Predictability, trading rule profitability and learning in currency markets

Abstract

This paper studies currency predictability over time. We assess predictability by testing for the presence of exploitable patterns in currency returns. To do so, we first generate consistent and parsimonious reduced-form estimates of currency expected returns and variances and then use these estimates to form dynamic trading strategies that maximize the multi-period Sharpe ratio. Our results show that currency predictability is time-varying and, for a number of currencies, has increased substantially in recent times, casting doubt on the widespread view that currency pricing may be on a path of convergence towards efficiency. We find, however, that currency markets learn in an efficient manner and a close relation between our strategies and indices that track popular technical trading rules, namely moving average cross-over rules and the carry trade, suggesting that the technical rules represent heuristics by which professional market participants exploit currency mispricing.