities. The
Nyangatom were in the ascendant: diereafler the Kara were obliged to
become the lesser partners in an alliance with their erstwhile opponents.
During this period there were sporadic attempts by the Ethiopian and
Kenyan governments to control the local warfare. This included paying
compensation for cross-border raids, negotiating settlements, undertaking
punitive patrols (four against the Dassenatch), and on one occasion, aerial
bombardment of villages (the Ethiopian government against the Hamar).
Between the mid-1970s and 1986, the level of violence was much lower,
with only 28 confirmed inter-tribal homicides.^'
These conflicts renuuned under the control of the leaders of the
respective ethnic groups. While involving regular vidlenoe and homicide,
the problems renuuned withm well-defined limits^ and the level of militaiy
technology was low. In the 1980s, with the intervention of regular armed
groups, notably the SPLA and the Kenyan army, and supplies of modem
weaponry from these sources and from the Sudan government, conditions
began to change, and bloodshed on a larger scale began to occur.
Alvarsson, 1989, p. 77.
342
Ihe Nyangaiom Massacre of the Muni
In the mid-1980s, both the SPLA and the Sudan government began
to distribute automatic weapons to a number of cattle-herding people close
to the Sudan-Ethiopia border. These groups then participated in cattle
raiding inside Ethiopia. One group that was heavily armed was the Toposa,
who were supported by the Khartoum government as an anti-SPLA militia.
Tb/i Toposa in turn distributed aims to local aUies inside Ethiopia,
prominent amcmg whom weie Ihe Nyangatom.
The Nyangatom were one group which benefitted from the Sudanese
suppljr of weapons. The Chai, as noted above, lost out heavily. The Mursi
were next in line. Mursi-Nyangatom conflict has been longstanding,
interspersed with periods of friendly relations and indeed interdependence
with mutual trade. A
typical incident of homicide occurred in April 1985,
when two Mursi boys were shot dead by Nyangatom. Cases such as this
were not considered exceptional; a cause for retribution but not for upsetting
a fundamentally equable relationship. In 1987, however, events occurred
out of all proportion to what had gone before.
In January or early February 1987, sxx. Nyangatom who were visiting
a Musi viUt^ to buy grain were killed by th& liDSi% using guns and
bush-knives. This was considered an outii^geoasvioktioa of local norms
of hoqiitality. In retaliation, tiie Nyaqg^rtom launched a massive raul on
February 21. Equipped with automatic weapons^ the destruction was
unprecedented. Aman who lost three family membeis m
the attadc
recounted what happened.
The Nyangatom crossed the Omo Kara village of Dus, south of

at the
the Omo-Mago moved northwards up
junction and, guided by Kara,
the east bank of the Omo, crossed the Mago and attacked the southern
Mursi, who were now sitting targets, from the east. Thus, when the
attack began at first light, the Mursi assumed that their attackers were
Hamar.^ It was only when they heard the sound of automatic rifles
that they realised they were Nyangatom.
The slaughter was mdiscriminate, most of those killed being women

and children. This was, firstly, because a good proportion of the men
were with the cattle, north of the Dara range and, secondly, because
it was easier for men and boys, unencumbered with young children.
^ The significance of this is that the Hamar were not in conflict with the Mursi,
so that the worst that could be expected was a small-scale cattle raid.
343
to scatter and hide In bush. The majority of people woe kOledwiii
speais, having been wounded in the rifle fire. 0m particularly reapecied
elder, who was well-known to the Nyangatom, was deliberately sought
out and speared to death. The hands of women and girls were chopped
off with bush knives so that their metal bracelets could be more easily
removed ..."
Another visitor to the area met a girl who survived despite having both
wrists severed with almost surgical neatness. Another Mursi described
die aftemiath of tfie mawane:
The vultures could not eat all the coqMes. Crocodiles pulled the bodies
into the water. Hie grass down tfaeie died became of aH the fill fkom
the bodies.^
Between 600 and 800 were killed —

over ten per cent of the entire
Mursi population. Almost the entire southernmost section of the Mursi
was annihilated.
The anthropologist Jan-Ake Alvarsson spoke with Nyangatom who
had participated in the massacre and recounted how it had come about.
Ibe attadc in question had been wdl planned in advance. Hie fofoe
was supposed to charge at dawn» the vaniguaidooBBisting of four people,
armed wifli one heavy and diree light madune gnns. They were aba
to carry four hand grenades. Hie second line was supposed to give
the first one cover. They were equipped with sixteen carbines. Hie
third line carried the ordinary (Austrian) rifles," and the rest were
intended to follow suit, equipped with spears or bush imives and to
finish off those shot down by the front lines.
In reality, things turned out differently.... It is unclear whether the target
[i.e. which Mursi section] or the day of the battle were the ones planned.
Furttiennore, the aimy croaaed the Qbw aioinid 8 ajB.t BBiKli la^
planned, and the attack was not assurprisiqg as intended as people were
Paraphrase based on an eyewitness account, in: Turton, 1988-89, p. 83.
^ hi: "The Land is Bad,** a film by Leslie WoodhBad, shown on Independent

Televisioii, UK, July 17, 1991.
These weapons were widely available during and after the second wodd
war, most brought by the Italians.
344
awake. The military ocder was soon transformed into an miocdered
and undisciplined row. At least eight Nyangatom warriors were killed
from behbd by their own foioes dozing the phase of wild shooting.^
The Mursi reprisal was taken weaker part of the

against the Kara, the
attacking alliance. On March 28, Kara (including two children) were
six
killed by Mursi, who retaliated by
killing seven Mursi. In November, two
Nyangatom were killed while working on a dug-out canoe. These were
however only short term responses. The Mursi were emphatic that a
counter-raid on a comparable scale was needed before an equitable peace
ooold be concluded with the Nyangatom.
Mursi plans for counter-attack verged on the suicidal they were —
heavfly ontnumbeied andpowessed noantomaric weapoos. much greater A
danger was fiuther Nyangatom raids, wMch if caiiied out oo a con^^
8cide» could have meant the end of the Mursi as a group.
Ihe Kenyan Massacre of the Nyangatom
These plans and fears were overtaken by events. The Nyangatom were
also engaged in raiding some of their other neighbors, such as the
Dassenatch. In July 1988, in alliance with the Toposa, they carried out
one such raid About 60 people were killed in the
in the Ilemi Triangle.
attack. Kenyan government had decided to annex
Earlier in the year the
the Triangle, and was fut developing a military presence ui the area. The
Sudan government was able to lodlg^ only diplomatic piolests, as all the
sunoundmg countryside was oontiolled fay ttie SPLA, which enjoys dose
relations with Kenya.
The Kenyan government hasaloqg-standing hostility to the pastoralists
who liveon its borders, who cross the international frontiers as if they did
not exist, and who engage in livestock raiding. The administration of these
nomads has long consisted of punitive expeditions interspersed with attempts
to persuade them to live a settled life, wear clothes and send their children
to school.
On July 28, the Kenyan police clashed with a group of Toposa or
Nyangatom raiders who had pieviously attacked the Dassenatch, and came
off worst, fifteen policemen were IdUed, and some taken hostage. The
Kenyan govenunent responded tiie foUowiog day witfi an attack using
helicopter gunships and paramilitary forces on the Nyangatom area of
Kibish, which stiaddks Ethiopia, Kenya and Sudan. About 200 Nyaqgatom
" AivanBcm, 1989, p 68.
345
raidersand a mmiimim of 500 dvflians woe kUled by llie KBaym fixoes
over the foUowiiig 18 days. In the attack, at leait five vffli^ inride
Ethiopia were partly destroyed, and the Swedish Philaddphfa Misiioa at
Kib^ was burned. Tlie Kenyan araiy abo mdertook aietaUatixy massacre
against the Toposa in Sudan.^^
Recent reports indicate that another round of killings may have started,
with a reliable account that Mursi raiders killed tens of Ari in May 1991
in retaliation for the killing of one MiusL
Conclusion
Successive Ethiopian governments centered m

the highlands have shared
a similar attitude to pastoral nomads to that of tlie Kenyan govemmeat
The power base of the EPRDF is, like its pfedeoessais, located hi Ae
highlands. The EPRDF avows an ideoiogy which gives equality to all and
the right of self-determination to peripheral people. However, the First
National Congress of the EPRDF, held in February 1991, adopted a political
program that included an item resolving "to settle nomads in settled
"''^
agriculture. This implies that if nomadic pastoralists do not agree to
settle, the state is entitled to settle them by force. The history of such
attempts indicates that the nomads will resist. Democratic rights and the
enforced settlement of nomads are incompatible.
The Lower Qmo vaUey presents a more general challenge to the
government Ethiopia (and mdeed those of Kniya aod Sudan). It is an
x>f
area where thehr writ scarbdy runs, and where central oontvol can on^ be
enforcedbyextraordinarily high levels of videnoe. Oovemment-mediated
settlements of local conflicts, the so-called "stranger's peace", can be
successful only when both sides to the conflict share an interest in a
settlement, and the terms of the settlement can be enforced. At least one
such negotiated "stranger's peace" broke down in the 1970s for these
reasons. A lasting peace and respect for human rights in this troubled area
can only be achieved through long and patient interaction with the
indigenous people, undertaken by all the governments concerned.
' See: Africa Watefa report, Kenya: Takb^ Utertfef, July 1991, pp. 334-6.
^ Revolutioiiary Democratic Program of the EPRDF adopted at the Fhst

National Congress, Political Program, Article 8(d).
346
19. DIVIDE AND MISRULE: THE EAST, 1984-91
In 1983-4, the Ethiopian government completed a bloody victory over
the insurgencies in the southeast. The Western Somali Liberation Front
(WSLF), Somali-Abo Liberation Front (SALF) and Sidama Liberation Front
were all defeated and the OLF was severely weakened and would no longer
pose a sigDificaiit mflitaiy threat. A key etemeot in this success was the
Ethiopian government's strategy of fostering divisions in the ranks of the
insurgents: its support for flie Somali Natkmal Movement (SNM) against
the WSLF was the clearest example. In this strategy of "divide and rule,"
the Ethiopian government was, ironically, agisted by the Somali
government, which was following exactly the same strategy.
As a result, the lines of conflict became more fragmented and complex.
As well as attacks on civilians by the Somali and Ethiopian armies, there
was an increasing level of inter-communal violence which extended
throughout eastern Ethiopia, Somalia and Djibouti. This chapter documents
some of the abuses of human rights that resulted from this fragmentation,
for which the Siad Barre and Mengistu governments share ultimate
reqMOsihility.
Unrest in the Ogaden
The Ogaden did not return to peace after the defeat of the WSLF; but
neither was there widespread rebelli(m. Instead there was a low level of
violence between the well-armed but impoverished, restricted and frustrated
herders, and the Ethiopian army, ix)lice and members of other communities.
A breakaway group from the WSLF, Ogaden National Liberation Front
the
(ONLF) was formed in 1984, but failed to make a military impact.^
The drought of 1984 together with government policies led to a number
of violent incidents in the Ogaden. The viUagization of the highland Oromo
involved the relocation of many communities in areas which had previously
been used as pasture by the herders. Otiier areas were aUocaled to resettlers
from WoUo. lliere were a number of violent diqmtes between the
pastoralists and fanners.
* The words "Western Somali" in "WSLF" indicate an attachment to the ideal
of a greater Sonudia; the "Ogaden National" title of the ONLF indicates the belief
that the Ogaden are a nationality, not merely a dan, and indicates no lelatioiiship
with the Somali state.
347
In September or October 1987, in retaliation for an ambush on a military
lorry at Dhakato, for which the ONLF claimed credit, army units moved
to Fiiq, south of Harer, confiscated animals and killed three people.
A series of incidents occurred in August 1988. Ogadeni herders had
delivered their animals to the government Livestock Exporting Company
agents in Gode and Degahabur, and were awaiting payment. The payment
did not come; and in reprisal some herders attacked a government geological
team in the area, killingone Soviet and three Ethiopian technicians. An
anny foice responded by surrouading the liveilock hods îîcfa had
congregated at a well at Bolale and hiding the admali unlfl kOlen
were handed over. The herders denied knowing who was responsible for
the killing, and the animals were conHscaled.
Hiis event led to increased tension between the army — still behaving
like an occupying force —and the local p>eople. Shortly beforehand, the
army had refused to let the herders use a valley between Degahabur and
Aware for grazing. It had been a traditional pasture area used especially
during times of drought, but had been occupied by the army as a training
area. Use of the valley had become more important since 1984 when
another valley at Tur, near Gode, had been allocated to a resettlement
scheme for farmers from Wollo. The members of a delegation sent by
local people to protest their caBclusion were detained. Some young Ogadeni
herdeis (aimed, as is usual) wmdered into the restricted area, iîeieupon
an army unit opened fire on them. The herders fired back. Aoooiaiqg
to reports, 21 of the herders and three soldiers were kiUed. hi another
incident at Dhanaan, soldiers kiUed 17 lierders in a punitive expedition
mounted in reprisal for the killing of a government offioer, befiore senior
army officers intervened and ordered them to halt.
While most of the violence was between Ogadeni clansmen and the
army, there were also clashes with member of other clans. One incident
occurred in 1985, following a fight between herders belonging to the Isaaq
and Abasguul clans. The governor of Jijiga ordered the army into the area
and 125 Isaaq herders were reported kUtod.
Arming Refugees hi SomaUa,
The refugee population Somalia played a central role in the

in northern
conflicts that developed after the demise of the WSLF in 1983. This
occurred because the Somali government used them for military purposes
against the SNM. At first, only members of the Ogaden clan were
conscripted or armed; later this included members of aU Somali clans and
also Oromo refugees. The arming of refugees was well under way before
the outbreak of full-scale war in northern Somalia in May 1988.
Immediately after its military defeat in March 1978, the Somali
government began recruiting Ogadeni refugees into the WSLF. In 1983,
this was changed, and refugees were instead conscripted directly into the
army. A
special Division (the 12th) was formed, initially comprising 5,000
refugee conscripts, the great majority of whom were from the Ogaden clan.
The conscription of refugees to the military is contrary to international law,
under the 4th Geneva Omventioo of 1949.
Fnim eady 1984, llie 12tfa Division was used against the local Isaaq
population. Some oi tiie abuses it committed include:
* December 6, 1984: Sbeikfa area: in retaliation for an ambush by the

