A Quest for More Reliable Estimates of Exchange Rate Exposure

The noticeable dichotomy between the research and practice of exchange rate exposure management may be partly due to the fact that the degree, the direction and the significance of the exposure to currency risk vastly depend on the method of estimation and the proxies used. In this paper, we examine the impact of incorporating a few stylized facts of financial time series -- namely, time-varying volatility, leptokurtosis and asymmetric volatility -- and a possible indirect exposure effect on the magnitude, significance and the sign of exchange rate exposure coefficients. To this end, an asymmetric GARCH-type model whose residuals are assumed to be t-distributed is employed to estimate exchange rate exposure. We observe that the negligence of these stylized facts may lead to seriously under/over-estimated exposure coefficients. Incorporating the indirect exposure effect offers some useful insights too