Implications of Slowing Growth for Poverty

Authors:
David Laborde, Senior Research Fellow, Markets, Trade and Institution Division
Will Martin, Senior Research Fellow, Markets, Trade and Institution Division
Research Support:
Tess Lallemant, Research Assistant, Markets, Trade and Institution Division

International Food Policy Research Institute

Washington DC

17 April 2016

Abstract
This paper examines the impact of the actual and projected slowdown in the world economy since
2012 on the poor and on the potential for achievement of the Social Development Goals (SDGS).
It builds on the changes between 2012 and late 2015 in the IMF’s World Economic Outlook
projections to provide the basic slowdown scenario. It then uses a global model to assess the
impacts of lower rates of productivity growth and consequent lower savings and investment on
key price and income variables. The productivity shocks are passed directly to the production
activities included in household microsimulation models for more than 300,000 households. These
households are also affected by the modeled changes in prices and wages. Simulations allow us to
assess the impacts of the slowdown on the real household incomes of the poor, and hence on
summary statistics such as the poverty rate. The results suggest that the productivity slowdown—
if continued—will slow down the reduction in the global poverty rate, increasing the importance
of policy interventions focusing directly on poverty reduction if the SDGs are to be achieved.
 TABLE of CONTENTS

1 Economic Growth: trend, drivers and projections .................................................................. 2

1.1. Developing Country Growth and Poverty: recent trends ................................................. 2
1.2. Drivers of economic growth and implications for the poor ............................................. 5
1.1 Recent projections ............................................................................................................ 7
1.1.1 GDP dynamics .......................................................................................................... 7
1.1.2 Savings Rates ............................................................................................................ 9
1.3. Current Account Dynamics ............................................................................................ 10
1.4. Commodity prices .......................................................................................................... 12
2 Methodology ......................................................................................................................... 14
2.1 Analyzing macroeconomic changes at the household level ........................................... 14
2.2 Moving to Broader Country Groupings ......................................................................... 16
2.3 Implementing alternative macroeconomic scenarios in the CGE .................................. 17
2.4 Reviewing policy options ................................................Error! Bookmark not defined.
2.5 Datasets .......................................................................................................................... 21
2.6 Regional and sectoral disaggregation ............................................................................. 22
3 Macroeconomic projections to 2030 ..................................................................................... 24
3.1 GDP projections ............................................................................................................. 24
3.2 Total Factor Productivity projections ............................................................................. 25
3.3 Real Unskilled Wages .................................................................................................... 26
4 Results ................................................................................................................................... 27
4.1 Evolution of world prices ............................................................................................... 27
4.2 Impacts on income ......................................................................................................... 28
4.3 Fiscal space for developing countries ............................................................................ 31
4.4 Impacts at the Household Level ..................................................................................... 31
5 Conclusions ........................................................................................................................... 40
6 Appendix ............................................................................................................................... 42
6.1 Country grouping ........................................................................................................... 42
6.2 Country level results....................................................................................................... 43
7 References ............................................................................................................................. 47
 LIST OF FIGURES

Figure 1. Growth rates of GDP per head in Developed and Developing Countries (3 year moving
average) ........................................................................................................................................... 3
Figure 2. Global Poverty Head Count Rate (% Population) ........................................................... 4
Figure 3 Growth rates of GDP per head (% Per year) in low income, middle income and high
income countries (3 year moving average) ..................................................................................... 5
Figure 4. Calorie consumption vs total Cereal Equivalent (CE) Consumption .............................. 6
Figure 5. Comparison of 2012 and 2015 country GDP growth projections for 2017 .................... 8
Figure 6. Comparison of 2012 and 2015 GDP growth projections for 2017 (selection of
countries)......................................................................................................................................... 9
Figure 7. World Commodity Price Projections (2015) ................................................................. 12
Figure 8. Analytical Steps ............................................................................................................. 14
Figure 9. Analytical Framework ................................................................................................... 17
Figure 10. China's Labor Force, 15-64 years ................................................................................ 21
Figure 11. Average annual growth rates for real GDP 2011-2030 ............................................... 24
Figure 12. Average annual growth rate for Total Factor Productivity between 2011 and 2030 .. 25
Figure 13. Real Unskilled Wage Rates from 2011 to 2030 .......................................................... 26
Figure 14. Global Poverty Headcount under alternative scenario, Percentage............................. 33
Figure 15. Net and gross movements into and out of poverty, Scenario 1, Percentage points,
Total Population. ........................................................................................................................... 34
Figure 16. Net and gross movements into and out of poverty, Scenario 2, Percentage points,
Total Population ............................................................................................................................ 35
Figure 17. Net and gross movements into and out of poverty, Scenario 2, Percentage points,
Farmer population ......................................................................................................................... 36
Figure 18. Relative contributions of different drivers in the change of household real income,
breakdown by poverty movement, Scenarios 1 and 2 .................................................................. 37
Figure 19. Changes in the number of poor people by country, thousands, compared to the 2030
baseline (S0).................................................................................................................................. 38
 LIST of TABLES

Table 1. Comparison of 2012 and 2015 Current Account Balance Projections (selection of
countries with larger current account imbalances) ....................................................................... 11
Table 2. Scenario Summary Table ................................................................................................ 20
Table 3. Country Nomenclature .................................................................................................... 22
Table 4. Sectoral Nomenclature .................................................................................................... 23
Table 5. World Prices - Changes in 2030 compared to the Baseline ........................................... 27
Table 6. Key Macroeconomic Results .......................................................................................... 29
Table 8. Effects of Different Poverty Lines in S0, S1, and S2 on Global Poverty Rate in 2030 . 39
Table 9. Cluster Composition (ISO Country Code)...................................................................... 42
 Implications of Slowing Growth in Emerging Market Economies for
Hunger and Poverty in Rural Areas of Developing Countries
Since the recent financial crisis, the global economy has suffered a sustained slowdown to levels
consistently below forecasts. For example, the IMF forecast for global growth in 2015 declined
from 4.8 percent in September 2011 to 3.1 percent in October 2015. Other international
organizations expect similar changes: global growth is expected to change quantitatively and
qualitatively. This report looks at the implications for poverty of the slowing growth of advanced
and emerging market economies, and particularly, how these less optimistic macroeconomic
projections may challenge the achievement of the Sustainable Development Goals (SDGs) by
2030.
The approach that we take in this project focusses on several of the key linkages between
growth in developing countries and the real incomes of the poor, and particularly the rural poor.
In particular, it considers: (i) the impacts of lower productivity growth on the incomes of the poor,
(ii) the impacts of changes in wage rates for unskilled workers on the incomes of the poor, (iii) the
impacts of changes in key commodity prices on the incomes of the poor, and (iv) the impacts of
changes in the cost of living on the incomes of the poor. By taking into account these key changes,
we aim to capture the economic impacts of the decline in global, regional and national growth
rates. We assess the impacts of changes in growth for individual households, check whether these
result in the household rising from below the poverty line to above, or from above the poverty line
to below, and use the resulting changes in the poverty headcount as an indicator of their impact on
the poor. We could use measures of the poverty gap and the poverty gap squared to provide
additional indicators of the impacts on the poor, but we face a plethora of results and have found
from experience that these measures usually respond in ways that are similar to the poverty
headcount.
In this report, we first consider the conceptual framework, focusing on the links between
growth in developing countries and the real incomes of the poor. A good understanding of the
ways in which the slowdown might affect people’s welfare is crucial to building a cohesive
analytical framework. We then describe this analytical framework: a combination of a global
computable general equilibrium model offering household disaggregation features developed by
the team (MIRAGRODEP) with a systematic household modelling approach to properly identify
short term vs long term consequences on poverty and hunger. Finally, we describe the different
growth scenarios we analyze and discuss the results.

1
1 Economic Growth: trend, drivers and projections
Changes in growth rates in developing economies can clearly have large impacts on the
welfare of poor people in rural areas. As we have seen during the past 15 years, during which many
developing economies have substantially outperformed the industrial economies, rural incomes
can rise rapidly and millions can move out of poverty. The projected slowdown in economic
growth in many developing countries is therefore a potentially serious source of concern. However,
to make any assessment of the implications of this change for the poor, we need to think very
carefully about the channels through which the slowdown might affect people’s welfare in
different countries.
In a first subsection, we examine the recent history of economic growth in developing
countries. Then we consider some key channels through which reductions in growth rates in
developing countries might be expected to affect the real incomes of the poor and near-poor, and
hence poverty rates. Then we review the most recent projections for growth over the next five
years. Finally, we turn to consider how we capture the impacts of the growth slowdown on poverty.

1.1. Developing Country Growth and Poverty: recent trends

There are good reasons to expect that poorer economies should grow faster than higher-
income economies. To the extent that the most advanced economies are generally using the best
available production techniques, they can only improve the economic efficiency and
systematically raise their incomes at the rate that the best-available technology improves. By
contrast, in poorer economies, it is possible to increase efficiency by catching up with the
economies at the frontier, as well as by taking advantage of improvements in the best available
technology. A number of initially-low-income economies have been able to increase their incomes
very rapidly in this way (Lucas 1988). Quah (1988) points to a very wide range of cases in which
lower-income countries or regions were able to catch up with their higher-income counterparts at
a rate of around 2 percent per year. Barro and Sala-i-Martin (1992) pointed to strong convergence
between US states, and Dowrick and Nguyen (1989) within the OECD set of developed
economies. Ben-David (1994) pointed to convergence between countries with strong trade links.
However, prior to the 1990s, most developing countries appeared unable to grow as rapidly
as their more-developed counterparts, with consequent widening of the distribution of income
across countries (Lucas 1988). In the 1990s, perhaps following the extensive policy reforms in
many developing countries, a pattern of higher growth in developing than in developed countries
emerged. As shown in Figure 1, with the exception of the Asian crisis period, average growth rates
of per capita incomes in developing countries have been substantially higher than in the high
income countries—with the gap in the order of 4 percentage points during much of this period.
This graph understates the outperformance of the developing countries to some degree because of
a composition effect—some middle income countries with very high growth rates move into the
high income group while poor performers initially in the high income group move into the
developing country group. The great turnaround in the performance of developing countries has,
justifiably, received enormous attention because of its importance in reducing poverty in

2
developing countries, and raising the geo-political importance of the larger developing countries
(O’Neill 2001).
Figure 1. Growth rates of GDP per head in Developed and Developing Countries (3 year moving
average)
7
GDP Growth Per Capita (%) 6
5
4
3
2
1
0
-1
-2
1973

2005
1967
1969
1971

1975
1977
1979
1981
1983
1985
1987
1989
1991
1993
1995
1997
1999
2001
2003

2007
2009
2011
2013
Year

Low and Middle Income Countries High Income Countries

Source: World Development Indicators, World Bank. Database version updated 24 September 2015

This increase in the growth rates of developing countries contributed to a dramatic

reduction in the global poverty headcount, from 44 percent of the world population in 1990 to 13
percent in 2012, the year for which the most recent estimates are available (see Figure 2).

3
Figure 2. Global Poverty Head Count Rate (% Population)

Source: World Development Indicators (World Bank) - Poverty headcount ratio at $1.90 a day (2011 PPP)
(% of population). Extracted 3/30/2016

However, as shown in Figure 3 we can see that, until the last decade, the low income countries
were lagging behind the other groups, increasing the development gap between the most
vulnerable economies and the advanced and emerging economies. In the last 10 years, the low
income countries have managed to outperform the rich countries despite the composition effect
described earlier and, since the last financial crisis, have even surpassed the growth performance
of the middle income countries giving, for the first time in decades, inclusive economic growth.
In this context, it is crucial to understand the extent to which the global slowdown may oppose or
confirm this trend, and, if convergence does occur, the extent to which its speed will be impacted.
This outperformance is all the more remarkable given the tendency for this measure to understate
the growth performance of the low-income country group as countries with the worst growth
performance fall into this group from higher-income groups and the countries with the best growth
performance rise out of the group.

4
Figure 3 Growth rates of GDP per head (% Per year) in low income, middle income and high
income countries (3 year moving average)

8
GDP Growth Per Capita (%)
6

-2

-4

-6

2002
1984

1986

1988

1990

1992

1994

1996

1998

2000

2004

2006

2008

2010

2012

2014
High Income Countries
Year Low Income Countries
Middle Income Countries

Source: World Development Indicators, World Bank.

