nach oben

Studies in Comparative International Development

Erschienen in:

01.06.2007

Reassessing the Effect of Economic Growth on Well-being in Less-developed Countries, 1980–2003

verfasst von: David Brady, Yunus Kaya, Jason Beckfield

Erschienen in: Studies in Comparative International Development | Ausgabe 1-2/2007

Einloggen

Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.

search-config

KI-gestützte Suche

Aus

Abstract

Development debates have been greatly influenced by the growth consensus: the conventional wisdom that economic growth should be the primary priority for less-developed countries (LDCs) because it most effectively improves the well-being of the world’s poor. We compare the impact of growth to other independent variables in an unbalanced panel analysis of up to 109 LDCs and 580 observations across six time points (1980, 1985, 1990, 1995, 2000, and 2003). Our dependent variables include caloric consumption, infant survival probability, one-to-five year survival probability, female life expectancy, and male life expectancy. First, we find that gross domestic product (GDP) has significant positive effects on caloric consumption, female life expectancy, and male life expectancy. Second, GDP does not have robust effects on infant and one-to-five survival probabilities. Third, fertility, urbanization, and secondary school enrollment have larger effects than GDP in the majority of models. The more powerful effects of fertility, urbanization, and secondary schooling cannot simply be attributed to an indirect effect of GDP. Fourth, we find that dependency variables do not have robust significant effects. Fifth, over time, GDP has become much less effective at improving caloric consumption and infant and one-to-five survival. We infer that there are serious limitations to concentrating exclusively on economic growth to improve well-being in LDCs.

Nächster Artikel Social Capability, History and the Economies of Communist and Postcommunist States

Sie haben noch keine Lizenz? Dann Informieren Sie sich jetzt über unsere Produkte:

Springer Professional "Wirtschaft+Technik"

Online-Abonnement

Mit Springer Professional "Wirtschaft+Technik" erhalten Sie Zugriff auf:

über 102.000 Bücher
über 537 Zeitschriften

aus folgenden Fachgebieten:

Automobil + Motoren
Bauwesen + Immobilien
Business IT + Informatik
Elektrotechnik + Elektronik
Energie + Nachhaltigkeit
Finance + Banking
Management + Führung
Marketing + Vertrieb
Maschinenbau + Werkstoffe
Versicherung + Risiko

Jetzt Wissensvorsprung sichern!

Jetzt informieren

Springer Professional "Wirtschaft"

Online-Abonnement

Mit Springer Professional "Wirtschaft" erhalten Sie Zugriff auf:

über 67.000 Bücher
über 340 Zeitschriften

aus folgenden Fachgebieten:

Bauwesen + Immobilien
Business IT + Informatik
Finance + Banking
Management + Führung
Marketing + Vertrieb
Versicherung + Risiko

Jetzt Wissensvorsprung sichern!

Jetzt informieren

Nur mit Berechtigung zugänglich

In sensitivity analyses in Appendix I, we test the rate of change in GDP (i.e., growth) and find identical results.

Much of Firebaugh and Beck’s focus was on challenging dependency/world systems claims about the damaging influence of unequal trade and exploitive FDI. They criticized semi-difference analyses stressing the lower well-being caused by trade and FDI (London and Williams 1990; Stokes and Anderson 1990; Wimberly 1990; Wimberly and Bello 1992).

Several have disputed Firebaugh (1992) and sought to show that FDI actually harms the growth of LDCs (Dixon and Boswell 1996), but, to our knowledge, no response to Firebaugh and Beck’s 1994 article has appeared.

Throughout, we are referring to material well-being (i.e., health), not psychological/subjective well-being.

We do not mean to imply that we are the first to problematize GDP as an unqualified predictor of well-being, as there is much literature in that vein (e.g., Morris 1979).

The most common strategy remains less than 75 countries with the dependent variable at one time point (e.g., 1995) as a function of the dependent variable and independent variables many years before (e.g., 1980). This is the semi-difference model criticized by Firebaugh and Beck, which suffers from both problems.

