Comparative Political Studies: Interests, Inequality, and Illusion in The Choice For Fair Elections

Comparative Political Studies
http://cps.sagepub.com/
Interests, Inequality, and Illusion in the Choice for Fair Elections

Nancy Bermeo
Comparative Political Studies 2010 43: 1119
DOI: 10.1177/0010414010370438
The online version of this article can be found at:

http://cps.sagepub.com/content/43/8-9/1119
Published by:
http://www.sagepublications.com
Additional services and information for Comparative Political Studies can be found at:
Email Alerts: http://cps.sagepub.com/cgi/alerts
Subscriptions: http://cps.sagepub.com/subscriptions
Reprints: http://www.sagepub.com/journalsReprints.nav
Permissions: http://www.sagepub.com/journalsPermissions.nav
Citations: http://cps.sagepub.com/content/43/8-9/1119.refs.html
>> Version of Record - Jun 28, 2010
What is This?
Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

Comparative Political Studies
43(8/9) 1119–1147
Interests, Inequality, © The Author(s) 2010
Reprints and permission: http://www.
and Illusion in the sagepub.com/journalsPermissions.nav
DOI: 10.1177/0010414010370438
Choice for Fair http://cps.sagepub.com
Elections
Nancy Bermeo1
Abstract
Why do actors in transitional governments choose to hold fair elections
when so many other options are available? The answer to this question is key
to understanding an essential element of democracy’s institutional collage.
This essay explores the choice of fair elections through the comparison of
two episodes in Portuguese history: the elections held at the founding of the
First Republic (which were unfair) and the elections held after the fall of the
Salazar–Caetano dictatorship (which were fair instead). The findings challenge
arguments strictly based on the socioeconomic and class-based determinants
of democratization: Although collective actors pursued outcomes on the
basis of the expected distributional consequences of their choices, the author
shows that cross-class political actors were more important than class actors
and that the distribution of institutional power was more important than the
distribution of wealth. The author also shows that illusions and misperceptions
were highly consequential for important institutional choices. If scholars seek
to explain democratization on the basis of structural realities alone, they risk
overrating the power of wealth and underrating the power of the imagined.
Keywords
democratization, Portugal, fair elections, episode of democratization, critical
junctures, political parties, military
1
University of Oxford, Oxford, UK
Corresponding Author:
Nancy Bermeo, Nuffield College, University of Oxford, New Road,
OX1 1NF Oxford, United Kingdom
Email: nancy.bermeo@nuffield.ox.ac.uk

1120 Comparative Political Studies 43(8/9)
A new democracy is more like a collage than a canvas. It will not emerge all
of a piece but rather bit by bit, with each component part shaped at different
times, by different hands. Democracies are inevitably a set of interconnected
institutions—each with a history and texture of its own. When our research
questions require us to draw neat distinctions between democracies and non-
democracies or to associate the emergence of democracy with a particular
date, these realities become obscured. The “historical turn” highlights them
instead. Recognizing that democracies are made “one institution at a time”
(Ziblatt, 2006), we see that an accurate rendering of democratization anywhere
requires a separate historical analysis of each of democracy’s constituent parts.
The institutional multiplicity of democracy has been highlighted by work
on “democracy with adjectives” (Collier & Levitsky, 1997) and, more recently,
by work on “hybrid regimes” (Levitsky & Way, in press). In pointing out the
serious shortcomings of the different regimes that we categorize as democracies,
these literatures force us to recognize the dramatic variation in the time it takes
for the many institutional components of democracy to develop. These litera-
tures also teach us a great deal about institutional deficiencies. But what factors
explain why and when a particular institutional component of democracy
takes a more laudable form? We should ask this question of a whole range of
democratic institutions. Here, I pose the question of electoral institutions or,
more precisely, elections.
Fair elections are a defining dimension of the democratic regime type. Yet
we know little about why actors in transitional governments choose to hold
fair elections when so many other options are available. History has shown
us that elections can be fixed, restricted, delayed, or cancelled altogether.
Why power holders choose to risk devolving power to the winners of fair,
and therefore unpredictable, competitions is far from obvious.
The Cases and the Argument

This essay explores the choice of fair elections through the comparison of
two electoral episodes in Portuguese history. The first occurs in 1911 at the
beginning of the First Republic and represents a choice for unfair elections.
The second occurs in 1975, after the fall of the Salazar–Caetano dictatorship,
and exemplifies a choice for fair elections instead.
Given the many decades of history and economic development that separate
these two cases, we might logically suspect that these different outcomes can
be explained with a strictly economic argument. Following the carefully
crafted leads offered by Boix (2003) and Acemoglu and Robinson (2006), we

Bermeo 1121
might reason that changes in the Portuguese economy led class actors to
weigh the risk of fair elections differently across time. More specifically, we
would expect to find that, between 1911 and 1975, economic development
advanced, levels of economic inequality dropped, asset mobility increased,
and the propertied classes, facing a relatively affluent median voter, came to
find fair elections more affordable.
Though this line of reasoning is logically appealing, it is revealed as
incomplete in the bright light of historical analysis. Though the main variables
in the rationalist-economic argument correlate perfectly with the outcomes in
both cases, the correlation is spurious. The explanatory variables do not work
as predicted, and two other sets of variables carry greater explanatory weight.
Each of these episodes thus represents what is known in the methodological
literature as a “typical case” meaning, one that seems to fit a widely accepted
theory but, after closer analysis, ultimately points to other casual mechanisms
(Seawright & Gerring, 2008, p. 299).
The mechanisms that do explain the puzzle of fair elections in Portugal
are not unrelated to the cognitive processes of class actors, but class actors
are decidedly less important than these particular rationalist-economic argu-
ments would lead us to predict. The key actors in the Portuguese drama are
not the rich, the middle class, and the poor but rather another set of political
actors with (mostly) cross-class identities and interests: namely, political par-
ties of varied sorts and the Portuguese military. The choice for fair elections
derives mostly from reasoning within these cross-class organizations. More-
over, the reasoning that led these actors to have fair elections (and hence to
start the process of democratization) focused less on material inequalities
than on inequalities of power and security. Portugal held fair elections
because leaders of cross-class political organizations came to see fair elec-
tions as beneficial to their organizations’ interests.
Figure 1 offers an overview of the cases and the argument to follow. As its
upper section indicates, both cases correspond in key respects with what our
rationalist-economic theories would predict. Portugal’s elections were unfair
when its economy was characterized by high inequality and low asset mobil-
ity and fair when inequality decreased and asset mobility increased. Yet his-
torical analysis enables us to see that changes in levels of inequality and
changes in asset mobility did not have the direct effects often posited. Classes
did not behave as unified actors in either episode, and threats to the resources
of propertied classes did not decrease as inequality dropped and asset mobil-
ity rose. On the contrary, threats to property were unmistakably greater
before the fair elections than before the unfair polls. Ultimately, the nature of

Rationalist-Economic Analysis
Founding
Election Economic Inequality Asset Mobility Predicted Outcome
1911 High Low Unfair Elections
1975 Moderate Moderate Fair Elections
Historical Analysis
Founding Economic Asset Incentives for Fair

Election Inequality Mobility Threats to Property Elections Outcome
Predicted Actual Parties in Military
By Transitional
Rationalist Government
Economic
Theory
1911 High Low High Low Weak Weak Unfair
Elections
1975 Moderate Moderate Low High Strong Strong Fair
Elections
Figure 1. Analytical approaches
elections was shaped much more by political parties and the military than by
any single class group. The lower section of Figure 1 provides an overview
of the historically accurate argument.
Why did the calculations of parties and the military change over time?
Changes in the structure of Portugal’s economy were only indirectly important.
Three very different sets of factors led cross-class political actors to see fair
elections as beneficial to their organizational interests in 1975, and each is
anticipated by Capoccia and Ziblatt (2010) in their argument for the “histori-
cal turn.” The first factors involved the creation of “undemocratic elite safe-
guards” (Ziblatt, 2006, p. 313) that made the outcome of fair elections less
consequential (and hence less risky) for both the military and pivotal political
parties. These safeguards took the form of a military–party pact.
The second set of factors leading to a positive view of fair elections was
essentially ideational. Though organizational interests varied across groups,
each group of actors drew on similar sorts of ideational materials to weigh the
choices before them. First, a cultural context (Capoccia & Ziblatt, 2010)
shaped by experience with mass emigration to democracies and with the dic-
tatorship’s own electoral fraud had led the great majority of the Portuguese

