Causes and consequences of corruption: What do we know from a cross-section of countries?

In recent years a growing body of empirical work has examined the causes and consequences of corruption. These investigation are mostly cross-country analyses, based on comparative assessments of the extent of corruption in various countries. Such assessment are sometimes compiled by privateagencies to determine country risks, and the data gathered are sold to investors. Other sources are surveys of the generalpublic or elite businesspeople. The data on corruption are thus based on subjective perceptions and expertise, and empirical work using these indices assumes that they are correlated with underlying real levels of corruption. With the exception of somemicro-level studies, perceptions data are the only information available on corruption levels. Efforts to use objective data, such as conviction for corruption or office, suffer from inherent biases that undermine their validity and are not available across a large cross-section of countries. Recognizing the limitations of perception-based indices, researchers have nevertheless been able to use them to significantly advance the study ofcorruption. These data allow empirical research on corruption to move beyond the anecdotal descriptions and purely theoretical considerations that previously dominated.