Despite using a semi-difference cross section of 59 LDCs, Shandra et al. (2003) conclude, “We need more longitudinal data both for our dependent and independent variables to understand the effects of these determinants on infant mortality. This would make it possible to pool cross-sectional data so as to increase the number of degrees of freedom making it possible to include more control variables in one model. We need data for more countries and for more time points.” The major reason that studies as encompassing as ours have not been done is that a sample this large has been almost impossible to assemble before recent improvements in international data reporting.

Nigeria’s per capita GDP in purchasing power parity dollars steadily declined: 1,220.52 (1980), 925.68 (1990), 890.60 (1995), 785.30 (2000) and 746.21 (2003). A two-time point approach would probably show a positive relationship between GDP and infant mortality. From 1980 to 2003, both GDP and infant mortality declined. In actuality, as GDP declined, infant mortality worsened from 1980 to 1995, and multiple time points would illustrate this.

We referred to 5 additional controls above. We discuss three in this section, and the 2 others (time and an interaction of GDP and time) in the next section.

Urbanization negatively correlates with agricultural employment as a percent of total employment (−0.8) and positively correlates with manufacturing employment (0.6). Agricultural employment negatively correlates with and manufacturing employment positively correlates with our well-being measures (calories −0.6 and 0.7, infant survival −0.8 and 0.7, one to five survival −0.8 and 0.7, female life expectancy −0.8 and 0.7, and male life expectancy −0.7 and 0.7) (World Bank 2005).

Some readers might be skeptical that one can sort out the influence of these factors from each other and/or from the effect of GDP. As we go to great lengths to show below, these variables are sufficiently independent that collinearity is not a serious problem. Indeed, omitted variable bias is more problematic than any potential collinearity or endogeneity.

Using a sample of LDCs and five-year time points from 1960–1985, Pritchett and Summers (1996: 852–853) show significant effects of year dummies. After controls, infant mortality fell 5% every 5 years.

We disagree with claims that panels like ours are an apparent population rather than a sample. The population should be viewed as all country-years between 1980 and 2003 (assuming 130 LDCs, this works out to roughly 3,120 possible cases). Our sample of 587 cases is roughly 19% of that population. We sample by choosing the intervals of 1980, 1985...2003 as opposed to all years, and we sample by choosing all countries that have data for at least two time points. Moreover, these data reflect sample statistics that estimate a population parameter (e.g., life expectancy) and were sampled at various points during a given year. As Bollen (1995) argues, more data can always be collected and what seems like a population is really the hidden compilation of sampling at various stages.

Firebaugh and Beck examined female and male life expectancy at age 1—“to avoid redundancy with the infant survival probability” (p.640). However, the World Bank provides data on life expectancy at birth. We would lose many cases if we recalculate life expectancy from age 1. Moreover, as we show below, since the results for life expectancy and the child mortality are quite divergent, we are not concerned by the redundancy between life expectancy and child mortality.

The bivariate correlations range from 0.63 between average calories and one-to-five years survival, to 0.93 between female life expectancy and male life expectancy.

Appendix I shows that BIC′ very strongly prefers our models over models with the natural log of all variables for 4 of 5 dependent variables.

In our sample, the only variables that have high skew values are debt service and inward FDI. While this might contribute to those variables’ insignificance below, logging them does not change any of our conclusions.

Firebaugh (1999) persuasively argues that purchasing power parity dollars are best for comparing LDCs, but Firebaugh and Beck used exchange rate standardized dollars. Firebaugh (1999) used the Penn World Tables, while we use the World Bank data. We do so because the Penn data stop in 2000, and have considerably more missing cases. Appendix I shows that if we substitute the Penn data, our conclusions would be consistent, but BIC′ strongly prefers the World Bank models.

We replicated our models while substituting FDI flows for FDI stock and the conclusions were consistent.

We experimented with several measures of debt: debt as a percent of GDP or exports, debt service as a percent of exports, and Przeworski and Vreeland’s (2000) measure of International Monetary Fund agreements. None of these was ever significant, so we present the debt service measure, which is straightforward and available most consistently.

Firebaugh and Beck (1994: 636) generally endorse panel models: “We argue that the use of panel data is appropriate for the questions raised by dependency theory, but the particular method used in dependency research is suspect.”

