This paper estimates the conditional hazard baseline (term-structure) of the hazard rate to default at the time of bonds' issuance by using two hazard models-one ignoring and another allowing unobserved heterogeneity (UH) in the hazard rate. Following Diamond (1989) one can predict a declining hazard rate to default due to adverse selection and moral hazard. After controlling for UH caused by adverse selection and time-series shocks, the hazard rate shows to be increasing over time and hence the moral hazard effect cannot be confirmed.