1.2. Drivers of economic growth and implications for the poor

Sustained growth in the measure of GDP used in Figure 1 can only be achieved through
improvements in productivity. The increases in productivity tend to be associated with reforms
that allow increased access to imported intermediates (Amiti and Konings 2007) and to other
economy-wide policy changes that allow increases in productivity in a wide range of sectors. To
the extent that these productivity gains raise the productivity of the business activities of poor
households—such as farms or other small businesses—they are likely to be strong contributors to
poverty reduction. Productivity gains may also raise the returns to poor households by raising the
wage rates earned by their household members.
Other potentially-important linkages between growth in developing countries and poverty
are likely to operate through changes in relative prices. Food prices are likely to be particularly
important because of their importance in both the consumption and the incomes of many poor
people (Ivanic and Martin 2015). One linkage with potentially important impacts on poverty works
through demand for food products. While demand for calories tends to flatten out at quite low
levels of income, the demand for agricultural output continues to grow quite rapidly through the
income range of today’s middle-income countries and to level out only at the income level of
today’s highest income countries. This growth in demand for agricultural output is not for greater
quantities of the same products or for more calories, but through shifts in the consumption of food
towards products that require greater productive inputs—and particularly through a shift from
plant-based to animal-based products (Fukase and Martin 2014). The relationships between
income levels and demand for final calories and for total cereal equivalents are shown in Figure 4,

5
 with actual figures for China. The cereal equivalent measures in Figure 4 take into account the
much higher agricultural output requirements for producing products such as meat and dairy
products more heavily consumed in richer countries.
Figure 4. Calorie consumption vs total Cereal Equivalent (CE) Consumption

China Calorie Consumption Fitted Calorie Consumption

China CE Consumption Fitted CE Consumption

2
15000

(tons/capita/year)
(kcal/capita/day)

1.5
10000

1
5000

.5
0

0
0 10000 20000 30000 40000 50000
Real GDP(PPP) per capita in 2005 int. $ 1980-2009

Source: Fukase and Martin (2016)

Given the concave relationship between food demand and incomes, an average rate of
global economic growth resulting from economic converge is likely to result in a larger increase
in food demand than one resulting from higher growth in the highest-income countries. These
linkages are incorporated in multi-sectoral computable general equilibrium models through higher
income elasticities of demand for non-staple foods, and through the high intermediate use
requirements for production of products such as meat and dairy products.
Another potentially important linkage between economic growth in developing countries
and relative prices operates through the demand for investment goods in high-growth economies.
If we take Doblin’s (1991) estimate of the capital-output ratio of around 2.5 and assume a
depreciation rate of 5 percent, then maintaining the capital stock at its initial level will require an
investment rate of 12.5% of GDP. If, consistent with Kaldor’s (1957) stylized facts of economic
growth, the capital-output ratio is to stay constant, an economy growing at 5 percent per year needs
an investment rate of 25 percent of GDP and one growing at 10 percent per year needs an
investment rate of 37.5% of GDP. The dramatic increase in average growth rates in developing

6
countries in the past 15 years was associated with sharp increases in the prices of investment and
energy products, particularly in the 2008-10 period.

1.1 Recent projections

The outlook for economic growth and poverty reduction in developing countries appears
to have deteriorated in recent years. Growth in a number of major developing countries has
decelerated sharply and the outlook appears less optimistic (The Economist 2016). In this section,
we review the different inputs based on the IMF World Economic Outlook 2012 and 2015 and
identify key drivers and narratives for our scenarios. The IMF projections are probably the most
comprehensive and carefully-scrutinized set of economic projections available. They provide
estimates of growth in national GDP, savings rates, and changes in the current account balances.
The changes in current account balances, given developments in savings rates, reflect changes in
investment rates of the type considered in the previous section.
It appears from the projections that the difficulties of high income countries in rebounding
from the financial crisis and the slowdown of emerging markets, in particular the rebalancing of
the Chinese economy, will have important consequences on global growth, commodity prices and
global investment dynamics.

1.1.1 GDP dynamics

The main measurement of economic growth is the evolution of GDP at constant prices.
The current slowdown has led to significant revision of projections for future growth at national
and global levels. Since we consider no change in population dynamics, an increase of one
percentage point in real GDP will imply an identical change in GDP per capita. Two important
factors are important to keep in mind, especially when looking forward and considering the 2030
targets (for the SDGs):
 The current projections involve lower growth rates for the future (e.g. 2016-2017 growth
rates);
 The actual starting point is lower today (i.e. 2015 GDP level) than expected a few years
ago.
Therefore comparing 2012 projections for the future with 2015 projections, it is critical to
see that both the slope and the level of the GDP trajectories have been affected.
For instance, global income growth between 2011 and 2015 was expected to be 17.5% in
2012 but turned out to be only 13.9%. By 2017, the GDP growth rate gap reaches 5% (the 2017
GDP growth rate as projected in 2012 is 5% above the 2015 projection for the same year) and the
growth gap (the GDP growth not achieved) reaches 22%. A simple extrapolation shows that global
GDP in 2030 will be 15% lower than initially expected. Our objective is to assess the implications
of this substantial decline in anticipated output for achieving the SDGs.
The growth projections have been reduced for 134 countries out of a total of 189, and an
overview is displayed in the following figure for the 2017 growth rate. Looking at global GDP,

7
the growth rate was revised from 4.66% to 3.80% (minus .86 points or 19%). For the country
sample, the median growth rate moved from 4% to 3.6% (minus .4 points or 10%).
Figure 5. Comparison of 2012 and 2015 country GDP growth projections for 2017

15

13

11

y = 0.8559x
2017 Growth rate projected in 2015

9 R² = 0.3118

-5 -3 -1 -1 1 3 5 7 9 11 13 15

-3

-5
2017 Growth rate projected in 2012

Source: World Economic Outlook (2015 and 2012) - IMF

In these projections, not all countries are affected in the same way, exporters of raw
commodities and crude oil are among the most affected (Gulf countries, but also Angola,
Equatorial Guinea, Democratic Republic of Congo) but also some key emerging economies: Brazil
and first of all, China with a decline from 8.5% to 6.3%. China plays a very critical role due to of
its size, its growth and its specific impact on world markets. In advanced economies, recovery is
slow but still expected for the most open economies (see Japan). The US case is also here very
relevant since its growth rate, among the highest among developed economies, is cut by one-third.

8
Figure 6. Comparison of 2012 and 2015 GDP growth projections for 2017 (selection of countries)

9.00

8.00

7.00

6.00
GDP Growth Rate (%)

5.00

4.00

3.00

2.00

1.00

0.00
BRA CAN CHN DEU FRA GBR IDN IND JPN NGA RUS USA

w2012 w2015

Source: World Economic Outlook (2015 and 2012) - IMF

1.1.2 Savings Rates

Saving rates are another important feature of the projections. Savings rates have an impact
on long term growth by conditioning the feasible level of investment. The latter can be funded
domestically or externally with implications for current account balances. We will discuss the
evolution of the current account (and the capacity to provide/require savings to/from the rest of
the world) in the next sub-section.
Long run saving rates are also impacted by demographic changes (such as the share of the
population in retirement or in their peak earning and saving years), but since demographic patterns
are constant across the scenarios, we focus on short and medium run variations in the WEO
projections. Three main features appear to be:
 An increase in the saving rates of some advanced economies facing significant public and
private debt challenges that require macroeconomic adjustment (Ireland, Hungary etc.);
 An increase or decrease in the savings rates of some oil exporting countries that face large
income shocks and may cut their savings rates to stabilize consumption (most of them), or
increase saving rates to support investment (Qatar, Equatorial Guinea)

9
  The specific case of China where the saving rate has declined sharply from the 50%
observed in 2011 and was expected to stay at this level in the WEO 2012 projections, while
it has now started to fall and, in the latest projections, reaches 39% by 2020.

1.3. Current Account Dynamics

An important driver for the world economy is the evolution of current account
surplus/deficits which, as noted above, reflect changes in both savings and investment rates.
Indeed, they are key drivers of the dynamics of world trade.
In Mirage, they are exogenous (but not fixed) indicators that allow us to define the
evolution of the real exchange rate. Therefore, the current account surplus/deficit may need to be
adjusted in the scenarios to be dynamically consistent with our narrative (internal vs external
adjustment, real exchange rate appreciation/depreciation). The following table summarizes the
WEO projections (2012 and 2015) for Current Account Balances (CAB) in billion USD for
selected economies (those with the largest surplus/deficit in 2012):

10
Table 1. Comparison of 2012 and 2015 Current Account Balance Projections (selection of
countries with larger current account imbalances)

WEO 2012 WEO 2015

Country 2013 2015 2017 2013 2015 2017 2018

Australia -84 -105 -118 -52 -50 -44 -46
Brazil -80 -97 -112 -91 -73 -67 -70
Switzerland 72 70 65 49 48 52 54
China 229 356 540 348 267 150 95
Germany 175 162 138 286 271 271 270
Great Britain -30 -12 -17 -136 -114 -83 -86
India -58 -64 -69 -30 -52 -78 -87
Japan 166 151 129 124 131 132 131
Korea 19 13 12 98 92 91 89
Kuwait 84 73 75 11 13 17 17
Netherlands 64 59 53 72 73 75 78
Norway 70 59 53 28 26 30 28
Qatar 58 36 20 10 -6 1 1
Russia 44 -19 -59 62 75 84 81
Saudi Arabia 152 113 107 -22 -13 4 -2
Singapore 60 61 62 61 55 54 54
Turkey -72 -79 -100 -33 -39 -49 -50
Taiwan, China 44 53 64 64 63 61 62
USA -499 -561 -696 -461 -629 -710 -747
South Africa -24 -27 -33 -14 -15 -16 -16
Total 390 241 114 376 122 -24 -144
Source: World Economic Outlook (2015 and 2012) - IMF

A few remarks can be drawn from the previous table:

 In both WEO projections, we see significant declines in global imbalances.
The overall surplus for the selected “leading” economies declines in both projections. This
decline is much larger in the 2015 version-- which includes the global slowdown. In this
case, we see that the surplus (that will finance the rest of the world here) disappears by
2017 and a sizeable deficit emerges in 2018.
 Important differences (but not at the aggregate level) exist for the first year,
2013, (see Qatar, Norway, Germany, China, Great Britain, Australia)
 Differences in trends are also noticeable especially China, Russia, Brazil
We consider three main drivers and outcomes when building our scenarios:

11
  The reduction in oil (and related commodity) prices. It leads to a significant
change in the initial (2013) trade balance of oil exporters/importers (very important for
Qatar, Saudi Arabia and countries in a similar situation).
 Shrinking of Chinese external savings (part of the rebalancing dynamics)
 Increase in US requirements for foreign savings

The combination of these three drivers will lead to a significant reduction in the amount of
foreign savings available for middle and low income countries, leading to a reduction in their
investment and growth.

1.4. Commodity prices

Even though commodity prices are endogenous to our model, we are interested to see how
some prices evolve in the IMF forecasts for two reasons:
 To see whether the CGE generates a time paths consistent with the commodity price
forecasts, especially in terms of price dynamics and cross price correlation;
 To assess whether there is a need to add some specific supply shocks, especially on the oil
and energy markets, to represent both OPEC supplies and supplies from non-conventional
sources
Figure 7. World Commodity Price Projections (2015)

140

120
Price Index 100 = 2011

100

80

60

40
2013 2014 2015 2016 2017
Year
Crude Oil (petroleum), simple average of three spot prices; Dated Brent, West Texas Intermediate, and the Dubai Fateh, US$ per barrel
Commodity Natural Gas Price Index includes European, Japanese, and American Natural Gas Price Indices
Commodity Coal Price Index includes Australian and South African Coal
Wheat, No.1 Hard Red Winter, ordinary protein, FOB Gulf of Mexico, US$ per metric tonne
Maize (corn), U.S. No.2 Yellow, FOB Gulf of Mexico, U.S. price, US$ per metric tonne
Rice, 5 percent broken milled white rice, Thailand nominal price quote, US$ per metric tonne
Soybeans, U.S. soybeans, Chicago Soybean futures contract (first contract forward) No. 2 yellow and par, US$ per metric tonne
Palm oil, Malaysia Palm Oil Futures (first contract forward) 4-5 percent FFA, US$ per metric tonne
Beef, Australian and New Zealand 85% lean fores, FOB U.S. import price, US cents per pound
Poultry (chicken), Whole bird spot price, Georgia docks, US cents per pound

12
 Source: World Economic Outlook (2015) - IMF

There were some major changes between the WEO projections (2015) for world
commodity prices and the 2012 forecasts, with the most striking being a sharp fall in oil prices,
followed by a slow recovery. In Figure 7, we see that the 2015 forecasts involve price declines of
between 25% and 40% for most commodities. Cereal prices decline by 36% on average and palm
oil prices by 45%. Agricultural raw materials, decline sharply, with cotton and wool prices falling
by 45%. Finally, meat prices are supported by higher demand.

13
2 Methodology
Regarding poverty outcomes, it is worthwhile to look at both short and long run effects.
Key impacts will be on commodity prices in the short run (affecting directly both the incomes and
the costs of living for households). In the longer run, changes in wage rates for unskilled workers,
will dominate. In order to disentangle the short and long term impacts of these changes, the
MIRAGRODEP approach is coupled with country-specific household models following Ivanic
and Martin (2014). This top-down approach, starting from broad macroeconomic changes and fine
tuning the analysis through the identification of key impacts is essential to be able to combine the
ambition of this project with the very tight constraints on time and resources. Figure 8 summarizes
our analytical steps starting from the macroeconomic scenario definitions (based on the IMF
outlook), their translation in the structural CGE model and the poverty and farm analysis.
Figure 8. Analytical Steps

Identification of macroeconomic changes to be analyzed

Design experiments that assess the implications of these changes for the world
economy and the prices that affect poor households

Global CGE Simulations: Identifty implications for the poor in different regions of the
different scenarios capturing different level of price transmission, macroeconomic
adjustment and global/regional pecuniar externalities

Household modeling: Simulate the implications for households of the prices that they
face and of the technological changes from which they benefit.

