The purpose of this paper is to discern the predictive power of money,interest rate and exchange rate with respect to real industrial production index and the consumer price index, respectively, in the context of a typical open economy: Korea. To accomplish the purpose, the error correction model and the standard vector autoregression (VAR) model is utilized with the monthly data covering the period 1973:7 - 1994:12. First, based on the Agmented Dickey-Fuller unit root tests and Johansen-Juselius cointegration test, an error correction model is selected. And then F-tests of Granger causality and variance decompositions are followed. For the purpose of comparison and robustness of the model, some other standard specificationf of VAR model were also examined. The rolling regression technique is also utilized to document thoroughly the subsample instability of the results in the VAR model. Our main findings are the following. Money predicts generally well the future movements both in the real industrial production index and the consumer price index. On the other hand, interest rate has almost no statistically significant predictive power with respect to either industrial production index or the consumer price index. In contrast, exchange rate predicts well the future movements only in prices.