SNM in which the commander of the 12th Division wn lolled, soldiers
rounded up teachers, merchants and others and traMfened tiiem to
Burao, where they were later shot. A total of 43 weie executed in this
incident.^
* January 1988: Alleybadey: about 20 civilians executed.
* March 14, 1988: Gebiley and Tug Wajale: 25 conmiunity leaders

executed in reprisal for an SNM attack.^
In 1985, as unrest intensified in aortfaem Somalia and more Isaaq
soldiers and ofGcers defected from the army to the SNM, the Somali
government increasingly used force to obtain conscripts to fill the gaps
in the ranks of other divisions. Between 1985 and 1987, refugee health
workers were taken more than ten times, and high level intervention was
needed by the refugee administration to obtain their release. At one point
in 1987, 140 conmiunity health workers were press-ganged; 50 were taken
in 1986, together with a number of hospital outpatients.** In May 1987,
the Somali army raided the refugee camp at Bihin near Berbera and took
200-300 refugees for military service, and also took conscripts from the
* See Africa Watch report,/! Government at War with its Own People: Somalia:
Testimonies abontAeKulb^midaieCoi^liet im iheNorth, January 1990, p. 65.
' Afrka Watch report, 1990, p. dS and appeodk 5.
^ Africa Watch interview with Mohamed Dahir Khaire, Refugee Health Unit,
September 1991.
349
neaiby Biyoley cacap. During Sqjtetnber and October 1987, 5,000-7,000
refugees were forcibly conscripted in the camps.^
TTie recruitment operations often included violence against refugees,
including some killings. UNHCR protests to the government met with
no response. However, these were not backed up with effective sanctions.
The US and other western donors continued to support the Somali
government with generous and unconditional aid, seeing it as a strategic
cold-war ally.
Starting in November 1984, the government also created camp militia.
The rationale for Ibis was that lefugee camps were often located in remote
places, far from anny and police posts, toad increasing SNM incunkms
meant that a self-dc^nse capacity was required. Membcn of the Ogaden
clan received arms preferenoaUy.
During 1984-6, there was a new influx of Ofomo refugees, fleeing
villagization and related atrocities. Some of the new refugee camps (for
example Bihin and Bioley) were sited in remote areas, unsuitable on account
of poor access and health hazards. The rationale behind their location was
that they would provide a military presence in these areas, where the SNM
was active.
In 1985, the Somali government began to arm some of the Oromo
refugees. Anorganization was created headed by Sheikh Ibrahim Belissa,
a religious leader who was foimerly a mendier of flie Somali-Abo
Liberation Front, who was living in exile in Haigeisa. Sheildi Belissa's
Oromo front was assisled with aims and intematiooany<-donated food lud,
with the intention tiiat it would infiltrate into northern Harerghe province
of Ethiopia to engage the Ethiopian army and the SNM. Before 1988, it
did not engage in militaiy actions or commit gigntf'™* atrocities inside
Somalia.
There was also a small group of Amhara refugees in Somalia, some
of whom were pressured into joining the Somali army — the alternative
was indefinite restriction to a refugee camp or detention.*
The refugee camps thus became military targets. One of the first SNM
actions agauost the camps was the kidnapping off a itaicfa medical team
in Tug Wajale camp on January 24, 1987. The eleven menibeis of the
team were released unharmed after ten days.
Âgence France Presse, October 27, 1987.
^ Hiram Ruiz, "Detained in Exile: Ethiopians in Somalia's

Shelembod Gonp,"
United States Committee for Refugees, Washington, D.C, 1987.
350
Other SomaU^backed Fronts
The Somali government also fostered two other small armed fronts from
among refugees, residents on northern Somalia, and fanner fighters of the
WSLF.
After the demise of the WSLF, one of its divisions, known as lil Tire,
remained active among members of the Issa clan in northern Harerghe.
This was l ater r enamed the Issa and Gurgura Liberation Front (IGLF), and
jolhed flie EPRDF. It enjoyed tiie support ci the Sooudi govenmait, and
dashed with the SNM on the Somali border. It is mictearwhedier dashes
between die Issa and Oromo comnronities near Dire Dawa and the Issa
and Afar in the late 1980s were related to the IGLF, or were due instead
to armed members of the clan unaffiliated to it.
In the late 1980s, the Somali government also assisted aGadabursi Front
(also known as the Somali Democratic Alliance). The Gadabursi Front
was hostile to the SNM
and is reported to have killed a mentally-
handicapped Isaaq civilian in a reprisal attack in 19897
The conscription and arming of refugees in southern Somalia also took
place, though there was no outright warfare in the region at the time.
Peace and War in Northern Somalia, April-May 1988
On April 1-2, 1988, just two weeks after the EPLF victory at Afabet,
President Meiigistu hurrfedly negotiated a formal peace with Somalia. His
aim was the transfer of troops from the Qgaden to Eritrea. The belated
peace agreement also allowed for the exchange of prisoners of war captured
in the war a decade before, and for the repatriation of refugees to be put
on a more formal basis. Finally, an essential part of the agreement was
a commitment by each country to end assistance to insurgent groups
operating out of each others' territory. While neither side followed this
policy fully, the promised expulsion of the SNM from Ethiopia had
immediate consequences.
In late May, the SNM launched surprise attacks on the major towns
of northern Somalia, and sncoeeded in occupying Bnrao and most of
Hargeisa.
The delivery of relief supplies to the Ethiopian refugees was cut off,
and the fighting also induded semal SNM
attacks on K^gee camps.
^ Robert Geraony, "Why Somalis Flee: Synthesis of AoocNints of Conflict

Experiences in Northern Somalia by Somali Refugees, Displaced Persons and
Others, "
Bureau for Refugee Programs, DeparUnent of State, Washington, D.C.,
1989, p. 43.
351
Accoiding to testimonies obtaised by Robert Gecsoay of the US State
Department, the SNM attacked nine refugee camps in thenQgioo and IdOed
240 civilians.^ While the killipg of civilians
at least is a grave abase, the
camps themselves also constituted genuine military targets.
Further abuses against refugees occurred in early 1989. The SNM
executed ten prisoners of war whom it described as "Ethiopian refugees
who had received weapons from the Somali Armed Forces in order to
combat the SNM."'' Another reported incident took place in March when
eleven Ogadeni refugees (including three children) were killed and 16
wounded in an SNM attack on a truck.
The Somali aimy lesponded with land and air attacks against flw SNM
and the towns of northern Somalia causmg an extremely hi^jb level of
devastation.^
Refugee soldiers and militia were prominent during the army counter-
offensive. This included Ogaden refugee units in the regular army (notably
the 12th Division), members of the camp militias, and — for the first time -
- Sheikh Belissa's Oromo front, which fought alongside the army. These
attacks witnessed an extraordinarily high level of indiscriminate violence
against the civilian population, including routine summary executions,
looting and rape as well as some large-scale massacres.
As a result of the war, about 20,000 refugees returned to Ethiopia and
about 400,000 northern Sranalis abo fled across the bt»der to take refoge
in camps near Jijiga*
Upheaval in Southem Somalia, January 1991
In 1989, the Ethiopian government also began to provide arms and

support to the United Somali Congress (USC), a lecently-fonned Oflywition
group active in central/southern Somalia.
In January 1991, the government of Siad Barre was finally overthrown
after a prolonged battle in Mogadishu. The USC was the immediate victor:
although a relative newcomer to the opposition, it drew its support from
the vicinity of the capital and so was able to occupy the seat of power —
or what was left of it. Ihe fall of the Siad Bane govemmeot did not,
pp. 50-2» 63. Gontmy to Genony^ claims^ tee wis

* Geisony, 1989,
widespread distribution of arms to ih» refugees dunng the preceding yean.
' Genony, 1989, p. 34.
See Africa Watch report, Somalia, 1990.
352
however, mean peace or stability. Fightii^ and insecurity oontiniies in
Mogadishu and the countryside.
The breakdown in security and the resulting cut-off in supplies to the
refugee camps made life extremely difficult for the refugees in southern
Somalia. There were also a number of attacks on refugees. On January
21-25, over 100 refugees were reported killed at Baladweyne by USC
forces. Other attacks on refugees occurred in the Hiran re^ôn.^^
Tliese &ctois cumnlatively led to a massive letom of refugees from
Somalia, plus an inflow of lefogees. About 140,000 refugees andietnmees
crossed the border in a matter of two months, and numbers continued to
increase. 90,000 crossed into northern Harerghe, about 100,000 into the
sonthero Qgaden, and 50,000 into the lowlands of Sidamo. This created
a humanitarian crisis, especially in the remote southern Qgaden camps.
The Oromo Fronts in Ethiopia
By 1984, Ethiopian military strategy was effective in counter-acting

the threat posed by the OLF and the Oromo Islamic Fronts (OIF)^^ in
Harerghe. The combination of villagization, continued military operations
and cutting off any possible support firom Somalia proved successful. Ibe
Somali government's strategy of using the Oromo populatioa fronts md
for its own purposes contributed to this military edipse.
The OIF, headed by Sheikh Jarer, emerged as a significant focce in
northeast Harerghe in the mid-1980s. Sheikh Jarer was a military
commander in the OLF until he split from them in 1978 to form his own
movement, which in about 1985 acquired the label "Islamic. "^^ It acquired
support from the Somali government and a link with Sheikh Belissa's front
inside northern Somalia, though the extent to which the two fronts
cooperated in military terms is unclear. The Somali government closed
the OLF office in Mogadishu in 1982 and prevented the organization
" A report by the OLF on February 11, 1991, that over 600 Oromo refugees
were killed by Ethiopian forces at Luuq, has not been confirmed and was almost
certainly inaccuiate (BBQ
Summary of World Broadcasts, M£/0995» February
13, 1991).
^ This front has used a variety of names during its eadstenoe; OIF is used
here for CQuvenience.
The front has no significant links with Islamic fundamentalist organizations

