A considerable body of theoretical and empirical literature has evaluated the credit channel of monetary transmission. This paper sets up an open-economy model under floating exchange rates with imperfect international capital mobility based on the Bernanke and Blinder model (1988). Employing our model, we show that a change in money supply has different impacts on the economy in many cases compared to the previous literature. The exchange rate puzzle may occur and when the exchange rate puzzle appears, the Fleming proposition is violated. Besides, by means of a cointegration analysis, we empirically verify the particular case of the exchange rate puzzle with the monthly data from May 1984 to January 2005 in Taiwan. Therefore, our empirical evidences can be matched with our theoretical derivations successfully.