Share This
June 2001 Cover Art

Search Finance & Development

Advanced Search

About F&D


Back Issues

Write Us

Copyright Information

Free Email Notification

Receive emails when we post new items of interest to you.

Subscribe or Modify your profile

Finance & Development
A quarterly magazine of the IMF
June 2001, Volume 38, Number 2

Macroeconomic Policies and Poverty Reduction: Some Cross-Country Evidence
Paul Cashin, Paolo Mauro, and Ratna Sahay

What is currently known about how countries' choices of macroeconomic policies affect their incidences of poverty, and what are the most promising directions for further investigation of this important relationship?

Reducing poverty is the key challenge facing the world community, and there is an important debate on the policies that may help attain that objective and on how the international financial institutions can help. This article looks at the Human Development Index (HDI), prepared by the United Nations Development Program (UNDP), as a measure of well-being and reviews the relationship between macroeconomic policies and poverty reduction in a sample of countries over recent decades. Its focus is on the interaction between macroeconomic policies—which are at the core of the IMF's mandate—and poverty.

Recent improvements in well-being

Over the past few decades, indicators of well-being have improved in the vast majority of countries, though with major variation both within and across countries. The HDI is defined as the arithmetic average of a country's achievements in three basic dimensions of human development. These include longevity (measured by life expectancy at birth); educational attainment (measured by a combination of the adult literacy rate and the enrollment ratio in primary, secondary, and tertiary education); and living standards (measured by GDP per capita in U.S. dollars at purchasing power parity). The HDI ranges between zero (low human development) and one (high human development). Chart 1: Human Development Index (HDI) and Gini coefficient in 1990

The HDI has a number of advantages: it moves beyond per capita income alone as a measure of well-being; it is compiled with uniform data sources and methodology over time and across countries; and it is available for 100 countries on a consistent basis for the period 1975-98. (For a critical assessment of the HDI, see Ravallion, 1997.) The HDI does not capture income inequality directly, though it is closely correlated with it (Chart 1). Moreover, for countries with a given per capita income, those where income is distributed more evenly tend to display greater average longevity and educational attainment, and therefore a higher HDI value. Chart 2: Developing countries: Human Development Index (HDI) and alternative poverty measures, 1998

Both the HDI and per capita income are highly correlated with other widely used measures of poverty (Chart 2), such as the share of the population with income of less than $1 per day (a World Bank measure) and the share of the population that is undernourished (a Food and Agriculture Organization measure).

Table 1 provides a list of selected countries for which 1998 HDI data are available, categorized by region and in descending order of their HDI values. In general, the African and Asian countries had relatively low values, while industrial, transition, and Latin American countries had relatively high ones. The HDI improved in almost all countries between 1975 and 1998, and the median HDI was significantly higher in 1998 (0.73) than in 1975 (0.62). At the same time, there was little change in the ranking of countries by HDI during this period: the correlation between countries' HDI ranks for 1975 and 1998 is 0.98.

Table 1
Human Development Index (HDI), selected countries, 1998

≤0.50 0.51-0.70 0.71-0.80 > 0.80

Sudan (0.48)
Mauritania (0.45)
Nigeria (0.44)
Congo, Dem. Rep. of the (0.43)
Zambia (0.42)
Côte d'Ivoire (0.42)
Senegal (0.42)
Tanzania (0.41)
Uganda (0.41)
Angola (0.40)
Mozambique (0.34)
Ethiopia (0.31)
Niger (0.29)
Sierra Leone (0.25)

Lao People's Dem. Rep. (0.48)
Nepal (0.47)
Bangladesh (0.46)

Middle East
Yemen (0.45)

Western Hemisphere
Haiti (0.44)
South Africa (0.70)
Botswana (0.59)
Gabon (0.59)
Ghana (0.56)
Zimbabwe (0.56)
Cameroon (0.53)
Kenya (0.51)
Congo, Rep. of (0.51)

Vietnam (0.67)
Indonesia (0.67)
India (0.56)
Pakistan (0.52)

Transition economies
Moldova (0.70)
Uzbekistan (0.69)
Tajikistan (0.66)

Middle East
Syrian Arab Republic (0.66)
Egypt (0.62)
Iraq (0.58)

Western Hemisphere
Bolivia (0.64)
Nicaragua (0.63)
Guatemala (0.62)
Thailand (0.74)
Philippines (0.74)
China (0.71)

Transition economies
Bulgaria (0.77)
Russia (0.77)
Romania (0.77)
Georgia (0.76)
Ukraine (0.74)
Azerbaijan (0.72)
Albania (0.71)

Middle East
Saudi Arabia (0.75)
Jordan (0.72)
Iran, Islamic Rep. Of (0.71)

Western Hemisphere
Mexico (0.78)
Colombia (0.76)
Brazil (0.75)
Peru (0.74)
Europe/Industrial countries
Canada (0.93)
United States (0.93)
Australia (0.93)
Japan (0.92)
United Kingdom (0.92)
France (0.92)
Germany (0.91)
Italy (0.90)
Spain (0.90)

Singapore (0.88)
Hong Kong SAR (0.87)
Korea, Rep. of (0.85)

Transition economies
Czech Republic (0.84)
Hungary (0.82)
Poland (0.81)

Middle East
Israel (0.88)
Kuwait (0.84)

Western Hemisphere
Argentina (0.84)
Chile (0.83)
Uruguay (0.82)

Source: UNDP, 2000.

