Social Spending, Human Capital, and Growth in Developing Countries: Implications for Achieving the MDGs

Using panel data from 120 developing countries from 1975 to 2000, this paper explores the direct and indirect channels linking social spending, human capital, and growth in a system of equations. The paper finds that both education and health spending have a positive and significant direct impact on the accumulation of education and health capital, and thus can lead to higher economic growth. The paper also finds that other policy interventions, such as improving governance, reducing excessive budget deficits, and taming inflation, can also be helpful in moving countries toward the Millennium Development Goals (MDGs). As such, higher spending alone is not sufficient to achieve the MDGs.

Abstract

Using panel data from 120 developing countries from 1975 to 2000, this paper explores the direct and indirect channels linking social spending, human capital, and growth in a system of equations. The paper finds that both education and health spending have a positive and significant direct impact on the accumulation of education and health capital, and thus can lead to higher economic growth. The paper also finds that other policy interventions, such as improving governance, reducing excessive budget deficits, and taming inflation, can also be helpful in moving countries toward the Millennium Development Goals (MDGs). As such, higher spending alone is not sufficient to achieve the MDGs.

I. Introduction

The role of human capital in fostering economic development is well recognized in the growth literature. Following Romer (1986) and Lucas (1988), human capital has been identified not only as a key determinant of growth and poverty alleviation, but as critical for human development more generally (Squire (1993), Ravallion and Chen (1997), Sen (1999), and Schultz (1999)). The growing focus on the Millennium Development Goals (MDGs) has further highlighted the importance of making tangible progress in indicators of human capital measured on the basis of key education and health indicators.

A crucial issue in this regard is the role of public policy in helping countries meet the MDGs. In most countries, the public sector plays a dominant role in providing the educational and health services necessary to build human capital. As such, the impact of this spending on social indicators, and the impact of higher spending versus other policy interventions (such as improvements in fiscal sustainability or improvements in governance) that might also help countries meet the MDGs (via their salutary effects on economic growth) are of great interest. While positive externalities or market failures may justify the involvement of the public sector in these areas, this does not, in itself, indicate that higher spending per se is the most effective or the only policy intervention for helping meet the MDGs.2

This paper seeks to contribute to this debate by providing an integrated assessment of the role of social spending and other policy interventions on human capital, economic growth, and social indicators. Building upon earlier studies, we analyze the dynamic direct and indirect effects of social spending on human capital and growth, while taking into account the interaction between education and health interventions. The empirical estimates are based on a panel dataset covering 120 developing countries from 1975 to 2000. The model examines the impact of different policy interventions on growth and social indicators, with a view to evaluate their implications for the MDGs.

The remainder of the paper is structured as follows. In Section II, a review of the existing literature is provided. In Section III, an explanation of the data and model is given. Section IV provides the empirical results, including robustness tests. Section V summarizes the simulated effects of different policy interventions on growth and social indicators. Section VI discusses policy implications and concludes the paper.

II. Literature Review

Despite the interlinkages between human capital and growth, most empirical studies have employed reduced-form equations that do not capture feedback effects. The literature often focuses on only one segment of the social spending-social indicators-growth nexus. That is, it either analyzes the growth effects of improving education or health indicators, or the impact of public spending on these indicators. Furthermore, research on the first stream has concentrated essentially on education capital, and has often focused on the impact of the initial stock of education capital on growth. Among these studies, Levine and Renelt (1992), Mankiw and others (1992), Barro and Sala-i-Martin (1995), Barro (1996a and 1996b), and Sala-i-Martin (1997) find a positive relationship between enrollment and/or schooling and growth. Also, using a more refined measure on skills, Coulombe and others (2004) find that a country with literacy scores 1 percent higher than the average experiences an increase in per capita GDP growth of 1.5 percentage points. Benhabib and Spiegel (1994) and Pritchett and Summers (1996), however, find that some macroeconomic evidence conflicts with the findings at the microeconomic level on the returns to education and conclude that the positive link from education attainment to output growth is, at best, weak.

The empirical literature on the effects of health capital on growth is relatively thin. Conceptually, a healthy person can not only work more effectively and efficiently but also devote more time to productive activities. Based on microeconomic evidences, Strauss and Thomas (1998) argue that health explains the variations in wages at least as much as education. Research at the macro level can better capture the potential externalities of health sector interventions, and the existing studies are supportive of the positive contribution of health capital to growth. Bloom and Canning (2003), Bloom and Canning (2004), and Gyimah-Brempong and Wilson (2004) find that health capital indicators positively influence aggregate output. They find that about 22 to 30 percent of the growth rate is attributed to health capital, and improvements in health conditions equivalent to one more year of life expectancy are associated with higher GDP growth of up to 4 percentage points per year.

Studies examining the impact of social spending on social indicators have produced mixed results. For example, based on cross-sectional data for developing countries, Baldacci and others (2003) and Gupta and others (2002b) find that social spending is an important determinant of education and health outcomes.3 They also find that education spending has a greater effect on social indicators than health outlays. The positive effect of social spending on social indicators is also supported by Anand and Ravallion (1993), Psacharopoulos (1994), Hojman (1996), Bidani and Ravallion (1997), and Psacharopoulos and Patrinos (2002). At the same time, a number of studies have found insignificant or very weak linkages between public education outlays and education indicators (Noss (1991), Mingat and Tan (1992 and 1998), and Flug, Spilimbergo, and Wachtenheim (1998)). Similarly, a number of studies find that the contribution of health spending to health status—as measured by infant mortality or child mortality—is either small or statistically insignificant (Kim and Moody (1992), McGuire and others (1993), Musgrove (1996), Pritchett (1996), Filmer and Pritchett (1997), and Filmer and others (1998)). In contrast, Gupta and others (2003) find a positive relationship between public spending on health care and the health status of the poor.

As key pillars in forming human capital, education and health are interlinked in their contribution to growth. Higher levels of education increase public awareness and the capacity of families to address their own health needs. At the same time, better health enhances the effective and sustained use of the knowledge and skills that individuals acquire through education (Schultz (1999)). Barro (1996b) further argues that better health can reduce the depreciation of education capital, and thus increases the favorable effect of education on growth. Few studies, however, have examined social spending, social indicators, and growth in an intergrated system. One exception is Arjona and others (2001), which studies the impact of social spending on growth in OECD countries. They find that although there is no clear impact of social spending on growth at the aggregate level, there exists a positive association between certain types of social spending (albeit excluding many forms of education and health outlays) and growth. A more recent study by Gyimah-Brempong and Wilson (2004) finds a positive and robust link between investment in health and growth in both sub-Saharan African and OECD countries.

Recent research has also highlighted the important role of institutions and governance in mediating the nexus between social spending, indicators, and growth. Poor governance has been identified as a key cause of ineffective social spending (e.g., Mauro (1998), Abed and Gupta (2002), Gupta and others (2002a), and Rajkumar and Swaroop (2002)). Therefore, the lack of control for governance could account for the weak relationship between spending, social indicators, and growth found in some previous studies. By explicitly incorporating governance into the model, our paper helps overcome this problem.