Fixed effects (FE) models and difference models explain historical variation within countries while removing the variation between countries. FE models perform OLS after including country-specific constants and subtracting all variables from their country-specific means. Between-effects (BE) models (and OLS of one cross section) explain the between-country variation while removing the variation within countries. The RE model is the matrix-weighted average of the within- (FE) and between-countries (BE) estimators (Greene 1990: 488; Hsiao 2003). RE models include a country-specific error term in addition to the general error term and, subtract a portion of the country-specific means. Cross-national differences in GDP and well-being are not constant over time, but relative stability exists in the cross-national ranking of countries for these variables — hence, FE and difference models effectively mask this crucial variation (Beck and Katz 2001: 487, 492). Further, trends are essential as well, and unfortunately, BE (and one time point cross section) models mask this essential within-country variation.

For every variable except one (including all dependent variables, excepting debt service), the standard deviation between countries is much more than twice as large as the standard deviation within countries.

In an earlier article, Firebaugh endorses our decision to preserve the cross-national variation. Comparing models of cross-national and historical variation, Firebaugh (1980: 342) writes, “This issue may be stated as follows: does covariance over time or covariance across nations better capture the underlying causal process? Determining which covariance is more appropriate is a theoretical question, not a methodological one. In short, the question of whether to use cross-national or historical methods in comparative study must be determined ultimately by theoretical considerations, not by methodological fiat.” Our RE-AR1 models include both cross-national and historical variation.

Only one alternative in Appendix I is preferred over the main models we present. This alternative would be an RE-AR1 model with the same variables, but also including regional dummies. Despite the virtues of such models, we chose to confine it to Appendix I since we want our results to be comparable with Firebaugh and Beck. Regardless, none of our conclusions would be different with this alternative specification.

As Appendix I reveals, the fit of the models predicting the dependency measures is far too weak to suggest collinearity is a problem and that their independent effects cannot be discerned.

Pritchett and Summers (1996) execute a variety of sophisticated techniques to report a significant effect of GDP for infant mortality, but they only control for years of schooling and year dummies. Our results suggest that all of their sophisticated techniques cannot substitute for a more complete model specification. These results suggest that their reported significant effects may be due to omitted variable bias (see also Appendix I).

Fertility has such large effects partly because it is linked with infant survival in the demographic transition. Infant survival also influences fertility. This relationship is complicated though hardly overpowering (e.g., the replacement effect is stronger in poorer societies and families, [Preston 1975a; Heuveline 2001; Bongaarts and Watkins 1996]), and any reverse causality is no worse than the reverse causality between average calories or life expectancy and GDP (Pritchett and Summers 1996). Also, see the discussion of the endogeneity test above.

Since GDP significantly influences secondary schooling and urbanization in Appendix I and those two influence infant survival in Table 2, one could conclude that GDP influences infant survival indirectly. However, GDP would have a very small indirect effect for infant or one to five years survival. In Appendix I, GDP has a moderate effect on secondary schooling (beta = 0.2) and a very small effect on urbanization (beta = 0.053). Multiplying those betas by the betas of secondary schooling and urbanization for infant survival (Table 2), GDP’s standardized indirect effect would be 0.024 through secondary schooling and 0.016 through urbanization. Setting aside the assumptions in these path analyses, the growth consensus has never advocated that GDP simply has indirect, small or conditional effects.

Some prefer the under-5 survival to our two measures, but the results are consistent with that alternative.

Since infant survival is a large component of life expectancy, much of the effect of fertility for life expectancy (male and female) may reflect the connection between fertility and infant mortality.

This finding is probably indicative of Goesling and Firebaugh’s (2004) conclusion that life expectancy declined in some LDCs (especially Sub-Saharan Africa) in the 1990s because of the AIDS crisis. Since the AIDS crisis is unobserved in this model, it manifests in a negative year effect. If we drop the Sub-Saharan African countries, year does not have a significant effect on female life expectancy (t = 0.83). Unfortunately, data on AIDS are not sufficiently available such that we could include them in our models (data for prevalence is only available in 1999, and data on death rates is only available in 2001 and 2003). We hope future research can accommodate this factor.

If we drop the Sub-Saharan African countries, year is nearly significantly positive (t = 1.7).

