Project data

Data: Evaluating policy reforms to reduce corruption

This paper exploits quasi-experimental variation in tariffs in southern Africa to estimate trade elasticities. Traded quantities respond only weakly to a 30 percent reduction in the average nominal tariff rate. Trade flow data combined with primary data on firm behavior and bribe payments suggest that corruption is a potential explanation for the observed low elasticities. In contexts of pervasive corruption, even small bribes can significantly reduce tariffs, making tariff liberalization schemes less likely to affect the extensive and the intensive margins of firms’ import behavior. The tariff liberalization scheme is, however, still associated with improved incentives to accurately report quantities of imported goods, and with a significant reduction in bribe transfers from importers to public officials.

Data details

The data sits under “Additional Materials” containing data and a replication do file.

From the ReadMe: The analysis is based on primary data on firms and bribes, and secondary data on trade flows and trade gaps obtained from the United Nations Comtrade Database (, for trade between Mozambique, South Africa and the Rest of the World (excluding South Africa) between 2006 and 2014.

Link to the data