Journal Issue

What Does It Take to Help the Poor?

International Monetary Fund. External Relations Dept.
Published Date:
June 2005
  • ShareShare
Show Summary Details

Spending on education and health can boost human capital in poor countries and help them reach the MDGs, but only if governments are held accountable

RECENT REPORTS by the Millennium Task Force—set up to measure progress toward the Millennium Development Goals (MDGs)—show that while economic growth has contributed to a rapid decline in extreme poverty in China, India, and other parts of Asia where the bulk of the poor live, little progress is being made in sub-Saharan Africa, where the incidence of extreme poverty is the highest. Progress toward meeting the other development goals has also been uneven, with gender equality, maternal mortality, and environment sustainability still lagging behind the targets. Vulnerability to pandemic diseases, including HIV/AIDS, remains elevated in most countries.

These trends underscore the need for better public policies to support growth and reduce poverty. But while the international community agrees that something needs to be done, how best to go about it remains the subject of vigorous debate. No one questions that human capital—in the form of better health status and higher levels of educational attainment—is a major building block for sustaining the productivity growth that would, in turn, spur broad-based economic growth in developing countries. But inefficiencies in the public provision of these services—due, for instance, to corruption or a lack of skilled workers—have led some to question whether just increasing public spending is the best route, especially given the role of other factors (such as income per capita) in determining social indicators. For that reason, we undertook a study to try to help policymakers evaluate the effects of different policies on social indicators and growth. This article examines our results, which show that while higher spending on health and education is worthwhile, poor governance and macroeconomic instability may offset the positive impact of social spending on growth and human development. But first it is helpful to review what past research has taught us.

Findings so far

What is the relationship between educational capital and growth? To date, researchers have mostly found a positive relationship between enrollment rates and/or years of schooling and GDP growth in developing countries. Moreover, a recent study (Coulombe, Tremblay, and Marchand, 2004), using a more refined measure of individual skills, found that a country with literacy scores above the sample’s average also experienced an above-average increase in annual per capita GDP growth. However, while results at the microeconomic level suggest that investing in education is an effective way to spur economic growth, macroeconomic evidence points to a weak relationship, at best, between education and growth.

How about building up health capital? Studies typically suggest that the health of a population matters greatly. Conceptually, a healthy person not only works more efficiently but can also devote more time to productive activities. Based on microeconomic evidence, many authors argue that health explains variations in wages at least as much as education. Research at the macro level also suggests that health capital positively influences aggregate output. Earlier studies have shown that up to one-third of annual GDP growth can be attributed to health capital, and an increase in life expectancy of one year is associated with an increase in the long-run growth rate of up to 4 percentage points in both developing and industrial countries (Bloom and Sevilla, 2004).

But it is not clear that higher government outlays on health and education always boost growth. Why might higher spending be ineffective? One reason is the macroeconomic effects of excessive public outlays. Empirical studies find a negative association between large fiscal deficits and growth in developing countries. If higher spending on health and education leads to ballooning fiscal deficits, the negative impact on macroeconomic stability and growth could more than offset the beneficial effects of such spending on social indicators. A second reason is poor governance. And a third is poorly targeted outlays. For example, spending on tertiary education may yield few benefits for children from low-income families who cannot even afford to complete secondary school.

It is also not clear that higher social outlays help improve social indicators. Why not? For one thing, poor institutions may reduce the quality of spending (for example, corruption may divert funds allocated for teaching supplies to “ghost” teachers). When this is the case, returns on education tend to be lower. Previous studies, however, generally fail to account for the impact of institutions on the effectiveness of social spending. Moreover, interactions among different types of social spending are important. Education spending, for example, is likely to be ineffective if students are in poor health. Such interactions have also not been captured in previous research.

Another shortcoming of the literature is that only a few studies have examined social spending, social indicators, and growth within an integrated system. Most focus on only one segment of the social spending-social indicators-growth nexus. That is, they either analyze the growth effects of improving education or health indicators or the impact of public spending on these indicators. But, as the examples above illustrate, capturing the potential feedback among these variables is key for predicting the likely impact of different policy interventions.

Using an integrated approach

Faced with these weaknesses in previous research, we decided to undertake a study—using a panel data set for 120 developing countries from 1975-2000—that would try to capture the potential feedback between social spending, social indicators, and growth. The building blocks of this approach can be represented by a simple economic model composed of three main relationships. The first describes output growth as a function of both physical and human capital inputs and labor; technology can be assumed to affect labor productivity. The second defines the accumulation of the stock of physical capital. The third describes the dynamic of human capital formation.

