【报告摘要】Many online retailers have recently adopted instant return credit (IRC)---offering a store credit immediately upon a return claim, without requiring the return to be received or verified. By improving consumers' temporary shopping budgets, IRC helps resolve the mismatches between products and consumer tastes and thus converts online returns into new sales. This process is, however, prone to costly fake returns from dishonest consumers. In this paper, we take the initiative to study IRC's fundamental dynamics and implications. We build a game theory model in which a retailer sells two horizontally differentiated products, facing consumers with heterogeneity in taste, budget level, and honesty type. We characterize the market risk condition using three factors: product cost, consumers' salvage value for a mismatched product, and proportion of dishonest consumers. In the base setting that the retailer offers IRC uniformly to all consumers, we identify three schemes to execute IRC: If the market risk is low, the retailer should offer partial IRC coupled with symmetric pricing; if the market risk is medium, the retailer should offer full IRC coupled with asymmetric pricing, although the two products are completely symmetric; if the market risk is high, IRC should not be offered. In the advanced setting that the retailer can distinguish between honest and dishonest consumers (e.g., with predictive analytics enabled by AI tools) and offer IRC contingently, full (partial or no) IRC should be offered to honest (dishonest) consumers; the retailer may still allow fake returns even when consumer types can be perfectly learned. We further demonstrate that uniform IRC is more valuable for low-risk markets, while making IRC contingent on consumer type is more valuable for medium-risk markets. Finally, uniform IRC may hurt both types of consumers for low-cost products, while making IRC contingent normally rewards honest consumers and penalizes dishonest consumers.