III. Data And Model

A. Methodology

The econometric approach is based on panel data regressions in a system of four equations for (1) real per capita income growth, (2) total investment, (3) education attainment, and (4) health status. The specification of this system is consistent with the literature and allows for the identification of the channels through which social spending and other policy interventions affect growth, as well as the full effects of higher spending over time. The model specification results in a recursive system.4 Under the assumption that the disturbances are not correlated across equations, standard least square techniques can produce unbiased and efficient estimates for each equation.5

1. Growth equation

Drawing upon Mankiw and others (1992), Barro (1996a and 1996b), Bassanini and Scarpetta (2001), Bloom and Sevilla (2004), and Gyimah-Brempong and Wilson (2004), the growth equation is based on a neoclassical growth framework augmented by both education capital and health capital. The per capita output equation is assumed to take the following form:

y=f(sk,he,ed,Ω),

where y is real per capita GDP; sk is the investment ratio; he denotes health capital; ed represents education capital; and Ω denotes the set of macro and institutional control variables, such as the fiscal balance, inflation rate, trade openness, and governance that augment the baseline specification of the model.

In addition, we assume there is a relationship between both the initial stock and increment in human capital with per capita GDP growth.6 As such, the growth of real per capita GDP g is given by:

g=f(sk,he,Δhe,ed,Δed,Ω).

Therefore, we consider the baseline growth equation as the following:

git=α1i+η1it+β11ln(yit1)+β12ni,t+β13ski,t+β14Edi,t1+β15Heit1+ β16ΔEdit+β17ΔHei,t+m=8nβ1m(Ω  itm)+uit[1]

where

  • git is real per capita GDP growth;

  • αi and η t denote the country-specific effect and period-specific effect, respectively;

  • log(yit– 1) is the lagged logarithm of per capita income to control for the expected reduction in growth rates as per capita incomes rise;

  • s k it denotes the investment ratio, measured in terms of gross capital formation (both private and public) in percent of GDP. As an increase in the investment ratio captures an increase in the stock of physical capital, its coefficient is expected to be positive;

  • Edit refers to the stock of education capital, proxied by the composite primary and secondary enrollment rate.7 ΔEd refers to changes in education capital. While the stock of human capital affects productivity growth, changes in education capital reflect an adjustment in the level of a productive input (i.e., educated labor). On the relationship between human capital and growth, Lucas (1988) emphasized that changes in human capital promote growth, whereas Romer (1986 and 1990) argued that the stock of human capital drives growth mainly via innovation. The derivation of the growth equation also leads to the inclusion of both the stocks and changes in human capital variables. 8 Therefore, both variables are included.

  • Heit refers to the stock of health capital, and Δhe refers to changes in health capital. The health condition of the population affects output growth. While it is desirable to account for both mortality and morbidity as a full measure of health capital, such data are not available. Instead, the logarithm of under-5 child mortality is used to proxy the stock of health capital.9 Following the derivation of the growth equation, as well as Bloom and Sevilla (2004) and Gyimah-Brempong and Wilson (2004), both stock and changes in health capital are included.

  • nit is the rate of growth of the total population.

  • m it consists of control variables. Trade openness (open), changes in terms of trade (dtot), fiscal balance (Balance), and inflation (p) were included, as they have been frequently identified as key determinants of growth (Levine and Renelt, 1992; Barro, 1996a; and Baldacci and others, 2004). Trade openness is expressed as the sum of total imports and exports in relation to GDP. Inflation rate is the logarithm of 1 plus the change in CPI. Following Fischer (1993) and Baldacci and others (2004), the high inflation dummy refers to cases where the inflation rate exceeds 20 percent per year. An interaction variable of low deficit fiscal balance (countries with deficits below 3 percent of GDP), LowdefBalance, is also created by multiplying fiscal balance with a low-deficit dummy10 to account for possible non-linear effect of fiscal policy. A dummy for poor governance (Lowgov) was also included, based on the anticorruption and democratic accountability index from International Country Risk Guide (ICRG) compiled by the Political Risk Services group. Including both indexes gives a more complete measure of governance. The dummy takes the value of one (zero otherwise) when the index value is lower than the mean. Also, we include both the level of the annual inflation rate (p) and the high-inflation dummy (Highp) to measure the possible nonlinear effects of inflation on growth.

  • uit is the error term.

2. Investment equation

Investment appears in the growth equation as a direct input to growth. It is also assumed to be influenced by education and health human capital, as well as the quality of governance, fiscal balance, and inflation. We estimate the following equation, building on the models used by Fischer (1993), Benhabib and Spiegel (1994), and Mauro (1996):

Iit=α2i+η2it+β21 log(yt1)+β22nit+β23Eduit+β24Heait+β25Balanceit+ β26LowdefBalanceit+β27pit+β28Highpit+β29Lowgovit+uit[2]

where all macro variables are defined as in the growth equation but only the levels of the education and health indicators are included.

3. Education equation

Eduit=α3i+η3it+β31 ln(yit)+β32Popu15it+β33Heait+β34Urbanit+ β35Qualityit+β36EduSpendingit+β37EduSpendingit1[3]+ β38EduSpendingit*Lowgovit+β39EduSpendingit1*Lowgovit1+β310femaleit+uit

This equation examines the direct impact of education spending on education capital, as proxied by the composite primary and secondary school enrollment rate.

  • School age population (Popu15). The age structure of the population can affect school enrollment (Mingat and Tan, 1992). Age structure is captured by the share of population below 15 years old.

  • Income level (y). Higher per capital income raises the demand for education. This variable is measured as the logarithm of real per capita GDP in PPP dollar terms.

  • Health capital. A healthier population is more likely to be able to afford to invest in education. The under-5 child mortality rate is used to proxy the stock of health capital.

  • Quality indicator (Quality). Following Gupta and others (2003), the quality of education is proxied by the school repetition rate. A higher repetition rate corresponds to a lower quality of education.

  • Education spending (EduSpending). Both the current five-year average and lagged five-year average of education spending, in percent of GDP, were used.11 A logarithmic functional form was used to achieve the best fit and control for diminishing returns to spending.

  • Poor governance (Lowgov). As shown in Abed and Gupta (2002), Baldacci and others (2003), and Gupta and others (2002b), poor governance has a strong adverse impact on the effect of public spending on social indicators.

  • Gender equality (Female) in education also affects aggregate education capital. As shown in Summers (1992), women are often the primary educators and nurturers of children, and therefore investing in girls’ education can reap significantly greater returns than investing in boys’ education. This is especially the case in developing countries, where gender inequality is more acute. This effect is captured by the share of female students in primary and secondary schools.

4. Health equation

Heait=α4i+η4t+β41ln(yi,t)+β42Fertilityit+β43Urbanit+β44HeaSpendingit+β45HeaSpendingit*Lowgovit+β46Femaleit+u4it[4]
  • Income level. Past studies indicate that this variable is a crucial determinant of health status (Carrin and Politi, 1996, and Filmer and Pritchett, 1997).

