We did a lot of reading during the lit review for our recent book, A Users’ Guide to Measuring Corruption, but there’s plenty of related work that we didn’t cite in that book. This 2007 paper, by Dilyan Donchev and Gergely Ujhelyi, tackles a central issue of our field head on.
Download the paper:
Do Corruption Indices Measure Corruption?
This paper presents empirical evidence that the most widely used indices to measure corruption might be biased in systematic ways. Evidence from the International Crime Victimization Survey suggests that actual corruption experience may be a weak predictor of reported corruption perception, and that some of the factors commonly found to “reduce” corruption, such as economic development, democratic institutions or Protestant traditions, systematically bias corruption perception downward from corruption experience. In addition, perception indices are influenced by absolute (as opposed to relative) levels of corruption, which tends to penalize large countries, and they exhibit diminishing sensitivity to both absolute and relative corruption, indicating that they may be a better proxy among countries with low levels of corruption than among highly corrupt ones. Individual characteristics, such as age, education, and employment status are also found to influence corruption perceptions holding experience constant.
— Jonathan Werve