We investigate the effect of distress risk premiums on the performance of structural models of credit default swap (CDS) spreads. The results show that structural variables inspired by theory are more likely to fail in accounting for the CDS spreads of firms with higher distress risk. We argue that the distress risk premium embedded in CDS spreads is culpable in hampering empirical studies using the structural approach because the distress risk premiums are unrelated to firm-specific default rates. Rather, the main driving forces of distress risk premiums are market-wide factors. Our findings point a new direction to resolve the credit spread puzzle.