What Can Firm and Household Surveys Tell Us About Expert Assessments of Corruption?

Cross-national comparisons of corruption levels exploit two kinds of data: expert assessments and surveys of firms and households. Examples of the former include the corruption ratings provided by Political Risk Services for its International Country Risk Guide (ICRG) and by Freedom House in its annual study of Nations in Transit (NIT). Examples of the latter, which generally but not exclusively focus on actual corruption experiences, include various surveys conducted by the World Bank and the European Bank for Reconstruction and Development (EBRD)—of which more below—and the World Economic Forum’s (WEF) Executive Opinion Survey. Two widely used corruption indicators aggregate information from multiple data sources in both categories: Transparency International’s (TI) Corruption Perceptions Index and the World Bank Institute’s (WBI) Control of Corruption Index. Most cross-national empirical work on corruption uses some combination of expert ratings and the composite indicators provided by TI and WBI, and the corresponding rankings of corruption levels often figure in public discussion of particular countries. The emphasis on expert assessments is a natural consequence of the relative infrequency and restricted geographic focus of many firm and household surveys. This limited coverage, in turn, implies that the composite indicators lean heavily on expert assessments to provide annual ratings of corruption in a large number of countries. To what extent do expert assessments of corruption levels correspond to actual corruption experiences? What do the country-level assessments miss? In this memo, I discuss the relationship between expert ratings and the picture provided by firm and household surveys.