Most of the region dummies were insignificant, with these exceptions (details available upon request): Sub-Saharan Africa has a significantly lower infant survival, one-to-five survival, female and male life expectancy; the Middle East and North Africa have a significantly higher average calories and male life expectancy; Eastern Europe and Central Asia have a significantly higher infant survival and female life expectancy; and Latin America and the Caribbean have a significantly higher male and female life expectancy (East Asia as reference).

Despite its anomalousness compared to other models, the significance of logged GDP in the RE-AR1 models for infant and one to five survival might vindicate Firebaugh and Beck. However, even though GDP is significant at the 0.05 level for both dependent variables, logged secondary schooling, fertility and urbanization are significant at the 0.001 level. More important, GDP’s beta for infant survival and one to five survival (0.086 and 0.075) is dramatically smaller than the betas for secondary schooling (0.159 and 0.165), fertility (−0.340 and −0.186), and urbanization (0.349 and 0.384).

On balance, GDP could be just as vulnerable to the charge of endogeneity. Endogenous growth theory claims education boosts GDP, so it could be that GDP is masking secondary schooling’s true effect. As Appendix I shows, GDP could be endogenous to secondary schooling, exports, raw material exports, and urbanization.

We would not dispute that fertility is associated with development generally (i.e., in the demographic transition). Bongaarts and Watkins (1996) show that fertility is negatively associated with the Human Development Index (HDI). The HDI is a broader measure than GDP, and includes many of the other dimensions of development that we argue are important (e.g., education) as well as some of our dependent variables. As Heuveline (2001) shows, the link between economic development and fertility decline is complicated (see also footnote 28).

It is important to acknowledge that Sen has made the point that economic growth may indirectly benefit the poor’s well-being through increased public revenues. Whether such increased revenue is turned into spending for the poor is an open question. Perhaps future research can test this indirect relationship between GDP and well-being, as well as the political question of how GDP translates to differential public spending.

Axinn WG, Barber JS. Mass education and fertility transition. Am Sociol Rev 2001;66:481–505.CrossRef

Beck N. Time-series-cross-section data: what have we learned in the past few years. Annu Rev Pol Sci 2001;4:271–93.CrossRef

Beck N, Katz JN. Throwing out the baby with the bath water. Int Organ 2001;55:487–95.CrossRef

Beckfield J. Does income inequality harm health? New cross-national evidence. J Health Soc Behav 2004;45:231–48.CrossRef

Bhagwati J. In defense of globalization. New York: Oxford University Press; 2004.

Blumberg RL. Engendering wealth and well-being. Boulder, CO: Westview; 1995.

Boehmer U, Williamson JB. The impact of women’s status on infant mortality rate: a cross-national analysis. Soc Indic Res 1996;37:333–60.CrossRef

Bollen K. Apparent and nonapparent significance tests. Sociol Method 1995;25:459–68.CrossRef

Bongaarts J, Watkins SC. Social interactions and contemporary fertility transitions. Popul Dev Rev 1996;22:639–82.CrossRef

Bose CE, Acosta-Belen E. Women in the latin American development process. Philadelphia: Temple University Press; 1995.

Bradshaw YW, Huang J. Intensifying global dependency: foreign debt, structural adjustment, and third world underdevelopment. Sociol Q 1991;32:321–42.CrossRef

Bradshaw YW, Noonan R, Gash L, Buchmann-Sershen C. Borrowing against the future: children and third world indebtedness. Soc Forces 1993;71:629–56.CrossRef

Bradshaw YW, Wallace M. Global inequalities. Thousand Oaks, CA: Pine Forge; 1996.

Buchmann C, Hannum E. Education and stratification in developing countries: a review of theories and research. Annu Rev Sociology 2001;27:77–102.CrossRef

Caldwell JC. Routes to low mortality in poor countries. Popul Dev Rev 1986;12:178–220.

Cutler D, Miller G. The role of public health improvements in health advances: the twentieth century United States. Demography 2005;42:1–22.CrossRef

Dixon WJ, Boswell T. Dependency, disarticulation, and denominator effects: another look at foreign capital penetration. Am J Sociol 1996;102:543–62.CrossRef

Dollar D, Kraay A. Growth is good for the poor. Washington, DC: World Bank; 2000.

