I’ve been reading news of corruption scandals in Brazil with a great deal of sadness. I lived in Brazil during its return to democracy and experienced first-hand the hope and optimism that came with that transition. Though such scandals are common around the world, I also find these scandals particularly troubling because they involve governments that contributed to unprecedented social and economic progress in Brazil. These scandals are not like those of leaders who have completely reversed progress in the countries they rule, like Mugabe in Zimbabwe or Chavez in Venezuela.
We hear over and over that foreign aid programs should have “zero tolerance” for corruption, but is that really so? I understand why foreign assistance to a country like Zimbabwe can be criticized for propping up a corrupt regime. But if we could turn back the clock, knowing now what we know about corruption during the presidencies of Lula da Silva and Dilma Rousseff, would we prefer to have stopped all foreign funding for Brazil? Were the World Bank’s loans for urban development, infrastructure, and agriculture just shams to hide pilfering? Was the $1 billion that Norway provided to reduce deforestation a waste of money?
In a recent policy paper, I argue that the answer to these questions is that decisions about funding projects in other countries should depend more on the results achieved by those countries than by formal actions meant to control corruption. Agencies that take a “zero tolerance” to corruption would have had to stop funding programs in Brazil. Yet Brazil reduced deforestation in the Amazon by 80 percent, improved educational opportunities for poor families, expanded access to health care, and supported substantial and widespread economic growth. Paying attention to these kinds of results of government actions, and the results of aid projects, would actually allow foreign aid agencies to combat corruption in a more effective and objective way than their current approaches.
If foreign aid agencies complemented their current anti-corruption efforts with greater attention to information about results, they would focus less on controlling how their funds are used and more on what governments achieve. Instead of financing public finance reforms that create the illusion of control, they would focus on what public spending actually accomplishes. Instead of expanding elaborate and costly ex ante procedures and reporting requirements, they would disburse against ex post achievements (duly verified of course). Instead of investigating allegations of corruption when they arise, they would focus on failing projects and investigate whether corruption is the cause. In other words, by using information on project results, agencies would have new strategies for limiting the impact of corruption on development. Here are the four examples I offer in my paper:
1. Measuring results would give agencies information to prioritize how they allocate their anti-corruption resources. If a program is successfully improving health or lowering transportation costs, any corruption involved must be pretty small and inconsequential. By contrast, a program which is failing to achieve its results is exactly the kind we should be investigating to distinguish whether or not corruption is a contributing factor.
2. With better information on results, agencies could test the effectiveness of their control strategies. They could learn whether project implementation procedures, risk management and detection strategies are related at all to project success. If more of this were done, agencies would develop evidence to help discard or refine their control strategies, making them less intrusive and more cost-effective.
3. Better results information would give agencies an opportunity to expand the use of pay-for-results programs. Programs that pay for results, if designed and implemented well, are less likely to be vulnerable to corruption than conventional aid projects. When corruption interferes with achieving program results, payments are automatically reduced — a self-correcting mechanism.
4. Results measurements can help aid agencies move away from subjective and arbitrary judgments about the trustworthiness of partner governments. Currently, agencies use subjective and relatively weak indicators to judge whether aid recipients are well governed. Measuring the trustworthiness of a partner government on the basis of actual results would be more objective and more useful for assuring that aid is applied effectively.
The difference between Brazil and Venezuela over the last 15 years stands out more clearly in their achievements than it does in their speeches and promises. While Brazil clearly had corruption — and in very high places — the abuses were either not large enough or not insidious enough to undermine a whole range of political, economic, and institutional initiatives that brought about positive change. In other words, we should not have a zero tolerance for corruption if that means spending time and money to combat corruption without regard for the impact on development. Rather we should embed anti-corruption within a systematic approach that relies on measuring the results that aid agencies are supposed to care about.
Ultimately, there is no way for foreign agencies to put an end to corruption in other countries. That is a domestic political struggle in the hands of peoples everywhere, whether they are Brazilian or American or Korean or Nigerian or Chilean or Chinese or Russian. What foreign aid agencies can do, however, is make sure that the aid they provide goes less to governments and projects based on their promises and more to those that demonstrate real positive change.