  • Gender equality. A number of studies have identified female education as an important determinant of the health status of infants and children, as well as the population in general (e.g., Schultz, 1993, and the World Bank, 1993).12 For example, in developing countries, women play a more important role in family health and sanitation. Furthermore, female education is positively associated with infants’ health and negatively associated with fertility rates. While it would be desirable to measure gender equality for society as a whole, such data are not available. Instead, the share of female students is used as a proxy.

  • Urbanization. Schultz (1993) finds that the mortality rate is higher for rural, low-income, agricultural households, and therefore urbanization can be an important determinant of the health status of the population. This effect is captured by including the share of the urban population in the regression.

  • Poor governance. This is measured in the same way as in the education equation.

  • Current health spending. The five-year average of health spending as a percent of GDP was used. A logarithmic functional form was adopted to achieve the best fit and control for decreasing returns to spending. This is consistent with Deaton (2004), which finds that the benefits from investment in health are stronger in countries at the initial stages of development.

5. Decomposition of short-run effects

In many cases, the variables in the system have both direct and indirect effects, and these effects can be separately identified in the following recursive system:

y=By+Γx+ξ, or in a reduced form y=x+ξ

where y is a vector of endogenous variables; x is a vector of exogenous variables; B is a lower triangular matrix containing coefficients for direct effects of each pair of endogenous variables i and j with elements β ij ≠ 0 and β ji =0; Γ is a matrix of direct effects of the exogenous variables on the endogenous variables, and ξ is a vector of the equations’ error terms. The total effect of x on y can be calculated by Π = (I – B)–1 Γ, and the indirect effect of x on y is then given by Θ = (IB)–1I.13

B. Data

A panel dataset for 120 developing countries from 1975 to 2000 was compiled for the purposes of the paper (see Table 1 for a description of the data and Appendix II for the list of countries). Five-year averages are used to smooth short-term fluctuations and minimize the measurement errors that can be present in annual observations. All macroeconomic variables are from the IMF’s World Economic Outlook 2003 database, while social indicators and social spending data are from the latest Barro-Lee dataset,14 the World Development Indicators 2003 database and a social spending database compiled from IMF staff reports and the IMF’s Government Financial Statistics. Data on governance are taken from ICRG index of corruption and democratic accountability.15

Table 1.

Summary Descriptive Statistics

article image
Sources: Various databases (see Appendix II for details).

The terms of trade variable compares the relative price of a country’s exports and imports. A positive change denotes an improvement in the terms of trade.

In this paper, output growth is measured by the five-year averages of the annual growth rate of real per capita GDP; education and health capital are proxied by education and health indicators; and social spending data are expressed as a percent of GDP.

We adopt a fixed-effect model (LSDV) as the baseline specification. This estimator is chosen in light of the fact that it provides consistent estimates in the context of a recursive system of equations. Nevertheless, to assess the robustness of the results for individual equations, we also provide results from alternative estimators to control for outliers, measurement error, autocorrelation, and endogeneity. These include a robust estimator,16 a Feasible Generalized Least Squares (FGLS) estimator,17 an error component two-stage least squares estimator (EC2SLS),18 and a fixed-effect instrumental variable estimator (2SLS).19 Moreover, as shown by Bond and others (2001), problems associated with endogeneity, measurement error, and omitted variables in dynamic growth models can be best addressed by a system GMM estimator. Therefore, in equations that involve lagged variables, results from a system GMM estimator are also provided as a further robustness check.20 Finally, we provide results from overidentification tests to show the validity of the instruments used.

Measurement errors are considered an important factor in empirical studies on human capital and growth. Temple (1999) and Krueger and Lindahl (2001) highlight the prevalence of such errors in human capital indicators and the significant impact they may have on the empirical estimates of the human capital-growth relationship. In this paper, measurement errors are minimized by using five-year averages of the data. The robustness of the results is also checked by using alternative indicators of health and education.

IV. Empirical Results

A. Main Results

The results for the growth, investment, education, and health equations are presented in Tables 25.21 In most cases the coefficients are statistically significant, and all equations have a good fit. Among the most salient results from the baseline (LSDV) model are the following:

  • Both education capital and health capital positively contribute to output growth, but through slightly different routes (Table 2). Both the stock and flow of education capital affect growth. The effects from the two channels are of similar magnitude. In contrast, the only direct effect of health capital on growth is through flows. The stock of health capital does, however, affect growth indirectly via investment (Table 3). Furthermore, the regressions in the education and health equations (Tables 4 and 5) show that education and health spending affects indicators on education and health capital. Thus, social spending also helps boost growth via its effects on human capital.22

  • Education spending has both an immediate and a lagged effect on education capital. About two-thirds of the direct effect are realized in the first five years, with the remainder realized over the next five-year span. 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 five years and of another 3 percentage points in the following period.23 Furthermore, the higher income brought by growth can generate additional improvement in education capital. The aggregate growth effect is further analyzed in Section V to illustrate the role of lagged effects and interactions between the key variables.

  • Health spending has a positive and significant contemporaneous impact on health capital. An increase in health spending of 1 percentage point of GDP is associated with a rise in the under-5 child survival rate of 0.2 percentage points, on average, in developing countries. However, lagged health spending has no further effect on health indicators, indicating that the impact of health spending is only contemporaneous.24

  • Education and health capital have strong interlinkages. Health capital contributes to the accumulation of education capital, with an elasticity of about 1.3. Moreover, education is also associated with health capital through the variable on gender equality.

  • Governance has significant direct effects on the nexus between social spending and social indicators, with health spending being particularly sensitive to governance. As shown in Tables 4 and 5, health spending has no effect on health indicators in countries suffering from poor governance. Similarly, the impact of education spending is also much reduced in such countries. The important role of governance—which has not always been incorporated in previous research—could help explain why some earlier studies have found a generally weak relationship between social spending and social indicators.

  • Gender equality influences both education and health indicators. The share of female students in primary and secondary schools is positively and significantly associated with education capital and negatively and significantly associated with child mortality. A 1 percentage point increase (e.g., from 40 percent to 41 percent) in the female share of enrolled students is associated with an increase of 2 percentage points in the composite enrollment rate, and a reduction of 0.3 percentage point in the child mortality rate. This strongly supports the view that in developing countries, gender equality is important in both education and health sector interventions.

  • Income levels matter for social indicators. GDP per capita is robustly and positively correlated with both education and health capital. This indicates that higher income levels and greater human capital reinforce each other and contribute to a virtuous circle of growth and higher human capital.

  • Poor governance reduces growth mainly through its impact on human capital and investment. Countries with poor governance, other things being equal, have growth of about 1.6 percentage points lower per year than other countries.25 Similarly, poor governance is associated with a rate of investment to GDP that is 2 percentage points lower. As shown in Annex Table 9, 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 income and regional country groups (Table 6). The impact of education capital on growth is more pronounced in low-income countries, where an increase of 1 percentage point in the composite enrollment rate is associated with 0.1 percentage point increase in per capita GDP growth. This effect is 1.5 times that in middle-income countries. Geographically, the impact is highest in sub-Saharan Africa and lowest in Eastern Europe and Central Asia. As for the impact of health capital on growth, the impact of a unit reduction in the under-5 child mortality rate is smaller in low-income countries than middle-income countries. However, this masks the fact that in terms of the elasticities, the impact of reducing child mortality is about 4 times as large when compared to middle-income countries, owing to high initial levels of mortality in the poorest countries.

