Political Power and Aid-Tying Practices in the Development Assistance Committee Countries

Abstract

This paper uses a panel of 22 OECD Development Assistance Committee countries to examine whether fragmentation of executive power and the degree of competition from the legislative branch of government increased the amount of tied aid over the period 1979–2009. Fragmentation is defined as the degree to which the costs of a dollar of aid expenditure are internalized by decision-makers and is measured as the number of decision-makers in government. Legislative competition is defined as the relative strength of the government in relation to the legislature. Three variables are used to capture this effect. The empirical results show tied aid, both in levels and as a percentage of total aid, increases as the number of decision-makers within the government increases, and decreases as the proportion of excess seats a governing coalition holds above a simple majority increases.

Notes

¹ The DAC has 24 members but the European Union was excluded because it is not a unified country and South Korea was excluded because it was not a DAC member until 1 January 1 2010. The DAC countries included are Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Japan, Luxembourg, the Netherlands, New Zealand, Norway, Portugal, Spain, Sweden, Switzerland, the UK and the USA.

² Huber et al. (1993) and Noel & Therien (1995) highlight this point by arguing that foreign aid is the international equivalent of domestic social spending, spending which is driven by government decision-makers. However, one significant way foreign aid differs from domestic budget allocations is that the beneficiaries and financiers of foreign aid live in different countries. This gap increases the likelihood that policymakers will put domestic interests before foreign interests in aid allocation (Svensson, 2006).

³ Senator Sessions has received direct campaign contributions from employees of Alatech Healthcare (http://www.opensecrets.org/index.php)

⁴ Practices zealously pursued pass into habit.

⁵ Since 1960, DAC countries have given 74.5% of total ODA.

⁶ The 1969 Pearson Commission and the 2005 Paris Declaration on Aid Effectiveness are two examples.

⁷ Bilateral aid is used because aid is fungible (Boone, 1996) and multilateral aid is difficult to tie. Therefore, bilateral aid is easier for the donor to control.

⁸ As a robustness check, Percentage Tied Aid is used as the dependent variable so Total Aid is dropped as an independent variable. Percentage Tied Aid is total bilateral ODA commitments tied divided by total bilateral ODA commitments.

⁹ Commitments rather than disbursements are used for three reasons. First, disbursements and commitments are highly correlated and estimation results are unlikely to be affected (Neumayer, 2003). Second, donors have complete control of commitments (Berthelemy & Tichit, 2004). Last, the data availability for commitments is greater than for disbursements.

¹⁰ Excess Seats is not collected for the USA because the bicameral system makes it difficult to determine the excess number of seats if the legislative chambers are controlled by different parties.

¹¹ For the data set, the average number of seats that constitute a simple majority is 148. Therefore, a 1% increase in the proportion of excess seats is an increase of 1.48 seats.

¹² The asterisks in the tables represent the statistical significance of individual t-values. When discussing regressions containing interaction terms, the text reports the statistical significance of joint hypothesis testing. This distinction is especially important when interpreting the results from Tables 3 and 4 since the individual t-statistics for the variables of interest are statistically insignificant but statistically significant in the joint hypothesis F-test (see Wooldridge, 2006, pp. 205–206).

¹³ Following Wooldridge (2006), the political variables are evaluated at the mean value of GDP per capita, a more meaningful value of GDP per capita than zero. Therefore, the coefficients in the text require an adjustment from the coefficients in the tables. For example, to calculate the $99.59 million value of Spending Ministers in the text, the coefficient of Spending Ministers × GDP per capita is multiplied by the mean value of GDP per capita and added to the coefficient of Spending Ministers. Before the coefficients are rounded in the tables ( − 0.0047037 × 27 343)+(228.21) = 99.59.

¹⁴ The year 1979 was dropped because it contained the fewest observations. Keeping 1979 and dropping 2009 did not alter the results.