In an earlier post I showed that Bill Gates’ supposition that only 2 percent of expenditures in development assistance projects were lost to corruption was wildly off the mark. I also asserted that such lowball estimates are a major hurdle to more effective aid programs: When corruption losses are lowballed, so are the resources devoted to combating corruption. If losses are 2 percent of the total budget, then it makes little sense to spend 4 percent of the budget trying to prevent them. But if losses are 20 percent, then 4 percent spent on audits and investigations is a miserly sum. If losses are closer to 40 percent, then spending 4 percent borders on criminal negligence.
So where did Gates get the 2 percent figure? It turns out that the likely source for that figure illustrates not only how casually influential people sometimes throw around baseless numbers, but also the perverse incentives that development programs sometimes face to downplay the seriousness of corruption in their projects.