“Let’s not mince words: we need to deal with the cancer of corruption.” So said former World Bank President James D Wolfensohn back in 1996. But according to Easterly (2007), little has changed since.
In 2002, Alesina and Weder (2002) pointed to donors granting aid indiscriminately to inefficient and dishonest governments. They argued that continuous aid inflows only led to further corruption, eroding governance and perpetuating the “need” for aid in a vicious cycle. The aid community tries to avoid this by making aid conditional on, among other things, more effective control of corruption, but according to critics such as Collier (1997), this conditionality has been nothing short of a failure.
Ex-ante conditions vs ex-post rewards
This is a fair verdict for the usual practice of conditionality. Traditionally, aid is disbursed once recipient governments promise to fight corruption or, more generally, implement reforms deemed necessary by donors. For instance, Kenya promised agricultural reforms in four rounds of negotiation with the World Bank within 15 years, but backtracked each time after having received World Bank aid. Such a breach of promises by recipients is rarely punished by donors, whose aid agencies are mainly interested in spending the funds allocated to them. Indeed, a 1997 parliamentary inquiry in Germany led the government to conclude that “no development cooperation contracts were annulled due to proof of corruption” (Cremer 2008: 122).
Time inconsistency problems might be overcome, however, if donors based their aid allocation on retrospective performance appraisals. Such a regime change to performance-based aid has been proposed by Collier et al. (1997) and Svensson (2003). Its proponents expect that the reform-mindedness of recipients will generally increase once aid is disbursed according to the relative performance of competing countries. But while sceptics admit that selectivity of donors may strengthen the incentives of recipients to improve local conditions prior to receiving aid, they insist that recipients may still reverse reforms after having been selected by donors as eligible for aid (Mosley et al. 2004).
Millennium Challenge Corporation: New evidence on conditional aid
The scarcity of performance-based aid schemes implies that conflicting hypotheses have hardly been subjected to systematic empirical tests. However, the Millennium Challenge Corporation (MCC) provides a major exception. The Bush administration announced the MCC as a new US aid agency in 2002 to prevent institutional legacies from undermining innovative allocation rules. The MCC would command over $5 billion annually and was deliberately shaped in a way to replace the traditional concept of conditionality by ex-post rewards for proven achievements. Moreover, effective control of corruption figures most prominently among MCC’s eligibility criteria. It is mandatory for countries striving for MCC compacts – the multi-year aid programs agreed upon between the MCC and eligible recipients – to score higher than the median of all candidate countries (64 in 2004) with regard to control of corruption. The ranking uses control of corruption as given by the World Bank’s Governance Indicators, ranging from -2.5 to +2.5 with higher values reflecting better control.
In recent research (Öhler et al. 2010), we analyse in detail whether this strict requirement induced MCC candidate countries to fight corruption more effectively. We employ a “difference-in-difference” approach which reveals whether the change in control of corruption before and after the announcement of the MCC was significantly stronger for a treatment group of MCC candidate countries, compared to a control group. In our baseline estimations, the treatment group consists of the quartile of candidate countries that scored slightly below the median with respect to control of corruption. The underlying argument is that the costs of compliance (i.e., surpassing the median by fighting corruption) are relatively low for this quartile, while the prospect of being rewarded by MCC aid is relatively favourable. The control group consists of all other candidate countries, either because their position far below the median implies high costs of compliance or because their position above the median does not preclude them from being eligible for MCC aid.
Figure 1 shows the results when comparing the changes in control of corruption for three sub-periods after the announcement of the MCC with the changes prior to the announcement (2002-2000). The first pair of bars in panel (1) points to strong anticipation effects. The treatment group, on average, controlled corruption more effectively immediately after the MCC had been announced (2004-2002) than before, in contrast to the control group; the difference between the two bars is statistically highly significant. In other words, the MCC had positive incentive effects even before becoming operational in 2004, probably because prospective candidate countries expected the Bush proposal to reward reform efforts with considerable amounts of additional aid. This suggests that, contrary to the failure of traditional forms of conditional aid, retrospective performance-based aid can be effective.
Figure 1. Change in control of corruption: Treatment vs control groups
Note: ***, **, * if the difference between the treatment group and the control group is statistically significant at the one, five, and ten percent level, respectively.