in Sudan or the Middle East. "Sheikh Jaier" is a mm de guerre; he is not a
religious leader.
353
operating from Somalia. There was also lack of unity between the OLF
and OIF and both failed to cooperate with the WSLF. These factors
combined with the success of the SNM
in controlling the northern Ogaden
border to create a great hindrance on the successful maintenance of an
Oromo insurrection in the southeast.
Throughout the period 1985-91, atrocities by the Ethiopian army
accompanied routine actions against suspected OLF sympathizers, and
reprisals after OLF actions. For example, in February 1987, there are
credible reports that the army killed up to 270 people at Qadridayah and
Dibleley, in reprisal for an attack by the OLF. The following month,
the army sealed off wells at Bullale, causing many ammab to die of Ifaiist,
and in July, the army rounded up an estimated 8,000 villageis and kqit
them for a while in four military camps. While most weie later released,
seven were reported shot dead, 41 remained in detention, and 23 young
women were kept for the soldiers* sexual gratification.
Some instances of OLF killings of civilians have also been reported,
including the selective killing of Amhara settlers in eastern Harerghe in
March 1990.
The Ethiopian government launched a number of military actions against
the OIF. It also engaged in a systematic attempt to destroy centers of
Moslem learning in Harerghe, which it saw as the source of anti-
government mobilization. Diuing 1985 and 1986, mosques. Islamic schools,
tombs of local Moslem holy men and pilgrimage sites were destroyed.
These included Fayaanbiro mosque and Sufi lodge (near Babile), the Sheikh
Sayed Ahmed lodge at Babile, and the tomb of Abdel Rahman Tîlahi,
between Degahabur and Qabridaharre.
The OIF clashed with the SNM on several occasions, particularly after
1988. In January 1991, an SNM unit kidnapped Sheikh Belissa and handed
him over to the Ethiopian government; he was detained until the fall of
the government in May. This incident reportedly led to violence between
the Oromo and Isaaq communities inside Ethiopia.
Instances in which OIF units engaged m ethnically-based violence
against Amhara settlers have also been reported.
On one or two occasions, the OIF also came into conflict with the OLF.
Some of the OIF leaders were defectors from the OLF, and there was deep
resentment between the two oiganizations. There were armed clashes
between the two fronts on several occasions, and OIF units ambushed and
killed some members of the OLF.
Africa Contemporary Record, 1987-8, p. B288.
354
The EPRDF Takeoyer, May-June 1991
Until 1991, all the aimed Qromo and SomaU groups in eastern Eddopia
had one thing in common: the knofwledge tfiat if their militaiy activities
became too successful, the army would engage in large-scale reprisals.
The sudden collapse of the Ethiopian army in May 1991 removed this fear.
The divisions that had deepened over the previous years were the cause
of an increase in inter-communal violence. Neither the OLF nor the OIF
had sufficient military strength or political mobilization to fill the power
vacuum that suddenly appeared.
In the days after the EPRDF entered Addis Ababa, the Third Army in
Harerghe was deeply split. Some officers proposed surrendering to the
EPRDF; others proposed resistance, and stfll omxs tried to flee the country,
hi Dire Dawa the small gairison put up no resistance and the town was
occupied on the evening of May 29. hi the brief mten^gnum there was
a looting spree, especially of the aiiport, and about six people were killed.
There were also ominous signs of mter-ethnic violence between the Issa
Somali and the Oromo communities. On May 30, there were anti-EPRDF
demonstrations in the town, which were dispersed when EPRDF fighters
fired over the heads of the crowd. No casualties were reported.
Two senior army officers who attempted to flee the country by helicopter
were apprehended by army units close to the border and later handed over
to the EPRDF.
The only significant resistance occurred around Harer, where a garrison
of about 10»000 troops was stationed. On May 28, the commanding officer
of the Hater gairison announced a suireader. IVro days later, middle-
ranking officers, encouraged by a spate of anti-EPRDF demonstrations
by students, mutinied and killed their superiors, and vowed to fight against
the EPRDF. The mutmy led to widespread violence in Harer, including
looting and killing. One employee of a UN relief program was killed
outside the Ras Hotel in the town center on May 31. By June 1, the
mutineers had gained control, and prepared defensive lines close to a
military base outside the town on the Dire Dawa road and distributed arras
to the civilian population with instructions to defend the town itself.
On June 2, the EPRDF force attacked the army positions. About 600-
800 combatants were killed in the engagement, probably includ ing some
former dvilians armed and miriiilized by the mutineers. The EPRDF foroes
won the batde. Most of the armed civilians refused the order to defend
Harer and instead let die EPRDF peacefully occupy the town the following
day. The occupation was disciplined and there are no reports of fighters
committing abuses against civilians.
355
The disintegration of theaimy and police led to a sudden upsurge in
banditry, indudiqg some robbery by desertiqg aoldieis. A relief convoy
was attacked at Kebri Debar on May 30, and two tmcks were stolen. Eight
other relief vehicles were stden in otber incideiils» and many nual roads
became impassable.
Continuing Violence Since the £PRDF Takeover
In the rural areas around Harer, there are numerous reports of violence
against Amhara civilians, including employees of the government and the
United Nations agencies, mainly by Oromo residents. Some of the violence
can be traced to members of the OIF. A
number of Amhara civilians fled
from die towns of Hareighe to Addis Aiidia, fearing Ib^ ScMne
reported that they had been oidered to leave the region within 24 horns.
There was a serious mcktent of inter-commnnal violeiioeiBlKreDiwa
on July 7. A large section of the Oromo conmiunity held a political rally
in the soccer staditmi, sponsored by the OLF. After the rally a number
of armed Oromo gathered and attacked Issa civilians in the town, killing
eight. EPRDF forces intervened to stop the fighting and disarm the two
groups; about 12-15 people were killed during this operation. In reprisal
for the deaths of the Issa civilians, between July 8 and 15, Issas in Djibouti
attacked Oromo refugees. Between ten and 15 were confirmed killed,
though some Oromo sources claimed that the number was as high as 200.
About 5(X) refugees fled to seek sanctuary in the cooqpoimd of the UNHCR,
and several thousand sought refuge among A£v groups in Djibouti,
particularly in Arbeha. There is fiear that this intBr-couMMinsl violeace
m
may intensify and bring about advil war Djibouti and ndgfiboring areas
of Ethiopia.
In the highlands of Harerghe and Bale, many rural communities also
resisted the re-introduction of central authority in the form of the EPRDF.
Community leaders argued that there was no need for an EPRDF military
presence, because the areas were "liberated" already. Incidents of violence
have occurred between residents and EPRDF forces, and a number of people
have been killed. The fighting has been sufficiently serious to close main
roads in Harerghe for weeks. It has caused friction between the EPRDF
and the OLF; the EPRDF demanded diat Ae OLF demobilize its fif^rteis
so as to restore law and order, or dose its oCGoes. However it is highly
questionable wfaedier die armed units are under te oonmand of the OLF.
An agreement was reached on August 27 which established zones of control
for both organizations, but did not address die fimdamental problems the
mutual suspicions between local Oromo villagers and the £PRDF units
controlling the towns and main roads.
356
The Oromo highlands of the southeast piesent a major challenge to the
new goveroment. The EPRDF needs to maintain law and order and to
allow the expression of popular aspirations without resorting to
indiscriminate violence. The OLF and OIF, for their part, needs to ensure
that the legitimate demands of the population are not channelled into violent
resistance while the options of demoaatic participation are open.
357
20. WESTERN POLICY TOWARDS ETmOPIA
Ethiopia has had close relations with a number of western countries
throughout the 20th century, including Belgium, France, Great Britain,
Israel, Italy, Sweden, and the United States. Successive governments have
turned selectively to different western countries for economic, military and
diplomatic support. During and immediately after the Second World War,
llie Ethiopian govemmenl was dominated by Britain. The British anny
had been instmmcntal in liberating the conntry from ItiJ^ fascist lule
in 1941, and British military administration continaed in Eritrea nntil 1952.
Thereafter, Haile Selassie began to cultivate closer ties with the United
States.
US Policy and Haile Selassie
Ethiopia under the Emperor Haile Selassie was the United States' closest
ally in Africa. This was due to several factor. One was that after the
Second World War the US was promoting a policy of decolonization in
Africa and Asia. Ethiopia, as the only indigenous independent state in
Africa, was symbolically central to this policy, which was duly encouraged
by Haile Selassie. The Emperor was instrumental in the formation ot the
Organization of African Unity (GAU) and its adoption of a pvo-westem
stance and conservative position on inherited boundaries. A
second reason
was Ethiopia's strategic position adjacent to the Red Sea, and
its possession
of the communications center at Kagnew, near Asmara. For more than

a decade after the US signed a 25 year lease on Kagnew base in 1953, it
was the most important "listening post" in the Middle East.
The strategic value of Ethiopia's Red Sea coastline, all of which fell
in the territory of Eritrea, and of Kagnew station, meant that the US was
an enthusiastic supporter of the unity of Eritrea and Ethiopia. In a now
famous statement. Secretary of State John Foster Dulles said to the UN
Security Council in 1952:
From tiie point of view of justice, the opinioiB of the Eritrean people
must receive consideration. Nevertheless, the strategic interest of the
United States in the Red Sea basin and considerations of security and
world peace make it necessary that the country has to be linked to our
ally, Ethiopia.
As a consequence, from 1953-76, Ethiopia was the largest recipient

of US aid in Africa. This included generous military assistance, based on
359
the 1953 Mutual Defense Assistance Agteenieiit. Ethiopia received mm
than half of the US's entire military assistance to Africa during this period.
Between 1953-76 the US trained and equipped an army of about 40,000
men and an air force of 2,000, at a cost of $280 million.^ The annual
subsidy to the army was about $12 million. About 4,000 officers and air
force personnel were trained in the US. However, when Haile Selassie
made repeated requests for increased arms deliveries to offset the threat
of Soviet-armed Somalia after 1960, the US was unwilling to supply all
the armaments and finance he requested. Such requests were made in 1967
and 1973, and turned down because the US conaideied them excessive.
Earlier, m recognition of US support, Haile Selassie had sent a battalion
to fight in Korea. However, be was no simple vassal of the US; for
example he Union and to discusA economic and military
visited the Soviet
assistance, and publicly supported the Arab states over against Israel in
1973.
The United States was not the only western supporter of Haile Selassie.
In the 1940s and '50s, and to a lesser extent afterwards, the Swedish
government had supplied training and material for the Imperial bodyguard
and the air force. The British trained the army until 1949, and provided
limited technical assistance thereafter. Ethiopia and Israel signed a military
pact in 1958, wheid)y Israel provided training at the Ifoleta mflitary
academy and (from 1964) counter-insuigency advisois in Eritrea. Hiis
continued untU Haile Selassie broke dijômatic rdations with Isnd in
October 1973, in deference to the OAU's support for die Arab states in
the 1973 Arab-Israeli war. Ethiopian-Israeli relations remained warm,
however, and Israeli military advisors returned during 1975-7.
The West and the Revolution
days of the revolution, the United States continued to be

In the early
the major foreign power with an interest in Ethiopia. For three years from
the outbreak of popular unrest in February 1974, the US attitude can be
described as conciliatory but concerned. Other western governments
followed comparable policies, whfle mtemal conditions led to scaled down
assistance to devetopment programs. Italy —
worried about the weUEue
—
of the large numbers of Italian citizens, especially in besieged Asmara
remained close to the government diplomatically, but net flows of economic
aid completely dried up by 1977. Thereafter there was a dramatic break
^
Lemmu Baissa, "Militarization and Foreign Fblicy in the Honi of Africa,"
inProceedings of the 5th Internatkmal Conference Cm ^HomcfAfrUa^ CUNY,
New York, May 26-28, 1991.
360
in US-Ethiopian relations, which cookd to the point of beoomiqg frozen
in cold-war hostility.
For its part, the Dergue both demanded armaments from the US, and
attacked "US Impeiialism" in official pronouncements, before instigatii^
the rupture itself.
During 1974-6, the US policy of supplying generous military aid to

Ethiopia continued.^ In tel, in 1975 and isn6^ aimi dUpacBls Incieased
substantially. H6wever,tfaeainoiiiitsdeliveied woe not enough to satisfy
the Defgne, wiiich sent missions to a variety of ooantries add^
assistance. The US expressed dismay at certain Ethiopian military policies,
notably the 1976 Peasants' March on Eritrea, and Secretary of State Henry
Kissinger went so &r as to say that if the march went ahead, continued
US military assistance would be jeopardized. On the whole, however, the
US sought to influence events by maintaining warm relations with the
government, including considerii^ inaeased aims requests. This was not
to last.
The inauguration of President Jimmy Carter in January 1977 was
followed within a month by the seizure of absolute power by Colonel
Mengistu Haile Mariam. Mengistu was not only the most determined,
nitfaless and militarislic of Hie Dergue members, but be was the USSM
closest aUy in the government. Pto-Soviet and anti-American statements
became moie fieqaent and vitridic, and aims shipments firara the USSR
started immediately. The Red Tenor ofHcially began. On February 23,
US Deputy Assistant Secretary of State Talcott Seelye visited Addis Ababa
and told the government that it could expect a teimination of militaiy
assistance, on human rights grounds.
Since 1977, when they were first issued, the Department of State's annual
human rights reports on Ethiopia have, with few exceptions, been thorough
and objective.
In April 1977, the US notified the Ethiopian government of the imminent
closure of Kagnew station the following September when the 25 -year lease
expired. The decision had in tet been made in 1973, on purely technical
groonds (the base had been snpeiseded by the use oif aerial and satellite
reconnaissance). On April 30, QdonelMeqgistn responded liyvnilatenUy
terminating the Mntual Defense Assistance Agreement, and ordering all
US military personnel to leave immediately. Shortly afteiwaids, he onteied
a cut in US Embassy staff of 50%.
^ The following discussion owes much to: David A. K£Bai,Ethkipki, Jhe United
States and the Soviet Union, London, 1986.
361
The US-Ethiopia breach was connrmed by the massive airlift of Soviet
arms to Ethiopiaduring the Ethiopia-Somalia war. In November 1977,
the US embargoed all arms deliveries to both Ethiopia and Somalia. Shortly
thereafter, the US made a military pact with Somalia, which had just
expelled the Soviets, accusing them of perfidy. (The naval base at Berbera
was the strategic prize that changed hands.) Ethiopia's other longtime
regional adversary, Sudan, also moved closer to the US at this time.
Meanwhile, between 1975 and 1980, Ethiopia's arms budget jumped tenfold.
More than $1 billion in mflitary assistance was provided by the USSR,
and 13,000 Cuban combat troops were stationed in the coontry. East
Gennan security advisors were to follow.
Despite Mengistu's evident preference for closer ties with the Soviet
Union, the US continued to make efforts to ensure that relations were not
broken off altogether. In 1978, the incoming US Ambassador, Frederick
Chapin, recommended a small program of economic assistance, and a small
amount of non-lethal military assistance was provided to the army.
However, any further attempts to upgrade economic ties were prevented
by the Hickenlooper Amendment to the Foreign Assistance Act, which
stipulates that all US development assistance must be cut off to a country
which nationalizes US-owned assets without providing compensation.
Some 25 US companies claimed compensation amounting to $30 million
following the Ethiopian government's nationalizatioiis of 1975, and the
Ethiopian government renised to pay this relatively insignificant sum.
Hie faflure to reach agreement on the compensation issue, together with
increasing cold-war tension, led to further mutual US-Ethiopian hostility.
USAID program was officially terminated, and in 1980, the
In 1979, the
US Ambassador was asked to leave; since then, the US has been represented
by a charge d'affaires. In 1981, Ethiopia signed more far-reaching
agreements with the USSR, allowing it use of military bases in the Red
Sea islands, and signed the Tripartite Agreement with South Yemen and
Libya. In 1982, the Reagan Administration's policy review for the Horn
of Africa determined to isolate Ethiopia, supporting its pro-westm
neighbors, but stopped short of supporting the xAei fronts.' In 1984, wlien
BandAid leader Bob Geldof visited Ethiopia and snggested that private
agencies pay the $30 million compensation due mider the Hickenlooper
' The conservative Ethiopian Democratic Union was the United States' favorite
rebel front, but after 1977 it was not an affective military force. In 1984-5 the
State Department considered stqiportmg the TPLF, but rejected it as too left-wing.
Military support for the Ethiopian People's Democratic Alliance, a conservative
organization of exiles, was entertained but rejected because of its lack of a military
presence in Ethiopia.
362
amendment, thereby removing the formal obstacles to the knger amounts
of official development assistance, the idea was given a cold reception
Other western countries' relations with Ethiopia cooled over the same
period, though none went so far as withdrawing fiill diplomatic accreditation
and suspending all development assistance.
Western Humanitarian Assistance 1980-84
Tliroughoiit the 1980s, all US assistance to Ethiopia was humanitarian

