In this study, we estimate term structures of default probabilities for private rms using Korean data comprising 1,440 default events from 29,894 rms between 1999 and 2011. We then study whether the reported interest expenses are re ective of the estimated default term structure. Each private rm's default likelihood is characterized by a forward intensity model employing both macro risk factors and rm-specic attributes derived from nancial statements. Although private rms have no traded stock prices, we devise a way of obtaining a public-rm equivalent distance-to-default by projection which references the distance-to-defaults of public rms with comparable rm attributes. Statistical tests indicate that the tted model provides accurate multiperiod forecasts of defaults for both nancial and non-nancial private rms. Our methodology can be directly applied by commercial lenders in charging appropriate interest rates upon lending decisions for dierent future periods.