In this paper, with access to the Korean automobile insurance data set, I test for the presence of the moral hazard in the automobile insurance market. So far, most empirical research on asymmetric information in the automobile insurance markets have explored the conditional correlation approach. However, in this methodology, it is almost impossible to distinguish moral hazard and adverse selection separately. Given this circumstance, there have been some recent empirical tests using dynamic data sets to attempt the separation of two phenomena. Among them, I employ the mixed proportional hazard model based on Abbring, Chiappori, Heckman and Pinquet (2002) to
distinguish the moral hazard phenomenon separately from adverse selection, controlling for unobserved heterogeneity. Different from the empirical result using the French data in Abbring, Chiappori, Heckman and Pinquet (2002), I have detected the existence of moral hazard in the Korean automobile insurance market. This would suggest that the market imperfections such as moral hazard are market-specific phenomenon.