aid.
Between 1977 and I960, the US
piovided cm aveiagp $15 mill^
assistance to Ethiopia each year, substantially down on previous years.
In 1981, that fell to $2 million and in 19^, to less than $1 million. It
was planned to phase out bilateral assistance altogether in 1983, but in fact
USAID responded to requests from humanitarian agencies, and assistance
rose to $6 million in 1983 and $18 million in 1984. Other western donors
such as Great Britain cut their programs in similar proportion in 1980/1,
though were quicker to increase them afterwards. By contrast, in 1985,
the US gave $142 million and in 1990, $177 million.
The United Nations was, comparatively speaking, much more generous.
Wiuie oveiall assistance stagnated between 1979 and 1982, the major UN
agmies (World Food Program, UN Hirih Oommissioner for Refingees,
JJS Development Program and UNIC^) all increased their assistance

substantially. From $13.1 million in 1977 and $22.6 million in 1979, Aeir
donations had trebled to $67.0 million by 1981. As shown in chapter 5,
some of this assistance was aimed at the "rehabilitation" of the war-affected
southeastern regions and the return of refugees. This assistance can be
government's war aims.
criticized for assisting the
While assistance from western Europe stagnated, aid from the European
Commission increased over the same period. The EC aid program to
Ethiopia was described by an ex-US Charge d'Affaires in these terms:
[It] seemed less a pfodoct of tlioaght than of bureancratic momentum
and badly oversimplified logic: the EEC had a fnnd to aid needy Hiird
World couitries, Ethiopia was a laig^ and exceptionally needy Third
World countiy, eigo the E£C program for Ethiopia.^
^ Figures from Organization for Economic CoopeiatioD and Development.
^ Kom, 1986, p. 58.
363
EC assistance to Ethiopia grew steadily from $4.2 million in 1977 to
$41.5 milUon in 1981. 0t reached $153 .4 million in 1988.) Like the UN
assistance, the program was little influenced by the human rights
considerations that affected US policy at its best. The only EC country
to increase its aid substantially over the period was Italy, wbkh donated
an average of over $30 million each year during 1980-2.
In 1983, all western donors increased their bilateral aid programs to
Ethiopia, and overall assistance rose from $221 million to $361 million.
An appeal by British volmitary agendes tat tUBom vdief donstioas in
March 1983 was considered a SDOOcsSkthoagJiteaiiKNinlB^ve^
compared to what was to oome later, la, 1984 ovenU aid leveb topped
$400 million and in 1985 reached nearly $800 million.
When the Ethiopian funine became a matter of domestic political inlERSt
in the West, one of the questions which was repeatedly asked was: why
had not the western countries responded sooner? The above account makes
it clear that there had been a response, albeit a selective one. For a famine
of 1970s proportions, the assistance would have been adequate. As it was
to turn out, the amounts given were far less than needed. The US
government was notably ungenerous.
Some donors claimed that their tardy re^nse was due to the Ethiopian
government concealiqg the Cuanie. Bnt, while it was certainly the case
that between April and September 1984 the Ethiopian government had been
preoccupied with the preparations for the tenth anniversary celebrations
for the revolution, and had been anxious to conceal the famine from its
own pecqde and from visiting journalists, this period was the exception
rather than the rule.^ Until March 1984 the government was quite frank
about the existence of a famine —
Colonel Mengistu even mentioned it
in his 1983 May Day speech — and at all times the RRC was publishing
figures of people in need and amounts of food required. There were no
fewer than 21 warnings of impending famine from March 1981 to October
1984 by RRC and icMef agencies. Western governments were well-
providcil with official information abont the developing famine.
Why was the response small relative to flic real need? Ttaearettree
reasons, which range from a genuine attempt to grapple with an inscdnble
moral dilemma, to behavii^ in a finmkly cynical aiid unethical manner.
^ See: Article 19, Suuving te SUaux: A Report am Fambm mid Cmaonklpt

London, 1990.
^ The Development of the

Peter Cutler, 19S3-85 Famine in Noctbem Ethiopia,"
PhD Thesis, London, 1988, pp. 365-7.
364
At the cynical end of the spectrum, Ethiopia was seen as a cold war
enemy that was engaging abuse of the west, and thus deserved
in gratuitous
to be shut off from any assistance. The attitude was that if Ethiopia needed
assistance, it should turn to its patron the USSR. Thus in early 1984, when
the US Congress drafted a bill mandating aid to Ethiopia, President Reagan
attached to it an amendment requesting military assistance to £1 Salvador
and the Nicaraguan contras.
This policy subseqaently came in for mncli ciitidam by the media and
liomanitarian agencies, and was at times even equaled with having caused
the fnnine. It did not of course cause the fiontne, thoug|i generous
assistance provided earlier would certainly have prevented much suffering.

A second reason for the lack of response was a degree of skepticism
about the figures produced by the RRC. The figures for people in need
were produced with an exactness that appeared spurious (with hindsight,
correctly so), and no opportunity for cross-checking. They were widely
held to have been inflated in order to attract western aid. As has been
shown, at times they were certainly mendacious. While the specifics of
the distortions of the figures were not always known, the general fact of
their manipulation was understood.
This justifiable skepticism combined with a lack of undeistanding of
the dynamics of £Eunine (specifically, few appredated the fwt that starvation
is the end result of a very long process of destitution) and with a general
cynicism to produce a climate of indifference in the upper echelons of the
aid community. At different times in the development of the famine, relief
officials were heard to say:^
Every year I have been here for the last four to five years they have
said that several million people were facing food shortages. If this was
true, at least a million would have died by now. (World Food Program
[WFP] official)
If we see the figures we tend to divide by ten —

maybe that is a very
cynical attitude. If there is a veiy bad year, we n^ght add ten per cent.
(European Community official)
The RRC says it will put on a show and it puts on a flop. (Canadian
official, after beii^ given a guided tour of a famine-stncken rôn)
" Quoted in: Cutler,

1988, pp. 362-3, and Peter Gill, A Year in the Death of
Africa: Politics, Bureaucracy and the Famine, London, 1986, p. 32.
365
This attitude even continuecl into 1985
How many do yon think died in 1984? would fervently hope that
I
it was less than one million. But, frankly, have believed those March
to
figures from the RRQ there should have been Uuee mfllion dead. (WFP
official)
Time and again, the RRC was accused of "crying wolf," This turned
out not to be the case. However, as this report has demonstrated, the major
reason why the famine developed into the cataclysm that it did was not
that western aid was tardy or inadequate, but that the Ethiopian government
insisted on puisumg aset of mililaiy and ecooomic policies Hut aerikmsly
aggravated tlie problem. Hie aid agency officials wm, te the most part,
not well-informed about these policies, and were thus unaware of their
likely consequences. This ignorance both led to an underestimalioa of te
severity of the impending famine (in 1983-4) and a failure to appiedste
what kind of response was really required (in 1985).
The third reason for withholding aid is the most legitimate. It assumes
that western governments were aware of the government's policies which
were creating famine, and were not prepared to underwrite them with funds.
The refusal of the US to support UN initiatives to repatriate and "rehabilit-
ate" refugees m
southeast Ethiopia during 1979-82 lends credence to this
view. The argument would be that te donors were presented witfi a
difficult moral dilemma: were tiiey to provide f'ms'^nm to muHdgito human
suffering, knowing that te underlying causes of that mtttm^ would not
be affected, and might even be strengthened or legitimized by their
assistance? Or should they refuse to intervene until basic conditions for
the neutrality and accountability of relief were met, and the policies creating
the suffering were changed? This is a genuine dilemma and one with no
easy solution.
There were undoubtedly some individuals in the major western donor
agencies who were aware of this dilemma. However, they were not the
ones who were dictating policy. As a result, the central issues of Ethiopian
government culpability m te fEonine woe never addressed.
In te first instance, te aUegstions of te abuse of hnnanitaiian aid
did not consist in documentipg its use as a counler-iBsurgency tool, but
instead consisted of claiming that it was bdqg Averted to tlie mililaiy.
Aid was, of course, being diverted to the mflitaiy (see chapter 10), but it
was relatively easy for RRC officials to impress any visiting ^ffjalt and
diplomats and "prove" that the allegations were unfounded.
In March 1983, the Canadian Ambassador to Ethiopia stated that there
was no evidence of any diversion of food. Britain's Overseas Development
366
Administiation endorsed this view in April. In May- June there was a visit
from a dd^tion from the European Parliamciit, which failed to uncover
any diversion. In March 1984, a bipartisan committee of the US Congress
also failed to find evidence of misallocation, and shortly afterwards a
mission from the US General Accounting Office pronounced the rate of
diversion to be within acceptable limits. None of these investigations visited
rebel-held areas or was able to travel independently of official guides.
Several were primarily concerned with the issue of whether relief food had
been re-exported to the USSR.
Hie sustained idnlity to pursue the nuugiiial issue d
food diversion,
and still not get to the bottom of it, oontinned into 1985/6. Hie western
donois would have been better advised to mvestigate whether the
government was pursuing a set of policies that created famine conditions.
Hie US was the only govenunent to do so, and then cmly in 1985, which
was too late.
Human Rights and Famine, 1985: UN and US Approaches
The sudden media interest in the Ethiopian famine in October 1984 and
afterwards radically changed the West's relations with Ethiopia.
In 1985, all western donois increased their assistance to Ethiopia
substantially. Assistance to Ethiopia was extremely generous. The US
became the country^ kogest donor. US assistance was all — with a small
exception — directed through non--goveinmeiital ofgamzatioos. Hie details
of the programs can be criticized, but not the fiict of the generosity and
the willingness to support a people whose government was eqgaged in
persistent hostility to the US. Other western donors were, however,
prepared to give support directly to the government RRC.
The US was the only donor government to undertake an investigation
human rights aspects of the famine. This investigation, mounted
into the
between July and September 1985, has to be seen in the light of the UN
role in the famine.
In late 1984, the UN
Emergency Office for Ethiopia (UNEOE) was set
up, ostensibly to coordinate rd^
dEforts, which were beoomiqg extremely
complex as a result of proliferation of donors and agencies. It was
firequeirtly turned to as an authoritative voice on issues to do with the
diversion of food, forcible resettlement, and other abuses. Rather tlian
investigating the abuses, UNEOE
consistently concealed disturbing evidence
produced by its own monitors. Its role has been described thus:
UNEOE's main function was to act as a "screening device", giving the

appearance of competent action in response to famine but not
367
compromisinig its actual position in Addis Abiba by uadaly antagonizing
the host government ... it would have been as embarrassing for the
donors who had entrusted resources to the Ethiopian government as
it was for the gqveniment itself to have aid misallocation exposed,^
Mr Kurt Jansson, head of UNEOE during 1984/5, continued to play

