In this paper we argue that the commonly employed exposure coefficient/beta is inadequate for capturing the entire impact of exchange rate changes on firms' future operating cash flows. Instead, we employ the bivariate Glosten–Jagannathan–Runkle generalized autoregressive conditional heteroskedasticity mean model to investigate four aspects of exchange rate exposure, including sensitivity of stock returns to exchange rate changes, sensitivity of stock returns to the volatility of exchange rate changes, sensitivity of conditional variance of returns to exchange rate volatility, and the dynamic conditional correlation between returns and exchange rate changes, respectively, using data from 10 industrial sectors in Japan. We find significant evidence of such exchange rate exposure which is not captured by the conventional measure. The diagnostic statistics confirm the adequacy of our model, and, hence, the robustness of the results.