This paper studies the dynamic effects of oil price fluctuations on business cycles and the unemployment rate in Japan. Much of the conventional research on relations between several economic variables has applied models based on the assumption of parameter invariability. However, when we consider a time series model over a period of several decades, assuming no structural changes is clearly unrealistic. In the present paper, we apply a time-varying coefficient vector autoregressive modeling approach based on a Bayesian method using smoothness priors. The results suggests that oil price fluctuations strongly influence business conditions. Oil price fluctuations also affect cyclical unemployment rate during the recession. Notably, the influence of oil price fluctuations on business conditions and unemployment has become large in recent years. Index Terms-Bayesian modeling, time-varying coefficient VAR model, oil price, unemployment rate, composite index.