this role even after leaving his position, as will be evident from reading
his account of the famine.
Some membas of the US CSopgress woe unhappy witii te UN
whitewashiiôftfaeEdiiopiangovenimeBt^cd In late July 1985,
Rqxesentative Toby Roth (R., ^âconsin) mlndiiced an amendment to
the foreign aid bill, the International Security and Cooperation Act of
1985.^^ President Reagan had to detennine within 30 days of signing
the bill whether the Ethiopian government was (1) conducting "a deliberate
policy of starvation of its people" and (2) failing to grant its citizens
"fundamental human rights." If the President found the Ethiopian
government guilty on both counts, the US would be obliged to impose a
trade embargo that would cover all exports save emergency humanitarian
aid. The bill was signed on August 8. It was to raise an insoluble moral
dilemma, and be the most aeveie teit off the US poliqr tawuds Ethiopia.
'
In early Augnat, UNEOE ideaaed a lepoit anthoied by Knit Janaaon,

consistmg of a "comprehensive review" of the relief operations in Eritrea
and Tigray. Jansson and other leading aid officials, including a US diplomat
and RRC director Dawit Wolde Giorgis had spent four days in their inquiry,
which was confined to a few garrisons in government held areas no —
contact was made with ERA and REST. (During its entire lifespan, UNEOE
made no effort to direct any assistance to ERA and REST at all.) The
report concluded that 75 per cent of the affected people in Eritrea and
Tigray were being reached from the government side. This was manifestly
untrue (see chapter 11). _
Rebuttals of the iqxMrt by ERA and REST were fogthcoming the m

following days. As htt been shown, it was prodooed at a tune when the
govermnem was directing oofy one tweotiedi of the relief to Tigray, whU
had one third of the afifbcted population. It had just completed a bloody
* Cutler, 1988, p. 408.
^°
Kurt Jansson, Miduwl Hanria and Aqgela Pemoae, IheEthicpiM FmKbut^
London, 1987.
" This account owes much to: Jonathan B. Tucker, "The Politics of Famine:
US Foreign Policy in Ethiopia, 1982-1985," mimeo, 1985.
368
offensive in Tigray and was preparing a massive offensive in Eritrea.
Nevertheless, UNEOE rejected these dissenting voices.
UNEOE had not addressed the dilemma of whether to channel assistance,
no (juestions asked, into the government side of the war zone. It was
prepared to endorse the ten per cent contribution to survival represented
by aid, without questioning the role of the government In attacking the
otber 90 per caA.
USAID administrator Peter McPhoson was asked Id piqNue the US
report for the President. For the moat part he accepted the UNEOE report's
conclusions, but with significant reservations. On August 24 he visited
Addis Ababa for three days, and then went on to Khartoum. He announced
that most of the needy were being reached from the government side. He
did not mention (and perhaps did not know — the subsequent report only
gives distribution figures up to July) that RRC deliveries to Eritrea and
Tigray had fallen 85 per cent since the UNEOE mission. McPherson also
mentioned that the US was supporting cross-border relief operations too.
The US support for the cross-border operation was indeed more generous
than any other donor, but it fell short of both promises made and real need.
The report was released on September 7.° It accused the govermnent
of deliberate policies that **liave no doubt caused vast and nmiecessary
suffering, including starvation," and documented a number of these policies.
But it also said that the evidence did not show that the government was
"at this time conducting a deliberate policy of starvation." The carefully-
chosen inclusion of the words "at this time" indicated that the investigator
was prepared to believe that the government had been pursuing such a
policy until recently— and implicitly trusted that the government's policy
had now changed for good. As chapter 10 has shown, the Ethiopian
government was at this very moment beginning to realize the asset that
it now had in the form of relief food, and was changing its counter-
insurgency strategy in Eritrea accordingly —though scarcely so in Tigray.
The essential problem with the inquiry was not that it was cursory, but
that it was essentisUy asking the wrong question. Hie aim of the military
policy was not to create starvation per se, but to create a population without
any independent means of livelihood — i.e. to create a choice between
starvation and submission for the civilian population living in areas

controlled by the rebel fronts. Without international aid, the policy would
be tantamount to starvation, but with that aid, it need not be, while still
meeting the same military and political goals. The deeper issue was: Was
Presidential Determination No. 85-20, "Determination with Respect to

Ethiopia," September 7, 1985.
369
the Ethiopian goveroment engaged m a war agunt the economic, social
and political fabric of Eritrea and Tigray and on what terns was it pmared
,
to provide assistance to the people? llie Ethiopian government could —

barely — escape the charge of using starvation as a weapon in August
1985, but it could not escape these other charges. The Presidential
Determination implicitly recognized this, concludii^:
[The amendment] does not call for any determination concerning the
past conduct and policies of the Ethiopian Govenunent concerning
staivation of its people. Nor does it call for an evaliiatkni of poliGies
that may have had political or military purposeSk but wliicii nevmieta
caused increased starvation. The Government's past conduct, and the
effects of its policies^ are matters of grave concern, evoi iioogli the
evidence on diese subjects cannot jaSdiy a determination nnder this
statute.
The US administration could justifiably have returned a verdict of

"guilty" and imposed an embargo on Ethiopia — and indeed gone further
and cut humanitarian assistance programs on the government side. This
would have also provided a post hoc justification for the earlier scaling
down of aid. However, the administration chose not to do so, and dso
not to defend its earlier policy on those grounds.
The pragmatic humanitarian rationale behind Ifae US decision to refect
the charge that the Ethiopian government was using slaivatiDn was that
at this point it was necessary to deliver aid to the hungry, no matter what.
The US was also providing cross-border assistance to ERA and REST,
and as the rebel-held areas expanded, USAID support to these organizations
grew to eclipse assistance given through the government side. (The same
was never true of other major donors such as the European Community,
let alone the UN.) The political rationale was a domestic one: since
BandAid and Live Aid, it was necessary in terms of domestic politics to
be seen to be giving generously. As television cameras and politicians
could not visit rebel-held areas (though they did go to refugee camps in
SudanX it was necessaiy for US assistance to have a hi^ media profile
in govemment-held areas.
Although itpossible to criticize the details of the assistance program
is
to Ethiopia after 1985, the essential fact was that the US gave WfffiiftmiT
generously, where it was needed, and limited that assistance to humanitarian
aid. 1990, Ethiopia was receiving the largest amount of US assistance

By
of any country in Africa —
without even full diplomatic representation
or a USAID mission.
370
The involvement of the OAU also deserves mention. The Chairman
of the OAU, President Julius Nyereie of Tanzania, was pressured to take
an initiative to negotiate a safe passage agreement for relief aid in the war
zone. President Nyerere declined, fearing that the precedent of giving even
a small degree of de facto recognition to insurgent groups (and in the case
of the EPLF, secessionists) would be contrary to the common interest of
African governments. The OAU, which has its headquarters at Addis
Ababa, took no further part in initiatives to relieve the famine, curtail human
rights abuses, or promote peace.
Western Assistance after 1985
During the 14 years of Mengistu's rule, western assistance to the country

grew from $105 million to over $1 billion. Most of this aid was disbursed
without the same scruples that informed the US assistance program at its
best. Because of the artificially low official exchange rate, an equivalent
or larger amount was contributed direct to the Ethiopian treasury.
In 1986 and 1987, western assistance to Ethiopia dropped slightly, but
by 1988 it surpassed 1985 levels, and in 1989 topped $1 billion. The main
stimulus to the increased levels of assistance was continued appeals for
hnmanitarian relief by the Ethiopian government oo account ck drought.
The appeal for 1987/8 was the most successful —
more than one million
tons of grain was donated, enough to fised the entire "at risk" population
with extra to spare. As only a dwindling proportion of the needy were
in government-held areas, especially after April 1988, the pattern continued
to be that the less-needy in government-held areas received priority over
the more needy in EPLF- and TPLF-held areas.
US assistance was consistently over $150 million per annum after 1987.
US relief aid was sent through both voluntary agencies working on the
government side and through ERA and REST. The US also assisted in
negotiating the Joint Relief Program for aid to TPLF-held Tigray in
1989/90, and for aid through EPLF-held Massawa to Asmara in 1990/91.
After Massawa was reopened in January 1991, the US stopped
committing further food to the cross-border opmtion, ostensibly on the
grounds that such food was no longer necessary because of the opening
of the port. This was not the case, because the Massawa qperati(» was
not able to provide all the necessary assistance. Suspicions were voiced
that USAID acted in this way to bring pressure to bear on the EPLF and
EPRDF to negotiate a ceasefire. After aid agencies publicly criticized the
May, it was reversed.
decision in early
The ban on US development assistance to Ethiopia on account of the
Hickenlooper Amendment remained in force. No further investigations
371
into human rights abuses and their idatiomUp to funlae wmmidertaken,
but the US continued to be higb^ critical of the Ethiopian govemoMafi
human nghts leoofd. In partkuhr, the US govenunent opposed giviqg
assistance to the tesettlement program.
European countries provided assistance to an even greater extent: in
1988, the EC and its member countries donated over $500 million, over
half of the entire assistance program to Ethiopia and twenty times the level
of 1977. Italian assistance alone was $246 million in 1988, including
generous aid to the "development" project of Tana-Beles, which was
coextensive with the Metekel resettlement project.^^ European countries
and the EC were much mote vehictant to support the rettef efibct in nbel"
held areas than the US, and the mat majority of the awistance was
provided to the government side.**^ In 1991, aevml European donon
followed the US in cuUing ciosB-boider assiitance to ERA
and REST.
This was not criticized in the same way by voluntary agencies becmae
agency staff recognized that these donors were unlikely to respond to such
criticism, and certainly would not respond rapidly —
unlike USAID, which
was much more sensitive to pressure.
European assistance included direct aid to the government, notably to
the RRC during 1985 and afterwards. There is no evidence for a significant
withholding of aid on human rights grounds. The only issue of internal
policy which was raised was economic liberalization in the years after 1988.
Assistance from the principal UN agendea to j^hiopia grevir fiom $107.6
mfllion m1985 to $155.8 mOlion m
1988. OoJ^ hi 1991 didUNasiistanoe
first flow to rebel-held areas, when the opemqg of the EPLF-held port
of Massawa was accompanied by an afftaaaA to allow half of the food

to be distributed by ERA. As in the case of European assistance, human
rights considerations do not appear to have influenced aid levels. On several
occasions, the UN appealed for assistance to be given to the resettlement
program, on the grounds of preventing suêring among the resettled
population.
From late 1988 onwards, the Ethiopian government engaged in a
concerted campaign to woo the West, in order to obtain increased economic
assistance. Vm led to Ghaqges among many WeilBin govenuMOli hi
return. The US, however, only hinted at an hnpioived idattonahip —
nothuig of substance changed.
''As detailed mcfaipler 12,in 1985,theliieldKl project was killing ordinary

people fiuler than the peiteit amuhi of the fiunine.
EC assistance was legally constrained by the pvovilions of the LOBW