Dreze J, Sen A. India: economic development and social opportunity. Delhi: Oxford University Press; 1995.

El-Ghannam AR. The determinants of social well-being, economic development, and development index in the third world countries. Perspectives on Global Development and Technology 2002;1:51–69.CrossRef

FAO (2005) Food balance sheets FAOSTAT-Agriculture. Available from <http://faostat.fao.org/faostat/form?collection=FBS&Domain=FBS&servlet=1&hasbulk=&version=ext&language=EN>.

Firebaugh G. Growth effects of foreign and domestic investment. Am J Sociol 1992;98:105–30.CrossRef

Firebaugh G. Empirics of world income inequality. Am J Sociol 1999;104:1597–630.CrossRef

Firebaugh G. The new geography of global income inequality. Cambridge, MA: Harvard University Press; 2003.

Firebaugh G, Beck FD. Does economic growth benefit the masses? Growth, dependence, and welfare in the third world. Am Sociol Rev 1994;59:631–53.CrossRef

Firebaugh G, Goesling B. Accounting for the recent decline in global income inequality. Am J Sociol 2004;110:283–312.CrossRef

Forsythe N, Korzeniewicz RP, Durant V. Gender inequalities and economic growth: a longitudinal evaluation. Econ Dev Cult Change 2000;48:573–617.CrossRef

Fosu GB. Access to health care in urban areas of developing societies. J Health Soc Behav 1989;30:398–411.CrossRef

Frey RS, Field C. The determinants of infant mortality in the less developed countries: a cross-national test of five theories. Soc Indic Res 2000;52:215–34.CrossRef

Goesling B, Firebaugh G. The trend in international health inequality. Popul Dev Rev 2004;30:131–46.CrossRef

Gortmaker, SL, Wise PH. The first injustice: socioeconomic disparities, health services technology, and infant mortality. Annu Rev Sociology 1997;23:147–70.CrossRef

Greene WH. Econometric analysis. New York: Macmillan; 1990.

Hales S, Howden-Chapman P, Salmond C, Woodward A, Mackenbach J. National infant mortality rates in relation to gross domestic product and distribution of income. The Lancet 1999;354:2047.CrossRef

Hannum E, Buchmann C. The consequences of global educational expansion: social science perspectives. Cambridge, MA: Academy of Arts and Sciences; 2003.

Hayek FA. The road to serfdom. Chicago: The University of Chicago Press; 1994 [1944].

Hertzman C, Siddiqi A. Health and rapid economic change in the late twentieth century. Soc Sci Med 2000;51:809–19.CrossRef

Heuveline P. Demographic pressure, economic development, and social engineering: an assessment of fertility declines in the second half of the twentieth century. Popul Res Policy Rev 2001;20:365–96.CrossRef

Hobcraft JN, McDonald JW, Rutstein SO. Socio-economic factors in infant and child mortality. A cross-national comparison. Popul Stud 1984;38:193–223.CrossRef

Hsiao C. Analysis of panel data, 2nd ed. New York: Cambridge University Press; 2003.

Jenkins JC, Scanlan SJ. Food security in less developed countries, 1970 to 1990. Am Sociol Rev 2001;66:718–44.CrossRef

Lall S. The technological structure and performance of developing country manufactured exports, 1985–1998. Oxf Dev Stud 2000;28(3):337–69.CrossRef

Li L, Ballweg J. Development, urbanization and fertility in China. Sociol Soc Res 1992;76:111–7.

London B, Williams BA. National politics, international dependency, and basic needs provision: a cross-national analysis. Soc Forces 1990;69:565–84.CrossRef

Mill JS. Utilitarianism. Toronto: University of Toronto Press; 1963 [1861].

Morris MD. Measuring the conditions of the world’s poor. Oxford: Pergamon; 1979.

Nussbaum M. Women and human development. New York: Cambridge University Press; 2000.

Nussbaum M. Women’s education: a global challenge. Signs 2004;29(2):325–55.CrossRef

Nussbaum M, Sen A. The quality of life. Oxford: Clarendon; 1993.

