A Dynamic Mechanism for Mitigating FraudulentReturns in E-Commerce

Fraudulent returns impose substantial costs on e-commerce marketplaces through refund losses, reverse logistics, markdowns, and the erosion of seller and customer trust. We study the design of a return-fraud mitigation mechanism in a three-player dynamic game among a customer, a seller, and a platform, with the platform acting as a Stackelberg leader. The analysis rests on a single economic condition: a customer refrains from fraud whenever his gain from a fraudulent return does not exceed the sum of the residual bond at risk, the expected audit penalty, the effort cost of fraud, and the discounted future value of a clean record. The platform's instruments a multi-stage refundable bond, probabilistic audits, risk-adjusted prices, and an aged reputation score are precisely the levers that move the terms of this condition. We characterize the conditions under which honest customer behavior and high seller quality form a Markov Perfect Stackelberg Equilibrium, and we show that (i)~bonds and audits are substitutes; (ii)~staged bond release lowers customer liquidity cost while preserving deterrence, provided a sufficient residual bond is held until verification; (iii)~score aging restores rehabilitative incentives without enabling score laundering; and (iv)~seller-adjusted scoring prevents customers from being penalized for seller-caused defects. The results provide design guidance for return policy, a central operational lever for e-commerce firms.