Results: Global, regional and national impacts on on real income, poverty, food Indicators
and macro-nutritient consumption

2.1 Analyzing macroeconomic changes at the household level

The impacts of macroeconomic changes on the incomes of the poor are more complex than
is commonly realized. These impacts are perhaps simplest when a decline in productivity lowers
the efficiency of production within an unincorporated firm (such as a family farm) managed by
their household. Another relatively simple case arises with a shock that results in a sizeable change
in a commodity price for which a poor household is a net seller or net buyer. In the short run, the
effect of this shock on real income can be identified with knowledge of the net seller or net buyer
status of the household (although recent evidence from household surveys suggests that our
knowledge of these net trade shares may be less reliable than we had formerly thought (Headey
14
and Martin 2016). The longer-term impact takes into account impacts of productivity changes on
the real wages obtained by household members from their sales of labor (and other factors) sold
outside any family-owned firm. It also needs to take into account the potential for changes in the
output and consumption patterns of the household, which may move from being a net seller of a
good whose price has fallen, to a net buyer. Major changes in income in different countries may
also be expected to result in changes in remittances from family members working abroad. So,
macroeconomic shocks can be expected to affect the real income of households through different
channels with macro and micro-economic implications:
i The direct impacts on the incomes of producers;
ii Through changes in the cost of living, and;
iii Through impacts on factor returns, particularly wage rates for unskilled labor sold outside
the household’s business activities.
iv Changes in remittances;

The methodology can be understood using the following approach. The impact of an external
shock on the real income of a household can be obtained using a money measure of household
welfare W at a given utility level, 𝑢, given by:

(1) 𝑾 = 𝝅(𝒑∗ , 𝑤, 𝝉) − 𝒆(𝒑, 𝑤, 𝒖),

Where 𝜋(𝒑∗ , 𝑤, ) is a profit function representing the profits generated by any unincorporated
household enterprise, such as a farm firm, specified as a function of effective commodity prices
(𝑝∗ ), factor prices (w), and technology (𝜏); 𝑒(𝒑, w, 𝑢) is a full cost function of the type used by
Deaton and Muellbauer (1981) for a household that consumes goods and supplies factors at a given
vector of commodity and factor prices, p, factor prices, w, and utility level, 𝑢. The distinction
between actual and effective output is discussed in Martin and Alston (1997). From the point of
view of the firm, quantity 𝑞 ∗ of effective output now translates into a larger quantity, q, of output,
where 𝑞 = 𝑞 ∗ 𝜏. The increase in the actual output from any given effective output results in an
increase in the effective price of output at any actual price, where the effective price is defined as
𝑝∗ = 𝑝𝜏

Note that the prices of goods and factors are frequently endogenous in the macro model but are
always exogenous at the household level. Therefore, we need a large scale CGE, here the
MIRAGRODEP model, to be able to provide the right information regarding price changes to the
household modeling component. Since a macroeconomic slowdown can include differentiated
evolution of technology (productivity) between agriculture, industry and services, this framework
allows us to tackle such heterogeneity at the household level.
The right side of equation (1) may usefully be rewritten as 𝒛(𝒑, 𝑤, 𝜏, 𝑢). With this
simplification, a second-order approximation of the welfare impact of changes in 𝒑, 𝑤 and 𝜏 is
given by:

𝛥𝒑 𝑧𝑝𝑝 𝑧𝑝𝑤 𝜋𝑝𝜏 𝛥𝒑

1
(2) ΔW = [𝑧𝑝 𝑧𝑤 𝜋𝜏 ] [𝛥𝑤 ] + 2 [𝛥𝒑 𝛥𝑤 ∆𝜏] [𝑧𝑤𝑝 𝑧𝑤𝑤 πwτ ] [𝛥𝑤 ] ,
∆𝜏 πτp πτw 𝜋𝜏𝜏 ∆𝜏

15
This quadratic form takes into account both the nonlinear (in this case quadratic) relationship
between prices, technological change and output levels in equation (2). The first term in (2)
includes the net sales of the household times the change in the price of the commodity, 𝑧𝑝 𝛥𝒑, the
measure of welfare change emphasized by Deaton (1989) for analysis of commodity price changes.
It also takes into account the impact of changes in factor prices, and especially wage rates, times
the net sales of the household outside any family-owned firm, 𝑧𝑤 𝛥𝑤. Finally, it takes into account
the direct impact of changes in technology on the profits generated by the family farm given a
change in technology, 𝜋𝜏 ∆𝜏. As shown in Ivanic and Martin (2016), the second order terms
generalize these familiar first-order impacts taking into account the changes in the output of farm
firms and changes in sales of labor off-farm when changes in prices and productivity are large.
Assessing impacts at the household level, for both rural and urban households, is much
more complicated than assessing aggregate impacts since household impacts depend not only on
the characteristics of the household as producers, but also impacts on the cost of living and on
household members’ participation in labor markets. Fortunately, a vast amount of work has been
done by Ivanic and Martin (2014) to collect data including income from agriculture and
expenditures on agricultural products (particularly food), and on household income received from
labor markets. We do not form “representative” households using these data, which necessarily
aggregates away a great deal of valuable information. Rather, we retain and use directly all of the
information available from household surveys on the income sources and expenditure patterns in
each of our 315,000 sample households.
These two-tiered approaches have already been implemented in both Bouet et al. (2011)
and Ivanic and Martin (2014) with some differences and strong complementarities. Merging these
approaches will generate significantly value added. The overall modeling framework is explained
in Figure 1. While Bouet et al (2011) use the MIRAGE model and Ivanic and Martin (2014) use
the GTAP model to determine the price changes that are fed to the household models, the
differences between these two models are relatively minor.
The second order effects identified in equation (2) are likely to be much more important
when looking at price changes than when the primary shock is, as in this study, changes in
productivity. However, even in this situation, Ivanic and Martin (2014) found that second order
impacts were considerably less important than the first-order impacts—which include both effects
resulting directly from price changes and those resulting from the impacts of price changes on
wage rates for unskilled labor. Since the primary impact of a productivity change is through first-
order impacts on income of producers and resulting changes in factor prices, we decided not to
incorporate second-order impacts resulting from induced price changes in our empirical analysis
at this stage.

2.2 Moving to Broader Country Groupings

Our sample of countries is widely-spread geographically and includes 76% of the world’s
poor people (Ivanic and Martin 2014) so it seems reasonable to assume that it can generate realistic
estimates of the global poverty impacts of changes such as the macroeconomic shocks under
consideration.

16
 We are interested to obtain estimates of poverty changes at the global level, and for
meaningful groupings of countries. To do this, we used clustering based on four variables: the
poverty headcount rate at $1.90 a day, the rural population ratio, the adjusted net national income
per capita (constant 2005 US$), and cereal yields (kg per hectare). If the income measure or the
poverty headcount ratio at $1.90 were not available, we used GDP per capita and PovcalNet’s
2005 PPP data respectively. Apart from the 2005 PPP PovcalNet data, these measures were
available in the World Bank’s World Development Indicators collection. Using these four
variables, we used a partitioning method - assigning each country to the group with the nearest
mean values while maximizing differences between each country clustersi.
Using this method, 5 country clusters were created, which were then divided by region
while making sure that each group contained at least 1 of the countries in the Ivanic and Martin
(2014) household data set. We then used the change in the poverty rate calculated by Ivanic and
Martin (2015) for the 31 countries, and applied it to the remaining countries in their respective
groups, allowing us to generate estimates of changes in worldwide poverty rates.

2.3 Implementing alternative macroeconomic scenarios in the CGE

This assessment utilizes a recursive dynamic multi-region, multi-sector model CGE model:
the MIRAGRODEP model. MIRAGRODEP (Laborde and al. 2013) is a Computable General
Equilibrium (CGE) model based on MIRAGE (Modelling International Relations under Applied
General Equilibrium, Decreux and Valin, 2007).
Figure 9. Analytical Framework

The multi country CGE model that we use allows a detailed and consistent representation of the
economic and trade relations between countries. International economic linkages are captured
through in international trade of goods and services, and capital flows. A dynamic, recursive
solution is obtained by solving the model sequentially and moving the equilibrium from one year
to another. In our study we assume perfect competition in all sectors which allows us to have a
17
detailed geographic and sector decomposition. Refer to Figure 9 for a more detailed analytical
framework.

In each country a representative consumer maximizes a CES-LES (Constant Elasticity

of Substitution – Linear Expenditure System (LES))1 utility function under a budget constraint to
allocate expenditures across goods. The LES system allows for different income elasticities of
demand, with those for food typically lower than those for manufactures and services. Once total
consumption of each good has been determined in the top level, the origin of the goods consumed
is determined by another CES (Constant Elasticity of Substitution) nested structure following the
Armington assumption.
On the production side, value added and intermediate goods are complements under the
Leontief hypothesis that intermediate inputs are needed in fixed proportion to output. Total value
added is represented as a CES function of unskilled labor and a composite of skilled labor and
capital: this allows us to specify a lower degree of substitutability between the last two production
factors. In agriculture and mining, production also depends on land and natural resources. New
capital is perfectly mobile across sectors while installed capital is immobile. Skilled labor is
perfectly mobile across sectors while unskilled labor is imperfectly mobile between agricultural
sectors and non-agricultural sectors. Investment is savings-driven and the real exchange rate
adjusts endogenously such that the current account is constant in terms of world Gross Domestic
Product (GDP). In some scenarios, we exogenously decrease (increase) the current account of
some countries, and restore global balance by proportional increases (decreases) to the current
account balances of other countries with initial imbalances. As is common in modeling long term
productivity growth or trade reform, the supply of labor is treated as exogenous. While the supply
of labor clearly adjusts to changes in wages in complex ways (Blundell and Macurdy 1998; Keane
and Rogerson 2012), any resulting welfare gains are second-order once the transitional dynamics
have died down, coming at the cost of reductions in leisure.
By contrast with MIRAGE and most global general equilibrium models, the government is
explicitly modelled as a different agent from the private agent in MIRAGRODEP. The income of
the government consists of taxes collected on production, on factors of production, on exports, on
imports, on consumption, and on households’ income. The government is assumed to maximize a
Cobb-Douglas utility function: government so that spending on each commodity remains a fixed
value share of total public expenditure on goods and services. The model structure ensures that
both domestic and external constraints are respected in each country.
As in Lakatos et al (2015), we begin our global macroeconomic analysis with an
assessment of the key macroeconomic adjustments under review. This builds on the analyses of
the macroeconomic situation by the IMF, comparing projections made in 2012 with the latest 2015
release. New trends are identified at the global, regional and national level. A very important task
is to identify in these new scenarios the drivers at play in the OECD economies, in the BRICs and
in other low and middle income countries. The case of China is particularly relevant for poor

1
The CES-LES is a variant of a CES function where minimal consumptions are introduced. It is equivalent to
replacing the Cobb-Douglas structure of the Stone-Geary function (i.e. LES) by a CES structure.

18
countries in Africa, Asia and Latin America and considers the direct impacts of the economic
slowdown in China but also the rebalancing of its economy between domestic and foreign demand,
final consumers and investments. An important element of this scenario is changes in the
opportunities for other countries to take up opportunities for production of labor intensive goods
currently produced in China, but expected to migrate to other developing countries. The turning
point in China’s labor force growth from growth to decline in 2015 reduces China’s overall growth
rate, and also reduces China’s ability to continue to expand exports of labor-intensive goods. These
elements of the projections will affect demand and supply on world markets in heterogeneous
ways.
For the purposes of this project, we consider 3 scenarios between 2011 and 2030:
 Scenario 0, our baseline, is based on the old, “optimistic” projections. We
consider the growth trajectory from the WEO 2012. Starting from our base year (dataset
2011) we implement the growth trajectory from WEO 2012 for 2012 to 2017.
Demographic projections are taken from the UN and the change in the economically active
population is used to define the growth of the labor force. TFP is computed at the
country/regional level to match the GDP trajectories. In order to build the growth trajectory
to 2030, we assume that the average annual TFP growth rate achieved in the 2015-2017
period is maintained between 2017 and 2030.
 Scenario 1 is our “first” change scenario looking at the reduced growth
projections. Starting from the base year, we implement the actual growth rate from 2011
to 2014 for all countries, and the WEO 2015 projections from 2014 to 2020 for our group
of “leading” economies defined as high income countries + Russia + Brazil + China. For
the group of leading economies, we assume that the average annual TFP growth rate for
the 2017-2020 period is maintained between 2020 and 2030. For the middle and low
income countries, we maintain the TFP growth rates computed in Scenario 0 between 2014
and 2030. In addition, we implement structural reforms in China leading to a “rebalancing”
of its economy through an increase in domestic consumption and a decrease in domestic
savings, leading to a reduction in domestic investment but also a reduction in the current
account surplus. Other changes from Scenario 0 include a reduction in energy prices
(savings rates in oil exporting countries, current account, natural resource endowments for
energy products).
 Scenario 2, includes all the elements of scenario 1 but considers additional
drivers for the LIC and MICs. Particularly, the TFP growth rate for these countries between
2014 and 2030 are readjusted to the new WEO scenario for these countries as well as for
the “driver” economies.