Parpart JM, Connelly P, Barriteau VE. Theoretical perspectives on gender and development. Ottawa: IDRC; 2000.

Preston SH. Health programs and population growth. Popul Dev Rev 1975a;1:189–99.CrossRef

Preston SH. The changing relation between mortality and level of economic development. Popul Stud 1975b;29:231–48.CrossRef

Preston SH. Population studies of mortality. Popul Stud 1996;50:525–36.CrossRef

Pritchett L, Summers LH. Wealthier is healthier. J Hum Resour 1996;31:841–68.CrossRef

Przeworski A, Vreeland JR. The effect of IMF programs on economic growth. J Dev Econ 2000;62:385–421.CrossRef

Raftery A. Bayesian model selection in social research. Sociol Method 1995;25:111–63.CrossRef

Ravallion M. Growth, inequality and poverty: looking beyond averages. World Dev 2001;29:1803–15.CrossRef

Sachs JD. Can extreme poverty be eliminated? Sci Am 2005;(August 22).

Schafer M. International non-governmental organizations and third world education in 1990: a cross-national study. Sociol Educ 1999;72:69–88.CrossRef

Sen A. Development as freedom. New York: Anchor; 1999.

Shandra JM, Nobles J, London B, Williamson JB. Dependency, democracy, and infant mortality: a quantitative, cross-national analysis of less developed countries. Soc Sci Med 2003;59:321–33.CrossRef

Shen C, Williamson JB. Child mortality, women’s status, economic dependency, and state strength: a cross-national study of less-developed countries. Soc Forces 1997;76:667–94.CrossRef

Stiglitz JE. Globalization and its discontents. New York: Norton; 2002.

Stokes RG, Anderson AB. Disarticulation and human welfare in less developed countries. Am Sociol Rev 1990;55:63–74.CrossRef

UNCTAD (2005) FDI Database. Available from <http://stats.unctad.org/fdi/eng/ReportFolders/Rfview/Explorerp.asp?CS_referer=>.

United Nations (2005) U.N. Statistical common database. Available from <http://unstats.un.org/unsd/cdb/cdb_simple_data_extract.asp>.

Wickrama KAS, Lorenz FO. Women’s status, fertility decline, and women’s health in developing countries: direct and indirect influences of social status on health. Rural Sociol 2002;67:255–77.CrossRef

Wimberly DW. Investment dependence and alternative explanations of third world mortality: a cross-national study. Am Sociol Rev 1990;55:75–91.CrossRef

Wimberly DW, Bello R. Effects of foreign investment, exports, and economic growth on third world food consumption. Soc Forces 1992;70:895–921.CrossRef

Wooldridge JM. Introductory econometrics: a modern approach mason. OH: South-Western College Pub; 2002.

World Bank (2005) World Development Indicators Online. Available from <http://devdata.worldbank.org/dataonline>.

Titel: Reassessing the Effect of Economic Growth on Well-being in Less-developed Countries, 1980–2003
verfasst von: David Brady
Yunus Kaya
Jason Beckfield
Publikationsdatum: 01.06.2007
Verlag: Springer-Verlag
Erschienen in: Studies in Comparative International Development / Ausgabe 1-2/2007
Print ISSN: 0039-3606
Elektronische ISSN: 1936-6167
DOI: https://doi.org/10.1007/s12116-007-9003-7

Springer Professional

Abstract

Bitte loggen Sie sich ein, um Zugang zu Ihrer Lizenz zu erhalten.

Sie haben noch keine Lizenz? Dann Informieren Sie sich jetzt über unsere Produkte:

Springer Professional "Wirtschaft+Technik"

Springer Professional "Wirtschaft"

Weitere Artikel der Ausgabe 1-2/2007

Participation in Transition: Mobilizing Ugandans in Constitution Making

State, Business, and Economic Growth in India

Autocracy, Elections, and Fiscal Policy: Evidence from Malaysia

Social Capability, History and the Economies of Communist and Postcommunist States

Dependency or Institutions? Economic Geography, Causal Mechanisms, and Logic in the Understanding of Development

State Involvement and Forest Co-Governance: Evidence from the Indian Himalayas

Premium Partner