Foreign aid and export diversification in developing countries

Abstract

This paper analyzes the effect of foreign aid on export diversification for a sample of developing countries while controlling for the effects of other factors that determine export diversification. We find that foreign aid not exceeding 20% of a country's GDP significantly promotes export diversification, while foreign aid in excess of 20% of GDP significantly impedes export diversification. The latter result corroborates evidence from related literature, which has shown that foreign aid can have an anti-export bias due to a Dutch disease effect. However, our results show that aid as a percent of GDP is below 20% in most low-income countries. This implies that in many low-income countries, varying amounts of additional aid can be used to enhance export diversification without causing a Dutch disease effect. As in the previous literature, we find that the level of development, infrastructure, transactions costs, and natural resources significantly affect export diversification. Our results are robust to the use of two different export diversification measures and different sub-samples.

Keywords:

• foreign aid
• export diversification
• developing countries

JEL Classifications:

• F35
• O35

Notes

1. See for example Andersson et al. (2005) and Bonaglia and Fukasaku (2003).

2. Aid/GDP could not be measured in logs because there are some negative aid numbers in our dataset. The decision to use log or level forms for the remaining independent variables was based on the form that yielded the most significant results. The dependent variable in each equation is measured in logs because this is the approach used by other studies in the literature. We also experimented with functional forms in which all the independent variables are in levels, but the results were not significant.

3. We also experimented with time fixed effects to capture global biases such as implementation of global trade agreements, but these did not improve our results, and were not used in our final estimation.

4.  where x_i is export of good i, and n is the total number of export products. When exports are evenly distributed over a large number of products, HI approaches zero; and when there is a single export product, HI is one. Thus, the more diversified the composition of exports, the smaller is the value of the Herfindahl index.

5. We also tried other measures of natural resource abundance such as dummies for oil exporters and resource rich countries, but the results were not significant.

6. Elbadawi (1999) for example uses a corruption index to measure transactions costs.

7. Given the small size of our sample, we were unable to perform time series analysis with the panel data.

8. The extent of export diversification can also affect aid. For example, in 2005, the WTO in partnership with the World Bank and IMF launched the aid for trade initiative, which was designed to provide aid that will accelerate export growth and diversification in poor countries.

9. All the instruments we use are significant in explaining aid.