Is Currency Exposure Puzzle Really a Puzzle? International Evidence
Accounting and Finance Dr. Chu-Sheng Tai
Motivated by the limited success in detecting significant currency exposure in the extant literature, in this paper I attempt to resolve the so-called “exposure puzzle” and thus contribute to the literature by investigating whether currency movements have any significant impact on stock returns for seven selected industries which are common across 12 diversified economies such as the G4 countries (Canada, Germany, Japan, the UK), as well as eight emerging markets, including four Asian markets (China, South Korea, Taiwan, and Thailand) and four Latin American markets (Argentina, Brazil, Chile and Mexico) if both conditional heteroscedasticity and asymmetric currency exposure are considered in the estimation. Based on the estimation of an asymmetric three-factor exposure model using multivariate GARCH (MGARCH) methodology, I find that currency exposure is highly significant and pervasive since 73 (86.90%) and 64 (76.19%) of 84 country- industry portfolios examined are significantly exposed to the bilateral exchange rate and trade-weighted exchange rate changes, respectively. In terms of the size of total currency exposures, the lowest mean total currency exposure is 0.131 for developed market industry portfolios when the trade-weighted exchange rate is used, suggesting that ceteris paribus a 1% appreciation of the trade-weighted USD exchange rate is, on average, associated with a 0.131% increase in the weekly excess returns of these portfolios. On the other hand, the highest mean total currency exposure is 1.241 for Asian emerging market industry portfolios when the bilateral exchange rates are used, implying that ceteris paribus a 1% appreciation of the local currency with respect to the USD is, on average, associated with a 1.241% increase in the weekly excess returns of these portfolios. As a result, the currency exposures are not only statistically significant but also economically important. These empirical results remain unchanged under several robust tests. This is an extremely important finding as it suggests that previous studies on currency exposure which rely heavily on the standard OLS or SUR method of estimation with the assumption of constant variance of firm’s or industry’s returns and symmetric exposure over appreciation-depreciation cycles are frequently mis-specified. Consequently, the methodological weakness from previous studies is the reason for the exposure puzzle, which may not necessarily due to the effective hedging as proposed by some researchers. However, this so-called “exposure puzzle” may not be a puzzle at all once a better methodology such as the MGARCH is utilized in the estimation.
Made with FlippingBook - Share PDF online