The effectiveness of aid on savings and investment in Sub-Saharan Africa: do volatility and institutional quality matter?

ABSTRACT

The effectiveness of aid is somewhat subject to its volatility. This study investigates the impact of aid volatility, especially on savings and investment, with a particular focus on the role of institutional quality in 45 Sub-Saharan African (SSA) countries over the period from 1990 to 2013. We used the Generalized Method of Moments (GMM) technique combined with a measure of volatility, namely an averaged data method involving six 4-year sub-periods. Our results suggest that while aid is positively associated with savings and investment, its volatility is harmful to savings and investment. However, when higher quality institutions exist, the volatility of aid has a less negative impact on savings and investment.

The policy implications are diverse; first, donors and SSA aid-dependent countries need to take into account the diversity of shocks to which aid can respond in mitigating its unpredictability. Secondly, SSA countries should be encouraged to establish the conditions for better quality institutions to mitigate the negative effects of the volatility of aid on macroeconomic aggregates such as savings and investment.

KEYWORDS:

• Aid
• savings
• investment
• volatility
• institutional quality
• Sub-Saharan Africa

JEL CLASSIFICATION:

• E2
• F35
• O55

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ The few small-scale studies that have tried to break down aid into various subcategories have either rejected parameter constancy across regions (White, 1992; and Mosley, Hudson, and Horrell, 1987 and 1992) or found the impact of aid to be insignificant in a cross-country set up (Mosley, Hudson, and Horrell, 1992). Individual country studies are also inconclusive (Pack and Pack, 1990 and 1993).

² Altruistic factors such as age dependency ratio and standard of living, and portfolio variables (interest rate volatility and exchange rate volatility).

³ In accordance with Lensink and Morrissey (2000), volatility is measured as a 4-year rolling standard deviation of the change in the variable and expressed in logarithmic terms.

⁴ Studies of aid volatility’s impact vary according to the welfare criterion selected as a dependent variable: Mosley and Suleiman (2006) take the poverty headcount, and Lensink and Morrissey (2000) the investment rate. Both find a negative impact of aid volatility on the criterion variable chosen.

⁵ See Lele & Goldsmith, (1989) for the case of aid to the agricultural sector.

⁶ The reduction of macroeconomic fluctuations in developing countries and emerging markets, the issue of whether opening up developing countries’ banking sectors to foreign entries might help significantly lower aggregate volatility.

⁷ Bulir and Hamann (2003, 2008) argued that the volatility of aid is greater than that of government revenue, which increases over time and is pro-cyclical (in other words, aid flows are directly correlated with the level of government expenditure in any particular year).

⁸ Aid effectiveness is also often reduced by aid volatility, both positive and negative, and abundant literature addresses optimal aid allocation, generally in the context of allocating aid between countries.

⁹ See also Harris and Sevestre (2008) for an overview of different dynamic panel models

¹⁰ Note that country-fixed effects are not explicitly included in system GMM estimation. The system includes equations in levels and differences. They are purged in the differenced equations and are not introduced in the level equations since they would introduce bias (Roodman 2009)

¹¹ Bulir and Hamann note that (i) the Hodrick–Prescott filter may create spurious serial correlations in de-trended data and (ii) end-period observations have greater mean square errors than observations in the middle of the sample (Cogley & Nason, 1995). Bulir and Hamann also note the relatively little difference between various methods of identifying residuals and that, for example, a first difference operator produces similar results to those using the Hodrick–Prescott filter

¹² Uganda, Kenya, Ethiopia, and Zimbabwe are good examples. For a comprehensive discussion of the democratic transition in Africa, see the collection of studies in the 2011 Special Issue ‘Democratization in Africa: Challenges and Prospects’ of Globalization, Vol. 18, issue 2.