Despite the basically unchanged ranking of countries, there is some evidence that low-HDI countries have been catching up, though slowly, with high-HDI countries. Considering the countries for which data are available for both 1975 and 1998, Table 2 shows that the countries that were in groups with relatively low HDI values in 1975 tended to display a greater improvement in HDI than other countries over the next two decades. The countries that displayed the greatest improvement in HDI from 1975 to 1998 are from Africa and Asia: Nepal (by 63 percent), Mali (53 percent), Pakistan (48 percent), The Gambia (47 percent), and Chad (45 percent). The countries with the least improvement were Guyana (5 percent), the Democratic Republic of the Congo (3 percent), Romania (3 percent), and Zambia (-5 percent).

Table 2
HDI transition matrix
(excluding industrial countries)1

  Change in HDI by 1998
  < 0.10 0.10-0.15 0.16-0.20

HDI in 1975

Low (0-0.5)
Burkina Faso, Burundi, Central African Republic, Dem. Rep. of the Congo, Rep. of Congo, Côte d'Ivoire, Guinea-Bissau, Kenya, Madagascar, Malawi, Niger, Togo, Zambia Bangladesh, Benin, Botswana, Cameroon, Chad, The Gambia, Ghana, Lesotho, Mali, Mauritania, Nigeria, Papua New Guinea, Senegal, Sudan Egypt, India, Indonesia, Morocco, Nepal, Pakistan

Medium (0.5-0.7)
Fiji, Guyana, Jamaica, Mexico, Nicaragua, Paraguay, Philippines, South Africa, Zimbabwe Bolivia, Brazil, Colombia, Dominican Republic, Ecuador, El Salvador, Guatemala, Honduras, Islamic Rep. of Iran, Mauritius, Peru, Sri Lanka, Swaziland, Syrian Arab Republic, Thailand, Turkey Algeria, China, Rep. of Korea, Malaysia, Saudi Arabia, Tunisia

High (0.7-0.8)
Argentina, Costa Rica, Hungary, Panama, Romania, Trinidad and Tobago, United Arab Emirates, Uruguay, Venezuela Chile, Hong Kong SAR, Malta Singapore

Source: UNDP, 2000.
1Industrial countries are excluded from the table because they almost invariably began with very high HDI values in 1975 and tended to display rather small improvements over the following two decades.

Macro policies, well-being, and HDI

Poverty in a given country can be reduced by fostering per capita GDP growth—that is, by increasing the total resources available to the population—and by increasing the share of those resources going to its poorer segments. A widely held view is that economic growth can be fostered by a set of policies aimed at promoting macroeconomic stability (low and stable inflation, low budget deficits, and sustainable external debt), openness to international trade, education, and the rule of law. The findings of many studies based upon cross-country evidence are consistent with that view, although the evidence on whether each individual policy among those listed above raises economic growth is typically far from conclusive (see Levine and Renelt, 1992).

Casual observation also broadly suggests an association between sound macroeconomic policies and rapid improvement in HDI. Table 3 shows that—within low-HDI, medium-HDI, and high-HDI groups of countries—lower inflation, lower fiscal deficits, lower variability of inflation, lower external debt, more effective rule of law, lower black-market foreign exchange premium, and lower frequency of financial crises were associated with greater improvement in HDI. As in the economic growth literature, though, it is difficult to show conclusively whether individual policies cause countries to experience more rapid improvements in well-being.

There is also a debate about the policies that improve the well-being of the poorer segments of the population for a given growth rate of GDP per capita and an even more fervent debate about whether certain policies imply a trade-off between increasing total available resources (increasing growth rates) and improving their distribution. In the latter respect, there seems to be broad agreement that policies aimed at improving basic education and health care can both increase economic growth and improve distribution, but, of course, there certainly is no consensus on the policies that will be most effective in improving education and health care.

