Intraday dynamics of foreign exchange rates in the presence of higher moments
In this study, we investigate the modeling of currency returns and volatilities in the presence of higher moments. The main objective is to examine whether the inclusion of higher moments improves the estimation performance of the currency market models. We use 15-minute exchange rates of the six most liquid currencies (i.e. the Australian Dollar, the British Pound, the Canadian Dollar, the Euro, the Japanese Yen, and the Swiss Franc) measured against the US Dollar over the period 2004 to 2014. We find strong evidence that the presence of higher moments improves the estimation performance of exchange rate models. Specifically, we show that inclusion of odd/even moments in modeling the return/variance increases the estimation performance of the models. We also propose a trading strategy to show that the forecasted returns obtained from our model can be used to generate significant profits for investors. Using a variety of sets of in-sample and out-of-sample periods, we discover that: (a) our proposed model performs significantly better than the traditional GARCH model in terms of offering the higher returns to investors; (b) returns increase as the in-sample (out-of-sample) periods decrease (increase) until they reach to their maximum level at 40% in-sample and 60% out-of-sample, and they decline afterward; (c) our findings on profitability (economic significance) are, generally, consistent with previous evidence on predictability (statistical significance). Finally, we test the robustness of our findings. We show that the superior performance of our proposed model maintains when we apply it on different sample periods (i.e. pre- and post- Global Financial Crisis). We conclude that our findings are insensitive to different sample periods.
JEL classification: C5; C58; F31; G15
Keywords: Foreign Exchange; High Frequency; Higher Moments; Modeling; Trading Strategy