Bermeo 1123
citizenry to expect fair elections under a new, democratic regime. Over time,
elite actors came to recognize that violating these expectations would impose
enormous reputational costs on those held responsible and damage the orga-
nizational interests these actors sought to maximize. Second, by the time the
elections were held, a broad range of party and military actors came to believe
that they had widespread popular support and thus that the elections would
work to their benefit. These actors saw elections as a means of both verifying
and building organizational legitimacy.
A third set of factors boosting the belief in the utility of fair elections was
essentially resource based. The idea of fair elections triumphed because a criti-
cal mass of political organizations came to see them as a means to acquire
important resources for political actors in later struggles (Capoccia & Ziblatt,
2010). Party leaders predicted that first their organizations would benefit from
an injection of material resources related to campaigning and that eventually
their votes could be translated into legitimacy resources that could be used as
political currency in the future.
The actors who organized the founding elections in 1911 faced a wholly
different set of incentives. First, conflict between military actors and the party
dominating the provisional government prevented the forging of any credible
pact and severely limited the probability that other “safeguards” would ever
be enforced. Second, the “cultural context” in which the First Republic
emerged meant that popular expectations of fair elections were low and, more
importantly, that the Republican Party had no illusions about its popularity.
Finally, neither the party leading the transitional government nor its military
sympathizers had any prospect of drawing on the material resources of for-
eign allies. Transitional government elites in both periods made choices based
on predictions of what elections might bring, but they did so in radically dif-
ferent contexts and with radically different sets of information.
The Argument and the Democratization Literature

My emphasis on elite perceptions and predictions constitutes a point of inter-
section with the class-based arguments I cite above, but my argument
diverges from these in several respects. The first concerns the greater weight
I accord to past experience. Like Capoccia and Ziblatt, I find that assump-
tions of unit independence are often difficult to sustain. Elite “preferences
over regimes” do derive from “predictions” regarding “economic and social
consequences” (Acemoglu & Robinson, 2006, p. 20), but predictions and
preferences are always shaped by interpretations of the past. Throughout the
long deliberations that preceded the holding of fair elections in 1975, actors

across the political spectrum used “past experiences of successful or failed

democratization” (Capoccia & Ziblatt, 2010, p. 940) as lenses to study which
options were most desirable. Portugal’s experience with fraudulent elections
during the First Republic left legacies that shaped elite preferences decades
later when the issue of fair elections was revisited.
My second difference with the economic arguments cited above derives
from the greater weight I accord to misperceptions, or illusions. The use of
past experience was often consequential in shaping choices, but it was some-
times misleading. Partial parallels often distorted elite perceptions and led to
predictions that were not borne out. Whether historical or contemporary, the
information actors used to make their choices was often poor. Cognitive
mechanisms certainly drove events forward as rationalist theories would predict,
but the information that fueled these mechanisms was tainted with illusions
of various sorts. Ironically, elites often choose to hold fair elections on the
basis of misinformation and misperceptions. Scholars who explain democra-
tization on the basis of calculations concerning structural realities alone risk
understating the power of the imagined.
The diminished weight I accord to class actors constitutes a third way my
argument diverges from strictly structuralist analyses. Of course, we cannot
explain democratization without attention to classes, but I argue that in any
democratization process there may be pivotal episodes when cleavages of
other sorts are determinant and when actors do not make choices on the basis
of their class identity alone.
Like rationalist and nonrationalist structuralists, I see cross-class alliances
as important, but I emphasize, first, that the incentives for alliances may not
be economic alone and, second, that the alliances that matter most are often
those among individuals with different class identities acting in cross-class
political organizations such as political parties and the military.
My emphasis on political parties and military organizations is a final factor
that distinguishes my argument from a purely structuralist perspective.
Whether rationalist or nonrationalist, structuralist theories often leave the
roles of the military and political parties undertheorized. When these organi-
zations are included in structuralist analyses, they are usually framed as
merely the vehicles of particular classes. The “historical turn” enables us to
see, first, that the organizational interests of parties and militaries can stimulate
choices that are autonomous of any single class and, second, that the prefer-
ences of party and military elites can be decisive in shaping democratic insti-
tutions. I illustrate these points in the comparison that follows and conclude
with a summary of how the comparison contributes to our understanding of
democratization.

Bermeo 1125
The First Republic and the Choice to Cheat

Fair elections involve impartial administration, freedom from intimidation,
honest ballot counting, reasonable media access, opportunities for a broad
spectrum of parties to compete, and, in contemporary times, universal suffrage
(Elkit & Svensson, 1997, p. 35). We can generally judge an election as fair if
it is based on “electoral rules that all parties accept and respect” and if all
significant “parties accept the process and respect the results” (Pastor, 1998,
p. 160).
By this definition, the regime that emerged in Portugal in 1910 began with
unfair elections, whereas the regime that emerged in 1974 began with fair
elections instead. The narratives that follow show that neither of these out-
comes is explicable with a class-based, economic argument alone. Cross-
class political actors in the form of political parties and the armed forces were
decisive in both cases, and they were moved less by calculations related to
property than by calculations related to organizational interests and power.
Portugal’s First Republic was born in October 1910 when armed groups
associated with the Republican Party ousted a constitutional monarchy that
had engaged in “oligarchic parliamentarism” (Costa Pinto, 1995, p. 92).
August elections had yielded the party only 14 seats in a 200-seat legislature
and caused an intolerable rise in frustration in its ranks. Although the party
had support among urban professionals, the petite bourgeoisie, and industrial
workers, the monarchy’s system of election rigging guaranteed that its repre-
sentation in parliament would never reflect its real popularity (Wheeler,
1978b, pp. 45-46). Armed insurrection seemed to be the only means of gain-
ing power within the state.
Promises of electoral reform had loomed large in Republican Party dis-
course, but the elections in the new regime were no fairer than those in the
old. Inspired by the Third Republic in France, the party leaders had promised
universal suffrage. Yet once in power, they actually eliminated illiterate voters
from the registration lists, making their first elections even less inclusive than
those held under the deposed monarchy. The fraction of the adult male popu-
lation eligible to vote dropped from 50% to 30% (Costa Pinto & Tavares de
Almeida, 2000, p. 6).
The unfairness of the May 1911 poll was multifaceted. Though electoral
turnout rose above previous levels, reaching a high of 61.4%, all the votes
cast went to Republican lists (Marques, 1975, p. 84). All the candidates for
the assembly elections were chosen by the Directorate of the Portuguese
Republican Party to ensure a complete Republican Party victory. No other
parties were permitted candidacies. Armed vigilantes, aptly dubbed the Ants,

swarmed the streets, intimidating voters and forcing most independents and
monarchists to abstain from voting altogether (Wheeler, 1978b, p. 78).
The risk-averse behavior of the Republican elites derived in part from the
high-stakes game at hand. The elections were for a Constitutional Assembly
that would shape the new order in fundamental ways. Several European
countries had made clear that they would not recognize the Republic until it
elected a Constitutional Assembly, and thus the regime was forced to hold
elections less than 8 months after the seizure of power. This short time frame
afforded the Republicans too little time to broker new deals with the local
patrons who had fixed elections under the old regime. Fixing the ballot was,
thus, the only way to ensure victory.
The unfairness of the May 1911 elections escaped neither the public nor
the political class. One of Lisbon’s most respected journalists immediately
concluded that they were not a “country’s elections but a party’s elections
instead” (Bobone, 2009). In July 1911, on the floor of the Constitutional
Assembly, a deputy spoke openly of the election’s unfairness, saying “with-
out an honest electoral law . . . the Parliamentary Republic, like the Constitu-
tional Monarchy, would be a lie” (Wheeler, 1978b, p. 80).
Republican electoral laws changed several times, but fair elections were
never held. Openly Catholic candidates eventually ran for office after 1915,
and after 1918 even monarchists entered electoral competition (Marques,
1975, p. 84), but the Republican Party, renamed the Democratic Party, simply
commandeered the old political machinery of the monarchist past, buying
votes through local patronage networks and silencing opposition through
threats and violence. The regime went on to have seven legislative elections
and eight presidential elections, but exclusion, violence, and fear defined
them all (Schmitter, 1978, p. 150). These conditions soon drove large sectors
of the Portuguese people toward abstention. As early as 1919, the abstention
rate in Lisbon rose to 80% (Marques, 1975, p. 82). By the Republic’s last
election in 1925, the situation was hardly better. As an eyewitness from the
U.S. State Department wrote at the time, “Only the more ardent and hardy
voters of the opposition parties appeared at the voting places, as others lacked
the temerity to face a possible physical encounter with the . . . poll workers”
(Wheeler, 1978b, p. 308).
Why did the leaders of the First Republic choose to have unfair elections?
Most of the answer lies with the organizational interests of the party control-
ling the government. The ruling party chose to cheat because it recognized
that fair elections meant sure defeat. The party had existed openly for over
three decades, and its leadership was painfully aware of the limitations of its
popularity. The monarchy had used unfair means to deprive the party of the

