Aid and Growth in Small Island Developing States

Abstract

Aid flows to small island developing states (SIDS) are enormous by international standards when compared to the size of their economies. Yet these countries face many severe economic challenges and many have experienced declines in the living standards of their citizens. This paper looks at the impact of aid on what is treated as a necessary precondition for improvements in living standards, typically defined. Specifically, it examines the impact of foreign aid on real per capita income growth in SIDS by econometrically analysing cross-country data for the period 1980 to 2004. A variety of econometric techniques and measures of aid are used. Results suggest that foreign aid is effective at spurring economic growth but with diminishing returns.

Acknowledgements

This paper has been prepared within the UNU-WIDER project entitled Fragility and Development, which is directed by Mark McGillivray and Amelia Santos-Paulino.

The authors gratefully acknowledge comments from anonymous referees and comments from participants of the UNU-WIDER Fragility and Development Project Meeting held in Fiji in December 2006. The usual disclaimer applies.

UNU-WIDER gratefully acknowledges the financial contribution to the project by the Australian Agency for International Development (AusAID) and the United Kingdom Department for International Development (DFID). Simon Feeny's work is partly supported by the Australian Research Council and World Vision Australia, grant LP0562486. The views expressed in this paper are those of the authors and not necessarily those of the funding organisations or their respective employers.

Notes

1. Feeny (2006) finds no evidence of foreign aid impacting on the rural sector in Melanesian countries, proxied by agricultural GDP growth. However, foreign aid is found to impact positively on overall economic growth.

2. An alternative approach would have been to take the panel datasets used in recent well cited aid effectiveness studies and augment the empirical model with a SIDS dummy variable and a multiplicative aid–SIDS interaction variable. Problems relating to data availability prevented this exercise. Datasets include only a very small number of SIDS due a paucity of relevant data for these countries.

3. Unlike many previous studies, a country's budget balance is not included since this variable is only available for a small number of SIDS.

4. This argument relies on an absence of temporal persistence in the aid variable. The high degree of volatility of aid to SIDS shown in Figure 1 lends support to the argument.

5. Strictly speaking it will be a lagged or non-contemporaneously endogenous variable, which equates to it being exogenous econometrically. However, it is noted that the aid variable could still be endogenous if there is an omitted variable that is correlated with both aid and growth.

6. This suggests that examining the dynamics of the impact of foreign aid warrants further attention. The dynamics of foreign aid has been largely neglected by the existing aid effectiveness literature.

7. The main results are robust to using disaggregated aid (see Table A3 and A4 in the Appendix). Foreign aid can be disaggregated into aid grants versus aid loans and bilateral versus multilateral aid. Results using annual data are provided in Table A3 in the Appendix. Again various lags of the aid variables are used. Results indicate that grants and bilateral aid are the categories which spur economic growth in SIDS. These findings are confirmed by results using averaged data and GMM estimations provided in Table A4.

8. Note that if all SIDS received a level of aid which accounted for 35 per cent of their GDP, there would be large absolute dollar increases to the larger SIDS.

9. Government finance statistics are rarely available for SIDS. Analysis of the data that are available for just a few countries over the sample period suggest that aid accounts for an average of 20 per cent of government expenditures in SIDS. However, there is a great deal of variation across countries with aid accounting for just 2.4 per cent of government expenditures in Barbados and as high as 80 per cent of government expenditures in Vanuatu.

10. Feeny (2007) finds that aid has led to lower tax revenues in Melanesian SIDS. Conversely, Clist and Morrissey (2010) find no evidence that aid is associated with lower tax/GDP ratios and indeed there is some evidence that aid encourages tax effort since the mid 1980s.