Convention, to which Ethiopia is a signatory.
372
The corollary of generous western assistance to Ethiopia was that aid
was given to neighboring countries such as Sudan and Somalia in a similarly
manner. In the case of Somalia, much of the assistance destined
uncritical
for refugeeswas used to feed soldiers and militiamen loyal to the Somali
government.
Ethiopia under Perestroika
Despite the anibilioiis of Soviet ocddnm slral^fa^

refused to be a simple vassal of the USSR. Certain policies, such as the
delayed formation of the Workers' Party of Etluopia, were opposed by his
Soviet advisors. In the late 1980s, Mengistu was also impervious to tiie
changes taking place in the USSR under President Gorbachev. He is
reported to have prohibited talk of both perestroika and glasnost, and
certainly he ignored advice from Soviet economists about the need for
economic reform in 1987. The USSR showed increasing impatience with
President Mengistu's unreconstructed Stalinism, insatiable appetite for
military hardware, repeated military disasters, and embarrassing famines,
and anDomioed that amis deliveries wonU be
expired in March 1991. Anns deliveries oontinDed up until that date, so
that the warning never actually impeded Meqgistu*s mflitary strengdi; he
was no straightforward "victim of perestroika.^
However, the Soviet declaration that the line of credit was finite struck
a psychological blow. Together with the military debacles in Eritrea and
Tigray, this encouraged the opposition and contributed to the attempted
coup in April 1989 and the "change in direction" whereby Marxism was
formally abandoned in March 1990. It also forced Mei)gistu to look
elsewhere for military and economic assistance.
Israel and the Falasha Issue
Under Haile Sdassie, Ethiopia had dose rekitions with Israel, kraers
primary interest at that time was a strategic one—it was OQUcemed about
die growth of mOitant Arab states in Sudan and on the shores of the Red
Sea. Israel was suppoftmg the Anyanya insuigents m
southern Sudan
through Ethiopia. A
secondary concern was the popuktkm of Ethiopian
Jews, the Falasha.
Hie Falasha are a small group who numbered about 50,000 in 1980.
They lived in northern Gonder and adjoining areas. They refer to
themselves as Beita Israel ("House of Israel") or Kayla, an Agau word of
uncertain interpretation. Their origins are uncertain and controversial; most
scholars logad them as belonging to the Agau ethnic group, who inhabited
373
northern Ethiopia before the arrival of the Tigrayans and Amhara. Their
Judaism incorporates the Pentateuch but neither the Mishnah nor Talmud;
their religious language is Ge'ez (ancient Ethiopic). The Israeli Rabbinate
was, until 1973, unsure of the legitimacy of their claim to Judaism. The
Falasha traditionally aspired to migrate to the Holy Land.
Until the revolution the Falasha were not allowed to own land, and made
their living primarily from occnpitioni wcfa as pottery and iaaflWuik; they
were also suljecled to various foims of diacriminatiM In oliierfeqpecli
they were relatively fortoaate; lor iastanoe their area of norftem Oonder
was little-affected by the draioglits (tf 1983-4 and oooliniied to pcodnoe
food surpluses.
In the 1970s, the pUght of the Falashas aroused concern among Israelis
and Jewish Americans, who lobbied for them to be allowed to migrate to
Israel.
After the revolution, the Dergue aligned itself with left-wing states in
the Middle East and north Africa, and became publicly hostile to Israel -
- though a low level of contacts between the two governments always
continued. From 1978 to 1988, Ethiopia's closest African ally was Libya,
which provided military and technical asaiitance, and tiie Nestine
UbenitionOiganization wasgiventhevseof an"einbiiaiy" in central Addis
Ababa. Israel, which had expfessed repeated coace m
0¥er the pliglht of
the Falashas was rebuffed several times.
In 1984/5, Israel spoiKored "Operation Moses** whereby Falasha were
encouraged to leave Ethiopia and migrate to Sudan as refugees, from where
they were airlifted to Israel. This clandestine operation was terminated
in early 1985 when it became public in Sudan, which is a member of the
Arab League and officially a supporter of the Palestine Liberation
Organization. After the exodus was made public, it was condenmed by
the Ethiopian government.
By November 1989, Ethiopia had reverMd its position: an brseli
Ambassador was accredited, and the Libyans were doae to hiang
expeUed.^
The emigration of Falasha direct from Ethiopia to Israel started in 1989,
In return for allowing the Falasha to emigrate, Ethiopia received weapons
and military instructors. Israeli officials at different times admitted to
having supplied small arms, non-kthal military technology, and trainii^
A bomb explosion in Addis Ababa was blamed on the Libyans. Later, the
Ethiopian government was to accuse the Libyans of supporting the EFRDF.
374
in counter-insurgency, and at other times denied giving any assistance.
However, a confidential congressional staff memo leaked to Washington
Jewish Week}^ confirmed that in late 1989 about 100 cluster bombs were
supplied, which the Ethiopian air force was particularly eager to have.
Fkesideiit Jimmy Curler said to bndi Knesset member Dedi Znctor: "You
don't need to sdl Mengistu fragmentation bonibs in order to pemiade him
to let your peqê go."^^ duster bombs were used in the bombing of
civilian targets in Eritrea induding Massawa, causing huge numbers of
civilian casualties.
Between 1989 and May 1991, the Ethiopian government repeatedly used
the Falasha population as pawns The US government
to obtain arms.
consistently opposed the delivery of arms to Ethiopia, and was particularly
hostile to the Israeli sale of cluster bombs, because these had been
developed with US technology. This opposition appears to have prevented
the further delivery of cluster bombs by the Israeli government from early
1990 onwards.
Almost the entire Falasha population was drawn from Gonder to Addis
AbiAia in 1990 in the expectation of resetdement in IstaeL They abandoned
m
their ftrms and livelihoods. Because the Ethiopian govern ent allowed
only a very slow rate of exodus, many were forced to remain in Addis
Ababa for a long period, without adequate shelter or food.
The US expressed particular concern for the Falasha population, and
repeatedly raised the question of their freedom to emigrate to Israel with
the Ethiopian authorities. The underlying reason for this concern was
pressure broughtby the pro-Israel lobby on congress and the administration,
and accompanying media attention.
In the dying days of the Mengistu regime, the Israeli government
mounted a dramatic airlift known as "Operation Solomon** to bring the
remaining Fahoha from Addis Ababa to Israel. This was snccesrfully
completed before the EPRDF occupation of the dty, which, it was feared,
would lead to disorder, reprisals against the Falasha, and restrictions on
their emigration by the new government. In the event, the EPRDF
occupation led to none of these things for the small number of Falasha
remaining behind.
For example: Israeli ambassador to Ethiopia, Mr Meir Joffe, interviewed

by Richard Dowden on British TV, Channel 4, The World this Week, June 2, 1990.
" July 12, 1990.
" Quoted in: Betum, Apiil/May 1990.
375
Peace NcgotiaUons
In 1990 and 1991, along with other intermediaries, the US was active
in convening a succession of negotiations between the government and
the rebel fronts. Given continued though declining Soviet patronage of
Ethiopia, and very limited US economic or strategic interest in the country,
this involvement was remarkable. The US agenda appeared to reflect its
cQncem with the Fabshas, the provisiao of lelief, and a commitmBni to
the resolution conflict by negotiation. With the ending of the odd war,
strategic Inteiests were fist waning; the only lemainiiig political commitment
of significance was the maintenance of Ethiopia*8 tenitorial integrity —
i.e., opposition to the independence of Eritrea.
During the first months of the Gulf crisis of August 1990-Febniary

1991, Ethiopia held a seat on the UN Security Council. In order to ensure
the votes were carried in its favor, the US cut deals with several other
members of the Security Council, notably China. Ethiopia was however
in a weak position to exact terms from the US: it declared its opposition
to Saddam Hussein in the very first days of the crisis (and attempted to
brand tiie EPLF as an baqi pawn), and was begging for US and taadi
assistance already. The Ethiopian goymment had some ncoeases: a
meetipg took place between Foreign Minister Tesfaye Dinka and US
Secretary of State James Baker, the highest level meeting between officials
of the two countries since Mengistu seized power, and the US softened
its opposition to some World Bank loans to Ethiopia. However, apart from
these developments and some reassuring diplomatic platitudes, US-Ethiopia

relations remained cool.
The US administration does, however, deserve criticism for a rather
clumsy attempt to portray the government and rebel fronts as equally
responsible for human rights abuses. Presumably this distorticxi of the fsOb
was made in order to further the peace negotiations. Aocofduig to the
Department of State Country Rqfort on Human KgfUs Practices ibr 1990:
Both government and tiie various insurgent movements, notably the

EPLF and TPLF, have practiced forced conscription, imprisonment
without recourse, violence against civilian populations, torture, and
extrajudicial killing. Women have fallen victim to rape and abuse by
government and rebel soldiers, as both sides loot and pillage.^^
^ Department of State, "Country Reports on Human Rights Practices for 1990^"

p. 118.
376
While the rebel fronts have certainly been gaOty of abuses, the failuie
to disdnguish between their record and that of the govenunent is a
significant shortconung. bonicaUy, in May 1991, the US government
praised the EPRDF for its treatment of civilians.
The US came to play a crucial role in the final days of the Mengistu
government. Assistant Secretary of State Herman Cohen was due to
convene a meeting between representatives of the government, the EPLF,
EPRDF and OLF in London in mid-May, in order to negotiate a ceasefire.
If this were successful, a second round of negotiations was envisaged, which
would establish the terms of a transitional government. The meeting was
postponed to May 27 at the instigation of the rebel fronts, almost certainly
because they recognized dieir military supremacy (which tin US and
E0iio|rian government had consistently underestnnated) and wanted to press
home their advantage before the talks heg^ Iheir victories cnised
Mengistu to flee on May 21, and the army to disintegrate over the following
week. In the final days, the US publicly appealed to the EPRDF not to
advance on Addis Ababa before the talks were concluded in an effort to
forestall the massive bloodletting that it was feared might occur if the
EPRDF encountered strong resistance.
As it happened, the peace talks convened in the Berkshire Hotel in
London on a morning when Acting- President Tesfaye Gebre Kidan had
lost effective control of the remnants of the army, and Addis Ababa was
almost completely undefended. The EPLF had taken Asmara and Assab
in the previous tluee days. After ooosnhmg witih the four parties, flie US
asked the EPRDF to take control of Addis Ababa, to prevent a breakdown
in law and order. The government delegation then withdrew in protest.
The US was later accused by some opposition groups which had been left
out of the ceasefire talks of having "given the kejfs** of Addis Ababa to
the EPRDF; the EPRDF replied that "the door was already open" — i.e.
that there were no military obstacles remaining to its victory.

Simultaneously, the US also dropped its longstanding demand for the
maintenance of Ethiopia's "territorial integrity," and accepted the EPLF
demand for a referendum on the issue of Eritrean independence.
The US made clear its conditional support for the transitional government
of the EPRDF and EPLF. The conditions were that a political conference
should be held by July 1 m which non-combatant political forces should
also be represented, and that multi-party elections be held witiiin one year
(this was later extended to two years). Herman Cohen described the
conditions as "no democracy, no cooperation." This principle can only
be applauded, though it remains to be seen if the US consistently applies
it in future dealings with Ethiopia and other African countries. If it had
been applied by the US, the European Community and the United Nations
377
over the previous thirty yeais many of the tragedies wliicfa fell Ethiopia
would have been avoided.
378
AFRICA WATCH'S RECOMMENDATIONS
L To the Ethiopian Government and Eritrean Administration:
1. Ratify international instruments safeguarding the rights of

Chilians in armed conflict
Ethiopia is a party to die Geneva Conventions of 1949. Common

Aitide 3 of these four Conventions addresses conflicts that are not of an
international chaiader and hence governs the relations between a
government and insurgent fraces.^ libis requires non-combatants to be
treated humanely by aU sides in an internal conflict, and prohibits under
any circumstances:
• Violence to life and person, in particular murder of any kind,

mutilation, cruel treatment and torture;
* Taking of hostages;
* Outn^ upon penooal dignity, m particular, humfliatmg and

degra^qg treatmrat;
* The passing of sentences and the carrying out of executions

without previous judgement pronounced by a regularly-constituted
court, affording all the judicial guarantees which are recognized as
indispensable according to international standards.
The Additional Protocol on the Protection of Victims of Non-

International Armed Conflicts (Protocol II), adopted in 1977, codifies in
greater eaqilidt delafl tiie inleniatioiially-acceptable standards for the

Ml] III 1of dvfl wais. Tlmigli Ethiopia is not a party to this protocol, the
1
WW which it lays down are geaenlly applicable to the conflicts in

that country under customary international law. Africa Watch calls upon
the Ethiopian govemment to sign and ratify this protocol and abide by all
its terms.
Customary international law prohibits attacks against the civilian
population in internal conflicts. This principle was adopted by unanimous
^ The other articles in the Convention apply expressly only to international

conflicts.
379
vote at the UN, in General Assembly Resoluticm 2444, December 1969.
Africa Watch calls upon the Ethiopian govcfimieiit to affirm its
determination to abide by this principle.
Africa Watch calls upon any future government of Eritrea to sign and
ratify the Geneva Conventions, the Additional Protocols, and affirm its
determination to abide by the principle of Resolution 2444.
2. Call to account those fnlndpaUy raônslMe fir bnauin rlg^

abases in war
The transitional government headed by the EPRDF has indicated that

it intends to try membeis of the fonner regime for the hmnaa rights
violations they have committed, in accordance with international
standards. Africa Watch welcomes this commitment. The EPLF
administration of Eritrea has indicated that has similar intentions, but
it
has not made a public commitment. Africa Watch urges it to do so.

