The Sustainable Development Goals’ (SDGs) new focus on fighting corruption and building institutions has generated quite a stir (including on this blog – see here, here, here, and here). But the Millennium Challenge Corporation (MCC) – a U.S. agency responsible for disbursement of assistance geared toward international development targets – has long been acting against corruption through its effort to achieve the SDG precursors, the Millennium Development Goals (MDGs). Institution-building does not appear among the substantive aims of the eight MDGs. Rather, the MCC made anticorruption central to its work by introducing corruption indices into its process for competitive selection of aid recipients. In brief, the MCC Board of Directors chooses aid-eligible countries by evaluating and scoring candidates countries’ “policy performance” on a number of measures. Crucially, in order to qualify for aid, countries must score above average for their income group on the Worldwide Governance Indicators (WGI) “Control of Corruption” score. The indicator is therefore known as the “hard hurdle.” The Board also assesses corruption trends in its analysis of a country’s ability to reduce poverty and generate economic growth, which, with policy performance, comprises the overall evaluation.

This strategy is known as performance-based lending, and the MCC has employed it to award over $10 billion in grants to nearly 40 countries over the past 12 years. Is the MCC approach a good one? Many critics say no. I say yes. Although it is a strategy that is still evolving, performance-based lending—including the corruption control “hard hurdle”—is not only innovative and effective, but important.

First things first: If we accept broadly that corruption is bad for development (for reasons why we should, see Matthew’s post), donors have an interest – even a responsibility – in seeing that their funds do not exacerbate graft. Yet studies from 2002 to 2015 indicate that more corrupt governments actually receive more aid than their cleaner counterparts. By providing substantially resources in environments where corruption is rife, donors can inadvertently augment the very harms they seek to combat–by directly enabling predatory leaders to stay in power, by symbolically supporting corrupt governments, or by financing projects with money that is itself lost to graft (as Rick has discussed). So the MCC’s hard hurdle serves a critical purpose and sends a strong message, all the more important for the fact that it is not a traditional tool.

But many analysts have strongly criticized performance-based lending, with varying degrees of persuasiveness. Here are their arguments, and why I think the MCC ought to forge ahead despite these objections:

• At first glance, performance-based lending evokes the roundly condemned conditional lending that characterized the structural adjustment period from 1981 to 1991. During that time, institutional lenders mandated specific policy reforms and sought additional commitments from recipient countries. The strategy had poor effects and imperialistic undertones. But performance-based lending differs in essential ways. Donors condition assistance on outcomes, not interventions. This approach has a dual benefit: it allows for creativity, flexibility, and autonomy on the part of the candidate state, and it depends upon retroactive assessment rather than (potentially empty) promises of future action.
• Another argument against the MCC’s model, heard in Congress (see here and here), is that it still supports corrupt governments. I don’t dispute the observation that informs this critique: most grant recipients – like Guatemala, the Philippines, and Indonesia – have a long way to go toward getting corruption under control. But development aid is inherently intended for high-risk areas and grants often support anticorruption efforts. The beauty of the performance-based approach is that it doesn’t impose arbitrary, finite standards but rather stimulates a “race to the top” among peer countries.
• A related, more serious allegation is that the “race to the top” does not actually happen. In other words, critics maintain that there is no “MCC Effect” through which performance-based lending leads to genuine policy change. At best, this claim is empirically disputed: while some studies have found little to no effect (see, for example, here), others have shown that the MCC’s approach yields results (see here and here). Unsurprisingly, the question of whether the MCC Effect exists is more nuanced than a simple “yes” or “no.” For example, the impact varies across countries and indicators. Scholars on both sides of the debate have found that policy effects are particularly pronounced in the areas of corruption and fiscal policy (see here and here). In individual instances, it is easier to see that the MCC has shaped outcomes. The Board put a $470 million power sector compact with Tanzania on hold pending improvement of the government’s anticorruption performance. (Tanzania achieved the desired corruption targets but has thus far been denied the grant on other grounds.) Interestingly, a well-known 2013 survey of policymakers and development practitioners in candidate countries showed the respondents to be far more approving of the MCC lending strategy, and conditional aid more generally, than empirical researchers have been. There may well be more to the MCC Effect than the numbers, such as its impact relative to other aid programs and its supporting influence on civil society anticorruption efforts.
• The most troubling complaint is that the entire MCC anticorruption analysis is premised on faulty indicators. An illustrative, trenchant critique contends that the “hard hurdle” is inadequate because it does not accurately measure levels of corruption (or perceptions thereof) and because slow-changing scores create distortionary lags that render the indicator useless as a measure of progress. Although the MCC Board, and the researchers who produce the WGI, believe that the Control of Corruption indicator performs well, they are transparent about the measure’s shortcomings. To its credit, the MCC has taken – and continues to take – strides to hone the hurdle (see a recent progress report here). Ultimately, in my view, it is better to improve on well-established methodologies than to start anew. Critics also challenge the rigid use of a measure that is perception-based. But perception-based indicators have advantages, particularly in the context of conditional lending – they are less amenable to manipulation by public officials willing to make superficial changes in order to secure funds.
• Much of the discomfort with the “hard hurdle” derives from the weight it carries – people would be more comfortable if the allegedly fuzzy numbers weren’t outcome-determinative. But it is important to recognize that the hurdle, though “hard,” is also padded. In recognition of the limits of the WGI, the MCC selection process provides for qualitative assessment in the form of examination of other index scores and analysis of corruption trends over time. In 2004, for example, positive trends in anticorruption efforts boosted Mozambique and Georgia over the hurdle even though they lacked the requisite WGI scores.

Fighting corruption in and with foreign aid should be more of a priority than it is. While the MCC system is by no means perfect, its performance-based model is a step in the right direction.