Reproducibility package for Which Data Do Economists Use to Study Corruption? A Cross-Section of Corruption Research

We examine a cross-section of published studies to inform our understanding of which data economists use to study corruption. For publication year 2022, we searched the EBSCO database and EconLit for articles with “corruption” in the abstract, and then focused on the subsample which identified Journal of Economic Literature (JEL) codes. We used the resulting dataset of 339 journal articles to examine the JEL codes used most often for corruption research, the most popular data sources for analysis, the type of data (e.g., survey, administrative, or experimental), the geographical foci, and whether the study examined the causes or consequences of corruption. Cross-country composite indicators remain the most popular measures, whereas single-country studies were more likely to use administrative data. Articles published in ranked journals were more likely to use administrative data and experimental data than those published in unranked journals. Studies examining the causes of corruption, while less numerous overall, were relatively more likely to be published in ranked journals. The full universe of 882 journal articles in a single year point to both the enormous academic interest in corruption and the larger literature on corruption in political science and public policy and public administration disciplines. The paper raises questions about inattention given to novel types of data and studies of the causes of corruption, as well as the need for a less-siloed approach within economics.