The renewed sense of urgency for faster, deeper poverty reduction has spawned a growing debate on the determinants of poverty and strategies for alleviating it. A key point of reference in the literature has been the impact of rapid economic growth on poverty reduction in East Asia. Indeed, recent empirical work has found that economic growth is a key driver of poverty reduction, with little or no direct role for economic policies (once account is taken of the effects of economic growth). The fact that recent research finds that public policies play only a small role, or no role, in directly lowering poverty motivated this study.
A major challenge for research in this area stems from the scarcity of poverty-related data for low-income countries. There are several ways of measuring poverty. For example, the United Nations, in setting its Millennium Development Goals, defines as poor those who live on less than $1 a day. In this study, poverty is defined as the average income of the lowest 20 percent of the population in the income distribu-tion—a measure used by other researchers for econometric analysis. We used data from a large set of developed and developing countries during 1950-99.
An initial, simple exercise of identifying correlations between variables (see box, below, for a summary of methodology and variables; see the Working Paper for details) indicates that, on average, countries in which the poor have incomes higher than those of their counterparts in other countries are characterized by higher macroeconomic stability, lower income inequality, better internal environments, more democratic political institutions and better governance, a better-educated population, more open trade regimes, and higher levels of financial development.
Methodology and variables
Econometric work with data for several countries spanning a number of years faces some daunting challenges: country-and time-specific effects, endogeneity of explanatory variables, omission of relevant variables, and uncertainty about the effectiveness of the underlying statistical model, among other things. In this study, we used two methodologies.
The first is a traditional framework to reproduce some of the existing results in the literature and to show that the methodology is perhaps inadequate for such analysis. In particular, while correcting for a number of the econometric problems noted, this framework cannot address model uncertainty, which arises because of the lack of clear theoretical guidance on the choice of explanatory variables.
The second methodology—a check for robustness—corrected for this problem along with the others. In simple terms, it attempts to account for all possible combinations of explanatory variables in the statistical regressions. We used 18 potential explanatory variables, basing the robustness check inferences on the results ofa very large number of regressions.
The 18 potential explanatory variables accounted for
- overall average income (which was kept in all the regressions);
- the internal environment or resources (including natural resources and ethnicity);
- institutions / governance (including rule of law and level of democracy);
- human capital (including educational outcomes and life expectancy);
- physical capital (including private and public investment);
- macroeconomic stability (including inflation and fiscal balance);
- government size (ratio of government consumption to GDP);
- the trade regime (including share of exports and imports in GDP);
- the external environment (including changes in the terms of trade); and
- financial development (including the ratio of broad money to GDP).
The econometric analysis confirmed the importance of economic growth in raising the incomes of the poor, although the elasticity of the incomes of the poor with respect to average income is sensitive to the variables that are included in the statistical regressions (an elasticity of one implies that a 1 percent increase in overall average income raises the average income of the poor by 1 percent). We also found that the econometric estimates derived from the traditional statistical regression framework were not robust. In particular, the traditional framework was unable to say anything about the direct impact of public policies on the incomes of the poor. As a result, one would tend to conclude that such policies may have no independent effect and alleviate poverty only through their influence on economic growth.
But, using a second framework to check for robustness (see box, page 128), we found that the data contained other interesting information, especially about the role of public policies. The findings confirmed the relationship between economic growth and poverty reduction, although the relationship is less than one to one. More specifically, for a given target for poverty reduction over a certain period of time, the economic growth rates required could exceed what can reasonably be expected (compared with what would be required if an increase in economic growth resulted in an increase of one to one or greater in the incomes of the poor). We also found that certain public policies have a direct impact on the incomes of the poor, even after controlling for the effect of economic growth. These include policies that lower inflation, shrink the size of the government, promote financial development, and raise the educational level. The policy-related variables are considered “super pro poor” because they raise the incomes of the poor directly, as well as indirectly, through economic growth. The direct and indirect effects are mutually reinforcing, and there are thus no identified trade-offs between growth promotion and poverty alleviation. The results also indicate that the poor are significantly vulnerable to adverse movements in the terms of trade.
