This work empirically investigates how commercial banks' aggregate credit supply is associated with business cycle over different regimes of the Korean economy. Linear empirical models employed in most of previous studies are subject to a potential missapecification problem because it is well known that both real GDP and credit supply reveal different dynamic properties over different regimes. This work finds that credit supply has asymmetric effect on business cycle for expansion and contraction phases when the Smooth Transition Autoregressive Vector Error Correction Model (or STAR-VECM) is employed. Our empirical findings are as follows. Firstly, we find that credit supply has procyclical effect on real GDP in all phases. Secondly, the procyclical effects are significantly intensified especially in contractionary phases which indicates asymmetry of its effect. In sum, this result supports 'Credit Acceleration Hypothesis' of Bernanke et al. (1999). Lastly, we further find that real GDP has asymmetric effects on banks' credit supply with countercyclical effect on expansionary regimes.