This paper examines how alternative coupon recovery assumptions—no-coupon recovery (NCR) versus full-coupon recovery (FCR)—affect default probability estimates and their empirical performance in the Korean corporate bond market. Specifically, we compare the NCR model of Hilscher, Jarrow, and van Deventer (2025) (HJD) with the FCR model of Jarrow and Turnbull (1995) (JT) to assess how differences in coupon recovery assumptions translate into variation in discriminatory power for actual credit events. Using daily corporate bond data for publicly listed firms from the FnPricing database over the period 2015–2024, we analyze the behavior of model-implied default probabilities around credit events. The results show that default probabilities derived from the HJD model—which excludes recovery on post-default coupons—respond more sensitively to market-implied measures of credit risk than those from the JT model. Overall, the findings provide evidence that coupon recovery assumptions play a critical role in determining the ability of default probability measures to identify realized credit events.