19
Table 2 provides additional details regarding our implementation strategy.
Table 2. Scenario Summary Table
Scenario 0 Scenario 1 Scenario 2

GDP Trajectory – Leading Exogenous Exogenous Exogenous

economies
(HIC+Russia+Brazil+China)
GDP Trajectory – MIC + Exogenous Endogenous Exogenous
LIC (unconstrained)
TFP – Leading economies Calibrated (WEO 2012) Calibrated (WEO 2015) Calibrated (WEO 2015)

TFP – MIC + LIC Calibrated (WEO 2012) Same value as in Calibrated (WEO 2015)
Scenario 0
Agricultural TFP Same dynamics as Same dynamics as Same dynamics as
national TFP national TFP national TFP
Savings rate Unchanged Exogenous change in As in Scenario 1
China 50% to 40% by
2020, 40% to 35% by
2030
Adjusted for oil
producing countries to
stabilize consumption
Current Account Balance Unchanged Surplus Reduction in As in scenario 1
China
Strong Surplus
Reduction in oil
exporting countries
Increase deficit in the US
Remittances Endogenous, expressed Endogenous, expressed Endogenous, expressed
as a percentage of as a percentage of as a percentage of
national (source country) national (source country) national (source country)
unskilled labor payments unskilled labor payments unskilled labor payments
Natural Resource Unchanged Calibrated for Gulf As in scenario 1
endowments countries, US, Russia,
Australia, China to target
oil, gas and coal prices
Labor Force Exogenous, follow As in Scenario 0. As in Scenario 0.
medium UN projections
for active population.

Population and active population growth rates follow the UN projections (central scenario).
This implies increases and decreases in the labor forces in particular countries. The case of China
is particularly interesting, with the workforce peaking in 2015, and declining by four percent over
the period to 2030. The labor force evolves over time but is constant across scenarios, so
differences between scenarios are unaffected by the projections.
Overall, the main exogenous driver of differences between the scenarios is TFP. It captures
both changes in per unit productivity at a constant utilization rate, or changes in the utilization rate

20
of productive factors (i.e. for the same population, a reduced number of effective working hours
or underutilization of the physical capital stock, etc.). The latter interpretation is particularly
relevant when addressing downturns in a macroeconomic cycle, with both having the same effects
on income per capita outcomes when considering homogenous changes in utilization rate across
factors and agents.
Figure 10. China's Labor Force, 15-64 years

1,020

1,010

1,000

990
Million Activie Workers

980

970

960

950

940
2011

2012

2013

2014

2015

2016

2017

2018

2019

2020

2021

2022

2023

2024

2025

2026

2027

2028

2029

2030
Year

Source: MIRAGRODEP baseline dataset, based on UN demographic projections

2.4 Datasets
The main source of data for the economy-wide projections is the GTAP 9 database. Tor
the household simulations we rely on the dataset developed by Martin and Ivanic (2014) that
includes 31 countries, representing about 80% of the world’s poor. This dataset cover large
emerging economies, including exporters and importers of food products (e.g. Brazil, China, India,
Indonesia, Nigeria, Pakistan, Cote d’Ivoire), and smaller or more vulnerable economies (e.g
Tanzania, Uganda, Rwanda, Zambia, Malawi Niger, Yemen, Albania, Moldova, Georgia,
Bangladesh, Nepal, Mongolia, Vietnam, Cambodia, Peru, Ecuador, Guatemala, Belize, Nicaragua,
Costa Rica). The consolidated dataset includes more than 300,000 representative households and
provides a basis for assessing the global impacts of price and productivity changes on poverty. The
individual countries for which household survey data are available and used are also identified in
appendix 6.2.

21
2.5 Regional and sectoral disaggregation

Table 3 and Table 4 summarize the regional and sectoral aggregations used in our analysis.
For the country aggregation, we proceed in two steps to address properly some numerical
constraints. First, we operate with the 22 key regions and countries, as listed above, to calibrate
global and leading economies trajectories as well as global commodity prices. Then we run specific
simulations where we single out key middle and low income countries needed for the poverty
analysis.
Table 3. Country Nomenclature
Region Code Region Label GTAP Regions
Oceania Oceania AUS, NZL, XOC,
CHN China CHN, HKG,
HICAsia High Income Asia JPN, KOR, TWN, XEA,
Easia East Asia KHM, IDN, LAO, MYS, PHL, SGP, THA, VNM, XSE,
Sasia South Asia BGD, PAK, LKA, NPL, XSA,
India India IND,
NAFTA USA and Canada CAN, USA, XNA,
MEXICO Mexico MEX,
LAC LAC wo Brazil and Mexico ARG, BOL, CHL, COL, ECU, PRY, PER, URY, VEN,XSM,
BRA Brazil BRA,
CRI, GTM, NIC, PAN, SLV, HND, XCA, DOM, JAM, PRI,
CAM Central America
TTO, XCB,
AUT, BEL, CYP, CZE, DNK, EST, FIN, FRA, DEU, GRC,
HUN, IRL, ITA, LVA, LTU, LUX, MLT, NLD, POL, PRT,
EFTA European Free Trade Association
SVK, SVN, ESP, SWE, GBR, CHE, NOR, XEF, BGR, HRV,
ROU, XTW,
ALB, BLR, UKR, XEE, XER, KAZ, KGZ, MNG, XSU,
CIS CIS countries wo Russia
ARM, AZE, GEO,
Russia Russia RUS,
Gulf Gulf Countries IRN, ARE, BHR, KWT, OMN, QAT, SAU, XNF,
MENA MENA countries TUR, ISR, JOR, XWS, EGY, TUN,
MAR Morocco MAR,
NGA Nigeria NGA,
WAF West Africa SEN, BEN, BFA, CIV, GHA, GIN, TGO, XWF,
CAF Central Africa CMR, XCF, XAC,
ETH, KEN, MDG, MWI, MUS, MOZ, RWA, TZA, UGA,
EAC Eastern Africa
ZMB, ZWE, XEC,
SACU SACU BWA, NAM, XSC,
ZAF South Africa ZAF

22
Table 4. Sectoral Nomenclature
Sector Code Sector Label GTAP Sectors
rice Rice PDR, PCR,
wheat Wheat WHT,
ocereals Other Cereals GRO,
v_f Vegetable and Fruits V_F,
osd Oilseeds OSD,
sug Sugar C_B, SGR,
pfb Pland Fibers PFB,
ocr Other Crops OCR,
cattle Cattle CTL, RMK,
otherAni Other Animal OAP,
wol Wool WOL,
forest Forestry FRS,
fish Fisheries FSH,
enepr Extraction - Energy COA, OIL, GAS,
miner Minerals OMN, NMM,
meatc Red Meat CMT,
meato White Meat OMT,
vol Vegetable Oil VOL,
dairy Dairy products MIL,
ofd Other Food OFD,
bevtob Beverage and Tobacco B_T,
tex Textile TEX
wap Wearing Apparel and Leather products WAP, LEA
paper Paper Products PPP, LUM
ffl Fossil Fuels P_C,
crp Chemicals CRP,
metal Other Mineral I_S, NFM, FMP,
TransEq Transport Equip. MVH, OTN,
electronics Electronics ELE,
capgoods Capital Goods OME,
omanuf Other Industries OMF,
utilities Utilities ELY, GDT, WTR,
construction Construction CNS,
trade Trade TRD,
trans Transportation OTP, WTP, ATP,
services Services CMN, OFI, ISR, OBS, ROS, OSG, DWE,

23
3 Macroeconomic projections to 2030
This section provides findings regarding the implications of our three scenarios.

3.1 GDP projections

The simplest indicator to illustrate our different growth trajectory is the real GDP growth rate
between 2011 and 2030. Figure 11 shows the annual average over the period and illustrates the
very substantial slowdown faced by emerging economies: from 8% to 6% in China, from 4.4% to
2% in Brazil and from 4 to 1.5% in Russia. Growth remains slow in advanced economies (EFTA
about 1.6%) or is reduced by one-third in some of the more dynamic OECD economies (eg the
USA, Canada, Japan, Korea). Middle income countries are less impacted (e.g. a reduction equal
to one tenth in the initial growth rate for Morocco). For West and Eastern African countries, the
projections are more optimistic today than three years ago and they manage (S2) to more than
compensate for the global slowdown (S1). When we consider the implications of the slowdown in
the leading economies (comparing S1 to S0), we find that the negative impacts are more
pronounced in West Africa (annual growth rate decline from 7% to 5.7%) and South Asia than in
East Africa and India, showing clearly different trade and investment linkages and regional
dynamics.
Figure 11. Average annual growth rates for real GDP 2011-2030

9.0

8.0
Average Annual Growth Rate, %

7.0

6.0

5.0

4.0
S0
S1
3.0
S2
2.0

1.0

0.0

Source: MIRAGRODEP model projections

24
3.2 Total Factor Productivity projections
Total Factor Productivity (TFP) is a key element in our analytical framework. Indeed, it is the main
driver of our calibration exercise to reproduce the GDP trajectories in the WEO scenarios since
the growth rates of factor endowments are defined either exogenously (labor) or endogenously
(capital) in response to income growth. At the same time, TFP is an essential driver in the evolution
of productivity for poor workers and opportunities to achieve real income increases. Given the
Solow-type framework that we use, changes in TFP are the primary exogenous source of long term
growth in any country. Other, relatively minor, changes in endowments include declines in land
per person and changes in endowments of natural resources.
Figure 12 displays the annual average growth rate of TFP for the different regions under our
different scenarios. They are lower than the GDP growth discussed in the previous section since
labor and capital accumulation also contribute to increases in the GDP growth rate. Of course the
gap between TFP and the GDP growth rate is very important for economies with rapidly increasing
workforces (demographics) and investment (e.g. Nigeria, extensive growth) and of more limited
importance when looking at mature economies where intensive growth (productivity) is the main
driver (e.g. EFTA).
By construction, TFP is identical between S1 and S2 for the leading economies, and S0 and S1 for
the others. Overall, low income countries achieve significant productivity gains in all scenarios,
especially in S2, with an average annual increase of 3% or more. Middle income economies are
the most impacted by the reduction in average productivity gains.
Figure 12. Average annual growth rate for Total Factor Productivity between 2011 and 2030

6.0

5.0
Average Annual Growth Rate, %

4.0

3.0

2.0

1.0

0.0

S0 S1 S2

Source: MIRAGRODEP model projections

25
3.3 Real Unskilled Wages
A first proxy for the impact of these alternative growth paths on the poor is the evolution
on the real unskilled wage, a major source of income for poor households. Results are displayed
in Figure 13 showing the increase between 2011 and 2030. In our initial scenario (S0), the
projections were quite optimistic ranging from +47% in East Africa to +343% in China. The most
rapid growth was expected to take place in Asia (South and East), Brazil and West Africa with a
doubling or more in the real incomes of unskilled workers. Less optimistic scenarios display highly
heterogeneous outcomes with countries being strongly impacted but having still impressive
performance (e.g. China); weakly affected but still improving the situation of their poor (e.g India);
those negatively impacted by the global slowdown (S1) but benefiting from positive domestic
dynamics (for example S2 is better than S1 for South Asia and West Africa as a bloc). For
individual and still poor countries such as Ghana, the improvement in unskilled wages is very
significantly impacted +43% in S2 vs + 75% in S0.
Figure 13. Real Unskilled Wage Rates from 2011 to 2030

400

350
Growth Rate 2011-2030, %

300

250

200
S0
150 S1
100 S2
50

Source: MIRAGRODEP model projections

26
4 Results
The first step in our analysis is to use the global model to estimate the implications of the growth
slowdown for key prices; for aggregate income and for key outcomes such as agricultural incomes
and unskilled wages. We first consider the impacts on world prices, and then turn to the other
variables.

