Patterns in corruption

Michael Moynihan writes about Ben Olken's work in 'The American' (via The Bayesian Heresy):
"In his paper “The Simple Economics of Extortion,” co-written with Patrick Barron of the World Bank, Olken again traveled to Indonesia to study “if the way in which we think about pric­ing for firms also applies to corrupt officials.” In other words, do the crooked respond to market forces in the way a corporation would? To test this, Olken and Barron looked at the number of roadside checkpoints—which act, essentially, as illegal toll booths where the motorist is required to pay a bribe—in Aceh, a region on the north­western tip of Sumatra long engaged in a guerilla war with separatist rebels.

“I looked at what happens when there is a change in the number of checkpoints along the road in Aceh. When I started data collection, you had to stop at, say, 90 checkpoints along a 600-kilome­ter route,” Olken says. “A lot of the checkpoints were associated with the military occupation. But when the peace agreement was signed [in 2005], the military pulled out, and the number of check­points declined. The question is: how do the prices at the remaining checkpoints change?”

There are three theories on how prices are set, he says. Is the bribe purely the product of cultural norms, influenced by customs specific to the region? Is the bribe simply an individual transaction, not affected by culture? Or is the bribe economy embedded within a bigger mar­ket? “Just as if there was a firm running a toll booth,” Olken says, “the firm would balance off how much revenue they are getting from the tolls, versus the fact that if they were charging more for the tolls there would be fewer cars driving through; they’ll be making an optimal trade-off. So, in fact, that bribe is the equi­librium of a firm’s price-setting decision.”

Olken and Barron found that when the number of checkpoints decreases, prices increase: “So it looks like the guys are behaving like a firm would behave.”

So what does all of this mean to institutions, like the World Bank, struggling against endemic corruption? There is often an understandable desire to go after “the big fish” in corrupt soci­eties, thus making an example of those whose pockets are fullest. “What this paper shows is that there is a potential cost of doing it that way. If corruption is decentralized, the model pre­dicts that the total amount of bribes is going to be higher than if there was a single, centralized person coordinating all the checkpoints.” Such results have broad implications for countries mired in corruption: “One policy that has been advocated in a lot of countries is simplifying the process of business registration. This the­ory predicts that that would be a good thing to do in reducing corruption because you would be moving from a decentralized corruption, where bribes are set independently, to a single, central­ized person, and the total amount you would pay would be less.”

Olken also has looked at graft and corruption in road projects in Indonesia. In his April 2007 paper, “Monitoring Corruption: Evidence from a Field Experiment in Indonesia,” he sought to determine the best method of reducing theft and graft in public works projects by doing controlled field experiments in 608 Indonesian villages.

Some village leaders involved in the building of roads were told that, upon completion of the project, they would be visited by government auditors, which increased “the probability of an external government audit in those villages from a baseline of about 4 percent to essentially 100 percent.” Other villages were chosen to partic­ipate in grassroots “accountability meetings,” during which project coordinators would pub­licly account for the use of government funds in a town-hall-like venue. Villagers would be offered anonymous forms to report graft.

Olken’s conclusion was that “increasing gov­ernment audits…reduced missing expenditures, as measured by discrepancies between official project costs and an independent engineer’s estimate of costs, by eight percentage points. By contrast, increasing grassroots participation in monitoring had little average impact…. Overall, the results suggest that traditional top-down monitoring can play an important role in reduc­ing corruption.” In short, to reduce the amount of corruption, it’s cost-effective to do more audits, not to trust the grassroots.

What’s the practical use of these results? Can they be applied to, say, corrupt countries in Africa? The study seems to suggest, Olken says, that, even in developing countries, “audit agen­cies might be more useful than people would have otherwise thought.” A widely held presumption was that auditors too were corrupt, and that they would only end up extracting more bribes, mak­ing the situation significantly worse.

The Indonesian study, he argues, might “sug­gest that that assumption is actually incorrect” and that when formulating an anti-corruption strategy, governments and institutions “shouldn’t dismiss the auditors out of hand.” Since mini­mizing corruption is often a key prerequisite in the development of democratic institutions, campaigners for clean government would do well to heed Olken’s advice."

Ben Olken has also written on whether leaders matter and on the effects of television on social capital. Links to Ben Olken's work discussed in popular press and Ben Olken's Home Page.