Working Paper: NBER ID: w31266
Authors: Gustavo J. Bobonis; Paul Gertler; Marco Gonzalez-Navarro; Simeon Nichter
Abstract: Does combating corruption reduce clientelism? We examine the impact of a prominent anti-corruption program on clientelism using a novel representative survey of rural Brazilians. Randomized audits reduce politicians’ provision of campaign handouts, decrease citizens’ demands for private goods, and reduce requests fulfilled by politicians. We investigate mechanisms by which audits may reduce clientelism, and find that audits significantly reduce citizens’ willingness to supply clientelist votes. Results also offer novel insights into audits’ dynamic effects: they have more pronounced effects in the short run, especially during electoral periods.
Keywords: corruption; clientelism; anticorruption audits; Brazil
JEL Codes: O10; P37
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| anticorruption audits (H57) | decrease in local politicians' provision of campaign handouts (D72) |
| anticorruption audits (H57) | decline in citizen requests for private goods from politicians (D72) |
| anticorruption audits (H57) | decline in the prevalence of requests fulfilled by politicians (D72) |
| anticorruption audits (H57) | decrease in interactions between citizens and politicians (D72) |
| anticorruption audits (H57) | deterioration of citizens' perceptions of politician reciprocity (D72) |
| anticorruption audits (H57) | undermine clientelist relationships (D72) |
| anticorruption audits (H57) | amplify declines in requests and fulfilled requests during electoral periods (K16) |