Table 2.

Growth Equation-Dependent Variable: Average Real Per Capita GDP Growth

article image
Absolute value of t statistics in parentheses.

significant at 10 percent;

significant at 5 percent;

significant at 1 percent.

Baseline regression.

Table 3.

Investment Equation-Dependent Variable: Average Annual Investment Ratio

article image
Absolute value of t statistics in parentheses.

significant at 10 percent;

significant at 5 percent;

significant at 1 percent.

Baseline regression.

Table 4.

Education Equation-Dependent Variable: Average Gross Composite School Enrollment Rate

article image
Absolute value of t statistics in parentheses.

significant at 10 percent;

significant at 5 percent;

significant at 1 percent.

Baseline regression.

Table 5.

Health Equation-Dependent Variable: Average Annual Under-5 Child Mortality Rate

article image
Absolute value of t statistics in parentheses.

significant at 10 percent;

significant at 5 percent;

significant at 1 percent.

Baseline regression.

Table 6.

Summary of Total Effect Coefficients by Country Group1/

article image
Source: Authors’ calculation.

The coefficient measures dy/dx, i.e., the impact of one percentage point change in column variables on row variables. These coefficients are calculated using estimates from the baseline model. If the original coefficients in the regressions are in y/lnx form, they are transformed into linear coefficients of dy/dx derived from the equation dy/dlnx=dy/(dx/x)=(dy/dx)/x.

Under-5 child survival rate is used to facilitate the interpretation of a positive effect, and the increase in the under-5 child survival rate matches exactly the reduction in the under-5 child mortality rate.

B. Robustness Checks

Our estimates generate elasticities that are similar to those found in earlier research. The econometric results imply that the elasticity of education indicators with respect to education spending is 0.25.26 This is close to the upper range of similar estimates on the returns to primary and secondary education in the literature (Psacharopoulos (1994), Psacharopoulos and Patrinos (2002), and Baldacci and others (2003)). The elasticity of health indicators with respect to changes in public health outlays is about 0.05, which is also consistent with the literature (Schultz (1999) and Gupta and others (2002b)).

The coefficients linking human capital indicators and growth are also consistent with the literature. On health capital and growth, Bloom and Canning (2003) and Bloom and Sevilla (2004) find elasticities of 1.3 and 1.9 respectively, compared with our estimates ranging from 1.4 to 3.2 for different country groups.27 On education capital and growth, Levine and Renelt (1992), Barro (1996a), Bassanini and Scarpetta (2001), Bils and Klenow (2000), and Sianesi and Van Reenen (2003) estimate that an additional year of schooling raises the growth rate by 0.3 to 3 percentage points per year. Our estimates, based on the composite enrollment rate, indicate that one year of additional primary and secondary schooling is associated with an increase in growth ranging between 0.4 and 0.8 percentage points per year, depending on the country group.28 Low-income countries and sub-Saharan African countries are on the high end, while Eastern Europe and Central Asia countries are on the low end.

The results are broadly consistent under a number of different estimators controlling for outliers (robust regression) and endogeneity (two-stage least squares and system GMM).29

Moreover, solving the growth, investment, education, and health equations simultaneously in a Seemingly Unrelated Regression framework does not lead to qualitatively different results. The Breusch-Pagan test fails to reject the independence of the residuals from the four equations. We also compared the recursive system with other specifications, including a full-fledged system of equations, and find that our model dominates other specifications based on the SIC.

Furthermore, the estimates on governance, inflation, and deficit are also broadly consistent with the literature. For example, Mauro (1996) finds that reducing corruption by one standard deviation is associated with a rise in the growth rate of 0.5 percentage point, while our results indicate that a discrete jump from a low-governance to a high-governance country is associated with an increase in the growth rate of 1.6 percentage points.30 Fischer (1993) and Burdekin and others (2000) find that lowering inflation by 10 percentage points raises growth by 0.4 and 0.831 percentage points respectively, while our results indicate the same inflation reduction is associated with higher growth of 0.5 percentage point. Baldacci and others (2004) find that lowering the deficit by 1 percentage point of GDP is associated with an increase in per capita GDP growth of 0.4 percentage point, while this paper shows an increase in per capita growth of 0.5 percentage point, also similar to that in Fischer (1993), Bleaney (1996), and Gupta and others (2004).

V. Simulation32 of the Impact of Policy Interventions and Implications for the MDGs

As shown in previous sections, social indicators and growth are affected by many interrelated factors. On the basis of the empirical estimates presented above, simulations can be conducted to assess the impact of different policy interventions for improving social indicators, economic growth, and the poverty headcount. As such, these simulations can provide input into the debate over the effectiveness of different policies in helping countries reach the MDGs. The simulations assess the impact of the following policy interventions: (1) an increase in education spending of one percent of GDP; (2) an increase in health spending of one percent of GDP; (3) an improvement in governance; (4) a reduction in the budget deficit of one percent of GDP; and (5) a reduction in inflation of 10 percentage points.

Each of the simulations assumes that the policy environment remains unchanged. This implies, for example, that increases in public spending do not result in an increase in the budget deficit or a deterioration in governance. As noted in the results below, the macroeconomic environment and governance are important determinants of growth and human development. As such, the effectiveness of spending initiatives would be greatly diminished if not accompanied by a deterioration of performance in these areas.

Increase education spending by 1 percent of GDP 33

Simulation results for raising education spending, and other interventions, are reported in Table 7. They suggest that raising average education spending by 1 percent of developing country GDP (and maintaining it at that higher level) would increase the sum of the primary and secondary enrollment rates (the composite rate) from an average of 154 to 173 over 15 years (Table 7).34 This implies that the net enrollment rate can increase from 90 to 100 percent. Thus, increases in spending of this magnitude would be greatly helpful in moving countries toward the MDG target for education, although not necessarily sufficient to achieve it in all regions. Increasing education spending could also contribute to health capital by lowering the child mortality rate from 76 per thousand in 2000 to 65 per thousand in 2015. In addition, the strengthening of human capital would have a beneficial effect on growth, and the per capita growth rate would rise by about 0.5 percentage point per year on average over the simulation period. Such an improvement in growth could reduce the initial poverty headcount by about 17 percent over a 15-year period.35

Table 7.

Simulation on the Impact on Growth and Social Indicators from Various Policy Interventions 1/

article image
Sources: World Development Indicators database 2003, World Economic Outlook database 2003, country authorities and authors’ calculations.

Simulations are based on coefficients obtained from the baseline model. The initial baseline values are averages for developing countries in the sample, and the policy interventions are assumed to take place at the end of baseline year. In addition, the increased social spending is assumed to be maintained over the simulation period. Furthermore, the elasticity estimates from the recursive model are assumed to hold for the period covered by the simulation while other variables stay at the same level. The effects of higher incomes on education and health capital are also captured.

The value of 1 percent of GDP of all developing countries was about $60 billion in 2000. The rise in spending is assumed to be deficit-neutral, and may be achieved through grants, increased revenue collection, and/or shifting of funding from unproductive outlays.