Bermeo 1127
parliamentary seats it might have earned in urban areas, but party leaders
knew that an end to monarchist vote-rigging would not mean Republican vic-
tory. On the contrary, they recognized that Republican sentiment in most
parts of the country was almost nonexistent. Republican elites considered rural
Portugal “hopeless” (Costa Pinto, 1995, p. 94). On the eve of the 1911 elec-
tions, nearly 60% of the party’s organizational units were in one of two cities,
yet the people—the potential popular majority—lived elsewhere, enmeshed
in dramatically different cultural networks (Costa Pinto & Tavares de
Almeida, 2000, p. 13). The few provincial offices that existed on paper did
not function in fact (Coutinho, 2009). Party elites could not escape knowing
that “it was rare to find a Republican 20 km from Lisbon or 10 km from
Porto” (Valente, 1997, p. 11).
Structural factors certainly played a role in limiting the party’s capacity to
build support. A rural illiteracy rate of 86% (Schwartzman, 1989, p. 139)
hampered the party’s ability to spread its ideology via newspapers (Costa
Pinto & Tavares de Almeida, 2000, p. 11). With only 16% of the Portuguese
population living in towns (Schwartzman, 1989, p. 139) and notoriously poor
transportation networks, new adherents were difficult to reach. Finally,
inequalities in the distribution of wealth meant that people in Portugal’s
countryside lived an extremely precarious existence (Costa Pinto & Tavares
de Almeida, 2000, p. 3). The meager resources local bosses traded for votes
had value precisely because the population was so poor.
Although economic structures hampered the dissemination of Republican
ideas, the party’s poor electoral prospects—and its resultant willingness to
cheat—derived much less from problems of exposure than from the party’s
ideology itself. The party had positioned itself at the intersection between
two cross-class cleavages of great depth: a religious cleavage related to
Catholicism and an institutional cleavage related to the monarchy. The party
preached republicanism and anticlericalism in a nation where monarchism
had a solid base and where the Catholic Church was fervently supported.
This was not a winning combination for electoral victory, and the Republican
Party elite had no illusions that it was. Indeed, the party was so conscious of
its frail foundations that even the success of its 1910 mobilization against the
king is rumored to have come as a surprise (Coutinho, 2009).
Support for the monarchy was weaker than support for the Catholic
Church but pervasive enough to make a fair election much too risky. The fact
that monarchists had pockets of support within the military made them espe-
cially threatening. In the period before the regime’s first election, the republic
was rife with rumors that monarchist military forces were amassing on the
Spanish border and moving westward to seize control (Coimbra, 2000, p. 99).

The Republicans had no way of knowing precisely how much support mon-
archists would gain in a fair election, but they knew that rural residents were
overwhelmingly monarchical and that the Crown had especially strong support
in Portugal’s north. Party leaders saw barriers to the free franchise as “necessary
evils to prevent the return of the monarchists” (Lloyd-Jones, 2003, p. 98).
Struggling to control and consolidate what was then one of only three repub-
lican regimes in all of Europe, the party took no chances in allowing monar-
chists the voice that a fair election would have required.
The party also took no chances competing with the Catholic Church, for
this, too, was a battle they would surely lose. The Church had been enjoying
a revival in the decades before the monarchy was deposed and was, without
doubt, the strongest single organization in Portugal. Catholic associational
life was vibrant and unambiguously cross-class. Its major mobilizational
arm, the Association of the Apostolate of Prayer, had a full 1,501 branches in
1909 and “claimed a membership of over 2 million—about half the total
population” (Robinson, 1977, p. 348). Some 12,000 workers belonged to
Catholic Workers Circles, an extensive network of Catholic rural associa-
tions dotted the north and center of the country, and Catholic University students
and graduates were mobilized in Academic Centers for Christian Democracy
(Robinson, 1977, p. 349; Wheeler, 1978b, p. 63). The nation’s Catholic party
had been founded in 1903, and though it was dwarfed in Parliament by the
two large parties supported by the monarchy, it was inspired and supported
by successful models in Germany and Belgium. Its organizational potential
meant that it was seen as a serious “challenge by the Republicans” (Robinson,
1977, p. 349).
The Republicans recognized that most Portuguese rejected the anti-Catholic
component of their ideology. Historians report that with the exception of the
landless workers in the sparsely populated Alentejo, most of the nation simply
“did not want to be secular” and that this generalization applied across classes
(Valente, 1997, pp. 10-11). The leadership of the Republican Party recog-
nized that “in the wake of the King’s removal” the Catholic Church was “the
principal enemy to be overcome” (Meneses, 2004, p. xiv). The possibility of
winning a fair electoral competition with such a powerful “enemy” was nil.
The Republican Party might have moderated its anticlericalism if it had
been forced to compete in a fair election, but it opted for an all out assault on
the Church instead. Just days after assuming power, the Republican Party
expelled the Jesuits, dissolved all religious orders, nationalized religious
property, decreed that priests could not vote, ended religious instruction in all
schools, and banned religious dress in public. At least one historian concludes
that the legislation passed before the May elections was “the most severe

Bermeo 1129
anticlerical action taken by a government in Europe up to that time” (Wheeler,

1978b, p. 68). The Republicans’ anticlerical position was so alien to Portuguese
cultural norms that it eventually precipitated a split in the party itself. Repub-
licans with a more moderate position formed two new parties, and the Repub-
licans renamed themselves the Democratic Party. The party’s war on the
Church eventually gave rise to a new militancy among believers (Valente,
1992), but in the short run it provided yet another reason to prevent fair elections.
The risks of an electoral loss to religious forces who would presumably take
revenge for their own victimization was too great to bear.
A final factor that led the Republicans to hold unfair elections was the
perception that no one would stop them. The Portuguese public had never
witnessed fair elections and had low expectations. More directly, the coercive
apparatus of the state was both unable, and unwilling, to ensure fair conduct.
The military simply “did not exist” as a unified organization when the monarchy
was overthrown (Ferreira, 1992, p. 29). It was “paralyzed by various contra-
dictions” (Ferreira, 1992, p. 34), and this paralysis was obvious as the monarchy
was ousted. The armed forces did not abandon the monarchy altogether, but
the great majority of military men simply remained neutral as the old regime
fell (Carrilho, 1985; Valente, 1992). The Republicans speculated that this
immobilism would last long enough for the founding elections to take place,
and this was correct, but the military’s immobilism was soon replaced by
armed factionalism.
The military’s factionalism enabled electoral cheating but eventually
undermined the regime that the cheating was intended to preserve. At first,
the Republicans formed an alliance with a large sector of the officer corps
who hoped that the new regime would provide the resources to modernize
and strengthen the military as an institution (Ferreira, 1992, p. 43). But the
concord was short lived. The war against the Church created problems of
public order that overwhelmed the military’s organizational capacities. When
the armed forces were prohibited from participating in religious ceremonies,
alliances were further strained (Coimbra, 2000, p. 26). To complicate matters
further, aristocrats, conservatives, and monarchists maintained their influ-
ence in military organizations of varied sorts. Only 43 officers left the mili-
tary when the regime changed, leaving the deep divisions within the
organization intact (Ferreira, 1992, p. 46). The republican authorities never
established a solid base with career officers, and by the end of the Republic,
military actors across the political spectrum had managed to incite some form
of military insurrection (Wheeler, 1978a, p. 870). After witnessing, and often
shaping, a full 45 different governments in less than 16 years, military forces
brought an end to the First Republic in 1926. They acted because factional-
ism, inadequate funding, and the burden of maintaining public order for a

