Working Paper: CEPR ID: DP3167
Authors: Jakob Svensson
Abstract: This Paper uses an unique data set on corruption containing quantitative information on estimated bribe payments of Ugandan firms. The data has two striking features: not all firms report they need to pay bribes; and there is considerable variation in reported graft across firms facing similar institutions/policies. To explain these patterns we construct a simple bargaining model. The model yields predictions on both the incidence and the level of graft. Consistent with the model we find that variation in policies/regulations (across industries) explain the incidence of corruption, while variation in profitability and technology choice explain the variation in bribes for the group of bribe paying firms. These findings suggest that public officials act as price (bribe) discriminators, and that prices of public services are endogenously determined in order to extract bribes.
Keywords: corruption; quantitative data
JEL Codes: D90; H20; K40; L10
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| industry-specific regulations (L59) | incidence of corruption (H57) |
| profitability (L21) | amount of bribes paid (H57) |
| technology choices (O33) | amount of bribes paid (H57) |