The implied net primary enrollment rate is calculated under the assumption that the ratio of the net primary enrollment rate over the composite enrollment rate remains what it was in 2000.

The initial value of the headcount ratio is standardized to be 100, and the implied headcount ratios are the projected end-of-period values after incorporating the incremental poverty reduction due to growth acceleration. The impact on the poverty headcount ratio is calculated by the growth elasticity of poverty of −2.6, an average based on estimates from Ravallion and Chen (1997), Bruno, Ravallion, and Squire (1998), and Adams (2003).

The impact on growth is estimated at the sample mean for deficit at 4 percent of GDP, and the reduction is assumed to be achieved through grants, increased revenue collection or cut in unproductive spending without affecting social spending.

Increase health spending by 1 percent of GDP

Raising average health spending by 1 percent of GDP would reduce the under-5 child mortality rate by 0.6 percentage point, that is from an average of 76 deaths per thousand live births in 2000 to 70 deaths per thousand in 2015. The simulations suggest that the MDG target—lowering the child mortality rate to 62 deaths per 1,000 live births (one-third of the 1990 level)—cannot be met under this scenario.36 In addition, there are also small but positive effects from health spending on school enrollment and growth. On average, the net enrollment rate in developing countries would rise by about 2 percentage points (i.e., from 90 to 92), and the growth rate would rise by a total of 0.5 percentage points over 15 years.

Improve governance to a level above the sample average

Enhancing governance is a powerful instrument to improve social indicators and growth. A change in the governance index from lower- to higher-than-average is associated with an immediate reduction of 0.5 percentage point in the child mortality rate, an increase of 6 percentage points in the composite enrollment rate, and a rise of 1.6 percentage points in per capita GDP growth. Through the reinforcing impact of higher income on human capital, this measure can lead to even better social indicators. For example, it can increase net primary enrollment by 10 percentage points in 10 years without additional resources. The positive impact from elevating a country from a below- to a higher-than-average level, therefore, is comparable to an increase in education spending of 1 percent of GDP.

Reduce inflation

The growth effects of lower inflation (and hence its effects on poverty) are substantial. Cutting the rate of inflation by 10 percentage points (e.g., from 40 percent to 30 percent) is associated with a 0.5 percentage point increase in annual growth. Moreover, countries that reduce their rate of inflation to below 20 percent would also experience an additional fillip to growth of 0.1 percentage point per year. The magnitude of this effect is consistent with the literature.

Reduce fiscal deficit

As discussed in previous sections, budget deficits have an adverse effect on growth. Improving the fiscal balance by 1 percentage point of GDP is associated with an increase in per capita GDP growth by 0.5 percentage point.37 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 of social spending. Furthermore, as shown in Table 2, the effects from increasing the fiscal balance in a low-deficit environment is no longer significant. This result is consistent with Gupta and others (2004), which finds no relationship between further deficit reduction and growth in low-income countries that have already achieved a modicum of macroeconomic stability.

VI. Conclusion And Policy Implications

The paper finds that a number of policy interventions could be effective in moving countries toward the MDGs. Both education and health spending have a positive and significant direct impact on the accumulation of education and health capital, and a positive and significant indirect impact on growth. An increase in education spending of 1 percentage point of GDP is associated with 3 more years of schooling on average and a total increase in growth of 1.4 percentage points in 15 years. Similarly, an increase in health spending of 1 percentage point of GDP is associated with an increase of 0.6 percentage points in the under-5 child survival rate and a rise of 0.5 percentage point in annual per capita GDP growth.

There is a significant time lag between increases in education spending and the realization of their full effects on social indicators and growth. Two-thirds of the direct impact of education spending is felt within five years, but the full impact materializes with a significant time lag of 10 to 15 years. Such a lag needs to be kept in mind when designing policy interventions. The impact of health spending, however, is immediate. The positive effects of both education and health spending are strongly influenced by the quality of governance. In countries suffering from poor governance, the positive effects of increased spending on education is reduced, and those of higher health spending can be completely negated.

There are substantial differences in the effects of social spending on social indicators and growth among different country groups. The positive effects are the highest in low-income countries and sub-Saharan Africa. This supports the view that social spending can be more effective in such countries in achieving MDGs, as the marginal returns to social spending tend to decline for countries with high levels of social outlays.

Other policy interventions may also achieve improvement of a similar size in social indicators and growth. In particular, strengthening governance can have a strong payoff for social indicators as well as for growth. Therefore, reducing corruption and increasing accountability for public spending are no less important than increasing spending. In addition, macroeconomic policies, such as reducing inflation and improving fiscal balances, also have a positive effect on growth and, in turn, on the poverty headcount.

The results have a number of implications for poverty reduction strategies aimed at meeting the MDGs. Given the importance of different policy interventions, efforts to meet the MDGs will need to be wide-ranging and include strengthening the macroeconomic environment and governance. Relative to the significant cost of raising spending, the moderate effects of social spending on indicators also confirm the important role of reforms aimed at improving the efficiency and targeting of these outlays. Furthermore, while improving human capital will have a salutary effect on growth, it will be far from a panacea for unlocking the more robust expansion in economic activity needed to achieve the MDGs. As such, additional research is needed to address the key policy interventions needed to achieve rapid economic growth.

APPENDIX I

Derivation of the Growth Equation with Human Capital 38

In a neoclassical growth model enhanced by human capital:

Y(t)=K(t)αH(t)β(A(t)L(t))1αβ

where Y is output, K and H are physical and human capital respectively, A is technology, L is population, and α and β are the partial elasticities of output with respect to physical and human capital. The right-hand side variables are assumed to take the following time paths:

k˙(t)=sk(t)A(t)1αβk(t)αh(t)β(n(t)+d)k(t)h˙(t)=sh(t)A(t)1αβk(t)αh(t)β(n(t)+d)h(t)[2]A˙(t)=g(t)A(t)L˙(t)=n(t)L(t)

where y = Y/L and k = K/L are output and physical capital in per capita terms, h = H/L represents average human capital, sk and sh are the investment rate in physical and human capital, n is population growth, g is the rate of technological change and d is the time-invariant depreciation rate.