regime with low legitimacy had caused a level of professional dissatisfaction

that could be corrected only with a regime change (Carrilho, 1985; Ferreira,
1992, p. 50).
Thus, the choice to have unfair elections in 1911 was not the consequence
of class actors making calculations about material redistribution. Despite low
development and high inequality, the new government was operating in an
environment that posed little or no threat to nonreligious property relations.
Class identities were trumped by cultural cleavages, especially those con-
nected with Catholicism. It would thus be inaccurate to frame the story of the
republic’s unfair elections as a story about classes fighting over redistribution.
The Republican Party had struggled since its foundation in 1876 to be a
cross-class party, and though it was doomed for ideological reasons to be a
small party, it had built a relatively heterogeneous base by 1910. Eyewitness
accounts of the Republican revolution described barefoot peasants with
pitchforks marching alongside men in suits toting rifles (Wheeler, 1978b,
p. 63). The party’s leadership was from the liberal professions, but there is no
evidence that the party was the tool of the liberal professional class or of any
other class grouping. The party’s attack on the Church disrupted all the social
spaces that liberal professionals required for their livelihood. The party did
not serve the interests of Portuguese workers either: Shortly after assuming
power, it sent its own militia marching through the streets of Lisbon to crush
strikes, claiming, with dubious logic, that the striking workers were doing the
bidding of the monarchists (Cabral, 1977, p. 448). It soon launched what
Valente called a “popular terror” to repress workers movements throughout
the country (Wheeler, 1978a, p. 867).
The men who organized the First Republic’s elections “did not like the
thought of interfering with the market or of reforming agriculture through
redistribution” (Meneses, 2004, p. xix), yet the Republicans were not a tool
of the propertied classes either. The party’s antimonarchical and anticlerical
ideology drove the vast majority of propertied elites into the ranks of conser-
vative parties. Ironically, it was Portugal’s cross-class conservative parties
who backed the broadest franchise. When conservatives managed to briefly
dislodge the Republicans from power in 1917, they immediately expanded
the franchise to include all Portuguese males because they imagined that uni-
versal suffrage would work in their organizational interests. The Republicans
reversed the law as soon as they regained power, (Ramos, 2004, p. 561), but
the incident illustrates the critical connection between organizational inter-
ests and institutional change. In this instance, as in so many others, key dem-
ocratic institutions were shaped by the organizational interests of cross-class
parties and not by classes.

Bermeo 1131
With classes divided along cultural lines, serious redistributive threats

posed formidable collective action problems and, thus, never materialized.
The cleavage related to Catholicism was far more salient than any cleavage
related to class. As R. A. H. Robinson (1977) aptly put it, “By 1910 . . . the
religious question had become the most important political issue and was
inextricably intertwined with the question of regime” (p. 351).
The Choice for Fair Elections

By the time of the Portuguese regime change in April 1974, divisions over
the Church had faded considerably and the Crown had few supporters.
Portugal had undergone substantial economic change, and by most statistical
measures levels of inequality had declined. Estimates suggest that between
1911 and 1975, Portugal’s literacy rate more than tripled, land inequality
dropped by 50% (Vanhanen, 1997), and GDP rose by a multiple of nearly 6
(Maddison, 1995).
Yet classes were still not the key players when the choice for fair elections
was made. The “historical turn” shows us that the organizational interests of
cross-class political actors were once again decisive. The new regime’s electoral
institutions were shaped not by class interests per se but by the cross-class inter-
ests of party and military organizations heading the transitional governments.
Portugal had crossed what scholars deemed a meaningful threshold of
modernization, but this dictated no necessary outcome in terms of either
political system type or level of material inequality. States with higher per
capita GDPs, such as Mexico, South Africa, and Iran, were still dictatorships
in the 1970s (Muller, 1998, p. 138), and states with nearly identical levels of
income inequality were turning away from democracy and toward dictator-
ship in the same time period. Table 1 presents the evidence.
The “historical turn” enables us to see that similar statistical profiles may
mask dramatically different sets of realities and that the noneconomic elements
of these realities may shape actors’ institutional preferences in powerful ways.
The preferences of military actors brought about the 1974 regime change,
for it began with a military coup. The 1910 regime change was brought about
by a political party and a paramilitary group instead. This contrast in origins
was highly consequential as actors made their predictions about the future.
The military dominating the transitional government in 1974 had the author-
ity and coercive capacity to make good on its promises. If the military men at
the head of the transitional government backed fair elections and democracy,
the people had a high probability of getting both. Moreover, a democracy

Table 1. Portugal’s Income Inequality in Comparative Perspective Circa 1970
Country Gini coefficient Income share top 20%
Portugal .40 49.1

Argentina .41 50.3
Chile .43 51.4
Uruguay .40 47.4
Source: Muller (1998, p. 138).
backed by the military was more likely to be an “iterative game”: The costs
of losing a first election could be offset by the prospects of winning in the
future. Political actors during the First Republic faced much more
uncertainty.
At the abstract level, the 1974 coup was surely a “critical juncture” for
Portugal: The revolutionaries faced a vast range of options, their choices
were momentous, and their time frame was short (Capoccia & Kelemen,
2007). In their first programmatic communication, the revolutionaries promised
that direct elections for a Constitutional Assembly would take place “within
a year” (de Almeida, 1978, pp. 366-368). The choice to hold the promised
elections and to ensure their fairness forged what a participant observer aptly
described as the “genetic code” of the future regime (Ferreira, 1983, p. 47).
The “momentous choice” the transitional elite made at the critical juncture
in 1974 set their regime on a path the leaders of the First Republic never took.
By creating credible electoral institutions, Portugal’s new elites gained a
level of popular legitimacy that the Republicans and their supporters never
attained. The Republicans’ choice to cheat undermined their right to rule. In
this case as in others, fraud robbed the victors of legitimation (Molina &
Lehoucq, 1999, p. 234).
The electoral institutions formed in 1974-1975 contrasted with those of the
First Republic from their very inception. The commission that the Movimento
das Forças Armadas (MFA) appointed to organize the elections produced
what a leading professor of constitutional law deemed “an extremely advanced
system of electoral legislation” (Moreira, 1999, p. 59) with no literacy require-
ment, proportional representation, and a 5% threshold. Multiparty election
monitors would supervise balloting, replacing the corrupt procedures of the
past. The press reported that the electoral law was accepted with little debate
among either the political elite or the public (Leão, 1974, p. 20).
Debate on other issues raged, however, as officers and civilians of varied
ideologies struggled to gain control of emerging institutions. Popular and
military mobilizations put the elections themselves in jeopardy, and a total of

Bermeo 1133
five provisional governments formed before the elections took place. In the
end, however, fair elections were held precisely as promised, within a year of
the coup. No one contested the results (Hammond, 1984, p. 269). What
explains this outcome?
Portugal’s fair elections cannot be explained with an argument focused on
calculations by the rich facing low material risks and a moderate median
voter. The 1975 elections took place when radicalization was high and
assaults on property were rampant. They did not occur when they would be least
threatening to the rich, as one might expect. They occurred less than 2 months
after a radical provisional government had nationalized banks, insurance
companies, and the commanding heights of the economy, when most of the
nation’s newspapers had come under state control and when large sectors of
the nation’s agricultural estates were being seized by landless workers. In
1975, Portugal’s upper class was under siege and its middle class was facing
not simply a “threat” of revolution but a revolution per se, yet the rich and
the middle class did not align to block either fair elections or democratiza-
tion. To understand this outcome we have to analyze all the key actors in the
Portuguese drama and the perceptions that shaped their institutional prefer-
ences. Here, as before, the outcome was determined largely by the calcula-
tions of party and military elites.
The calculations of both groups were greatly affected by the generalized
expectations of fair elections among Portuguese of all classes. Two factors
contributed to high expectations. The first came from massive emigration to
liberal democracies—an experience directly or indirectly shared by an esti-
mated one third of the Portuguese population. Emigration to Western Europe
and the United States enabled Portuguese citizens to see contemporary democ-
racies firsthand and to expect democratic elections at home (Brettell, 1984).
Historical experience within Portugal itself also had profound effects on
popular expectations. The dictatorship’s unfair elections had long served as a
focal point for both communist and noncommunist opposition activities. In
the 1940s, the opposition worked together across party lines to form the
Movement for Democratic Unity. Movement candidates contested the unfair
elections on a regular basis, and in 1958 its presidential candidate might even
have won the presidency had the dictatorship not manipulated balloting. In
1973, on the eve of the revolution, opposition groups managed to unite to
fight the dictatorship at the ballot box once again, and Captain Melo Antunes,
the author of the political components of the MFA program, was part of this
alliance. The fact that he included the promise of free elections in the MFA’s
first substantive address to the nation was the legacy of decades of struggle
for fair balloting. The opposition to dictatorship had literally defined itself in