4.1 Evolution of world prices

Key results for price changes relative to the US CPI are shown in Table 5.
Table 5. World Prices - Changes in 2030 compared to the Baseline
S1 S2
Agricultural World Prices
Vegetable and Fruits 2.22 3.53
Oilseeds 4.81 6.32
Plant Fibers 2.18 1.72
Other Crops 2.88 2.88
Wool 0.17 1.14
Vegetable Oils 1.57 4.05
Processed Food 1.97 2.67
Rice 1.88 2.96
Wheat 3.75 4.00
Other Cereals 2.52 3.90
Sugar 5.36 5.89
Cattle 2.77 2.71
Other Animal Products 3.01 3.36
Fisheries 1.48 2.30
Red Meat 2.59 3.16
White Meat 3.16 3.45
Dairy 1.51 1.79
Non Agricultural World Prices (Selected)
Energy Commodities -26.96 -24.36
Mineral Products -9.30 -8.20
Metal Products -0.52 0.24
Capital Goods 1.65 1.79
Aggregate World Prices
Agriculture 2.39 3.19
Manufacturing goods 0.77 1.16
Services 0.82 1.00
Extraction -22.34 -20.04
Source: MIRAGRODEP model projections

The most striking change in Table 5 is the sharp fall in the prices of mineral products. This reflects
the constraints on output of these products created by mineral and oil resource constraints, and
their consequent price volatility in response to substantial changes in demand. The situation for
agricultural products is quite different. In this case, the impact of lower productivity is to increase
prices. The reason for this seemingly paradoxical result is that we are assuming a homogenous
reduction in productivity across all sectors as a key driver of growth changes. When agricultural
productivity rises in line with productivity in other sectors, the increase in national income results
in a smaller-than-proportional increase in demand for agricultural products simply because of
Engel’s Law. With equal productivity growth across sectors, higher income results in a decline in

27
the relative price of agricultural products. Thus, a slowdown in the rate of growth results in a
modest increase in agricultural prices relative to the S0 baseline, the effects being stronger for
staple foods (low income elasticity compared to high value products such as meat and dairy
products)
The prices of manufactures rise slightly relative to the numeraire. The slowdown in growth reduces
the demand for these goods more than for agricultural products, tending to push their prices up
more than for agricultural products, but the decline in the prices of minerals and fuels helps to hold
them down. The average price of services rises under the lower productivity scenario for a similar
reason.
Therefore, we see that the global commodity price effects of the economic slowdown captured in
scenarios 1 and 2 have contrasting outcomes for the poor. Agricultural prices are more resilient to
the crisis and will generate higher incomes for farmers; however they also mean higher costs for
food consumers. However, we are going to see that these price effects can be dominated by the
consequences of the different scenarios on poor people of declines in productivity.

4.2 Impacts on income

The key variables presented in Table 6 show a wide range of responses to the slowdown. The most
direct effect is a sharp decline in real income in the countries most seriously affected by the
slowdown. Thus, under the S1 scenario, output declines sharply in leading economies such as
Brazil, China, Russia, the US and Canada. The impacts of this scenario on other economies tend
to be small, and to depend upon whether they are suppliers of minerals or agriculture, with
suppliers of minerals, such as Morocco, tending to lose while suppliers of agricultural products
tend to gain.
Regions depending on global demand and still oriented to OECD markets are particularly
negatively affected by these pecuniary externalities are located in: West Africa WAEMU countries
(-18 percent of real income compared to the baseline in 2030), Ghana (-8 percent), South Asia (-
18 percent except India that is very resilient to the external shock) and Central America (-9
percent). It is also important to underline that these countries are also affected by a reduction in
availably of global savings to fund their growth as they were used to do. Under scenario 2, real
incomes fall in a wider range of countries where growth prospects have been marked down, such
as Ghana (-23 percent), Nigeria (-26 percent), South Africa (-21 percent) and South America.
Economies such as East Africa, where prospects have been upgraded, experience an increase in
real income, while India sees essentially no change in income growth rates. For most of the regions
already negatively affected the outcome is worse (except for West Africa that benefits from
updated productivity gains), especially for South American countries (up to -38 percent of
projected real income) and the MENA region (-18.5 percent)

28
Table 6. Key Macroeconomic Results
Real Income Agriculture VA Unskilled Wages Unskilled Wages - Real Government Real Inbound
Rural Income Remittances

S1 S2 S1 S2 S1 S2 S1 S2 S1 S2 S1 S2

Brazil -35 -35.1 -27.3 -27 -34.4 -34.7 -28.8 -28.7 -35.9 -36.2 -12.1 -17.8
China -30.5 -30.7 -18.4 -18.5 -31.1 -31.3 -24.8 -24.9 -27.5 -27.7 -21.2 -24
Russia -38.2 -37.8 -32.1 -32 -32.5 -32.6 -33.1 -33.1 -42.1 -41.1 -2.2 -17.5
Oceania -10.4 -10.5 -1.2 -2.3 -8.9 -9.2 -3.8 -4.7 -10.4 -10.5 -9.1 -10.5
High Income Asia -10.4 -10.7 -4.9 -5.1 -10.2 -10.5 -5.8 -6.1 -11.5 -11.9 -16.6 -17.7
USA and Canada -19.4 -19.6 -6.8 -7 -19.9 -20.2 -10.9 -11.2 -20.2 -20.5 -5.3 -9
EFTA (inc. EU28) -3.2 -3.7 1.3 1.1 -2.9 -3.4 0.6 0.5 -4.3 -4.8 -6.6 -9.6
Ghana -8 -23.1 4.4 -11 -3.6 -18.4 2.5 -12.6 -11 -25.9 -3.2 -9
Morocco -3.2 -8.4 1.9 -2.7 -3.4 -8.3 0.4 -4.1 -5.8 -10.9 -1.5 -4.5
Nigeria -1.5 -26.3 3 -17.7 2.7 -20 2.8 -18.2 0.3 -24.7 -9.8 -8.6
South Africa 1.1 -21.8 6.8 -11.5 1.4 -21 7.1 -12.1 -0.2 -22.4 -5.5 -2.1
Central Africa -1 -17.8 19 0.4 1.5 -15.2 10.4 -6.7 2.7 -14.5 -5.4 -15.1
East Asia (exc. leaders) 3.9 -15.5 4.8 -11.3 5.1 -14.1 4.8 -12.4 2.9 -16.1 -4.7 -11.2
Rest of South Asia -18.8 2.1 -9.2 8.8 -16.5 4.1 -12.8 6.2 -23.1 -4.3 5.4 8.3
India 2.1 0 2.9 1.1 2.2 0 2.7 0.9 -2.1 -4.2 4.4 7.2
Mexico 3.1 -1 4.8 1.9 3.8 -0.2 5.9 2.8 0.2 -2.5 -17.7 -18.6
South America (ex. Brazil) 2.3 -38 5.6 -27.3 3.7 -36.4 5.4 -29.5 1.8 -38.3 -6.9 -22.6
Central America (ex. Mexico) -8.9 -8.8 -0.2 -2.8 -9.6 -9.7 -3.2 -3.3 -11.8 -11.8 -16.2 -16.4
CIS 8.5 -12.9 11.2 -7.6 13.2 -8.8 11.5 -8.1 8 -14.1 -11.8 -17.8
MENA -2.9 -18.5 3.3 -10.2 -1.2 -16.7 2.4 -11.8 -4.8 -20 8.3 4
West Africa -18.8 -3.6 -7.8 7.3 -11.6 3.3 -7.4 7.4 -27.5 -15.2 -7.1 -1.7
East Africa -6.6 28.6 -1.5 30.1 -4.7 29.2 -2.2 28.5 -8.8 24.9 -2.4 11.3
Rest of SACU 8.4 0.1 7.9 0.7 8.9 0.6 7.9 0.5 7.6 -0.2 -1.7 -18.1
Source: MIRAGRODEP model projections

29
Under Scenario 1, real agricultural value added falls in the countries directly affected by the shock
(leading economies), even despite the rise in real agricultural prices. In the countries not directly
affected by the productivity shock, however, agricultural prices and incomes rise modestly. For
these economies, the situation depends on their exact specialization. If they mainly sell products
to advanced economies and are not in competition with the products originating in these countries
(e.g. cocoa), they are net losers with shrinking markets and stable or declining prices. On the other
side, if they produce similar products, they benefit from the deceleration of productivity in
developed economies that pushes prices up. Gains are substantial in middle income countries such
as Ghana (+4.4 percent), South Africa (+6.8 percent), Mexico (+4.8 percent) and South East Asia
(+4.8 percent) and South American countries (+5.6 percent). A few countries, such as West Africa
and South Asia (except India), have net losses in real agricultural value added, but these results
are mainly driven by the overall macroeconomic crisis. In all cases, Agriculture remains more
resilient to the crisis and its value added much less affected than national income. For instance in
Central America, real value added in agriculture remains stable (-0.2 percent) while national
income falls by 8.9 percent.
Once the adverse shocks are extended to a broader set of countries in Scenario 2, real agricultural
value added also falls there. The productivity effects dominate and the gains are reduced or losses
increased (e.g. from +5.6 to -27 percent between Scenario 1 and 2 for South America). Only in
regions with upgraded productivity projections are farmers better off (West Africa, East Africa).
But in all cases, agriculture appears to be more resistant to the crisis, with much lower losses than
the economy as a whole.
The third variable presented in Table 6 is the average unskilled wage rate, which is a weighted
average of the rural and urban wage rates distinguished in the model. Average real wages follow
a similar pattern to real incomes, declining sharply when productivity falls. In scenario 1, the
average unskilled wage rage is strongly negatively impacted in West Africa (-11.6 percent in
average for WAEMU countries, -3.6 percent in Ghana), in South Asia (except India,-16.5 percent)
and Central America (-9.6 percent) limiting the opportunity for poverty reduction. Other poor
countries are not as strongly impacted and some even experience higher real wages (South East
Asia, SACU countries).
In rural areas, the same pattern appears but with better outcomes: losses are smaller (-7.4 in
WAEMU, -12 percent in South Asia, -3.2 in Central America) or substantive gains emerge (+7
percent in SACU, +2.5 percent in Ghana, + 5.4 percent in South America, 2.4 percent in MENA,
+10 percent in Central Africa). When the updated productivity projections are applied to all
countries, these positive effects tend to vanish (except in WAEMU countries, South Asia and East
Africa were productivity is updated upward). Real rural unskilled wages are cut by more than 10
percent – and up to 30 percent in South America—in many middle income economies such as
Ghana, Nigeria, South Africa, and MENA.
Another important source of income for many poor households in developing countries is
remittances from people working in other countries. These tend to respond heavily to changes in
income in the originating countries. As a consequence, inbound remittances to Mexico decline by
almost 18 percent under S1, even though real domestic income rises by 3.1 percent under this

30
scenario. For smaller economies more dependent on this resource in Central America, real inbound
remittances fall by 16 percent in S1, affecting more several families depending on them and
contribute to the average fall of income nation-wide (-8.8%).
For other regions, the fall of remittances becomes very significant (between 10 and 20 percent)
when the productivity reduction affects the developing countries too (S2). The crisis in South
Africa, Nigeria and Brazil leads to reduction of 18 percent of this source of income for other SACU
countries, of 15 percent in Central Africa, and 23 percent for the rest of South America.

4.3 Fiscal space for developing countries

Many potential responses to economic downturns involve government spending so it is important
to examine their impact on government revenues, for which results are given in the fifth column
of Table 6. In most cases, the decline in real income is strongly linked with the decline in the real
value of GDP. However, there are some cases, such as Russia, where the decline in government
revenues is greater than the decline in real income, because of the strong dependence of Russian
real income on revenues from energy.
For lower income countries, governments in small countries in Central America will face some of
the strongest constraints on their investment and social policies requiring government spending,
with government revenues 12 percent lower than initially expected in all scenarios. For other poor
countries, the situation is challenging—see, for example, West Africa in scenario 1 (-27.5 percent).
However, for most of the low income countries the main challenges appear in scenario 2, when
declines in productivity growth rates occur in these countries too. Africa, except East Africa, will
face declines in public resources relative to initial projections (-26 percent in Ghana, -25 percent
in Nigeria, -22 percent in South Africa) leading to difficult trade-offs in a situation where needs
will be increasing (see previous section). This illustrates the key role of maintaining long term
reform and structural investment policies to maintain productivity improvements in these
countries.
In this context, it also important to discuss the fiscal situation of leading economies where fiscal
resources will also be scarcer and foreign aid money might decline relative to the 2012 income
projections. While the European fiscal revenues are relatively stable (-4.5 percent on average),
North America is more impacted (-20 percent) and newer players such as China, may become less
likely to emerge as key donors (-27 percent).

4.4 Impacts at the Household Level

Figure 14 shows the estimated global poverty rate in 2030 under each of our three scenarios. In
each case, the real income of each household changes over time in response to: changes in business
incomes (driven by productivity growth); to changes in wage rates for households selling labor; to
changes in the cost of living and the evolution of income from household business activities.
Because our long-run scenarios include substantial economic growth over the period to 2030, the

31
poverty rates under all of these scenarios are substantially below current levels.2 The first bar in
each set of three shows the poverty rate, defined as the share of the population below the 1.90
dollar, a day, using the 2011 PPP, definition, under the baseline assumptions. As expected the
poverty rate among farmers is higher than the rate among the population at large. Interestingly,
however, the poverty rate is higher among the rural population more broadly than among
farmers—with higher poverty rates amongst groups such as the landless poor outweighing lower
poverty rates among groups like teachers working in rural areas.
Net Global Poverty Impacts
Turning to our comparison of the scenarios, we see that Scenario 1 involves a sizeable increase in
global poverty—with the poverty headcount rising from 4.79 percent to 5.21 percent, or by 8
percent of its initial level, representing 34 million people reverting to extreme poverty. For farmers,
the increase is larger in absolute terms, but slightly smaller as a share of the initial poverty level.
For the rural group, dominated by farmers, the increase is from 7.15 percent to 7.74 percent. When
we turn to Scenario 2, the overall poverty rate goes up slightly at the whole-country level (38
million people, reverting to extreme poverty compared to the baseline), but declines slightly for
the farmer-headed and rural groups. This decline reflects the improvements in the growth outlook
in a number of African and South Asian countries noted in Figure 12, which contribute to stronger
employment growth and poverty reduction in these countries, together with the modest increases
in agricultural prices presented in Table 5.