Under the assumption of decreasing returns to scale (α+β < 1), this system of equations can be solved to obtain steady-state values of k* and h* defined by:

ln k*(t)=ln A(t)+1β1αβlnsk(t)+β1αβlnsh(t)11αβln(g(t)+n(t)+d)ln h*(t)=ln A(t)+α1αβlnsk(t)+1α1αβlnsh(t)11αβln(g(t)+n(t)+d)[3]

Substituting these two equations into the production function and taking logs yields the expression for the steady-state output in per capita terms. The human capital component can be expressed either as a function of sh and the other variables or as a function of h* and the other variables (see Mankiw and others, 1992). In this paper, the human capital component is expressed in terms of human capital stock, proxied by the gross enrollment rate. Therefore, the steady-state path of output can be written as:

ln y*(t)=ln A(t)+α1αlnsk(t)+β1αln h*(t)α1αln(g(t)+n(t)+d)[4]

h* is unobservable, but a relationship between h* and the actual level of human capital can be established by solving the system of differential equations in [2] and substituting out the respective investment rates by means of the equations in [3]:39

d lnkAdt=(n+g+d)e(1α)lnkk*eβlnhh*d lnhAdt=(n+g+d)eαlnkk*e(1β)lnhh*[5]

The linearized solution for ln h is the following:

ln(h(t)/A(t))=ψ(ln(h*(t)/A(t))+(1ψ)ln(h(t1)/A(t1))[6]

where ψ is a function of (α, β) and the term (n+g+d). Rearranging equation [6] yields an expression for ln h*:

ln h*(t)=ln h(t)+1ψψΔ ln(h(t)/A(t))[7]

By replacing the term ln h* with its expression [7] into equation [4], we obtain an expression for the steady state output as a function of the investment rate and the actual human capital stock. Assuming countries are in their steady-states or their deviations from the steady states were independent and identically distributed, the transitional growth dynamics can be expressed as follows (Mankiw and others, 1992):

d ln(y(t)/A(t))dt=λ(ln(y*(t)/A(t))ln(y(t)/A(t)))[8]

where λ = (1–α–β)(g(t) + n(t) + d).

Substituting the expression for y* and h* into the solution of this differential equation yields the following:40

Δ ln y(t)=ϕ(λ)ln (y(t1))+ϕ(λ)α1αlnsk(t)+ϕ(λ)β1αln h(t)+1ψψβ1αΔln h(t)ϕ(λ)α1αln (g+n(t)+d)+(1ϕ(λ)ψ)g+ϕ(λ)ln A(0)+ϕ(λ)gt[9]

Since g is not observable and therefore cannot be distinguished from the constant term empirically, the estimated baseline growth equation can be expressed as follows:

Δln y(t)=a0ϕ ln y(t1)+a1ln sk(t)+a2ln h(t)a3n(t)+a4t+bΔln h(t)+ε(t)[10]

In this paper, we separate human capital further into two components: education capital— ed, and health capital— he, and assume human capital h(t) takes the following form:

h(t)=ed(t)γ*he(t)η or ln h(t)=γ ln ed(t)+η ln he(t).

We also assume that the time paths for education capital, e, and health capital, h, are as follows:

edt=edt1+Δedthet=het1+Δhet

Hence, [10] becomes [11]

Δlny(t)=a0ϕlny(t1)+a1lnsk(t)+a2lnhe(t)+a3lned(t)a4n(t)+a54t+b1Δlnhe(t)+b2Δlned(t)+ε(t)[11]

Or, in a simplified representation after adding a matrix of control variables, Ω,

g=f(sk,he,Δhe,ed,Δed,Ω)[12]

APPENDIX II

List of Countries Included in the Sample and Data Sources

The countries included in the study are Albania, Algeria, Angola, Argentina, Armenia, Azerbaijan, Bangladesh, Barbados, Belarus, Belize, Benin, Bolivia, Bosnia and Herzegovina, Botswana, Brazil, Bulgaria, Burkina Faso, Burundi, Cambodia, Cameroon, Cape Verde, Central African Republic, Chad, Chile, China, Colombia, Comoros, Democratic Republic of Congo, Republic of Congo, Costa Rica, Côte d’Ivoire, Croatia, Czech Republic, Djibouti, Dominican Republic, Ecuador, Egypt, El Salvador, Eritrea, Estonia, Ethiopia, Fiji, The Gambia, Ghana, Guatemala, Guinea, Guinea-Bissau, Guyana, Haiti, Honduras, Hungary, India, Indonesia, Islamic Republic of Iran, Jamaica, Jordan, Kazakhstan, Kenya, Kyrgyz Republic, Lao People’s Democratic Republic, Latvia, Lebanon, Lesotho, Libya, Lithuania, FYR Macedonia, Madagascar, Malawi, Malaysia, Maldives, Mali, Mauritania, Mauritius, Mexico, Moldova, Mongolia, Morocco, Mozambique, Myanmar, Namibia, Nepal, Nicaragua, Niger, Nigeria, Oman, Pakistan, Panama, Papua New Guinea, Paraguay, Peru, Philippines, Poland, Romania, Russia, Rwanda, Saudi Arabia, Senegal, Sierra Leone, Slovak Republic, Solomon Islands, South Africa, Sri Lanka, Sudan, Swaziland, Syrian Arab Republic, Tajikistan, Tanzania, Thailand, Togo, Trinidad and Tobago, Tunisia, Turkey, Uganda, Ukraine, Uruguay, Venezuela, Vietnam, Yemen, Zambia, Zimbabwe.

All data series are five-year averages of annual data. Data on real per capita GDP, inflation, investment, terms of trade, trade openness, total government expenditure, and fiscal balance are based on the World Economic Outlook (WEO) database. In particular, real GDP growth rates are calculated by the change of GDP in constant prices; the rate of inflation is calculated by the rate of increase in CPI index; investment, government expenditure, and fiscal balances are calculated as a ratio to GDP at current prices; and trade openness is calculated by the value of total exports and imports of goods and services as a share of GDP. Data on child mortality, the school repetition rate, ratio of female students in school, population growth, the shares of the under-15 population and urban population, and fertility rates are taken from the World Bank’s World Development Indicators (WDI) database; data on the enrollment rate is taken from the 2003 Barro-Lee dataset; and the composite enrollment rate is the sum of primary and secondary enrollment rates. Child mortality and enrollment rates are not available for all years, but five-year averages allow the construction of a consistent panel dataset nonetheless. Data on education spending as a share of GDP is taken from the WDI database, but health spending as a share of GDP is taken from an IMF database that has better country coverage. Data in the IMF database is compiled from IMF staff reports and the IMF’s Government Financial Statistics. Data on governance are calculated as the sum of the simple annual averages of two indices on corruption and democratic accountability, which are two components of the ICRG rating produced by the Political Risk Service Group.

Appendix Table 8.

Short-Run Total Effects Decomposition: Selected Endogenous Variables

article image
Appendix Table 9.

Short-Run Total Effects Decomposition: Selected Exogenous Variables

article image

References

  • Abed, George T., and Sanjeev Gupta, eds., , 2002, Governance, Corruption, and Economic Performance (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Adams, Richard, 2003, “Economic Growth, Inequality and Poverty: Findings from a New Data Set,” World Bank Policy Research Working Paper No. 2922 (Washington: World Bank).

    • Search Google Scholar
    • Export Citation
  • Arjona, Roman, Maxime Ladaique Mark Pearson ,2001, “Growth, Inequality and Social Protection,” OECD Labor Market and Social Policy Occasional Paper No. 51 (Paris: OECD).

    • Search Google Scholar
    • Export Citation
  • Anand, Sudhir, and Martin Ravallion, 1993, “Human Development in Poor Countries: On the Role of Private Incomes and Public Services,” Journal of Economic Perspectives, Vol. 7 (Winter), pp. 13350.

    • Search Google Scholar
    • Export Citation
  • Baldacci, Emanuele, Maria Teresa Guin-Sui, and Luiz de Mello, 2003, “More on the Effectiveness of Public Spending on Health Care and Education: A Covariance Structure Model,” Journal of International Development, Vol. 15, pp. 70925.

