Cultures of Corruption or Selection Bias (wonkish)?

In a 2006 paper called "Cultures of Corruption: Evidence from Diplomatic Parking Tickets" Ray Fisman and Ted Miguel, two economists at Columbia and Berkley respectively, gave an intriguing answer to the question: is corruption by nurture or by nature? Are people corrupt because they have incentives to behave corruptly (nurture) or because they are born in environments where respecting the rule of law is an option (nature)? For those of you who did not read either the paper or their bestseller Economic Gangster yet, here is a summary of what they do:

"We exploit a natural experiment, the stationing of thousands of diplomats from around the world in New York City. Diplomatic immunity means there was essentially zero legal enforcement of diplomatic parking violations, allowing us to examine the role of cultural norms alone. This generates a revealed preference measure of government officials’ corruption based on real-world behavior taking place in the same setting. We find strong persistence in corruption norms: diplomats from high corruption countries (based on existing survey-based indices) have significantly more parking violations, and these differences persist overtime.”

The idea of the natural experiment sounds intriguing and the authors explain very neatly why this setup is useful for their purpose. However, when reading their entertaining explanation, I started to feel a little problem with their experiment: are we sure that diplomats are a representative un-biased sample of their population of origin? The selection process of "government officials" is likely to be biased toward "very honest" people in low corruption countries, and "very corrupt" people in high corruption country. Why? Very Corrupt People in Corrupt Country realize ex-ante how entering into politics will allow them to live above the law. They have more to gain from the practice of politics (think about BS somewhere down the alps...) Viceversa, Honest people are more likely to practice politics in Low Corruption Country, because of multiple reasons: 1) they live in a system with high public scrutiny; 2)conformity to the law is for honest people a public good and a personal benefit 3) people don’t want their country to be run by other people of dubious qualities; 4) if they are caught misbehaving, they are probably arrested. So high cost for dis-honest people, and low gains.

The bottom line is that the observations the authors deal with are a biased sample of the population of origin they want to study. This being said, this is not to say that their results are totally invalidated. Indeed, their econometrics result holds when correlating parking tickets with corruption. The problem I find is that parking tickets certainly do not represent an unbiased measure of a country corruption as much as survey-based measures of corruption fail the purpose. Diplomats are not anyone. Their selection process is long and costly. I don't think it's hard to imagine that only those who will benefit more from its practice will likely sort into it. And this sorting goes along the overall perception of how the system works. So although I find their conclusion somehow misled, I think the paper is terrific for the following conclusion: it is an international shame to give immunity from the law to people who don’t deserve it at all.