2
If we keep constant the poverty rate for each country, but factor in the heterogeneous demographic growth pattern,
the global average poverty headcount in 2030, without considering economic growth, would reach 20 percent,
compared to the current level of 13%. This illustrates the importance of rapid population growth in the poorest areas
of the planet. However, economic growth reduces this global rate to 4.79% with no update in the poverty line
definition.

32
Figure 14. Global Poverty Headcount under alternative scenarios, Percentage

Source: Authors’ calculations.

Note: Poverty is defined by the $1.90 PPP 2011 threshold.

Gross Global Poverty Impacts

Figure 15 presents the net changes in overall poverty between the original forecast values and
Scenario 1, both at the global level and for four regions that were found useful in organizing the
data. It contrast these net changes with the gross movements into and out of poverty for the same
regional groups. While economists tend to focus on the net movement of people into or out of
poverty, policy makers may—given the well-known tendency for people (and voters) to focus
more on losses than on gains—be more concerned with the gross movements of people into
poverty than with the net changes. A key finding is that, for the world and for three of the four
groups, the net change in poverty is very similar to the gross flow, with few people rising out of
poverty. However, for poorest country group, there are important offsetting changes, with 0.61
percent of the population falling into poverty and 0.16 percent rising above the poverty line for a
net increase of 0.45 percentage points.

33
Figure 15. Net and gross movements into and out of poverty, Scenario 1, Percentage points, Total
Population.

1.40%
Changes in Poverty headcount (%) between S0 and S1,

1.20%

1.00%

0.80%
Total population

0.60%

0.40%

0.20%

0.00%

-0.20%
World Middle Income Poorest Poor Poor
Countries countries Countries, Countries, Low
High Rural Rural
Total Population INTO poverty OUT of poverty

Source: Authors’ calculations.

Note: Poverty is defined by the $1.90 PPP 2011 threshold.
Country grouping refers to our cluster analysis. See details in Appendix.

When we consider changes in the broader set of countries covered by Scenario 2, in Figure 16, we
find a strikingly different picture. While the change in the overall poverty rate is in the same order
of magnitude for the population as a whole (0.47 vs 0.42 percent), some of the gross flows into
poverty are substantially larger. In particular, the movement into poverty in the poorest countries
is, at 4.85 percentage points, much larger than anything seen in previous analyses. This is offset
by a 1.76 percentage point gross movement out of poverty, but the result is still a net increase in
poverty in this country group of over 3 percentage points.

34
Figure 16. Net and gross movements into and out of poverty, Scenario 2, Percentage points, Total
Population

6.00%
Changes in Poverty headcount (%) between S0 and S2,

5.00%

4.00%
Total population

3.00%

2.00%

1.00%

0.00%

-1.00%
World Middle Income Poorest Poor Poor
Countries countries Countries, Countries, Low
High Rural Rural
Total Population INTO poverty OUT of poverty

Source: Authors’ calculations.

Note: Poverty is defined by the $1.90 PPP 2011 threshold.
Country grouping refers to our cluster analysis. See details in Appendix 6.2

When we focus on households headed by farmers, in Figure 17, we see a broadly similar picture,
with the biggest impacts on poverty in the poorest countries, and the (substantial) net increase in
poverty being a consequence of a large gross movement into poverty (over 4 percentage points)
and a substantial group (over 2 percentage points) moving out of poverty. In middle income
countries, there is a substantial increase in the net poverty rate (1.5 percentage points) with almost
no farm families moving out of poverty.

35
Figure 17. Net and gross movements into and out of poverty, Scenario 2, Percentage points,
Farmer population

6.00%
Changes in Poverty headcount (%) between S0 and S2, Farmer

5.00%

4.00%

3.00%
population

2.00%

1.00%

0.00%

-1.00%
World Middle Income Poorest Poor Poor
Countries countries Countries, Countries, Low
High Rural Rural
Total Population INTO poverty OUT of poverty

Source: Authors’ calculations.

Note: Poverty is defined by the $1.90 PPP 2011 threshold.
Country grouping refers to our cluster analysis. See details in Appendix 6.1.

Heterogenous drivers
To help understand the cause of the large movements into and out of poverty it is useful to look at
the sources of the changes in income that resulted in households moving the way that they did. In
Figure 18 , we first examine the changes in income that helped households move out of poverty,
and then turn to the changes in income that resulted in households initially above the poverty line
falling into poverty. For the sizeable group rising out of poverty in the poorest countries, the most
important change was an increase in the business profits of smallholders. In some cases, this would
have resulted from households increasing their sales on the market. However, another element of
this rise in income would be from households reducing their net purchases from the market because
of the increase in their farm output. In the other group of countries in which lifting households out
of poverty made an important contribution to the global poverty outcome—the low income
countries, rising farm incomes and wages for labor sold outside the farm firm were important.

36
When we turn to the households falling into poverty, the largest change is in the poorest countries,
and the largest identifiable contributor to this decline is a reduction in sales by smallholders.

Figure 18. Relative contributions of different drivers in the change of household real income,
breakdown by poverty movement, Scenarios 1 and 2

Source: Authors’ calculations

Main countries contributing to global poverty changes

Figure 19 shows that in scenario 1, the country with the most people falling into poverty is Pakistan
with over 22 million people, while the country with the most people escaping from poverty is
Nigeria. However, due to its population size, the size of the shock, and the vulnerability of many
Nigerians to poverty (see Table 6), Nigeria is the country with the most people moving into poverty
in scenario 2 with 29 million people moving into poverty. Ethiopia, Indonesia and Kenya are also
hit hard under this scenario. By contrast, poverty falls substantially in Pakistan, Tanzania, Uganda
and Zambia. Detailed country level results are available in Appendix 6.2

37
Figure 19. Changes in the number of poor people by country, thousands, compared to the 2030
baseline (S0)

Others AFG

BRA CAF

S1 CHN ETH

IDN KEN

MDG MOZ

MWI NER

NGA NPL

S2 PAK RWA

TZA UGA

ZMB

-30 -20 -10 0 10 20 30 40 50 60

Poverty Headcount, million people
Source: Authors’ calculations
Note: 3 letters ISO codes are used to identify key countries

Changes in the Poverty Line

All of the results presented to this point have retained the internationally-standard benchmark of
$1.90 per person per day at 2011 prices currently used by the World Bank. This means that poverty
rates fall to very low levels in many countries by the end of the simulation period. Since this
poverty rate is based on the poverty rates chosen by fifteen poor countries (Ferreira, Jolliffe and
Pryoz 2015), it seems likely that this internationally-accepted poverty rate will change as the
incomes of the poorest countries rise from today’s levels. To assess whether our results are

38
sensitive to the choice of poverty line, we considered a poverty line double the current international
standard. The resulting poverty lines and poverty rates are presented in Table 7.
These results show that the poverty rate is strongly affected by the poverty line. The 2030
levels of poverty are roughly double those at the current poverty line. The effects of the
productivity growth slowdown (in percentage point terms) also roughly double when we move to
the higher poverty line. However, there are no changes in the qualitative nature of the impacts.
Table 7. Effects of Different Poverty Lines in S0, S1, and S2 on Global Poverty Rate in 2030
Scenarios $1.90 Poverty line 2030 poverty line
defined at $3.8
Total Population
S0 4.79% 10.19%
S1 5.21% 11.02%
S2 5.26% 11.29%
Farmer Population
S0 6.79% 13.63%
S1 7.38% 14.82%
S2 7.18% 14.76%
Rural Population
S0 7.15% 14.72%
S1 7.74% 15.85%
S2 7.52% 15.77%
Source: Authors’ calculations

39
5 Conclusions
The recent sharp downward revisions in the prospects for global economic growth, and for growth
in many developing countries, add up to a substantial long-term deterioration of the outlook for
many developing countries. Even countries that are not directly affected by productivity changes
are likely to be affected by the changes in commodity prices and in current account balances
associated with this productivity slowdown.
To assess the likely consequences of this growth slowdown for the achievement of SDG 1 by 2030,
we compare the most recent IMF forecasts for economic growth to 2017 with those from an earlier
and more benign environment—in 2012—and assume that the changes in projected growth rates
over that period are continued to 2030. The resulting changes in the level of GDP are very
substantial. They point to sharply reduced growth prospects for many developing countries,
particularly relative to the period of more rapid economic growth in developing countries and of
income convergence between poor and rich countries experienced since the 1990s.
In this paper, we assess the implications of these changes for the poor—and particularly the rural
poor, who tend to be the poorest of the poor—of this sharp slowdown in economic prospects. We
do this by first projecting the implications of the global growth slowdown at a national level—
including reductions in productivity growth rates and the consequential shifts in economic
balances and changes in relative prices. Then, we use household models for over 300,000
households to assess the implications of these changes for the poor. These models let us assess the
impacts of productivity changes within households’ farm firms, and the real income changes
associated with changes in real wage rates for labor sold by households and changes in the prices
received and paid for food.
We find that the changes in poverty rates associated with this growth slowdown are particularly
sharp in the poorest countries. In these countries, there are also very substantial gross changes in
poverty, with over 5 percent of the population falling into poverty, while another 2 percent are able
to rise out of poverty. When we focus on the households headed by farmers, the most striking
change is a substantial increase in poverty in middle income countries, with over 1.5 percent of
the farm population at serious risk of falling into absolute poverty as a consequence of this growth
slowdown.
This analysis highlights the potential power of the framework utilized for this analysis to capture
the implications for the poor of a wide range of shocks, such as changes in economic growth,
changes in weather and climatic conditions, changes in food prices, and changes in technology.
By using economy-wide models to capture the implications of the original shocks for key variables
affecting households—such as changes in productivity, changes in commodity prices and changes
in factor prices—and then passing these changes to household simulation models, this modeling
approach allows us to capture the impacts on poor people at national and global scales, rather than
to focus only on national aggregates which are likely to be of little interest to policy makers.
We should emphasize that the growth of TFP plays an important role in driving key results since
it impacts both incomes and relative prices. Our results should be looked at keeping in mind that
we assume a uniform changes across sectors. In general, we think it sufficiently likely that a

40
productivity slowdown will affect all sectors to use this as our base case. If, for instance, the cause
of the slowdown is a rise in interest rates, it is likely to raise production costs in all sectors.
Similarly, a decision to raise import barriers is likely to reduce innovation by restricting access to
imported inputs.
However, there are also potential reasons for interest in differential rates of productivity
growth. If, for instance, the extent of adoption of available innovations differs between sectors,
then productivity may be growing at different rates. Since productivity growth in agriculture may
be influenced by increased investments in public research and development following the recent
food price crisis, agricultural growth may prove more robust. Since, agricultural growth is known
to have strong impacts on poverty, estimates of the effects of changes in agricultural productivity
growth may also be policy relevant. Lower productivity in agriculture compared with the rest of
the economy will lead to higher relative prices for agricultural products but lower quantities
produced, and weaker wages. So, lower productivity has a negative outcome for the poor in most
cases, but some farmers may still be better off if the price effects dominate. In our setting, we can
also identify what happens if the asymmetric shock (across sectors) happens only in leading
economies (scenario 1) or if it happens globally (scenario 2).
The adverse impacts of slower growth on the poor highlight the need for careful analysis of
the sources of this slowdown at the country level and for policies to ameliorate the impacts on the
poor. If the slowdown reflects shocks such as adjustment to a decline in the terms of trade, and if
the country then fiscal stimulus might be appropriate. If there are opportunities to raise agricultural
productivity, then this is likely to have particularly strong favorable impacts on poverty in both
urban and rural areas, but particularly in rural areas, over the medium to long term (Ivanic and
Martin 2016; Loayza and Raddatz 2010). Development and enhancement of social safety nets
remains important to help poor and vulnerable households to deal with short term shocks from a
variety of sources.

