The Dynamics of Bribery: Evidence from Tunisian Customs Transactions (Pre-Recorded)

• In Event: State Capacity and State Building

Abstract

Social scientists have made great progress in identifying the conditions that give rise to corruption in state institutions. But one underexplored complexity is that bribe-paying citizens and bribe-receiving officials often have repeated interactions. These dynamics are very likely to affect how often bribes change hands, the characteristics of those bribes, and downstream economic outcomes like trade volumes and tax revenues. Our paper therefore focuses on two fundamental questions: What is the effect of repeated interactions between citizens and officials on the frequency and magnitude of bribery? And what is the effect of officials’ incentives evolving over their “time on the job” on the frequency and magnitude of bribery? To answer these questions, we built an original dataset on Tunisian customs transactions using an audit-study design to directly observe bribery, to our knowledge the first of its kind in the Middle East and North Africa. We also leveraged a natural experiment in which customs inspectors are randomly assigned to import shipments using a computer algorithm. Our results suggest that increasing the number of past interactions between clearing agents and customs inspectors decreases both the frequency and the magnitude of bribery. The reason is that, while more past interactions builds trust that makes it easier to engage in tax evasion, the government more aggressively audits those transactions in which the parties involved have longer-term relationships. Our results also suggest that more time on the job increases bribes, likely due to a combination of officials being “captured,” having shorter time horizons, and building bribery-relevant human capital.

Authors

• Nate Grubman, Stanford University

• Samuel Leone, University of California - Berkeley

• Jawaher Mbarek, Independant