Table 3
Macroeconomic performance and the HDI, 1975-98

Table 3: Macroeconomic performance and the HDI, 1975-98

Sources: UNDP, 2000; World Bank, World Development Indicators; and IMF, International Financial Statistics.
10-100 index for 1998 with a higher number indicating stronger property rights.
2Defined as [(parallel exchange rate/official exchange rate)-1] x 100.
3Countries include Botswana, Burkina Faso, Burundi, Cameroon, the Central African Republic, the Democratic Republic of the Congo, the Republic of Congo, Côte d'Ivoire, Ghana, Guinea-Bissau, Kenya, Lesotho, Madagascar, Malawi, Mauritania, Niger, Papua New Guinea, Senegal, and Togo.
4Countries include Bangladesh, Benin, Chad, Egypt, The Gambia, India, Indonesia, Mali, Morocco, Nepal, Nigeria, Pakistan, and Sudan.
5Countries include Brazil, Colombia, Dominican Republic, Ecuador, El Salvador, Fiji, Guyana, Jamaica, Mauritius, Mexico, Nicaragua, Paraguay, Peru, the Philippines, South Africa, Sri Lanka, and Zimbabwe.
6Countries include Algeria, Bolivia, China, Guatemala, Honduras, the Islamic Republic of Iran, Korea, Malaysia, Saudi Arabia, Swaziland, the Syrian Arab Republic, Thailand, Tunisia, and Turkey.
7Countries include Argentina, Costa Rica, Hungary, Panama, Romania, Trinidad and Tobago, the United Arab Emirates, Uruguay, and Venezuela.
8Countries/economies include Chile, Hong Kong SAR, Israel, Malta, and Singapore.

To examine whether macroeconomic policies have a direct impact on poverty, we attempted, for a sample of countries, to estimate the relationship between economic policies and improvements in HDI values (or other indicators of well-being, such as infant mortality and life expectancy), for a given rate of growth of GDP per capita. The rationale for doing this is that when particular policies bring about greater improvements in life expectancy in a country than would be expected based on its economic growth rate alone, those policies are likely to be especially beneficial to the poorer segments of its population. This makes it possible, in principle, to estimate how much particular policies can contribute to the improvement in well-being that is unrelated to economic growth.

We examined a large set of economic variables related to economic policies. It included many of the variables that previous researchers have used to analyze the determinants of economic growth, such as inflation and its variance; budget deficits, government spending, and foreign aid as shares of GDP; indicators of openness, such as the ratio of foreign trade to GDP and the black-market foreign exchange premium; and indices of the strength of the rule of law. Other variables we included have received less attention in previous work, such as the occurrence and length of exchange rate or banking crises, and initial external debt as a share of GDP.

Using a cross-country regression approach, we did not find strong evidence that any of the variables is individually associated with either pro-poor or anti-poor economic growth. Of course, this finding does not by any means constitute proof that these policies do not affect a country's success in reducing the incidence of poverty. On the contrary, it suggests that alternative research approaches are needed to find strong and convincing evidence on the direction and strength of the effects these policies have on the poor. Other studies have relied on panel regressions, which use the information contained in the variation both over time (for individual countries) and between countries. These studies have generally not found significant evidence of links between particular policies and improvements in the well-being of the poor, with the possible exception of a significant association between lower inflation and improved well-being (see, for example, Easterly and Fischer, 2001).


On the basis of cross-country studies, the current state of knowledge is that economic growth is associated with improvements in indicators of well-being. Little has been conclusively proved, however, about which macroeconomic policies help raise economic growth, and even less is known about the individual policies that help reduce poverty for a given rate of economic growth. Of course, examining a wide range of country experiences has made it possible for policymakers to develop some expertise about how they can go about achieving these important economic objectives. Nonetheless, the effectiveness of specific policies still needs to be confirmed by systematic empirical studies, which leaves economists with an important and comprehensive research agenda. Further cross-country studies appear unlikely to yield much more useful information about the effects of macroeconomic policies on poverty unless the dynamic effects of these policies are appropriately considered. Greater payoffs are likely to be obtained by conducting studies based on regional or national survey data (for households or firms) for periods encompassing clearly identifiable macroeconomic shocks. However, the number of developing countries for which reliable surveys are currently available is relatively limited. Data-collection efforts undertaken to remedy this deficiency may greatly contribute to our knowledge about the links between macroeconomic policies and poverty reduction.

William Easterly and Stanley Fischer, 2001, "Inflation and the Poor," World Bank Policy Research Working Paper No. 2335 (Washington), forthcoming in the
Journal of Money, Credit and Banking.
Ross Levine and David Renelt, 1992, "Sensitivity Analysis of Cross-Country Growth Regressions,"
American Economic Review, Vol. 82 (September), pp. 942-63.
Martin Ravallion, 1997, "Good and Bad Growth: The Human Development Reports,"
World Development, Vol. 25 (May), pp. 631-38.
United Nations Development Program (UNDP), 2000,
Human Development Report 2000 (New York and Oxford: Oxford University Press for the UNDP).

Paul Cashin and Paolo Mauro are Economists in, respectively, the Commodities and Special Issues Division and the Developing Country Studies Division of the IMF's Research Department.

Ratna Sahay is Advisor to the IMF's First Deputy Managing Director.