    • Search Google Scholar
    • Export Citation
  • Baldacci, Emanuele, Arye L. Hillman, and Naoko C. Kojo, 2004, “Growth, Governance, and Fiscal Policy Transmission Channels in Low-Income Countries,” European Journal of Political Economy, Vol. 20, No. 3, pp. 51749. Also reprinted in Sanjeev Gupta, Benedict Clements, and Gabriela Inchauste, eds., Helping Countries Develop: The Role of Fiscal Policy (Washington: International Monetary Fund, 2004).

    • Search Google Scholar
    • Export Citation
  • Baltagi, Badi, 2001, The Econometrics of Panel Data, second edition (New York: John Wiley & Sons).

  • Barro, Robert J., 1996a, “Determinants of Economic Growth: A Cross-Country Empirical Study,” NBER Working Paper No. 5968 (Cambridge, Massachusetts: National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation
  • Barro, Robert J., 1996b, “Health, Human Capital and Economic Growth,” Paper for the Program on Public Policy and Health, Pan American Health Organization and World Health Organization (Washington: Pan American Health Organization).

    • Search Google Scholar
    • Export Citation
  • Barro, Robert J., and Jong-Wha Lee, 2003, “International Data on Educational Attainment: Updates and Implications,” Center for International Development Working Paper No. 42 (Cambridge, Massachusetts: Harvard University).

    • Search Google Scholar
    • Export Citation
  • Barro, Robert, and Xavier Sala-i-Martin, 1995, Economic Growth McGraw-Hill.

  • Bassanini, Andrea, and Stefano Scarpetta, 2001, “Does Human Capital Matter for Growth in OECD Countries? Evidence from Pooled Mean-Group Estimates,” OECD Economics Department Working Paper No. 282 (Paris: OECD).

    • Search Google Scholar
    • Export Citation
  • Benhabib, Jess, and Mark Spiegel, 1994, “The Role of Human Capital in Economic Development: Evidence from Aggregate Cross-Country Data,” Journal of Monetary Economics, Vol. 34, pp. 14373.

    • Search Google Scholar
    • Export Citation
  • Bidani, Benu, and Ravallion Martin, 1997, “Decomposing Social Indicators Using Distributional Data,” Journal of Econometrics, 77, No. 1, pp. 12539.

    • Search Google Scholar
    • Export Citation
  • Bils, Mark, and Peter J. Klenow, 2000, “Does Schooling Cause Growth?American Economic Review, Vol. 90, No. 5, pp. 116083.

  • Bleaney, Michael F., 1996, “Macroeconomic Stability, Investment and Growth in Developing Countries,” Journal of Development Economics, Vol. 48, No. 2, pp. 46177.

    • Search Google Scholar
    • Export Citation
  • Bloom, David, and David Canning, 2003, “The Health and Poverty of Nations: From Theory to Practice,” Journal of Human Development, Vol. 4, No. 1, pp. 4771.

    • Search Google Scholar
    • Export Citation
  • Bloom, David, and David Canning, and Jaypee Sevilla, 2004, “The Effect of Health on Economic Growth: A Production Function Approach,” World Development, Vol. 32, No. 1, pp. 113.

    • Search Google Scholar
    • Export Citation
  • Blundell, R., and Stephen Bond, 1998, “Initial Conditions and Moment Restrictions in Dynamic Panel Data Models,” Journal of Econometrics, Vol. 87, No. 1, pp. 11543.

    • Search Google Scholar
    • Export Citation
  • Bond, Stephen, Anke Hoeffler, and Jonathan, Temple, 2001, GMM Estimation of Empirical Growth Models, Nuffield College Economics Paper No. 2001-W21 (Oxford: Oxford University).

    • Search Google Scholar
    • Export Citation
  • Bruno, Michael, Martin Ravallion, and Lyn Squire, 1998, “Equity and Growth in Developing Countries: Old and New Perspectives on the Policy Issues,” in Vito Tanzi and Ke-young, Chu eds., Income Distribution and High Growth (Cambridge, Massachusetts: MIT Press).

    • Search Google Scholar
    • Export Citation
  • Burdekin, Richard C.K., Arthur T. Denzau, and Manfred W. Keil, 2000, “When Does Inflation Hurt Economic Growth? Different Nonlinearities for Different Economies,” Working Paper in Economics No. 22 (Claremont, California: Claremont McKenna College).

    • Search Google Scholar
    • Export Citation
  • Carrin, Guy, and Claudio Politi, 1996, “Exploring the Health Impact of Economic Growth, Poverty Reduction and Public Health Expenditure,” Macroeconomics, Health, and Development Series, No. 18 (Geneva: World Health Organization).

    • Search Google Scholar
    • Export Citation
  • Castro, Claudio de Moura, and Philip Musgrove, 2001, “Why Education and Health Are More Different than Alike,” Economics of Education Series (Washington: World Bank).

    • Search Google Scholar
    • Export Citation
  • Coulombe, Serge, Jean-François Tremblay, and Sylvie Marchand, 2004, “Literacy Scores, Human Capital and Growth Across Fourteen OECD Countries,” International Adult Literacy Survey Monograph Series (Ottawa: Statistics Canada).

    • Search Google Scholar
    • Export Citation
  • Deaton, Angus, 2004, “Health in an Age of Globalization,” paper prepared for the Brookings Trade Forum (Washington: Brookings Institution).

    • Search Google Scholar
    • Export Citation
  • Duncan, Otis Dudley, 1975, Introduction to Structural Equations Models (New York Academic Press).

  • Filmer, Deon, and Lant Pritchett, 1997, “Child Mortality and Public Spending on Health: How Much Does Money Matter?World Bank Policy Research Working Paper No. 1864 (Washington: World Bank).

    • Search Google Scholar
    • Export Citation
  • Filmer, Deon, Jeffrey Hammer, and Lant Pritchett, 1998, “Health Policy in Poor Countries: Weak Links in the Chain,” World Bank Policy Research Working Paper No. 1874 (Washington: World Bank).

    • Search Google Scholar
    • Export Citation
  • Fischer, Stanley, 1993, “Role of Macroeconomic Factors in Growth,” Journal of Monetary Economics, Vol. 32, pp. 485512.

  • Flug, K. A. Spilimbergo, and E. Wachtenheim, 1998, “Investment in Education: Do Economic Volatility and Credit Constraints Matter?Journal of Development Economics, Vol. 55, pp. 46581.

    • Search Google Scholar
    • Export Citation
  • Gemmel, N., 1996, “Evaluating the Impacts of Human Capital Stocks and Accumulation on Economic Growth: Some New Evidence,” Oxford Bulletin of Economics and Statistics, Vol. 58, pp. 928.

    • Search Google Scholar
    • Export Citation
  • Gupta, Sanjeev, Hamid Davoodi, and Rosa Alonso-Termé, 2002a, “Does Corruption Affect Income Inequality and Poverty?Economics of Governance, Spring.