Solving these equations yields an expression for per capita output growth as a function of the initial income level, the stock of and new investment in human capital (separately for education and health), and the stock of and new investment in physical capital. If we combine specific expressions for the accumulation of physical and human capital with this growth equation, we obtain a system that links social spending to both human and physical capital accumulation and growth. Furthermore, we used a number of different techniques to address problems associated with endogeneity, measurement error, and omitted variable biases, and were able to find a consistent set of results.

Our results show that:

  • Both education capital and health capital contribute positively to output growth, but through slightly different channels. While both the stock and flow of education capital affect growth by a similar magnitude, the only direct effect of health capital on growth is through flows. The stock of health capital does, however, affect growth indirectly via its positive effect on physical investment.
  • Education spending has both an immediate and a lagged effect on education capital (see Chart 1). About two-thirds of the direct effect are realized in the first five years, with the remainder realized over the next five years. For example, the direct effect of an increase in education spending of 1 percentage point of GDP is associated with an increase in the composite enrollment rate of 6 percentage points within a five-year period and of another 3 percentage points in the following five-year period.
  • Health spending has a positive and significant immediate impact on health capital. For example, an increase in health spending of 1 percentage point of GDP is associated with a rise in the survival rate of children under age 5 of 0.2 percentage point, on average, in developing countries. However, lagged health spending has no further effect on health indicators. Such a result can be explained intuitively: first, higher health spending immediately translates into better health outcomes, while the full gains from education involve a longer gestation period as students complete their schooling. Also, unlike education outcomes, health conditions are not cumulative and have to be maintained by regular care. This constrains the effects of health spending to the short term.
  • Education and health capital have strong links. Health capital contributes to the accumulation of education capital, with an elasticity of about 1.3. This means, for example, that an increase in health capital of 10 percent would raise education capital by 13 percent. Therefore, in developing countries, good health conditions help promote better educational outcomes to a significant extent.
  • Improvements in gender equality improve health and education capital through higher access to basic services. For example, a 1 percentage point increase in the female share of enrolled students is associated with an increase of 2 percentage points in the composite enrollment rate, and an increase of 0.3 percentage point in the survival rate of children under age 5.
  • Higher income levels and greater human capital reinforce each other and contribute to a virtuous circle of growth and higher human capital.
  • Governance has a significant direct impact on the links between social spending and social indicators, with health spending being particularly sensitive to governance. Poor governance reduces growth mainly through its impact on human capital and investment. Countries with poor governance tend to have growth of about 1.6 percentage points lower per year than other countries. Similarly, poor governance is associated with a rate of investment to GDP that is 2 percentage points lower. The effect of governance on growth is transmitted through indirect channels via social indicators and investment.
  • The impact of education and health capital on growth varies in different country groups (see Chart 2). The impact of education capital on growth is more pronounced in low-income countries. Geographically, the impact of incremental improvements in education is highest in sub-Saharan Africa and lowest in Asia. Similarly, the impact of improving children’s chance of survival on growth is about 11 times as large in low-income countries when compared with middle-income countries, owing to high initial levels of child mortality in the poorest countries.

Chart 1.Positive impact

Both current and past education spending affect today’s human capital, but only new health spending matters.1

Source: IMF staff estimates.

1 Effects on an average developing country. Both current education and health spending have significant effects. Lagged education spending (expenditure that occurred five years earlier) is also found to have a significant effect, but not lagged health spending.

2Increases in the sum of the gross primary and secondary enrollment in percent of the corresponding school-age population.

3In numbers of additional children who survived until age 5, per 1,000 live births.

Chart 2.Low-income countries benefit the most


Investing in health and education yields the largest returns in low-income countries.1

Source: IMF staff estimates.

1 Data refer to elasticity estimates that measure the percentage change in real per capita GDP growth in response to a 1 percent change in the variables on the x-axis.

2Increases in the sum of the gross primary and secondary enrollment in percent of the corresponding school-age population.

3In numbers of additional children who survived until age 5, per 1,000 live births.

Policy implications

What are the policy implications of these findings? Using the model’s results, we conducted a set of simulations to assess the impact of different policy interventions for improving social indicators and economic growth, and reducing the poverty headcount. The simulations assess the impact of an increase in education spending, an increase in health spending, an improvement in governance, a reduction in the budget deficit, and a reduction in inflation (see Chart 3). Each of the simulations assumes that the policy environment remains unchanged (except, of course, in the case of simulated changes in inflation and the budget deficit).