Another interesting result relates to the importance of secondary school (and not primary school) enrollment in directly raising the incomes of the poor. This result appears puzzling in view of the notion that primary education plays a central role in lowering poverty. Our finding does not contradict this view; instead, given that social indicators for health and education are highly correlated, this result tends to highlight the importance of public policies that enhance social outcomes. In addition, to the extent that the secondary school enrollment ratio represents the quality of the accumulation of investment in basic education, health, and nutrition, this result points to the importance of the quality of investment in human capital in poverty reduction efforts.
Some of the statistically nonsignificant results are also noteworthy. For example, a number of variables—such as trade openness, the investment rate, the extent of democracy, life expectancy at birth, and the extent of civil wars—that have been shown in the empirical literature to affect economic growth do not directly influence the incomes of the poor (once the level of overall average income has been accounted for). A corollary is that, to the extent that it is closely related to globalization, trade openness does not appear to hurt the incomes of the poor but, instead, has a positive effect by boosting overall economic growth.
We found that raising overall economic growth is indeed key to poverty reduction. The bottom line is that growth—and lots of it—is needed so that the extra income generated can also benefit the poor. But, while growth is a necessary condition for poverty reduction, it is by no means sufficient. Countries need to implement policies that focus on creating an enabling environment for the poor to participate in, and benefit from, the growth process.
Our results also suggest that the IMF has a key role to play in helping to alleviate poverty. The IMF’s primary mandates include assisting countries in the design of policies that lead to low and predictable inflation; reorienting government resources to productive outlays (including health and education) by, for example, lowering unproductive government consumption; and deepening the financial sector. In promoting such policies successfully, the IMF can do a lot to lower poverty in the world, and it needs to continue focusing on this issue in sub-Saharan Africa, Latin America and the Caribbean, and parts of Europe and Asia where large segments of the populations live in absolute poverty. Coordination with the World Bank and other development agencies is important, as they also have a key role in focusing on, for example, social policies that lead to higher social outcomes in low-income countries, as well as on strategies to help these countries diversify their export bases (to lower their vulnerability to adverse terms of trade shocks).
The pro-poor public policies described in this paper enable the poor to participate in the growth process through the labor market or self-employment and to safeguard the purchasing power of their incomes. A stable macroeconomic environment, characterized by low and predictable inflation, makes it possible for the poor to do the latter. One argument in the economic literature is that inflation is a “harsh tax” on the poor because they are less likely than the rich to have access to financial hedging instruments protecting the real value of their wealth. Financial sector development also benefits the poor by facilitating access to credit and improving risk sharing and resource allocation. Educational achievement, facilitated by government investment in health and education, allows the poor to participate in the economic growth process through employment. The reduction of government consumption (used as a proxy for government size in this study) allows more scarce resources to be devoted to investment in health and education.
The existing empirical literature on poverty reduction leaves a number of questions unresolved. First, how can public policies have an independent effect on the incomes of the poor when the underlying framework controls for overall income as an explanatory variable, which itself incorporates the poor’s income? We believe that the independent effect occurs through a distributional channel, but how exactly? For such an evaluation, it would be necessary to specify relevant transmission mechanisms and to rigorously test their empirical relevance. It would also be important for more empirical work to be undertaken on policies that allow the poor to participate fully in, and benefit from, the economic growth process.
Second, it is not clear how relevant the recommendations derived from cross-country regression studies are for individual countries. There is, therefore, a need for more country-specific studies of experiences with poverty reduction, despite the dearth of data. Third, the impact of economic growth and the “super pro poor” policies on the incomes of the poor does not necessarily imply causation from one to the other. An examination of the issue of causation would require continuous data over long periods of time for a given country—but such data rarely exist.
Finally, the issue of how long it takes for public policies to make a significant dent in poverty needs careful consideration. Again, more poverty-related data are needed. An important challenge for the international community is to finance the collection and dissemination of more household-level poverty data over time in low-income countries.
Copies of Working Paper No. 02/118, “Is Growth Enough? Macroeconomic Policy and Poverty Reduction,” by Dhaneshwar Ghura, Carlos Leite, and Charalambos Tsangarides, are available for $15.00 each from IMF Publication Services. See page 131 for ordering information. The full text of the Working Paper is also available on the IMF’s website (www.imf.org).