    • Search Google Scholar
    • Export Citation
  • Gupta, Sanjeev, Marijn Verhoeven, and Erwin Tiongson, 2002b, “The Effectiveness of Government Spending on Education and Health Care in Developing and Transition Economies,” European Journal of Political Economy, Vol. 18, No. 4, pp. 71737. Also reprinted in Sanjeev Gupta, Benedict Clements, and Gabriela Inchauste, eds., Helping Countries Develop: The Role of Fiscal Policy (Washington: International Monetary Fund, 2004)

    • Search Google Scholar
    • Export Citation
  • Gupta, Sanjeev, Marijn Verhoeven, and Erwin Tiongson, 2003, “Public Spending on Health Care and the Poor,” Health Economics, Vol. 12, pp. 68596.Also reprinted in Sanjeev Gupta, Benedict Clements, and Gabriela Inchauste, eds., Helping Countries Develop: The Role of Fiscal Policy (Washington: International Monetary Fund, 2004).

    • Search Google Scholar
    • Export Citation
  • Gupta, Sanjeev, Benedict Clements, Emanuele Baldacci, and Carlos Mulas-Granados, 2004, “Fiscal Policy, Expenditure Composition, and Growth,” Journal of International Money and Finance, forthcoming. Also reprinted in

    • Search Google Scholar
    • Export Citation

    Sanjeev Gupta, Benedict Clements, and Gabriela Inchauste, eds., Helping Countries Develop: The Role of Fiscal Policy (Washington: International Monetary Fund, 2004).

    • Search Google Scholar
    • Export Citation
  • Gyimah-Brempong, Kwabena, and Mark Wilson, 2004, “Health Human Capital and Economic Growth in Sub-Saharan African and OECD Countries,” The Quarterly Review of Economics and Finance, Vol. 44, No. 2, pp. 296320.

    • Search Google Scholar
    • Export Citation
  • Hamilton, Lawrence C. 1991, “How Robust is Robust Regression?Stata Technical Bulletin, Vol. 2 July, pp. 2126.

  • Hojman, David E., 1996, “Economic and other Determinants of Infant and Child Mortality in Small Developing Countries: The Case of Central America and the Caribbean,” Applied Economics, Vol. 28, pp. 28190.

    • Search Google Scholar
    • Export Citation
  • Kim, K., and Moody, P.M., 1992, “More Resources, Better Health? A Cross-National Perspective,” Social Science and Medicine, Vol. 34, pp. 83742.

    • Search Google Scholar
    • Export Citation
  • Kaufmann, D., Kraay, A., and M. Mastruzzi (2003), “Governance Matters III: Governance Indicators for 1996–2002.” World Bank Policy Research Working Paper 3106 (Washington: World Bank).

    • Search Google Scholar
    • Export Citation
  • Krueger, Alan B., and Mikael Lindahl, 2001, “Education for Growth: Why and For Whom?Journal of Economic Literature, Vol. 39, No. 4, pp. 11011136.

    • Search Google Scholar
    • Export Citation
  • Levine, Ross, and David Renelt, 1992, “A Sensitivity Analysis of Cross-Country Growth Regressions,” American Economic Review, Vol. 82, No. 4, pp. 94263.

    • Search Google Scholar
    • Export Citation
  • Lucas, Robert E. Jr. 1988, “On the Mechanic of Economic Development,” Journal of Monetary Economics, Vol. 22, pp. 342.

  • Mankiw, Gregory N., David Romer, and David N., Weil 1992, “A Contribution to the Empirics of Economic Growth,” Quarterly Journal of Economics, Vol. 107, ppp. 407– 37.

    • Search Google Scholar
    • Export Citation
  • Mauro, Paolo, 1998, “Corruption and the Composition of Government Expenditure,” Journal of Public Economics, Vol. 69, No. 2, pp. 26379.

    • Search Google Scholar
    • Export Citation
  • Mauro, Paolo, 1996, “The Effects of Corruption on Growth, Investment, and Government Expenditure “, IMF Working Paper No. 96/98 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • McGuire, A., D. Parkin, D. Hughes, and K. Gerard, 1993, “Econometric Analyses of National Health Expenditures: Can Positive Economics Help Answer Normative Questions?Health Economics, Vol. 2, pp. 11326.

    • Search Google Scholar
    • Export Citation
  • Mingat, Alain, and Jee-Peng Tan, 1992, Education in Asia: A Comparative Study of Cost and Financing (Washington: World Bank).

  • Mingat, Alain, and Jee-Peng Tan, 1998, “The Mechanics of Progress in Education: Evidence from Cross-Country Data,” Policy Research Working Paper No. 2015 (Washington: World Bank).

    • Search Google Scholar
    • Export Citation
  • Musgrove, P., 1996, “Public and Private Roles in Health: Theory and Financing Patterns,” World Bank Discussion Paper No. 339 (Washington: World Bank).

    • Search Google Scholar
    • Export Citation
  • Noss, Andrew, 1991, “Education and Adjustment: A Review of the Literature,” PREM Working Paper WPS 701 (Washington: World Bank).

  • Political Risk Services, International Country Risk Guide (CD-ROM version). Access via the Joint Bank-Fund Library Network Website: http://jolis.

    • Search Google Scholar
    • Export Citation
  • Pritchett, Lant, 1996, “Where Has All the Education Gone?World Bank Policy Research Working Paper No. 1581 (Washington: World Bank).

    • Search Google Scholar
    • Export Citation
  • Pritchett, Lant, and Lawrence H. Summers, 1996, “Wealthier is Healthier,” Journal of Human Resources, Vol. 31 (Fall), pp. 84168.

  • Psacharopoulos, George, 1994, “Returns to Investment in Education: A Global Update,” World Development, Vol. 22, No. 9 (September), pp. 132543.

    • Search Google Scholar
    • Export Citation
  • Psacharopoulos, George, and Harry Anthony Patrinos, 2002, “Returns to Investment in Education: A Further Update,” World Bank Policy Research Paper No. 2881 (Washington: World Bank).

    • Search Google Scholar
    • Export Citation
  • Rajkumar, Andrew Sunil, and Vinaya Swaroop, 2002, “Public Spending and Outcomes: Does Governance Matter?World Bank Working Papers No. 2840 (Washington: World Bank).

    • Search Google Scholar
    • Export Citation
  • Ravallion, Martin, and Shaohua Chen, 1997, “What Can New Survey Data Tell Us About Recent Changes in Distribution and Poverty?World Bank Economic Review, Vol. 11, No. 2, pp. 35782.

    • Search Google Scholar
    • Export Citation
  • Romer, Paul, 1986, “Increasing Returns and Long-Run Growth,” Journal of Political Economy, Vol. 94, No. 5, pp. 100237.

  • Romer, Paul, 1990, “Endogenous Technological Change,” Journal of Political Economy, Vol. 98, No. 5, pp. S71S102.

  • Sachs, Jeffrey and members of the Commission on Macroeconomics and Health, 2003, Macroeconomics and Health: Investing in Health for Economic Development (Geneva: World Health Organization).

    • Search Google Scholar
    • Export Citation
  • Sala-i-Martin, Xavier, 1997, “I Just Ran Two Million Regressions,” American Economic Review, Vol. 87, pp. 17883.

  • Sen, Amartya, 1999, Development as Freedom (Alfred A. Knopf Inc.: New York).

    • Search Google Scholar