The Impact of Dual-agency Leniency Policy on Cartel Detection

We study cartel detection when two public authorities operate separate leniency programs within the same jurisdiction, as in Korea. We develop a simple repeated-game model to compare single-agency enforcement with dual-agency enforcement, to distinguish independent operation from cooperation, and to examine how the structure of leniency relief affects reporting incentives. When the two programs operate independently and do not recognize each other's leniency status, firms may have weaker incentives to self-report, and reporting can become concentrated in only one program. Cooperation that recognizes leniency rank across authorities restores a race to report and can make self-reporting attractive under a broader range of enforcement environments. The analysis also shows that cooperation is most reliable when early applicants receive comparable treatment across authorities: when second-in-line relief is available only in the administrative program, stronger criminal exposure can reduce the effectiveness of cooperation by raising the residual risk borne by non-first applicants. The policy implication is that effective dual-agency leniency can be achieved through a narrow form of coordination that verifies marker status and aligns the relief structure across authorities while preserving confidentiality.