opposition to unfair elections (Fernandes, 2006; Martins, 1969), and this

made public expectations of fair elections after the regime change extremely
high. As a participant politician observed, “For dozens of years, a strong
ideological current of society grew up around the desire to exercise democratic
liberties and this had an autonomous role in the development of the pre-
constitutional political process” (Ferreira, 1983, p. 212). Tampering with the
elections would have been an “affront to the Portuguese people” (Maxwell,
1995, p. 112) with enormous reputational costs.
Cheating on election day would have required the complicity of either the
military or political parties, and neither set of actors found it in their organi-
zational interests to either stop the elections or to cheat. Facing a transitional
government dominated by the military, the noncommunist political parties
backed the elections because votes were the only currency that they could
exchange for power. They also backed the elections because they were opti-
mistic about securing a meaningful number of seats in the assembly. Mário
Soares, the leader of the Socialist Party (PS) was confident of a strong showing
from the earliest days of the provisional governments when he saw his party’s
membership jump from 3,000 to 150,000 in a single month. Though Soares
and his party were constantly challenged from actors on their left, the leaders
of the major socialist parties in Europe assured him that victory was within
his grasp (Avillez, 1996, pp. 276-279). On the eve of the elections, he was
confident enough to insist, in a conversation with Vasco Gonçalves, that the
socialists would win (Avillez, 1996, p. 423). The leaders of the Popular Dem-
ocrats (PPD) were not as optimistic but still confident. They had approxi-
mately 100,000 members on the eve of the election, and though they
apparently did not expect a plurality (as the PS did), they projected a solid
second place position far ahead of the communists (de Abreu, 1975, pp. 549-
551). The leaders of the moderate right-wing CDS (Democratic and Social
Center) had the least information about their popular support because violent
attacks prevented the party from holding rallies in much of the country. As
one explained, “We don’t have grand rallies or public demonstrations because
we know that our members and sympathizers will be victims of aggression”
(de Abreu, 1975, p. 458). Yet the CDS, too, imagined that it would fare respect-
ably in an open contest. The party’s leader, Freitas do Amaral, expected that
the party would win 10% to 12% of the vote, whereas his more optimistic
colleagues expected the party to win 15% (do Amaral, 1995 p. 372).
Portugal’s public opinion polling was in its infancy in this period. Fearing
polarization, the government decreed that no poll results could be reported by
any public media for the 2-month period prior to the balloting. Parties reportedly
commissioned their own polls, but these were thought to lack rigor. To add

Bermeo 1135
to the uncertainty, the results of the first national, nonpartisan, election poll
were stolen during the tumult following a failed coup in March. This poll,
conducted late in 1974, concluded that the PS had the most support and that
the PPD and Communist Party (PCP) held second and third positions, respec-
tively, but it also found that nearly 50% of all respondents were undecided
(Bacalhau, 1994, pp. 162-163). This uncertainty worked in the interests of
the fledgling democracy because it allowed political parties to speculate longer
about their own public support. Poor information allowed noncommunist
party leaders to convince themselves that fair elections were sure to advance
both their party’s interests and democracy.
Noncommunist parties also backed elections because democratic actors
abroad were offering them the material resources to do so. Parties during the
First Republic had no such patrons, but in the 1970s European governments
and party foundations actively sought to keep Portugal’s democratization
project alive. Funds poured in from throughout Europe. West German party
foundations were especially engaged, as was the Socialist International.
Because these organizations were still reeling from the tragedy of Allende’s
Chile, Portugal’s promised experiment with democratic socialism took on a
level of extraordinary symbolic importance. The U.S. ambassador, drawing a
parallel with postwar Italy, furnished covert support as well.
The PCP and the ultra left never fully backed fair elections, but they came
to tolerate them. The communist and ultra-left parties made this choice largely
because they became convinced that the outcome of the elections could be
framed in a manner that would serve their organizational interests. Illusion
and the creation of undemocratic safeguards made these choices possible.
The mostly young, middle-class, urban residents who dominated the ultra
left sought to establish a regime based on “popular power.” Their leader, a
charismatic major named Otelo Saraiva de Carvalho, had been given com-
mand of the military region that included Lisbon, and their early triumphs
with mobilization in the capital gave them the impression that they had a
broad urban base. Outside the capital, ultra-left initiatives to mobilize the
citizenry on a whole range of issues related to the revolution were sometimes
received with hostility and often ignored. Drawing on a suspicion of the rural
poor that dated from the First Republic, ultra-left leaders mistook disinterest
for ignorance and indecision and began a campaign for blank balloting based
on this misperception. They argued that voters who were either undecided or
supporters of the MFA should vote blank and came to believe that some 50%
of the electorate might follow their lead (Maxwell, 1995, p. 113). If voters
acted as they predicted, the elections could be framed as a rejection of liberal
democracy and thereby advance the ultra-left project. With leading figures

from the MFA promoting the blank ballot initiative, the ultra-left was opti-
mistic and let the elections run their course (de Carvalho, 1991, p. 126;
Rezola, 2006, p. 180).
The PCP tolerated fair elections for a variety of reasons. Its Soviet advisers
had issued guidelines drawn from the recent coup in Chile that the party
should focus “on a tactical alliance with the Armed Forces” and “on securing
control of local government” (Ferreira, 1983, p. 98). When the MFA decided
to back fair elections, the party had strong incentives to offer at least passive
support. The goal of “securing control of local government” required support
for elections too. Because the Soviets showed no willingness to support
armed struggle inside a NATO country, elections were the only route to local
power (Graham, 1984, p. 251).
The PCP also supported elections because at least some of its leaders came
to believe that the party would perform well at the polls. They guessed that
they would have strong support in Lisbon (where most of the votes were con-
centrated) and in the rural south. The party’s key MFA ally, Vasco Gon-
çalves, predicted that the election would be won by the MDP/CDE Portuguese
Democratic Movement/Democratic Elections Committee, which the Com-
munists controlled) and that the PCP would come in second (Gomes & Cas-
tanheira, 2006, p. 74). If this were true, fair elections might be worth the risk.
But what if this prediction were incorrect? A nondemocratic institutional
safeguard was forged to make the risks of elections tolerable to all. The safe-
guard took the form of a pact between the MFA and all the major political
parties. It was forged just 2 weeks before the actual balloting took place and
was fundamental to securing not simply the support of the PCP but also the
support of elements of the MFA. The MFA–party pact of April 11 guaranteed
that the military would play a role in governing Portugal for at least 3 years.
It relegated the provisional government to a subordinate position under the
Council of the Revolution and made the choice of president a joint decision
between the MFA Assembly and any future National Assembly created by
the Constitution (Maxwell, 1995, p. 112). Guaranteeing the MFA’s continuing
role in government, the pact diluted the meaning of the elections to come.
Portugal would have fair elections, but the assembly they produced would
not be a fully sovereign civilian authority.
It is very unlikely that fair elections would have been tolerated by such a
broad range of actors without this guarantee. The hegemonic forces in the mili-
tary were too uncertain about the voters and about the fate of the MFA as an
institution to move ahead without this safeguard. A victory of the Right or even
the Center might mean not simply a rollback of revolutionary achievements

Bermeo 1137
but a counterpurge of the military. The fact that the elections were for a Con-
stitutional Assembly was little consolation. During the First Republic, the
Constitutional Assembly had simply turned itself into a national assembly and
then passed legislation deleterious to the military—including a law depriving
military personnel of the right to vote.
The MFA was the obvious veto player in the drama leading up to Portugal’s
founding elections, and its leaders backed fair elections only when they had
the assurance that fair balloting would not compromise either their revolu-
tionary achievements or their organizational interests. But it is important to
underscore the fact that the MFA was not a singular actor. Its membership
was far from homogeneous, and thus its organizational interests were hotly
contested. Leftist forces sympathetic to the PCP had control of the most power-
ful MFA institutions as the elections approached, but more moderate actors
were numerically dominant in the MFA as a whole. Indeed, moderates had
succeeded in winning several key military elections in January and February
of 1975 (Graham, 1984, p. 241). Military moderates backed fair elections
too, and though their incentives for doing so overlapped with those of the
radicals in one sense, they differed dramatically in another.
The moderates and the radicals overlapped in that they both saw fair elections
as a means for the military to maintain popular legitimacy. The fact that the
Armed Forces had been an instrument of the dictatorship weighed heavily on
the minds of military elites in both camps (Rodrigues, Borges, & Cardoso,
1974, p. 249). As a result, the MFA exhibited a “constant preoccupation with
maintaining the support of the people” (Ferreira, 1983, p. 212). President
Costa Gomes argued against postponing the elections because the military’s
promise was “a question of honor” (do Amaral, 1995, p. 358). This argument
won the day precisely because the popular legitimacy of the military as an
institution was valued across the MFA’s political spectrum.
The MFA used the elections as a means of distinguishing their military
intervention from the despised military interventions taking place in Chile and
elsewhere. As Melo Antunes, the author of the MFA’s political program put
it, “Our idea was to overthrow one regime and establish another . . . another—
and this had to be clear—on a western [democratic] model” (Avillez, 1994,
p. 14). Costa Gomes had also made an unambiguous commitment to holding
elections from the earliest days of the first provisional government (Avillez,
1996, p. 278). Though mainstream political party leaders were convinced
that MFA leader Vasco Gonçalves wanted to scotch the elections after a cen-
trist countercoup in March, he too endorsed fair elections in the end. Having
said, from the early days of the revolution, that the military “staked its honor”
on free balloting, an about-face would have been extremely hard to