Chart 3.Comparing policy choices

Poverty reduction can be achieved through a combination of policy measures as shown by the simulations below.

Source: IMF staff estimates.

Note: The assumptions for the simulations are: (1) the initial values for the indicators are averages for developing countries in the sample in 2000. For example, the initial real per capita GDP growth is 1.3 percent per year; (2) elasticities obtained in the regressions persist across the simulation period; and (3) the initial poverty headcount ratio is standardized to be 100 percent.

1Reduction in the poverty headcount ratio is measured in percent of the initial poverty headcount ratio.

Based on the simulation results, an increase in education spending of 1 percentage point of GDP is associated with 3 more years of schooling, on average, and an increase in annual growth of 1½ percentage points of GDP in 15 years, translating into a cumulative reduction of the initial poverty headcount ratio by about 17 percent. Similarly, an increase in health spending of 1 percentage point of GDP is associated with an increase of ½ percentage point in the survival rate of children under age 5 and a rise of ½ percentage point in annual per capita GDP growth, which corresponds to a cumulative reduction of the initial poverty headcount ratio by about 12 percent.

Enhancing governance is a powerful instrument to improve social indicators and growth. A change in the governance index from lower- to higher-than-average (implying reduced corruption) is associated with an immediate reduction in the child mortality rate, an increase in the composite enrollment rate, and a rise in per capita GDP growth of a magnitude similar to the spending increases described above. Through the reinforcing impact of higher income on human capital, this measure can lead to even better social indicators.

The growth effects of lower inflation (and hence its effects on poverty) are also substantial. Cutting the rate of inflation by 10 percentage points is associated with ½ percentage point increase in annual growth. Improving the fiscal balance by 1 percentage point of GDP is associated with an increase in per capita GDP growth by ½ percentage point when the deficit is high. However, while the initial impact on growth is comparable to that achieved with increased social spending, it does not bring additional lagged positive effects, as in the case with social spending. Furthermore, the effects from improving the fiscal balance in countries that have already achieved a modicum of macroeconomic stability are no longer significant.

No panacea

What are the implications for strategies aimed at meeting the MDGs? Given the positive effects of a number of different policies, efforts to meet the MDGs should be wide-ranging, as recently proposed in the reports of the Commission for Africa (an independent advisory panel set up by British Prime Minister Tony Blair) and of the UN Millennium Project (an independent advisory body to the UN Secretary-General). Spending increases should be accompanied by efforts to improve both the efficiency and targeting of public spending.

However, while improving human capital will have a salutary effect on growth, it is not by itself a panacea for unlocking the robust expansion in economic activity that is needed to achieve the MDGs. Social spending will be more effective in countries with better governance and lower levels of government expenditure, as the marginal returns to social spending tend to decline for countries that already spend substantially on these areas.

Emanuele Baldacci is a Senior Economist in the World Bank’s Human Development Department of the Latin American and Caribbean Region; Benedict Clements is an Advisor in the IMF’s Western Hemisphere Department; Qiang Cui is a Senior Research Officer in the IMF’s Fiscal Affairs Department; and Sanjeev Gupta is an Assistant Director in the IMF’s African Department.

This article is based on their recent working paper: Baldacci, Emanuele, Benedict Clements, Sanjeev Gupta, and Qiang Cui, 2004, “Social Spending, Human Capital and Growth in Developing Countries: Implications for Achieving the MDGs,” IMF Working Paper, 04/217 (Washington: International Monetary Fund).


    BloomDavid and JaypeeSevilla2004“The Effect of Health on Economic Growth: A Production Function Approach,”World DevelopmentVol. 32No. 1 pp. 1-13.

    • Crossref
    • Search Google Scholar
    • Export Citation

    The Commission for Africa2005Our Common Interest: Report of the Commission for Africa

    • Search Google Scholar
    • Export Citation

    CoulombeSergeJean-FrançoisTremblay and SylvieMarchand2004“Literacy Scores Human Capital and Growth Across Fourteen OECD Countries” International Adult Literacy Survey Monograph Series (Ottawa: Statistics Canada).

    • Search Google Scholar
    • Export Citation

    United Nations Millennium Project Task Force2005“A Practical Plan to Achieve the Millennium Development Goals” and the interim reports for selected countries. and

    • Search Google Scholar
    • Export Citation

Other Resources Citing This Publication