41
6 Appendix
6.1 Country grouping
Table 8. Cluster Composition (ISO Country Code)

Cluster 1: Middle Income Countries ABW, ARG, ATG, BRA, BRB, BRN, COL, CRI,
CUB, DJI, DMA, DOM, DZA, ECU, GRD, GUY,
IRN, IRQ, JOR, KNA, LBY, MAR, MEX, MHL,
MNG, MYS, OMN, PAN, PER, SLV, SYR, TTO,
TUN, VEN, WBG
Cluster 2: Poorest countries AFG, BDI, BEN, BFA, BGD, CAF, COM, ERI, ETH,
GIN, GNB, HTI, KEN, LAO, LBR, LSO, MDG,
MDV, MLI, MOZ, NGA, PNG, SEN, SLE, SWZ,
TCD, TGO, TLS, TZA, UGA, ZAR, ZMB, ZWE
Cluster 3: Countries with no significant ADO, ARE, AUS, AUT, BEL, BHR, BHS, BMU,
poverty level CAN, CHE, CHL, CYP, CZE, DEU, DNK, ESP, EST,
FIN, FRA, GBR, GRC, GRL, HKG, HRV, HUN, IMY,
IRL, ISL, ISR, ITA, JPN, KOR, KSV, KWT, LBN,
LTU, LUX, LVA, MAC, MCO, MLT, NLD, NOR,
NZL, PLW, POL, PRI, PRT, QAT, SAU, SGP, SMR,
SWE, TUR, URY, USA
Cluster 4: Poor Countries, High Rural AGO, BIH, BTN, BWA, CHI, CIV, CMR, COG,
CPV, EGY, FSM, GAB, GHA, GMB, GNQ, IND,
KGZ, KHM, LIE, LKA, MRT, MUS, MWI, NAM,
NER, NPL, PAK, RWA, SDN, STP, SYC, TJK, TON,
UZB, VNM, VUT, WSM, YEM, ZAF
Cluster 5: Poor Countries, Low Rural ALB, ARM, AZE, BGR, BLR, BLZ, BOL, CHN, FJI,
GEO, GTM, HND, IDN, JAM, KAZ, KIR, MDA,
MKD, MNE, NIC, PHL, PRY, ROM, RUS, SRB,
SUR, SVK, SVN, THA, TKM, TUV, UKR
Source: Authors’ calculations.
Note. These clusters were created using the method described in Section 2.2. Specifically, we used STATA’s kmeans
command. In bold are the countries used in household survey analysis.

42
6.2 Country level results
Poverty Rate, Total Poverty Rate, Farmer Poverty Rate, Rural Changes in
Population Population Population number of poor
people (1000)
ISO Country S0 S1 S2 S0 S1 S2 S0 S1 S2 S1 S2
ABW Aruba 0.1% 0.1% 0.1% 0.1% 0.1% 0.2% 0.1% 0.1% 0.2% -0.01 0.06
AFG Afghanistan 23.7% 24.1% 27.1% 22.7% 23.2% 24.6% 23.6% 24.1% 25.5% 226.46 1,810.94
AGO Angola 9.8% 11.5% 8.9% 8.6% 10.6% 8.0% 9.3% 11.4% 8.7% 524.43 -292.80
ALB Albania 0.3% 0.3% 0.4% 0.4% 0.4% 0.5% 0.3% 0.3% 0.3% 0.00 3.20
ARG Argentina 0.2% 0.2% 0.3% 0.3% 0.3% 0.7% 0.3% 0.3% 0.6% -6.84 67.34
ARM Armenia 0.3% 0.3% 0.4% 0.4% 0.3% 0.4% 0.3% 0.3% 0.4% -1.16 2.76
ATG Antigua and Barbuda 0.1% 0.1% 0.1% 0.1% 0.1% 0.2% 0.1% 0.1% 0.2% 0.00 0.04
AZE Azerbaijan 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.10 0.78
BDI Burundi 36.2% 36.8% 41.4% 34.6% 35.4% 37.6% 36.0% 36.7% 38.9% 74.28 593.96
BEN Benin 21.1% 21.4% 24.1% 20.2% 20.6% 21.9% 21.0% 21.4% 22.7% 55.28 442.05
BFA Burkina Faso 19.8% 20.2% 22.7% 19.0% 19.4% 20.6% 19.7% 20.1% 21.4% 103.65 828.90
BGD Bangladesh 1.0% 0.9% 1.2% 0.3% 0.3% 0.5% 1.1% 1.1% 1.4% -80.91 424.81
BIH Bosnia and Herzeg. 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.04 -0.02
BLR Belarus 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.02 0.15
BLZ Belize 6.5% 7.0% 6.7% 8.7% 9.3% 8.7% 7.1% 7.6% 7.1% 2.02 0.57
BOL Bolivia 0.6% 0.6% 0.8% 0.6% 0.6% 0.9% 0.6% 0.6% 0.8% 4.49 34.92
BRA Brazil 1.3% 1.2% 2.2% 2.0% 1.9% 4.4% 1.8% 1.7% 3.9% -215.1 2,119.05
BRB Barbados 0.1% 0.1% 0.1% 0.1% 0.1% 0.2% 0.1% 0.1% 0.2% -0.02 0.15
BRN Brunei Darussalam 0.1% 0.1% 0.1% 0.1% 0.1% 0.2% 0.1% 0.1% 0.2% -0.03 0.28
BTN Bhutan 1.8% 2.2% 1.7% 1.6% 2.0% 1.5% 1.8% 2.1% 1.6% 2.88 -1.61
BWA Botswana 5.6% 6.6% 5.1% 4.9% 6.1% 4.6% 5.4% 6.6% 5.0% 22.95 -12.82
CAF Central African Rep. 28.0% 28.5% 32.0% 26.8% 27.4% 29.0% 27.8% 28.4% 30.1% 31.93 255.30
CHN China 0.0% 0.1% 0.1% 0.0% 0.1% 0.1% 0.0% 0.1% 0.1% 601.56 601.56
CIV Cote d'Ivoire 7.7% 10.1% 8.0% 11.2% 14.0% 11.1% 10.2% 12.6% 10.2% 710.61 78.40
CMR Cameroon 1.7% 2.0% 1.6% 1.5% 1.9% 1.4% 1.6% 2.0% 1.5% 86.35 -48.21
COG Congo, Rep. 9.8% 11.5% 8.8% 8.5% 10.5% 7.9% 9.3% 11.4% 8.7% 104.63 -58.42
COL Colombia 1.7% 1.5% 3.0% 2.7% 2.6% 5.8% 2.4% 2.3% 5.2% -73.73 726.18
COM Comoros 20.5% 20.9% 23.5% 19.6% 20.1% 21.3% 20.4% 20.8% 22.1% 4.27 34.15
CPV Cabo Verde 3.8% 4.5% 3.4% 3.3% 4.1% 3.1% 3.6% 4.4% 3.4% 3.88 -2.16
CRI Costa Rica 0.6% 0.6% 1.1% 1.0% 1.0% 2.2% 0.9% 0.9% 2.0% -2.82 27.80
CUB Cuba 0.4% 0.4% 0.8% 0.7% 0.7% 1.5% 0.6% 0.6% 1.4% -3.72 36.62
DJI Djibouti 3.9% 3.6% 6.8% 6.2% 6.0% 13.4% 5.5% 5.3% 12.0% -3.78 37.24
DMA Dominica 0.3% 0.3% 0.5% 0.5% 0.5% 1.0% 0.4% 0.4% 0.9% -0.02 0.15
DOM Dominican Republic 0.5% 0.4% 0.8% 0.7% 0.7% 1.6% 0.7% 0.6% 1.4% -4.29 42.28
DZA Algeria 1.4% 1.3% 2.5% 2.2% 2.2% 4.8% 2.0% 1.9% 4.3% -46.91 462.05
ECU Ecuador 0.8% 0.8% 2.3% 2.9% 2.8% 7.5% 2.0% 1.9% 5.3% -7.53 275.40
(to be continued on next page)

43
 Poverty Rate, Total Poverty Rate, Farmer Poverty Rate, Rural Changes in
Population Population Population number of poor
people (1000)
ISO Country S0 S1 S2 S0 S1 S2 S0 S1 S2 S1 S2
EGY Egypt, Arab Rep. 0.3% 0.4% 0.3% 0.3% 0.3% 0.2% 0.3% 0.4% 0.3% 56.43 -31.50
ERI Eritrea 18.9% 19.3% 21.6% 18.1% 18.5% 19.7% 18.8% 19.2% 20.4% 28.51 228.00
ETH Ethiopia 17.3% 17.6% 19.8% 16.6% 17.0% 18.0% 17.2% 17.6% 18.7% 368.61 2,947.73
FJI Fiji 0.2% 0.2% 0.3% 0.2% 0.2% 0.3% 0.2% 0.2% 0.3% 0.12 0.94
FSM Micronesia, Fed. Sts. 5.6% 6.6% 5.1% 4.9% 6.1% 4.6% 5.4% 6.5% 5.0% 1.26 -0.70
GAB Gabon 0.9% 1.0% 0.8% 0.8% 0.9% 0.7% 0.8% 1.0% 0.8% 3.26 -1.82
GEO Georgia 0.6% 0.6% 0.8% 0.6% 0.6% 0.8% 0.5% 0.6% 0.8% 1.23 9.59
GHA Ghana 5.2% 6.1% 4.7% 4.5% 5.6% 4.2% 4.9% 6.0% 4.6% 327.49 -182.85
GIN Guinea 19.3% 19.6% 22.1% 18.5% 18.9% 20.0% 19.2% 19.6% 20.8% 55.17 441.19
GMB Gambia, The 6.1% 7.1% 5.5% 5.3% 6.5% 4.9% 5.8% 7.1% 5.4% 29.71 -16.59
GNB Guinea-Bissau 21.8% 22.1% 24.9% 20.8% 21.3% 22.6% 21.6% 22.1% 23.4% 8.83 70.63
GNQ Equatorial Guinea 0.0% 0.1% 0.0% 0.0% 0.1% 0.0% 0.0% 0.1% 0.0% 0.09 -0.05
GRD Grenada 0.3% 0.3% 0.6% 0.5% 0.5% 1.1% 0.5% 0.4% 1.0% -0.03 0.26
GTM Guatemala 2.8% 3.7% 3.3% 2.8% 3.7% 3.3% 5.0% 6.3% 5.8% 195.32 119.09
GUY Guyana 1.8% 1.6% 3.1% 2.8% 2.8% 6.2% 2.6% 2.5% 5.5% -1.10 10.82
HND Honduras 0.7% 0.7% 1.0% 0.6% 0.7% 1.0% 0.6% 0.7% 0.9% 4.10 31.91
HTI Haiti 27.5% 28.0% 31.4% 26.3% 26.9% 28.5% 27.3% 27.9% 29.6% 61.72 493.56
IDN Indonesia 2.0% 1.9% 2.9% 2.5% 2.4% 3.8% 2.3% 2.1% 3.3% -236.4 2,530.64
IND India 3.9% 3.9% 4.0% 4.2% 4.2% 4.2% 4.7% 4.7% 4.8% -350.7 1,034.82
IRN Iran, Islamic Rep. 0.3% 0.3% 0.5% 0.5% 0.5% 1.0% 0.4% 0.4% 0.9% -19.46 191.64
IRQ Iraq 0.6% 0.5% 1.0% 0.9% 0.9% 2.0% 0.8% 0.8% 1.8% -24.76 243.90
JAM Jamaica 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.01 0.10
JOR Jordan 0.0% 0.0% 0.0% 0.0% 0.0% 0.1% 0.0% 0.0% 0.1% -0.16 1.58
KAZ Kazakhstan 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.04 0.35
KEN Kenya 19.3% 19.6% 22.1% 18.5% 18.9% 20.0% 19.2% 19.6% 20.8% 228.27 1,825.39
KGZ Kyrgyz Republic 1.1% 1.3% 1.0% 1.0% 1.2% 0.9% 1.1% 1.3% 1.0% 13.02 -7.27
KHM Cambodia 0.2% 0.2% 0.4% 0.1% 0.1% 0.3% 0.2% 0.2% 0.5% -1.78 46.02
KIR Kiribati 0.5% 0.5% 0.8% 0.5% 0.6% 0.8% 0.5% 0.5% 0.7% 0.04 0.31
KNA St. Kitts and Nevis 0.1% 0.1% 0.1% 0.1% 0.1% 0.2% 0.1% 0.1% 0.2% 0.00 0.03
LAO Lao PDR 15.1% 15.3% 17.2% 14.4% 14.8% 15.6% 15.0% 15.3% 16.2% 20.97 167.70
LBR Liberia 37.3% 37.9% 42.6% 35.7% 36.5% 38.7% 37.1% 37.8% 40.1% 43.68 349.33
LBY Libya 3.9% 3.6% 6.9% 6.2% 6.1% 13.5% 5.6% 5.4% 12.2% -23.59 232.31
LKA Sri Lanka 0.3% 0.6% 0.4% 0.5% 0.8% 0.6% 0.4% 0.7% 0.4% 63.91 13.22
LSO Lesotho 19.3% 19.7% 22.1% 18.5% 18.9% 20.1% 19.2% 19.6% 20.8% 8.89 71.11
MAR Morocco 0.5% 0.5% 0.9% 0.8% 0.8% 1.8% 0.7% 0.7% 1.6% -15.02 147.92
MDA Moldova 0.0% 0.0% 0.1% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.00 0.15
(to be continued on next page)

