A few years ago, Scott Guggenheim, a social development sector coordinator at the World Bank, hiked to a village in North Sumatra, Indonesia, to see its new water system. But he couldn’t find the apparatus. Eventually, he located its PVC pipes: The village head had fashioned them into lawn furniture, which he was selling in the nearby market town.
“Corruption is rampant in Indonesia, and it acts like a tax on development projects and business,” attests Benjamin A. Olken, an economist and junior fellow of the Harvard Society of Fellows. To find out how best to stem the swindling, Olken conducted a sophisticated experiment in 608 Javanese villages. His results, which appear in the April 2007 Journal of Political Economy, challenge current wisdom: Send in the outside auditors, rather than rely on local monitors.
“Prioritizing grassroots involvement is the key to a lot of development efforts,” says Olken – and with good reason. Communities have a higher stake in, say, building a safe bridge than do outside auditors, and so have more incentives to stave off corruption. Entire villages are also harder to bribe and hoodwink than are a few government officials.
Yet Olken discovered that villages expecting a government audit at the end of a road-resurfacing project could account for more of their expenditures than could villages that used community-based checks.
All of the villages in Olken’s study were part of the Kecamatan Development Project (KDP), which is funded by a loan from the World Bank. Roughly half of the villages were randomly assigned to an audit condition, in which they learned that Indonesian government officials would inspect their road-building project’s financial records and construction activities, and then report their findings at a village meeting. (Ordinarily, the Indonesian government audits only about 4 percent of its development projects.) The remaining villages – the control condition – executed the KDP’s grassroots monitoring program, which requires project leaders to account for their spending at village meetings.
As villages finished their projects, a team of researchers took core samples of the roads to estimate what materials were in them and how much they cost. The team also interviewed villagers to find out how much they were paid. Using these numbers, the researchers estimated how much each village actually spent on its project.
Overall, the researchers’ estimates added up to only 76 percent of the villages’ reported costs. In other words, the villages could not account for 24 percent of their expenditures. But villages in the audit condition had 80 percent less in missing expenditures than did those that relied on grassroots monitoring.
“I don’t think anyone would have guessed that Indonesia would be a place for auditors to have a particularly favorable effect,” Olken says, noting Indonesian officials’ reputation for well-greased palms. But because the specter of audit did reduce graft in this corruptible climate, he says, it is likely to work in many other developing nations.
Although Olken showed that guaranteeing audits reduces corruption, he does not think that governments or NGOs need to check up on all projects. “Just sending people a very salient letter saying that they have a 20 percent chance of being audited” would make loan recipients and grantees more accountable, he says.
Read more stories by Alana Conner.