Despite this, the enthusiasm among candidate countries about possible rewards appears to have cooled in the course of MCC operations since 2004. The difference between the pair of bars in panel (2) of the figure is still significant at the 1%. Yet it is striking that the incentive effects slightly weaken when comparing the first two years of MCC operations with those prior to its announcement. This could be attributed to the “rough start” of the MCC and the resistance of US Congress to provide the funding requested by the Bush administration (Rieffel and Fox 2008).
Yet the prospects of reform efforts being rewarded with additional US aid were diluted further with ongoing operations (panel 3). The number of signed compacts remained fairly small (20 until early 2010), and it became increasingly obvious that the actual MCC budget would persistently fall short of the originally proposed $5 billion per annum. Lancaster (2008) observes that the MCC has been “extraordinarily slow” in disbursing the available amount of funding, “raising questions about the efficacy of this new model of performance and ownership-based aid giving.”
Why the Millennium Challenge Corporation works imperfectly
Redefining the treatment and controls groups reveals further flaws of performance-based MCC aid. The incentive to fight corruption weakens considerably when classifying all candidate countries below the median into the treatment group. For instance, the previously strong anticipation effect decreases considerably in panel (4), compared to panel (1) in the figure. This effect dwindles further and proves completely insignificant when restricting the treatment group to those candidates with the most unfavourable starting conditions (1st quartile; not shown). This is probably because the costs of compliance are relatively high, and aid rewards may appear to be out or reach for candidates with more unfavourable starting conditions.
The incentive effect also weakens when the treatment group includes candidates scoring above the median (3rd quartile), in addition to those slightly below the median (2nd quartile; see panel 5). In other words, candidates above the median have not been incentivised to step up efforts of controlling corruption in order to “defend” their eligibility for MCC aid against reformers scoring immediately below them. This seemingly myopic behaviour may actually be rational. In contrast to the strict performance appraisal at the gate-keeping stage, it was open to question whether the MCC would be equally strict in suspending eligibility of countries once they had been selected. MCC’s policy statements were vague in that regard (“remediation letters [are sent] to our partner countries that exhibited certain slippages in our criteria”); its credibility to sanction the ex-post violation of conditions does not appear to be greater than that of traditional aid agencies.
A promising alternative?
In summary, performance-based aid provides a promising alternative to the failed traditional approach whereby donors make aid conditional on reform promises of recipients. However, uncertainty about the timeliness and amount of aid rewards must be avoided for incentive effects to persist. Broken promises on the part of donors may undermine the sustainability of reforms, similar to insufficient stamina on the part of recipients. Automatic suspension should complement strict selectivity for performance-based aid to have broader and lasting incentive effects. At the same time, offering fruits that hang too high prevents reforms exactly where they are most needed. The incentive effects for countries with pervasive corruption could be strengthened by designing the threshold of aid eligibility relative to its own previous performance, rather than relative to other candidate countries.
Alesina, A and B Weder (2002), “Do Corrupt Governments Receive Less Foreign Aid?”, American Economic Review, 92(4):1126-1137.
Collier, P (1997), “The Failure of Conditionality”, in C Gwin and JM Nelson (eds.), Perspectives on Aid and Development. Washington, DC (John Hopkins University Press):51-77.
Collier, P, P Guillaumont, S Guillaumont, and JW Gunning (1997), “Redesigning Conditionality”, World Development, 25(9):1399-1407.
Cremer, G (2008), Corruption and Development Aid: Confronting the Challenges. Boulder (Lynne Rienner Publishers).
Easterly, W (2007), “Are Aid Agencies Improving?”, Economic Policy, 22(52):633-678.
Lancaster, C (2008), George Bush’s Foreign Aid: Transformation or Chaos? Washington, DC, Center for Global Development.
Mosley, P, J Hudson, and A Verschoor (2004), “Aid, Poverty Reduction and the ‘New Conditionality”, Economic Journal, 114(June):F217-F243.
Öhler, H, P Nunnenkamp, and A Dreher (2010), “Does Conditionality Work? A Test for an Innovative US Aid Scheme”, Working Paper 1630, Kiel Institute for the World Economy.
Rieffel, L and JW Fox (2008), “The Millennium Challenge Corporation: An Opportunity for the Next President”, Brookings Global Economy and Development Working Paper 30. Washington, DC.
Svensson, J (2003), “Why Conditional Aid Does Not Work and What Can Be Done About It?”, Journal of Development Economics, 70(2):381-402.