Aid and growth regressions

https://doi.org/10.1016/S0304-3878(00)00150-4Get rights and content

Abstract

This paper examines the relationship between foreign aid and growth in real GDP per capita as it emerges from simple augmentations of popular cross-country growth specifications. It is shown that aid in all likelihood increases the growth rate, and this result is not conditional on ‘good’ policy. There are, however, decreasing returns to aid, and the estimated effectiveness of aid is highly sensitive to the choice of estimator and the set of control variables. When investment and human capital are controlled for, no positive effect of aid is found. Yet, aid continues to impact on growth via investment. We conclude by stressing the need for more theoretical work before this kind of cross-country regression is used for policy purposes.

Introduction

The past 30 years have witnessed the publication of a large number of studies on aid effectiveness. The methods employed to study the subject range from detailed case studies at the project level to regression analyses of the growth impact of aid in samples of almost a hundred countries. The micro and macro analyses have generally been perceived to yield different results; many successful projects leaving no lasting imprint on economy wide growth rates. We concentrate in this paper on the aid–growth relation at the macro level as it emerges from cross-country regressions based on large panel data sets.

We compare recent growth studies in which the relationship between aid and growth is modeled as non-linear. On one side, Burnside and Dollar (2000) have found evidence of threshold effects by the introduction of an aid–policy interaction term in the growth regressions. In their model, aid contributes positively to growth, but only in good policy environments. On the other side, Hadjimichael et al. (1995), Durbarry et al. (1998), and Lensink and White (1999) have found positive, but decreasing marginal returns to aid flows, by the introduction of aid squared. As none of the studies provide models that encompass both possibilities we start by formulating a unified empirical model where quadratic aid and policy terms appear alongside the aid–policy interaction. A reduction to the Burnside–Dollar specification is not supported by the data. This is important in view of the widespread attention paid to the Burnside–Dollar results.

Concern about simultaneity bias in aid–growth regressions, caused by potential endogeneity of aid, is another common feature in recent studies. To our surprise, they all find negligible bias. Due to the importance of this finding for the conclusions drawn about the impact of aid, we re-visit the endogeneity issue and suggest an explanation for why others conclude that aid is exogenous. There are strong signs of biased results once the model is enlarged to take account of country specific effects. We argue that such effects are important and proceed to estimate a fairly standard growth model using ordinary least squares as well as a generalized method of moments estimator that yield consistent estimates, also in the presence of both endogenous regressors and country specific effects. Compared to the other studies mentioned, we find very different (and positive) estimates of the impact of aid.

Turning to the question why aid increases growth, we include investment and human capital in the growth regression. The result is that, conditional on these variables, aid has no effect on growth. We interpret this finding as supportive of the view that aid affects growth via capital accumulation. An equation for gross domestic investment as a share of GDP is also estimated. By introducing aid squared we allow for level effects by which the fraction of aid that is actually invested can change with the aid dependency ratio. It emerges that in the majority of countries, there is a one-to-one relation between increased aid flows and increased investment.

The remainder of the paper is organized as follows: In Section 2, we compare the latest cross-country aid–growth regressions. In Section 3, we give results for the model when endogeneity of aid and country specific effects are taken into account. The model is respecified to include the main determinants of capital accumulation in Section 4, in which we also examine the impact of aid on investment. Section 5 concludes.

Section snippets

Cross-country aid effectiveness studies

Dalgaard et al. (2000) provide a survey of empirical analyses from the last 30 years that make use of cross-country regressions in assessing the effectiveness of foreign aid. From 131 such regressions, a reasonably consistent pattern emerges: (i) aid increases aggregate saving, although not by as much as the aid flow, (ii) aid increases investment, and (iii) aid has a positive effect on the growth rate whenever growth is driven by capital accumulation.

In the 1990s, a new generation of aid

Endogeneity of aid and country specific effects

Boone, 1994, Boone, 1996, Hadjimichael et al. (1995), and Burnside and Dollar (2000) explicitly consider simultaneity bias due to endogeneity of aid. Boone and Burnside and Dollar briefly discuss the reasons for the possible endogeneity of aid in the growth regressions. The main reason is that it is difficult to perceive of aid as a lump-sum transfer, independent of the level of income. Empirically, a negative relation between aid and income per capita is well established.9

Aid, capital accumulation, and growth

Even though the theoretical model underlying modern empirical aid–growth work has moved beyond the Harrod–Domar model, aid is still meant to impact on growth via capital accumulation. To analyze whether aid works through the investment link, it is necessary to show that (i) investment impacts on growth, and (ii) aid impacts on investment. Accordingly, we reformulate the growth regression and include investment and human capital explicitly as shown in Table 4. If aid has an effect on growth,

Conclusion

Aid effectiveness is likely to remain a contentious area of debate. Substantial resources are involved, and the widespread perception that aid has been ineffective in fostering growth at the macro level has led to aid fatigue in many donor countries. In this paper, we have investigated what modern cross-country growth regressions can tell about the effect of aid on aggregate growth. We find that aid increases the growth rate, and this conclusion is not conditional on the policy index

Acknowledgements

This paper was completed while Henrik Hansen was visiting the European University Institute, Florence, Italy. The hospitality of the EUI is gratefully acknowledged.

References (36)

  • R.J Barro et al.

    Economic Growth

    (1995)
  • D.A Belsley et al.

    Regression Diagnostics: Identifying Influential Data and Sources of Collinearity

    (1980)
  • Boone, P., 1994. The impact of foreign aid on savings and growth. Mimeo, London School of...
  • C Burnside et al.

    Aid, policies, and growth

    American Economic Review

    (2000)
  • Chang, C.C., Fernandez-Arias, E., Servén, L., 1998. Measuring aid flows: a new approach. Working paper. The World Bank....
  • R.D Cook et al.

    Residuals and Influence in Regression

    (1982)
  • Dalgaard, C.-J., Hansen, H., 2000. On aid, growth, and good policies. Working Paper, Institute of Economics, University...
  • C.-J Dalgaard et al.

    Aid effectiveness disputed

    Journal of International Development

    (2000)
  • Cited by (547)

    • Aid and social cohesion

      2023, Quarterly Review of Economics and Finance
    View all citing articles on Scopus
    View full text