rationalize (Soares, 1974, p. 18). The words he used to justify the military’s
position on the eve of the elections were especially significant. He stated
clearly, “We will have the elections, despite the risks they may bring, because
they will demonstrate that we are true to our word” (de Abreu, 1975, p. 497).
Demonstrating that the military could now be trusted was essential to the
future of the military as an institution.
Although military radicals agreed to elections despite what they might
bring, moderates actively sought elections because they would enable the
armed forces to at least partially disengage from politics and therefore restore
the integrity of the military as an institution. The dangers of “partisanship
and fragmentation” had grown more disruptive over time (Carrilho, 1994,
p. 49). A substantial minority, if not a majority, of officers found active politics
“alien” and associated a politicized military with the “infighting and govern-
mental instability characteristic of the [First] Republic” (Graham, 1984, p. 233).
Infighting had weakened the military as an institution by disrupting channels
for promotion and job security. As moderate officers were transferred,
bypassed for promotion, or even jailed, the number of actors seeking a less
politicized role for the military grew (Carrilho, 1994, p. 41; Porch, 1977, p. 236).
When the Council of the Revolution decreed on March 21 that “any officers
who did not support the principles espoused by the MFA would be placed on
reserve” (Costa Pinto, 2008, p. 315), the future for moderates looked omi-
nous indeed.
The election of a civilian Constitutional Assembly would mean the creation
of an alternative decision-making structure that would enable the military to
focus more attention on professional concerns. As one observer put it, “The
Armed Forces needed the elections for the Constitutional Assembly” to
restructure its own institutions (Ferreira, 1983, pp. 212-213). Noncommunist
officers were not fearful that the elections would be won by the reactionary
Right. This may have been because they had more accurate information on
the Portuguese electorate. Many of these officers came from the countryside
themselves, and many were the sons of the small farmers the communists
feared most (Rodrigues et al., 1974, pp. 343-347). Moreover, the vast majority
of moderates were stationed outside of Lisbon, and this too gave them more
constant contact with the core of Portuguese society (Porch, 1997, p. 195).
Noncommunist officers predicted that the Center-Left and Center-Right
would win the elections and thus that elections were a low-risk means of
advancing their institutional interests. With a critical mass of military officers
backing them, the elections proceeded as planned.
The outcome of the April 25 elections proved surprising to many of the
actors involved. MFA leaders calculated that no more than 80% of the

Bermeo 1139
electorate would vote and that at least 40% of all voters would cast blank
ballots (Porch, 1977, p. 181). In fact, nearly 92% of the electorate exercised
their right to vote and only 7% cast blank ballots. The PS triumphed with
nearly 38% of the vote whereas the PPD won more than 26%. The Commu-
nists polled only 12.5% of the vote whereas their ally, the MDP/CDE, polled
only 4.1%. The CDS polled 7.6%, confirming that even the moderate Right
was not particularly popular.
The elections allowed the major noncommunist parties to demonstrate
beyond doubt that they were, in fact, cross-class institutions. The PS drew
voters from across the class spectrum and from all regions. The PPD drew
support from the professional classes, the progressive business interests, and
the more traditional business elites and peasantry of the interior. The CDS
garnered the support of religious Catholics of various classes, northern small-
holders, and segments of the urban bourgeoisie (Goldey, 1983). The cross-
class nature of party support confirmed estimates made before the electoral
campaign began (Bacalhau, 1994, p. 162). Neither class nor religiosity
proved as divisive as some predicted. The Catholic Church allied unambiguously
with democratic forces and urged people not only to shun violence but also
to vote their conscience (Salgado de Matos, 2001).
The contrast with the founding elections of the First Republic could not
have been starker. Within months, other elements of Portugal’s democratic
project emerged and forecast a dramatic contrast in the outcome of the
democratization project as a whole. An attempted left-wing military coup in
November 1975 led moderate officers to seize control of key MFA institu-
tions and begin a concerted effort to cede political power to the civilian lead-
ers who had won the fair elections.
Significantly, these moderate officers made no attempt to reverse the massive
redistribution of property that had taken place in the months preceding their
seizure of power. This is precisely because they did not act on behalf of the
rich or any particular class. The Portuguese military had changed dramatically
since the tumultuous days of the First Republic. Its long experience fighting
the colonial wars and its membership in NATO had turned it into a highly
professionalized organization. Its officer corps was a reflection of main-
stream Portuguese society. Its links with the aristocracy or with the landed
elite had faded. Contradicting what class-based theories of democratization
would predict, the military did not serve the interests of the rich. It was
divided ideologically. Yet, ironically, its divisions on the left–right spectrum
in the 1970s were not nearly as consequential as divisions over religion and
the monarchy decades before. Discounting its left–right ideological divisions,
the Portuguese military acted primarily in defense of its own organizational

interests—inspired by the conviction that the Armed Forces and the nation
would be best served by a Western-style pluralist democracy. Despite mas-
sive property transfers, Portugal’s fledgling democratic institutions endured.
Implications for Democratization Theory

This analysis suggests lessons that may apply more generally. The first con-
cerns the conditions under which transitional governments will choose to
have fair elections. Our contrasting cases suggest that the choice will be
made when a critical mass of transition elites come to believe that fair elections
will either protect or enhance their group’s organizational interests. Ideology
will shape their beliefs, but their predictions will be based on historical situations
deemed parallel, on assessments of popular support, on resource incentives,
and on the weight of the elections in the overall distribution of power.
In its basic logic, this answer to the puzzle of fair elections is in keeping
with the rational choice perspective that informs so much of our contemporary
work on democratization. However, the comparison as a whole highlights
four additional cautionary lessons.
The first lesson is that we should be cautious about exaggerating the relative
power of the rich. Our latest economic theories of democracy assume that the
rich make the choice among democracy, partial democracy, and repression
(Acemoglu & Robinson, 2006). But history shows us that the rich often lack
the coercive power to make this choice. The rich are certainly powerful but
are neither all-powerful nor powerful at all times. No class, regardless of its
economic or numerical strength, can choose to change a political regime
without the implicit or explicit support of the actors who dominate the coercive
apparatus of the state.
To the extent that we can make assumptions about the choices of a single
class at all, the empirics of the Portuguese case suggest that the regime
choices of the rich are realized only if they coincide with the choices made
by hegemonic groups within the military. In both 1910 and 1974, Portugal’s
rich might have chosen “repression,” but they lacked the capacity to be
repressive as long as the armed forces failed to choose the repression option
as well. During the First Republic, the military’s regime choice was initially
unclear. Different factions of the armed forces formed coalitions with differ-
ent civilian groups. But the tumult lasted a full 15 years and killed more than
4,000 people precisely because no single class had the capacity to impose its
choice on the military (Wheeler, 1978a, p. 870). The capacity to “choose”
repression lay, in the end, with the armed forces. In 1974, the military was

