Direct versus indirect regression estimates of foreign exchange cash flow exposure

Abstract

To estimate foreign exchange (FX) cash flow exposure, one may choose between direct and indirect regression approaches, where the direct approach uses accounting-based cash flow data and the indirect approach uses equity returns as a cash flow proxy. The indirect approach typically includes one or more additional independent variables to control for the impact of FX changes on the required rate of return. Frequently, the control variable is an equity index. We propose that using a bond return control variable instead of equity returns addresses several theoretical problems inherent in the indirect estimation approach. In our empirical analysis we find that using the bond-based control variable results in FX cash flow exposure estimates that are more highly correlated with direct measures than using an equity index as a control variable.