44
 Poverty Rate, Total Poverty Rate, Farmer Poverty Rate, Rural Changes in number of
Population Population Population poor people (1000)
ISO Country S0 S1 S2 S0 S1 S2 S0 S1 S2 S1 S2
MDG Madagascar 36.2% 36.8% 41.4% 34.6% 35.4% 37.5% 36.0% 36.7% 38.9% 229.30 1,833.64
MDV Maldives 2.5% 2.5% 2.8% 2.4% 2.4% 2.6% 2.5% 2.5% 2.7% 0.17 1.37
MEX Mexico 0.2% 0.2% 0.4% 0.4% 0.4% 0.8% 0.3% 0.3% 0.7% -24.75 243.72
MHL Marshall 0.2% 0.1% 0.3% 0.3% 0.2% 0.5% 0.2% 0.2% 0.5% -0.01 0.08
Islands
MLI Mali 22.4% 22.8% 25.7% 21.5% 22.0% 23.3% 22.3% 22.8% 24.2% 107.83 862.31
MNE Montenegro 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.00 0.01
MNG Mongolia 4.5% 3.9% 6.3% 4.2% 3.1% 6.3% 5.6% 4.6% 7.9% -21.22 61.84
MOZ Mozambique 26.5% 27.0% 30.3% 25.4% 25.9% 27.5% 26.4% 26.9% 28.5% 170.79 1,365.75
MRT Mauritania 4.2% 5.0% 3.8% 3.7% 4.6% 3.4% 4.0% 4.9% 3.8% 38.19 -21.32
MUS Mauritius 0.1% 0.1% 0.1% 0.1% 0.1% 0.1% 0.1% 0.1% 0.1% 0.23 -0.13
MWI Malawi 44.6% 47.8% 31.3% 50.1% 53.3% 35.3% 48.1% 51.3% 33.9% 899.22 -3,734.74
NAM Namibia 5.8% 6.8% 5.2% 5.0% 6.2% 4.7% 5.5% 6.7% 5.1% 30.44 -16.99
NER Niger 9.4% 13.8% 10.4% 9.0% 13.1% 9.7% 10.9% 16.0% 12.0% 1,370.40 305.66
NGA Nigeria 29.1% 28.8% 40.6% 31.4% 31.1% 44.2% 33.8% 33.4% 46.5% -861.66 29,679.82
NIC Nicaragua 3.7% 4.8% 4.4% 6.9% 8.9% 8.0% 7.6% 9.6% 8.9% 77.61 50.42
NPL Nepal 3.8% 6.7% 5.2% 3.7% 6.6% 5.0% 2.9% 5.8% 4.2% 1,141.53 535.01
OMN Oman 0.2% 0.2% 0.4% 0.4% 0.4% 0.8% 0.3% 0.3% 0.7% -0.66 6.48
PAK Pakistan 5.1% 14.6% 3.2% 5.1% 14.6% 3.2% 5.7% 15.5% 3.4% 22,359.31 -4,331.48
PAN Panama 2.7% 2.9% 2.9% 8.2% 8.9% 8.6% 5.3% 5.8% 5.6% 11.24 7.59
PER Peru 0.9% 0.8% 2.2% 1.6% 1.5% 4.6% 2.1% 2.0% 5.5% -14.43 465.71
PHL Philippines 0.7% 0.7% 1.0% 0.7% 0.7% 1.0% 0.6% 0.7% 0.9% 50.01 388.75
PNG Papua New 15.9% 16.2% 18.2% 15.2% 15.6% 16.5% 15.8% 16.2% 17.1% 29.10 232.71
Guinea
PRY Paraguay 0.3% 0.3% 0.4% 0.3% 0.3% 0.4% 0.3% 0.3% 0.4% 1.33 10.37
RWA Rwanda 38.1% 40.7% 30.5% 41.0% 43.8% 32.7% 42.8% 45.6% 34.4% 456.40 -1,338.99
SDN Sudan 3.6% 4.2% 3.2% 3.1% 3.9% 2.9% 3.4% 4.2% 3.2% 415.00 -231.71
SEN Senegal 14.9% 15.2% 17.0% 14.3% 14.6% 15.5% 14.8% 15.1% 16.0% 53.39 426.94
SLE Sierra Leone 12.8% 20.6% 13.7% 14.4% 21.6% 15.1% 14.7% 22.2% 15.3% 663.28 82.07
SLV El Salvador 1.8% 1.7% 3.2% 2.9% 2.8% 6.4% 2.6% 2.5% 5.7% -10.11 99.58
SRB Serbia 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.05 0.41
STP Sao Tome and 5.1% 6.0% 4.6% 4.4% 5.5% 4.1% 4.8% 5.9% 4.5% 2.07 -1.16
Principe
SUR Suriname 0.6% 0.6% 0.8% 0.6% 0.6% 0.9% 0.5% 0.6% 0.8% 0.20 1.56
SVN Slovenia 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.00 0.02
SWZ Swaziland 18.1% 18.4% 20.7% 17.3% 17.7% 18.8% 18.0% 18.4% 19.5% 4.74 37.92
SYC Seychelles 0.0% 0.1% 0.0% 0.0% 0.0% 0.0% 0.0% 0.1% 0.0% 0.01 0.00
(to be continued on next page)

45
 Poverty Rate, Total Poverty Rate, Farmer Poverty Rate, Rural Changes in
Population Population Population number of poor
people (1000)
ISO Country S0 S1 S2 S0 S1 S2 S0 S1 S2 S1 S2
SYR Syrian Arab Republic 0.4% 0.3% 0.6% 0.6% 0.5% 1.2% 0.5% 0.5% 1.1% -7.57 74.57
TCD Chad 27.6% 28.1% 31.5% 26.4% 27.0% 28.6% 27.4% 28.0% 29.7% 91.17 729.06
TGO Togo 17.2% 17.5% 19.7% 16.5% 16.8% 17.9% 17.1% 17.5% 18.5% 26.82 214.45
THA Thailand 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.58 4.53
TJK Tajikistan 0.9% 0.8% 1.2% 1.0% 0.8% 1.6% 1.1% 0.9% 1.4% -10.60 28.30
TKM Turkmenistan 0.9% 1.0% 1.3% 0.9% 1.0% 1.4% 0.9% 0.9% 1.3% 3.29 25.56
TLS Timor-Leste 9.2% 7.1% 5.9% 10.7% 8.2% 7.0% 11.6% 9.1% 7.7% -40.64 -64.51
TON Tonga 0.3% 0.4% 0.3% 0.3% 0.4% 0.3% 0.3% 0.4% 0.3% 0.07 -0.04
TTO Trinidad and Tobago 0.9% 0.8% 1.5% 1.4% 1.3% 2.9% 1.2% 1.2% 2.7% -0.90 8.82
TUN Tunisia 0.3% 0.3% 0.5% 0.4% 0.4% 1.0% 0.4% 0.4% 0.9% -2.62 25.81
TUV Tuvalu 0.2% 0.2% 0.3% 0.2% 0.2% 0.3% 0.2% 0.2% 0.3% 0.00 0.01
TZA Tanzania 53.7% 55.8% 47.1% 60.6% 62.5% 53.4% 60.5% 62.4% 53.5% 1,741.17 -5,395
UGA Uganda 18.3% 19.9% 11.8% 20.5% 22.3% 13.4% 20.6% 22.5% 13.3% 965.96 -3,880
UKR Ukraine 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.05 0.41
UZB Uzbekistan 0.1% 0.2% 0.1% 0.1% 0.2% 0.1% 0.1% 0.2% 0.1% 8.37 -4.67
VEN Venezuela, RB 1.4% 1.3% 2.4% 2.2% 2.1% 4.7% 1.9% 1.9% 4.2% -39.03 384.39
VNM Vietnam 0.6% 0.5% 1.0% 0.7% 0.6% 1.2% 0.8% 0.7% 1.4% -70.80 423.86
VUT Vanuatu 2.8% 3.3% 2.5% 2.4% 3.0% 2.3% 2.6% 3.2% 2.5% 1.79 -1.00
WSM Samoa 0.1% 0.2% 0.1% 0.1% 0.2% 0.1% 0.1% 0.2% 0.1% 0.05 -0.03
YEM Yemen, Rep. 2.2% 2.4% 3.8% 2.0% 2.3% 3.9% 2.6% 2.9% 4.6% 100.35 699.00
ZAF South Africa 2.5% 2.9% 2.2% 2.2% 2.7% 2.0% 2.4% 2.9% 2.2% 236.19 -131.8
ZMB Zambia 51.5% 54.1% 42.0% 70.4% 72.9% 59.4% 67.3% 69.7% 56.6% 640.75 -2,337
ZWE Zimbabwe 27.1% 27.6% 31.0% 25.9% 26.5% 28.1% 26.9% 27.5% 29.1% 85.64 684.81
Source: Authors’ calculations.
Note: Poverty is defined by the $1.90 PPP 2011 threshold.
Countries for which detailed household survey information is available. These countries are used as
representative for their respective clusters.

46
7 References
Amiti, M. and Konings, J. (2007), ‘Trade liberalization, intermediate inputs, and productivity:
evidence from Indonesia’ American Economic Review 97(5): 1611-34.
Barro, R. and Sala-i-Martin, X. (1992), ‘Convergence’ Journal of Political Economy 100(2): 223-
51.
Ben-David, D. (1994), ‘Income disparity among countries and the effects of freer trade’ in
Pasinetti, L. and Solow, R. (eds.), Economic Growth and the Structure of Long Run
Development, London: Macmillan, 45-64.
Blundell, and Macurdy, T. (1998), ‘Labor Supply: a review of alternative approaches’ Chapter 27
in Ashenfelter, O. and Card, D. eds. Handbook of Labor Economics, Volume 3,
Amsterdam: Elsevier Science.
Bouet, A., Estrades, C. and Laborde, D. (2011), Household heterogeneity in a global CGE model:
an illustration with the MIRAGE poverty module, International Food Policy Research
Institute, Washington DC.
Deaton, A. (1989) ‘Rice prices and income distribution in Thailand: a non-parametric analysis’
The Economic Journal 99 395:1–37.
Deaton, A. and Muellbauer, J. (1981) ‘Functional forms for labor supply and commodity demands
with and without quantity restrictions’ Econometrica, 49(6):1521–32.
Decreux Y. and Valin H. (2007) MIRAGE, Updated Version of the Model for. Trade Policy
Analysis. Focus on Agriculture and Dynamics. CEPII Working Paper. No 2007 – 15.
October. Paris, France.
Dervis, K. (2012), ‘World economy: convergence, divergence and interdependence’ Finance and
Development, September 12-14.
Doblin, C. (1991), ‘Capital Stock and Capital Output Ratios: A Comparison between the United
States, United Kingdom and the Federal Republic of Germany, IIASA Working Paper WP-
91-29
Dowrick, S. and Nguyen, D.T. (1989), ‘OECD Comparative economic growth 1950-85: catch-up
and convergence’ American Economic Review 79(5):1010-30.
Ferreira, F., Jolliffe, D., and Prydz, E. (2015), ‘The international poverty line has just been raised
to $1.90 a day, but global poverty is basically unchanged. How is that even possible?’ Let’s
Talk Development Blog, World Bank.
Fukase, E. and Martin, W. (2016), ‘Who Will Feed China in the 21st Century? Income Growth
and Food Demand and Supply in China’ Journal of Agricultural Economics Open access
doi: 10.1111/1477-9552.12117.
Headey, D. and Martin, W. (2016), ‘Food Prices, Poverty and Food Security’ Annual Review of
Resource Economics, forthcoming.

47
Ivanic, M. and Martin, W. (2014) ‘Short- and Long-Run Impacts of Food Price Changes on
Poverty’ Policy Research Working Paper 7011, World Bank, Washington DC.
Ivanic, M. and Martin, W. (2016), ‘Sectoral productivity growth and poverty reduction: national
and global impacts’ Mimeo, International Food Policy Research Institute, Washington DC.
Kaldor, N. (1957). ‘A model of economic growth’ The Economic Journal 67:268: 591–624.
Keane, M. and Rogerson, R. (2012) ‘Micro and Macro Labor Supply Elasticities: A Reassessment
of Conventional Wisdom’ Journal of Economic Literature 50(2):464–76.
Lakatos, C., Maliszewska, M., Osorio-Rodarte, M. and Go, D. (2015) ‘China’s Slowdown and
Rebalancing: Potential Growth and Poverty Impacts on Sub-Saharan Africa’ Mimeo,
World Bank.
Loayza, N., & Raddatz, C. (2010). ‘The composition of growth matters for poverty alleviation’
Journal of Development Economics 93:137–51.
Lucas, R. E. (1988), ‘On the mechanics of economic growth’ Journal of Monetary Economics
22:3-42.
Martin, W., and Alston, J. M. (1997). ‘Producer surplus without apology? Evaluating Investments
in R&D’. Economic Record, 73(221):146–58.
O’Neill, J. (2001), ‘Building Better Global Economic BRICs’ Goldman Sachs Global Economic
Paper 66.
Quah, D. (1995), ‘Empirics for Economic Growth and Convergence’ CEPR Discussion Paper 253,
July.

48