Unmasking the Relationships Between Exchange Rate Exposure and Its Determinants: A More Complete Picture from Quantile Regressions

Abstract

Largely ignored, empirical evidence is mixed regarding the sign of the relationships between exchange rate exposure and its determinants. Employing quantile regressions, we resolve the conflicting findings for S&P 500 firms by demonstrating that the culprit is the inherent limitations of the ordinary least squares (OLS) approach used in previous studies. Instead of the conflicting partial views of the picture of the mean depicted in previous OLS results, we present a clearer fuller picture of the spectrum of the relationships. Specifically, across the range of exchange rate exposure — extremely negative to extremely positive — the sign of the relationships change systematically for foreign sales, size, and debt ratio, while it remains consistently negative for quick ratio and positive for book-to-market. Test results further show significant differences in the coefficient estimates between high and low quantiles, indicating an asymmetric effect of the determinants on the exposure.