The cost of corruption

Today is anti-corruption day, and today is also when Daniel Kaufman gives his farewell speech at the World Bank. He is leaving the World Bank Institute for the Brookings Institute frustrated about the lack of progress in recent years in the fight against corruption. This obviously begs the question: is corruption really that bad?

The classic paper in this regard is that of Paolo Mauro, which documented in cross-country regressions how corruption was significantly and negatively affecting GDP and investment. Such an exercise is, however, subject to concerns of endogeneity that this paper does not address properly, as documented by Philip Shaw, Marina-Selini Katsaiti and Marius Jurgilas. The latter find that with a statistically adequate instrument, corruption has no effect on investment or GDP. Zvika Neeman, Daniele Paserman and Avi Simhon also show a no impact result, this time limited to financially closed economies, while corruption is bad in financially open economies. Toke Aidt, Jayasri Dutta and Vania Sena show that corruption does not matter if institutions are of low quality to start with.

Costas Azariadis and Amartya Lahiri argue that the causality runs the other way: rich countries elect good and honest governments. Cesar Calderon and Alberto Chong make a similar argument. Jac Heckelman and Benjamin Powell claim that corruption is in fact growth enhancing when economic freedom is limited.

Obviously, this survey is heavily tilted to demonstrate that corruption is not bad. There is also good evidence that shows the opposite. My point is, do not take for granted that corruption is necessarily bad.