In all cases, we call for great care to see to it that the manner in
which such prosecutions and conducted embodies the rule of law.
It is a right of victims of hmnan rights abuse tet those reiyoosibte for
committiqg the abuses be brought to justice. Only the victuns of abuses
have the night to forgive the abuser; it is not for a govemment to do so.
It is essential that prosecutions for human rights abuses should be
conducted in a mamier that respects both the substance and the

appearance of justice and therefofo is likely to streqgthen rule of law
and the legal process itself.
Those principally responsible for gross human rights abuses should be
prosecuted for the crimes they have committed according to the relevant
domestic law. The trial may take place either in the ordinary courts or
before a tribunal specially convened for the purpose. In either case, it is
essential that international standards of doe process of law be adhered to.
The defendant should be allowed a defease conaBd of his or her own
choice, should have free access to tfiat cooasel, sfaoidd be allowed
sufficient time to prepare a defense, and should be allowed to Mimmnw
and cross^3Eamine witnesses.
It is important that the desire to bring as many as possible of those
principally responsible for gross abuses to court should not lead to the
adoption of a lower standard of proof, or the abrogation of due process.
A few prosecutions of those with the highest level of responsibility for
the most grave abuses, conducted according to internationally-recognized
standards, are more important than many conducted against lesser
offenders according to lower standards.
380
3. Award compensaUon to the victims of humaii rights abuses and
their families
Those who have been subject

to human rights abuse, or their families
in the case of the victim having been killed, should be entitled to
compensation. The payment of compensatiQn afaould be oonditiaiial only
on the proof that an abuse has been committed, itshould not be necessary
to identify the individual responsible or obtain a criminil conviction ot
that mdividuaL
4 Set up a Human Rights Commission
Watch urges both the transitional government headed by the

Africa
EPRDF and the EPLF administration of Eritrea to set up Human Rights
Commissions. While the precise structure and mandate of these
commissions is a subject open to discussion, Africa Watch believes that
die following conditions should be met:
(1) The Commission should take as its fiist task the investigation of
past human rights abuses. The purpose of this is two-fold. One
is to obtain evidence that can help to obtain criminal convictions
against those who have committed such abuses. The second is to
investigate the general background to the abuses, so that
recommendations can be made to the government for institutional
and legal safeguards for the future protection of iwman rights.
(2) The Commission should be an independent body, composed of

citizens who are widely respected for their integrity and their
commitment to human rights, with wide powm to decide nihn
issues and incidents to investigate, to launch and cany duough
investiigitfimiB of its own, to send investigative wissiows to different
parts of the country and abroad, to summon witneases and to take
necessary precautions to ensure the protection and/or anonymity of
witnesses. It should have the staff and budget to carry out these
tasks. should be able to recommend the criminal prosecution of

It
any individual it finds to be responsible for abuses.
(3) The proceedings of the Commission should be public, so far as is

possible and is consistent with the security of witnesses and the
need to ensure that the rights of flie accused are protected. The
findings of the Commisskm and the reaaom for its decisions
381
should be regularly aired on television and radio and in the print
media.
(4) A permanent Commission should operate as a standing body to

human rights abuses. It should be composed and
investigate
empowered according to the same criteria as the Conunission to
investigate past abnses.
5.Gnanntee aod nfeguard frMdom of bovcbmbC nd prohlbR ima-

vohmtary relocatloB
The government headed by the EPRDF has lifted

transitional
restrictionson internal travel. Those restrictions must not be re-imposed.
The EPLF should lift any such restrictions. Programs of non-voluntary
relocation should not be implemented, and people must be allowed to
leave resettlement sites and government villages where they have
previously been relocated without hindrance or threat of reprisal.
In the exceptional situations when overriding considerations of
physical security wanant the idocatum of a population, this must be done
only nnder the foUowiqg oooditioiis:
(1) Compensatory economic and health assistance is provided to

mitigate the adverse ejects of the lelocatioa and pnvcat
unnecessary suffering.
(2) The relocation is strictly temporary and is followed by a program

of assisted voluntary return to the previous habitations as soon as
possible. The decision to return and its timing should be made in
consultation with representatives of the relocated population, and
in accordance with their wishes as to as possible.
6. Affirm the principle of free aoocss to Imnnnltaritti eiilitiacc Ihr

all people
The new government in Ethiopia and the administration in Eritrea

should affirm the [înciple that all civilians should have free access to
humanitarian relief.
7. Adopt and enforce guidelines on recruitment to the Armed Forces
Africa Watch takes no stand on conscription per se and has no

objection m principle to the National MilitarySerWce Frodamatian of
382
1983. However, the implementatioii of that proclamation fell far short of
the principles of the Proclamatioii, and involved widespread abuses of
human rights. Africa Watch uigies the establishment of an independent
body (perhaps part of the Human Rights Commission) to monitor the
practices of recruitment to the armed forces. This body should have the
power to intervene when violations of the conscription guidelines occur,
in order to release under-age boys or those who have been conscripted
in an arbitrary and violent manner. If necessary, such a body should
have the power to recommend publicly that conscription be suspended
until it can be earned out in accofdance with the law.
8. Set up mcdiatloii and dispute-scttlnncnt coniBiiaBions to negotiate

peaoeftil sohitions to local conflicts la the floathwcat
Many of the conflicts in southwest Ethiopia did not consist in the

previous government fighting against local resistance, and therefore are
not likely to be resolved by the fall of that government. External forces
and the ready though selective availability of firearms are in many cases
more responsible for the violence. However, the new government has a
responsibility to all Ethiopian citizens to ensure that local conflicts are
settled in a peaceful manner. Government intervention is warranted.
Military intervention to make the warring parties submit by force of arms
is not only lilcely to M
but will, as in tfie past, involve human rights
abuses. Africa Watch therefore calls upon the new government to set up
conciliation committees, which will £Kilitate a peaceful resolution 6t
tiiese conflicts, and will make recommendations for the maintenance of
law and order and the continued respect for basic rights.
Africa Watch also calls upon the new government not to undertake
programs of forcible settlement of nomads. Any such programs must be
strictly voluntary.
9. Set up a Commission of Inquiry into the causes of famine
Famine has clearly been shown to be closely associated with abuses

of human rights. Understanding famine is essential to preventing it. This
requires an mvestigation that encompasses a wide range of dSciplines,
including inier aUa meteorology, environmental science, agriculture,
economics, social anthropology, medicine, demography, political science,
military strategy and human rights. Such an inquiry should include an
investigation of the manner in which inappropriate assistance, or
assistance in the wrong context, can contribute to famine.
383
This Commission of Inquiry should be empowered to hear evidence
from expert witnesses in the disciplines mentioned above, tbose with a
professional involvement in relief and development, employees of
international organizations, ordinary Ethiopian citizens, and anybody else
it deems appropriate. It should be empowered to launch investigations
of own into neglected aspects of the problem.

its
This Commission may produce evidence of a type which would

warrant a criminal prosecution against certain individuals or institutions
for deliberately pursuing a policy intended to cause starvation or other
forms of suffering amounting to fiumne. It is likdy to pnidnoe evidence
which shows that certam institntionB acted in an ineaponaible and
counter-productive manner. This should not however be its main
purpose, which should be to pcodnce recommendations about the fatme
strategy for preventing fonune, and for oombattu^ it should it occur.
U. To the Aid Donors and Aid Agencies
The following recommendations apply to United Nations organizations,

the European Community and bilateral donors such as USAID and
Britain's Overseas Development Administiation, and Id private voluntary
agencies.
10. Cooperate with the hMiairy faito the canscs of luDinc
International aid institutions should cooperate with the Commission of

Inquiry into famine. They should be prepared to reveal previously-
confidential information about their conduct towards the Ethiopian
government and insurgent movements, and, where appropriate,
neighboring countries with refugee populations and the former rebel
fronts.
Voluntary organizations should also be prepared to coopeiate with the
mquiry in a similar manner. The International Committee of fbt Red
Cross is the only organization that should be exempt from the
responsibility to testify, because fulfillment of its humanitarian mission
is closely linked to adherence to standards of confidentiality and
impartiality in its statutes.
384
11. Accept culpability for complicity in human rights abuses where
appropriate
a United Nations agency, or other aid donor or agency is found by

If
iht Human Rights Commission to have acted knowingly in a way likely
to contti^te to the contumation or extension of human rights abuses, it
shoidd be pvqMied to pay compensation in accordance with the princifdes
determined by the Commission. A plea of "acting in good faith'* may be
taken into consideration in mitigatiqg penalties, but should not be
sufficient to indicate lack of guilt.
12. Draw up principles for the conduct of relief operations in

situations ofhuman rights abuse and armed conflict
in the 1980s has clearly shown that the

The experience of Ethiopia
ethic humanitarian intervention to relieve suffering under any
of
conditions and without preconditions is deeply problematical, and may
lead to more sutfermg in the long nm. Africa Watch hu no easy
answers to the dilemmas raised. However, numy of the difBculties can
be resolved if humanitarian agencies agree on the human rights
preconditions for their worlc Such preconditions should include:
(1) A duty to make public information concerning gross human rights

abuses.
(2) A duty to examine the strategic context of any humanitarian

intervention that is taking place in or near to armed conflict or
government or hisurgent policies of forced retocation or restriction

on free movement. The examination should cover such issues as
whether the humanitarian assistance is allowm^ the ooinbatants to
consolidate or unprove their mUitary or political position. The
findings of any such examination shoidd be made public promptly
and reguhuly.
(3) A public commitment to assist the civilian population of an area,

provided there is reasonable security, whatever armed force is in
de facto control of the area. Thus, if a humanitarian organization
has a program in an area which changes from government to
hisurgent control, the program should contimie, subject to the
cooperation of the insurgent group (and vice versa). It is not
necessary for an oiganizatioa ahvayg to wtA
on bofii sides of a
conflict, merely to accept the principle that under these
385
ciicumstances it will do so, and to ensure that its aid to one side
does not become the basis for denyiqg aid to the otlier aide.
Africa Watch also urges humanitarian organizations and their donors

to impose humanitarian conditions on humanitarian aid. A principal
condition is that no assistance will be given by the international
community to facilitate the displacement of civilians, as displacement
itself is an abuse of human rights and a cause of unnecessary suffering,
and it should not be aided and abetted.
386
AFRICA WATCH BOARD OF DIRECTORS
WiUiam Carmichael, Chair
Alice Brown, Vice Chair
Roland Algranl
Robert L. Bemstem
Julius Chambers
Michael Clough
Roberta Cohen
Carol Corillon
Alison L. DesForges
Adrian W. DeWind
Thomas M. Franck
Gail Gerhart
Jack Greenberg
Alice Henkm
Richaid A Joseph
Thomas Karis
Russell Karp
Stephen Kass
Randal Kennedy
John A. Marcum
Gay McDougall
Toni Morrison
James C.N. Paul
Robert Preiskel
Nomum Redlich
Randall Robinson
David S.Titel
CSaude E Welch, Jr.
PUBUCATIONS
AFRICA WATCH REPORTS:
• Kenya: Taking Liberties^ July 1991, 432 pages.
• Academic Freedom <Sc Human Rights Abuses in Africa, April

1991, 153 pages.
• The Killings in South Africa: The Role of the Security Forces and
the Response of the Siate, Jaouaiy 1991» 85 pages.
• Malawi: Where Silence Rules —

The Siqfpression of Dissent in
Malawi, October 1990» 104 pages.
• Liberia: Flight From Terror — Testimony of Abuses in Nimba

County, May 1990, 30 pages.
• Denying the "Honor of Living" — Sudan: A Human RiÛs

Disaster, March 1990, 152 pages.
• Somalia: A Government at War with its Own People —

Testimonies About the Killings and the Conflict in the North,
January 1990, 268 pages.
• Zimbabwe: A Break with the Fast? Human Rights and Political

Unity, October 1989, 109 pages.
• No Neutral Ground: South Africa's Confrontation with the

ActMst Churches, Angnst 1989, 145 ptges.
• Angola: VU>Uaions of the Lam of Warty BoOtSideSjApn^

148 pages.
NEWSLETTERS:
ANGOIA: CM^ DgmMtted I9 iS Year War,
CAMEROON; Attacks Against Independent Press, February 12, 1991.
Government Cracks Down On Demands For HAiUparty Slystem: Seven

fBUei; JommBsIt mid Soidtm Anttted; DomUi Tern' IHed, hm
15, 1990.
HI -Treatment of Somali Rejugees: Denial of Refuge, Deportadoaa

and Harsh Coaditums of Deientkm, October 30, 1989.
ETfflOPIA Human Rights Crisis as Central Power Crumbles: Killings,

Detentions, Forcible Conscription and Obstruction of Relief, April 30,
1991.
a^HglMiu^ En^ty Democraq^: OhaYoart^^lttfiirmltAmumieti

Nifbvramimat lnCMmdFo^leamt^,UnA
200 Days in the Death of Asmara: Starvation as a Weapon and
Violations of the Humanitarian Laws of War, September 20, 1990.
Violent Suppression of Student Protest, August 30, 1990.
"Mengfm has Deddtd Bum Zft W» Wood': Bombing ofOvlUmu

and Chfilkm Targets by the Air Force, July 24, 1990.
Conscription: Abuses of Human Bl^ts daring Reendimem to the

Armed Forces, June 1, 1990.
GHANA Government Denies Existence of Political Prisoners; AGnister Says

Detainees Safer" in Custody, August 12, 1991.
Olffkial Attacks on ReOgious Fttedam, Ifiy 18» 1990.
Lawyers Detained for Commemorating Jhtdge's Murder; Human

Rights and the law mder the FNDC govemmem, July 14, 1989.
KEN YA Illegal Expulsion of More than 1000 Refugees, December 11, 1990.
Screening of Ethnic Soma lis, the Cruel Consequences of Kenya's

Passbook System, September 5, 1990.
PoUlicml Omckdemm bnensilfim, llagr 2, 1990.
Once Again, a Critical Magazine Faces Threat of a Banning Order:
The Nairobi Law Monthfy and its Editor Under Fire, Apiil 5, 19SX).
Harassment of Ethnic Somalis, December 6, 1989.
&q)pression of Proa R^tedom, Banning of CrUeai Papers and