Bermeo 1141
key once again. Despite the fact that property relations were challenged in
unambiguous ways, Portugal’s rich could not “choose” repression because
hegemonic forces in the military had chosen democracy.
The rich could not even “choose” to cheat in the elections, as they had
done historically in Portugal and elsewhere (Posada-Carbo, 1996). Wealth
can subvert formal democratic structures through “pre-existing asymmetries
of resources” (Ziblatt, 2009, p. 3), but asymmetries can be eroded. In Portu-
gal, land seizures undermined the foundation of informal clientelism (Ber-
meo, 1986) while access to formal positions in electoral administration was
blocked: The rich were not able to capture local electoral institutions because
elites in the transitional governments had made the purging of local electoral
institutions a top priority. They had learned from Portugal’s past and changed
local procedures and officials before the elections were held. We do not
know if Portugal’s landed elites even wanted fraudulent elections in 1975,
but we know they lacked the capacity to engage in significant electoral fraud.
The rich depend on formal and informal institutions to wield their power, and
these institutions can be shaped and controlled by others.
A second lesson emanating from the case studies concerns how the long
drama of democratization is affected by the connections between parties and
classes. Classes are profoundly relevant to democratization, but their role in
either preventing or supporting a democratic regime change is inevitably
mediated by perceptions of party landscapes. Modern democracy is “unthink-
able” without political parties (Schattschneider, 1942, p. 1), and thus class (or
other social) actors who weigh the costs of democracy inevitably do so with
thoughts of real and imagined political parties in mind. What sorts of parties
will win democratic elections? What sorts of nonelectoral institutions will
affect party capacity? Even if we believe that economic interests are the
prime determinants of political preferences when making the choice for, or
against, democracy, these questions must inevitably be asked. Yet their
answers will be determined by historical and cultural factors that purely
class-based arguments cannot embrace. Class structures rarely map on to
party systems in any straightforward way. Parties are often cross-class orga-
nizations, and some sort of party system is usually in place (or emerging)
when class actors weigh the merits of any particular regime choice. These
two facts mean that class actors do not evaluate the merits of democracy or
partial democracy in a vacuum but in the context of real-world institutional
landscapes shaped by religious, regional, and ideational factors that may be
autonomous from class structures. Theories of democracy that neglect the
role of parties and focus on classes and economic structures alone are likely

to mislead. A focus on class calculations alone might have led us to predict

unfair elections in 1975. Powerful redistributive threats should have meant a
choice to cheat or to scotch the elections altogether.
A third lesson suggested by this analysis is that we should be wary of
exaggerating the role of declining economic inequality in explaining regime
change. How historical inequality trends should be measured and which
trends matter for regime change are subjects of great debate. For Portugal,
the Theil index suggests that income inequality actually rose between the
1920s and the 1970s (Rafecas, 2009, p. 20). But even assuming that overall
inequality decreased between 1910 and 1975, there is little evidence that this
change accounted for either the fair elections or the democratization that fol-
lowed. On the contrary, within Portugal, references to decreased inequality
were absent from political discourse in 1974-1975. Political actors emphasized
continuities in inequality instead. These perceived continuities were the
foundation of communist and far-Left skepticism about the political sophis-
tication of rural voters. Extremes in inequality were also the rationale for the
profoundly redistributive platforms of both the PS and the PPD. The fact that
both these parties sought membership in the Socialist International indicates
how salient perceptions of economic inequalities were. We can speculate that
democratization is triggered by declining levels of inequality, but if real-world
political actors perceive different trends, scholarly attempts to attribute cau-
sality to inequality seem risky, at best.
Portugal had what social scientists might deem a “middle level of inequality”
at the time of its regime change. Its Gini coefficient for income in 1970 was
.40, putting it squarely in the medium range (Boix, 2003, p. 215). In isolation,
this fact bode well for democratic regime change because intermediate levels
of inequality are supposed to generate the choice behaviors that make democ-
racy most likely. As Acemoglu and Robinson (2006) forecast, “Democracy
has the best chance to emerge in societies with middle levels of inequal-
ity . . . the citizens are not totally satisfied and the elites are not so averse” (p. 37).
This logic is compelling in the abstract. Yet the people of Portugal were
either unaware of or unmoved by their position on our abstract inequality
scale. As a result, the nature of the redistributive conflicts they generated
bore no resemblance to the model’s prediction. A middle level of inequality
did not prevent a serious threat to property, and fair elections emerged despite
serious threats to the material interests of the upper classes. A middle level of
inequality coincided with successful democratization, but this happened
despite material conflicts that the models do not anticipate. The outcome
occurred because economic actors and economic considerations were not
paramount.

Bermeo 1143
The final lesson our historical comparison offers us concerns the critical
importance of illusion in the crafting of democratic institutions. As Hanson
(2010) and Ertman (2010) illustrate, political actors at critical junctures are
often mistaken about the effects of institutional reform. The illusions of politi-
cal actors merit close attention because the institutions they shape endure long
after the illusions themselves. In Portugal, the institutional arrangements and
the social coalitions that supported fair elections outlasted Portugal’s critical
juncture. Yet the arrangements would not have been possible without the illu-
sions of military and especially party elites. Here, once again, the historical turn
brings us insight into how the institutional preferences of key actors are shaped.
Most of the parties that created the electoral institutions of Portugal in
1974-1975 were new. Their lack of experience enabled them to fantasize
about their popularity—and to imagine either an immediate victory or a victory
in the near future. The PCP had a longer history, and its experience contributed
to its wariness about elections. In the end, however, the heroic status the
party had attained because of its opposition to the dictatorship gave at least
some of its leaders and military allies the illusion of broad support. These
illusions helped make fair elections possible. In 1910, the Republican Party
had no illusions regarding its popularity. They knew that fair elections meant
a loss of power, and they crafted unfair institutions as a result.
This exploratory essay focused on only two cases. Whether they are rep-
resentative of other episodes in the history of democratic institution making
remains to be explored, but the assembled evidence strongly suggests the fol-
lowing conclusion: Structural changes may generate demands for democracy,
but the establishment of individual democratic institutions depends on values
and illusions as well as structural realities, and thus on factors and actors that
many rationalist-economic paradigms do not embrace.
Acknowledgments
The author thanks Jorge Miguel Alves Fernandes for research assistance and Nuffield
College, Oxford University for research support.
Declaration of Conflicting Interests

The author declared no potential conflicts of interests with respect to the authorship
and/or publication of this article.
Financial Disclosure/Funding
The author received financial support for research from Nuffield College at Oxford
University.

References
Acemoglu, D., & Robinson, J. A. (2006). The economic origins of dictatorship and
democracy. New York, NY: Cambridge University Press.
Avillez, M. J. (1994). Do fundo da revolução. Lisbon, Portugal: Público.
Avillez, M. J. (1996). Soares: Democracia. Lisbon, Portugal: Público.
Bacalhau, M. (1994). Atitudes, opiniões e comportamentos políticos dos portugueses,
1973-1993. Lisbon, Portugal: Ediçoes Heptagono.
Bermeo, N. (1986). The revolution within the revolution: Workers’ control in rural
Portugal. Princeton, NJ: Princeton University Press.
Bobone, C. (2009). As primeiras eleiçooes da República vistas de perto. Retrieved
from http://www.centenariodarepublica.org/centenario/2009/09/04/do-liris
Boix, C. (2003). Democracy and redistribution. New York, NY: Cambridge University
Press.
Brettell, C. (1984). Emigration and its implications for the revolution In L. S. Graham
& H. Makler (Eds.), Contemporary Portugal (pp. 281-298). Madison:
University of Wisconsin Press.
Cabral, M. V. (1977). O operariado nas vesperas da implantaçao da República.
Análise Social, 7(50), 419-448.
Capoccia, G., & Kelemen, R. D. (2007). The study of critical junctures. World Politics,
59, 341-369.
Capoccia, G., & Ziblatt, D. (2010). The historical turn in democratization studies:
A new research agenda for Europe and beyond. Comparative Political Studies,
43, 931-968.
Carrilho, M. (1985). Forças armadas e mudança política em Portugal. Lisbon, Portugal:
INCM.
Carrilho, M. (1994). Democracia e defesa. Lisbon, Portugal: Dom Quixote.
Coimbra, A. F. (2000). Paiva Couçeiro e a contra-revolução monarquica. Braga,
Portugal: Universidade do Minho.
Collier, D., & Levitsky, S. (1997). Democracy with adjectives: Conceptual innovation
in comparative research. World Politics, 49, 430-451.
Costa Pinto, A. (1995). Salazar’s dictatorship and European fascism problems of inter-
pretation. New York, NY: Columbia University Press, Social Science Monographs.
Costa Pinto, A. (2008). Political purges and state crisis in Portugal’s transition to
democracy, 1975-76. Journal of Contemporary History, 43, 305-332.
Costa Pinto, A., & Tavares de Almeida, P. (2000). On liberalism and the emergence
of civil society in Portugal. In N. Bermeo & P. G. Nord (Eds.), Civil society before
democracy: Lessons from 19th century Europe (pp. 3-22). Lanham, MD: Rowman
& Littlefield.

