Perception vs. experience: Explaining differences in corruption measures using microdata

Most corruption research suffers from one common problem: There is no objective measure of public-sector corruption for a cross-section of countries. Studies on the determinants or the effects of corruption typically rely on indicators of corruption perception. In recent years, a second type of indicator reflecting experience with bribery has become available. If corruption perception is primarily informed by experience with corruption, these two types of indicators should be very highly correlated, which they are not. This study examines the variation in individual corruption perception that cannot be explained by individual corruption experience alone. We find that both respondent characteristics and country characteristics affect corruption perception beyond what can be explained by individuals' first-hand experience of corruption. Some of these biases may force us to reevaluate results of corruption research that is based on perception data, as well as the anti-corruption policies designed in response to these results.