Imolenmce ofDbsent, December 6^ 1989.
Forcible Return of Somali Refugees: Denial of Refuge, Deportation

and Harsh Conditions of Detention, November 17, 1989.
LIBERIA: Liberia: A Human Rights Disaster; Violations of the Laws of War by

All Parties to the Conflict, October 26, 1990.
Nina Years of Doe's tbOa: Afiica Watch Aaatm the Mecord, April
11, 1969.
MALAWI: ExOad JoumaM AamstHmted in laMa, October 2^ 1989.
A Poet Out Of Bounds: Jack Mapanfe Detained without Trial for Two
Years, September 20, 1989.
Deaths in Custody, Detentions and Discrimination, April 24, 1989.
MAURITANIA: Slavery: jiUve and 10 Years After it was Last MfoUsked, June
29,1990.
Persecution of Black Mauritanians - Summary Executions,

Deprivatitm of CitizenA^ JUegal Ejûlsions and Arbitrary Arrests,
September 7, 1989.
MOZAMBIQUE: New Constinaion Protects Basic Rights, But Political Prisoners still
Suffer VnfOr WOs, Wbamy 11, 1991.
NIGERIA: 27 New Executions for Coiq> Attempt; Government IMbans Academic

Union, Bars Caeferanoe, Septmbu 20, 1990.
42 Executed After Unfair Trial, July 31, 1990.
Post-Coup Crackdown Continues: Press, Lawyers, Government
Critics Targeted, June 18, 1990.
Aftermath of Abortive Military Coup: Suppression of Press Freedom,

Arbitrary Arrests, Fear of Unfair Trials, Torture and Summary
Eseeydau, May 10, 1990.
SOMALIA: Evading Reality: Government Announces Cosmetic Changes As

Abuses Continue mid Chtdkngu to Bngmm tnmuify, .ScpteaibMr 12»
1990.
An Update on Human Rights Developments Since Mid-July;

Government Attempts to Divert Attention from Abuses and Internal
Oppoation to Us Rule, September 22, 1989.
The Opvenmienti hmieqmttw BtaponM to itt KUOng tmd Armt €i

Dtmoiutmton, hitf 25, 1989.
Tiananmen Square Revisited: Soldiers Shoot Demonstnttors: R^priaal

KUlings and Mass Arrests FoUaw, Jtdy 21, 1989.
Women in Prison, June 13, 1989.
Hmnm Kffktt Abtaa In Ao Nat^' A Rq/ort finm the US GamtU

Aeeomtfng Office, May 26, 1989.
Govemmau ROeaaet 21 FutÔeal Ptbonen, Much 17, 1989.
SOUTH AFRICA: Out qfSi^' The Misety in BofikuOiatnmm, Stptuabu 16, 1991.
Forced Incorporation in Ciskei Bantustan: One Community Obtains

Land in South Africa; Other Communities Suffer for Resisting
hicarporatkm, January 12, 1990.
Bopkmhammm: RettOkalon Agnkut CmmmmMtM FlgkUng

IteorpomtSom Into Biminaimi: HmtnUgjkliWbHtm Banned and S7
VUkigen Charged wM Muder, Scipl«lw 1, 1989.
AiassAnests in Vrnda, auaxltActivisuUnderFln,Aag/aat4, 1989.
Chronicle of a Death Foretold: David Webster Writes on the South

African Death Squads, June 12, 1989.
SUDiAN: Inside Al Bashir's Prisons: Torture, Denial of Medical Attention and
Poor Conditions, Febniaiy 11, 1991.
WiMi ^rmt Iht SMr

iÛfe": Tim DmtHAon ofMahjoub Sker^,
ôtt 4mI ISsKter* DtocjBbtt 5^ 1990*
Miit0ii-«Me Faadne: CulpabU Nt^figence in Ac agement cf Mm

Food Security, War, and the Use of Food «f « FoUHeai Wtapon,
Novembei 7, 1990. AiU Repoit, $Sm
Suppression of Information: Curbs on the Press; Attacks on
Journalist^ Writers and Academics, August 30, 1990.
Lest Tkty Be ForgoUm^ Umn Vnm SktMa Primm, Uxf %, 1990.
Ogken Ex&emi mid Doctor Tortund to Duak, Apdl 26, 1990.
Threat To Women's Status From Fundamentalist Regime: Dismissal^

Atresia and Reslrkliom om Woman's Activiims, Afdl 9, 1990.
The Forgotten War in Darfur Flares Again, Apdl 6^ 1990.
The Maatocre at el Jebelein, Januaiy 23, 1990.
BoHiico! Dtkdimet kt Sndmr Lawyers, January 22, 1990.
Political Detainees in Sudan: Jhide Utdotdm, Jaanny 22, 1990.
BoUHcal Detauma in Sudan: Academics, Jaauaiy 12, 1990.
Political Detainees in Sudan: Medical Doctors, Jaouaiy 12, 1990.
Political Detainees in Sudan: Journalists, Poets and Writers, January

8, 1990.
Recent DevtiopmeiOs in Khtatatmn: Ah Update, DecBMbM 13, 1969.
Khartoum: Government to Execute Peaceful Protesters; The

ProwKot: MOitiotSllkig^mtdStarmkM Return, December 7, 1909.
Political Detainees in Sudan, October 27, 1989.
Destruction of the Independent Judiciary; Military Government

Clamps Down on Press Freedom, September 25, 1989.
Jhe June Coup d'Etat: F\f^ Days On, Augpst 22, 1989.
TANZANIA: Executive Order Denies Land Rights, Banbaig suffer heatings, arson
md crtmM charges, Ifnck 12> 1990.
ZANZIBAR: DetenUm Without Trial and CrtKdmi Charge* AgdnM Oommmt

Opptmmts, May 19, 1989.
ZIMBABWE: Government Moves to Curb Academic Freedom; Constitutioruil Rights
under Threat, November 20, 1990.
Jfejbue ofDelakiei MUdaa, May 18^ 1990.
jêr the General Election: OpposUitm polUicimi demined aid

beaten, Apdl 12, 1990.
Harassment of Opposition Party Members: Africa Watch Calls for

Election Safeguards, Maic^ 21, 1990.
Govemmtnt Defies Cam to HoU Trada Students;IMMk

IMmOt^ Reapmad Bat Shiduit Activity Cmhad, Odribtt 25, 1989.
Trade Umomsts and Politicians Detained; GovemmeM CUmea

IMimsUy, Arrtais Student Ltaderg, October 12, 1989.
Release of Detainees, My 4, 1989, updates News Fzom Zimbabwe,

June 13, im.
Lawyer and Opposition Politician Detained, June 13, 1989.
JOINT PUBUCAHONS with HUBflAN RIGHTS WATCH,

1990 and 1991
• Human Rights Watch, a quarterly newsletter published by Human

Rights Watch.
• Human Rights Watch World Report --An Annual Review of

Human Rights Developments and the Bush Administration's
Policy on Human Bights Worldwide, January 1991.
Human Rights Watch
Human Rights Watch is composed of Africa Watch, Americas Watch, Asia
Watch, Helsinki Watch and Middle East Watch, and the Fund for Free
ExpKssioiL
inD CDBCmiVO HIIIIIIIIIBO Mlllf MBO OK MIHNn JU MKBrnUBf GBUCS APOUk
DeWimi* vke dMk; Rolaid Atyanl; lin Aadanoiw Mer Bdl, Alke Bamn,
IVilUim OmfcinBl, DonMlijr GtaUnuBw Imbb Dinood, JonadHn Fntan* Jack
Greenberg, Alice H. Henkin, Stephen Kass, Marina Kaufman, Jeri Laber, Arydl
Neier, Bruce Ribbb Kwmcth RoUv Otvilto SdMll, Qtay Suk, and Robot
Wedgeworth.
The staff includes Aryeh Neier, executive director; Kenneth Roth, deputy
director; Holly J. Burkhalter, Washington director; Ellen Lutz, California
director; Susan Osnos, press director; Jemera Rone, counsel; Joanna Weschler,
Prison Ftoject director; Dorothy Q. Thomas, Women's Rights Project director.
Executive Directors
Africa Watch Americas Watch Asia Watch

Juan R|yfc%MkiT Subey Ji
Helsinki Watch Middle East Watch Fund for Free EaqiiessioD

Jeri Laber Andrew Whitley Gara LaMarche
Addresses for Ihamm Eights IKridk ami its Divislau
485 Fifth Avenue 1522 K Street, NW, Suite 910

New York, NY
10017 Washington, DC 20005
Tel: (212) 972-8400 Tel: (202) 371-6592
Fax: 0,12) 972-09QS Fax: (202) 371-0124
10951 West Pico Blvd., #203 90 Borough High Street

Los Angeles, CA 90064 London SEI ELL, UK
Tel: (213) 475-3070 Tel: (071) 378-8008
Bn: (213) 475^13 Fne (071) 378-8029
For the past thirty years—under t)oth Emperor
Halle Selassie and President Mengistu Haile
Manam—Ethiopia has suffered cxjntinuous war and
intermittent famine until every single province has
been affected by war to some degree. Evil Days,

documents the wide rmqe of violations of basic
numan rights committ^ by alt sides in the conflict,
especially the Mengistu government's direct
responsibitty for the deaths of at least half a million
Ethiopian civilians.
The Ethiopian amfty and air force have killed tens

of thousands of civilians. The notorious urban ''Red
Terror **
of 1977-78 was matched by indiscriminate
violence against rural populations, especially tn
Eritrea and Tigray. Counterinsurgency strategies
involved forcibly relocating millions of rural people
and cutting focKf supplies to insurgent areas. These

military policies were instrumental in creating
famine. The government used sup^l^ as
relief
weapons to further its war aims. There is now a

prospect of lasting peac^, but concerns remain
such as the d^mmd for justice and the future
protection of human rights.
ISBN 1-56432-038-3

Ethiopia Big Book

Uploaded by

Document Informationclick to expand document information

Document Informationclick to expand document information

Copyright:

Available Formats

Ethiopia Big Book

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ethiopia Big Book

Uploaded by

Copyright:

Available Formats

30 YEARS OF WAR AND

A Division of Human Rights Watch

Thirty Years of War and Famine in Ethiopia

An Africa Watch Report

Human Rights Watch

Cover design by Dan NdH and Debonh Thomas.

Libiaiy of Coafftas Catalog-in-Fnhlicatiflp Data

Evil days: thirty years of war and famine in Ethiopia,

I.Ethiopia— and government — 1974-1991.

Glossary of AGronynis vii

X. Background to War and Famine in Ethiopia 19

2. Scorched Earth in Eritrea, 1961-77 39

3. Rebellion and Famine in the North under Haile Selassie 55

4. Insurection and Invasion in the Southeast, 1962-78 65

5. The Secret Wais to Crush the Southeast, 1978-84 81

6. The Red Terror 101

7. Total War in Eritrea, 1978-84 113

8. Counter-Insurgency and Famine in Tigray

9. "Economic War" on the Peasants and Famme 157

10. War and the Use of Relief as a Weapon

11. Starving Tigray, 1984-88 195

12. Resettlement 211

13. Vfflagization, 1984-90 231

14. Eritrea under Siege, 1988-91 237

15. Aimed Decision: The North, 1988-91 255

16. The Politics of ReUef, 1989-91 277

18. Wars within Wars: The Western and

19. Divide and Misrule: Tte East, 1984-91 347

20. Western Policy towards Ethiopia 359

AMca Watch's Recooimendations 379

* Richard Pankhurst, The Ethiopian Famine: Cultural Survival's Report

^ The designation TPLF-EPRDF is intended to encompass the TPLF from

This report contains dfacusskms of tfie available statistics for tiie

Ethiopia is beset by coofliciiqg teiritarial daims; the status of Eritrea

This report was researched and written by Alex de Waal, Associate

ALF Afar Liberation Front

Why this Report Now?

The principal wars in Ethiopia are now

intervention. This is a tough ethical dilemma with no easy solutions, but

becoming their unwitting tool.

Violence against Civilians

The most characteristic feature of the war has been indiscriminate

Counter -Insurgency Strategy

Throughout the war, the government followed a more-or-less

(2) Hie systematic lestriction of food supplies (botfa ooBunacial and

(3) The use of indiscciminate violence and exemplary tenor against

The strategies were implemented of the

The Creation cf Famine

One consequence of the government's military policies, particularly

The Treatment of Soldiers

Inflaming Local and Neighboring Conflicts

Deliberately or inadvertently, the Ethiopian government inflamed other

The Record of the Rebel Fronts

The committed abuses against Chilians and prisoneis of

The Impact on Civil Politics

^ See: "Ethiopia: Mengisln's Empty Democracy," News jum Africa Watch,

The International Role

Hie international community has played a mixed ide in Ethiopia's