Bermeo 1145
Coutinho, L. P. (2009). Historia do Partido Republicano Português. Retrieved from

http://www.centenariodarepublica.org/centenario/2009/09/04/do-liris
de Abreu, D. (1975). Eleições em abril, diário da campanha. Amadora: Liber.
de Almeida, D. (1978). Ascensão, apogeu e queda do M.F.A. Lisbon, Portugal:
Edições Sociais.
de Carvalho, O. S. (1991). Alvorada em Abril vol 1. Lisbon, Portugal: Alfa.
do Amaral, D. F. (1995). Antigo regime e a revolução: Memórias políticas (1941-1975).
Lisbon, Portugal: Bertrand.
Elkit, J., & Svensson, P. (1997). What makes elections free and fair? Journal of
Democracy, 8(3), 32-46.
Ertman, T. (2010). The Great Reform Act of 1832 and British democratization.
Comparative Political Studies, 43, 1000-1022.
Fernandes, T. (2006). Nem ditadura, nem revolução. Lisbon, Portugal: Dom Quixote.
Ferreira, J. M. (1983). Ensaio histórico sobre a revolução do 25 de abril. Lisbon,
Portugal: INCM.
Ferreira, J. M. (1992). Comportamento político dos militares Lisbon, Portugal:
Estampa.
Goldey, D. B. (1983). Elections and the consolidation of Portuguese democracy:
1974–1983. Electoral Studies, 2, 229-240.
Gomes, A., & Castanheira, J. P. (2006). Os dias loucos do PREC. Lisbon, Portugal:
Edimpresa.
Graham, L. (1984). The military in politics: The politicization of the Portuguese
armed forces. In L. S. Graham & H. Makler (Eds.), Contemporary Portugal
(pp. 221-256). Madison: University of Wisconsin Press.
Hammond, J. L. (1984). Electoral behavior and political militancy. In L. S. Graham
& H. Makler (Eds.), Contemporary Portugal (pp. 257-280). Madison: University
of Wisconsin Press.
Hanson, S. E. (2010). The founding of the French Third Republic. Comparative Polit-
ical Studies, 43, 1023-1058.
Leão, M. (1974, September 28). Lei eleitoral: A discussão que não houve. Expresso, p. 20.
Levitsky, S., & Way, L. (in press). Competitive authoritarianism: The origin and
evolution of hybrid regimes. New York, NY: Cambridge University Press.
Lloyd-Jones, S. (2003). The slow death of the First Republic. Portuguese Studies
Review, 10, 81-100.
Maddison, A. (1995). Monitoring the world economy 1820-1992. Paris, France:
Organisation for Economic Co-operation and Development.
Marques, A. H. O. (1975). A primeira república Portuguesa. Lisbon, Portugal: Livros
Horizonte.
Martins, H. (1969). Opposition in Portugal. Government and Opposition, 4(2),
250-263.

Maxwell, K. (1995). The making of Portuguese democracy. New York, NY: Cambridge
University Press.
Meneses, F. R., de. (2004). Portugal 1914-1926. Bristol, UK: University of Bristol.
Molina, I., & Lehoucq, F. E. (1999). Political competition and electoral fraud: A Latin
American case study. Journal of Interdisciplinary History, 30, 199-236.
Moreira, V. (1999). A Instituição da democracia. In F. Rosas (Ed.), Portugal e a tran-
sição a democracia (pp. 21-59). Lisbon, Portugal: Edições Colibri.
Muller, E. N. (1998). Economic determinants of democracy. In M. Midlarsky
(Ed.), Inequality, democracy and development (pp. 133-155). New York, NY:
Cambridge University Press.
Pastor, R. A. (1998). Mediating elections. Journal of Democracy, 9(1), 154-163.
Porch, D. (1977). The Portuguese armed forces and the revolution. Stanford, CA:
Hoover Institution Press.
Posada Carbó, E. (1996). Elections before democracy: The history of elections in
Europe and Latin America. London: Macmillan.
Rafecas, J. G. (2009). The evolution of wage inequality in Spain and Portugal over
the twentieth century. Barcelona, Spain: Universitat de Barcelona, Instituto de
Ciências Sociais.
Ramos, R. (2004). Para uma política da cidadania. Análise Social, 39(172), 547-569.
Rezola, M. I. (2006). Militares na revolução de Abril. Lisbon, Portugal: Campo da
Comunicação.
Robinson, R. A. H. (1977). The religious question and the Catholic revival in Portugal,
1900–30. Journal of Contemporary History, 12, 345-362.
Rodrigues, A., Borges, C., & Cardoso, M. (1974). O movimento dos capitães. Lisbon,
Portugal: Moraes.
Salgado de Matos, L. (2001). A igreja na revolução em Portugal (1974-1982). In J. M.
Brandão de Brito (Ed.), O Pais em Revolução. Lisbon, Portugal: Editorial.
Schattschneider, E. E. (1942). Party government. New York, NY: Farrar and Rinehart.
Schmitter, P. (1978). Elections in authoritarian Portugal, 1933-1974. In G. Hermet,
R. Rose, & A. Rouquie (Eds.), Elections without choice (pp. 145-168). New York,
NY: John Wiley.
Schwartzman, K. C. (1989). The social origins of democratic collapse: The First
Portuguese Republic in the global economy. Topeka: University Press of Kansas.
Seawright, J., & Gerring, J. (2008). Case selection techniques in case study research.
Political Research Quarterly, 61, 299-300.
Soares, M. (1974, October 12). Perguntas. Expresso, pp. 18-19.
Valente, V. P. (1992). Revoluções: A república velha. Análise Social, 115(1), 7-63.
Valente, V. P. (1997). Ed. A República Velha 1910-1917. Lisbon, Portugal: Gradiva.
Vanhanen, T. (1997). Prospects for democracy. London, UK: Routledge.

Bermeo 1147
Wheeler, D. (1978a). A Primeira República Portuguesa. Análise Social, 14(56),

865-872.
Wheeler, D. (1978b). Republican Portugal. Madison: University of Wisconsin Press.
Ziblatt, D. (2006). How did Europe democratize? World Politics, 58, 311-338.
Ziblatt, D. (2009). Shaping democratic practice and the causes of electoral fraud: The
case of nineteenth-century Germany. American Political Science Review, 103(1),
1-21.
Bio
Nancy Bermeo taught at Princeton University until 2007 and now holds the Nuffield
Chair in Comparative Politics at the University of Oxford, where she directs the
Centre for the Study of Inequality and Democracy. She is the author of Ordinary
People in Extraordinary Times, which won the best book award from the Comparative
Democratization Section of the American Political Science Association, and many
other works on the causes and consequences of both regime change and inequality.

Comparative Political Studies: Interests, Inequality, and Illusion in The Choice For Fair Elections

Uploaded by

Copyright:

Available Formats

Comparative Political Studies: Interests, Inequality, and Illusion in The Choice For Fair Elections

Uploaded by

Document Information

Original Title

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

Comparative Political Studies: Interests, Inequality, and Illusion in The Choice For Fair Elections

Uploaded by

Copyright:

Available Formats

Comparative Political Studies

Interests, Inequality, and Illusion in the Choice for Fair Elections

The online version of this article can be found at:

Email Alerts: http://cps.sagepub.com/cgi/alerts

>> Version of Record - Jun 28, 2010

Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

The Cases and the Argument

Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

1911 High Low Unfair Elections

1975 Moderate Moderate Fair Elections

Founding Economic Asset Incentives for Fair

Figure 1. Analytical approaches

Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

The Argument and the Democratization Literature

Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

across the political spectrum used “past experiences of successful or failed

Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

The First Republic and the Choice to Cheat

Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

anticlerical action taken by a government in Europe up to that time” (Wheeler,

Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

regime with low legitimacy had caused a level of professional dissatisfaction

Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

With classes divided along cultural lines, serious redistributive threats

The Choice for Fair Elections

Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

Table 1. Portugal’s Income Inequality in Comparative Perspective Circa 1970

Country Gini coefficient Income share top 20%

Portugal .40 49.1

Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

opposition to unfair elections (Fernandes, 2006; Martins, 1969), and this

Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

Implications for Democratization Theory

Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

to mislead. A focus on class calculations alone might have led us to predict

Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

Declaration of Conflicting Interests

Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

Coutinho, L. P. (2009). Historia do Partido Republicano Português. Retrieved from

Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

Wheeler, D. (1978a). A Primeira República Portuguesa. Análise Social, 14(56),

Downloaded from cps.sagepub.com at Uni Lucian Blaga on October 26, 2011

You might also like