Do carry trade returns show signs of long memory?

Abstract

Abstract Using Hurst coefficient approaches within rescaled range, detrended fluctuation, periodogram regression and average wavelet coefficient frameworks, this article examines the dynamics of currency carry trades. Specifically, we look for evidence of long-range dependence (LRD) in carry trade returns (CTR) of thirteen developed and nine emerging markets. Our analysis provides two important insights. (a) We find strong and robust evidence of LRD for CTR in developed markets and somewhat weaker evidence in emerging markets. Considering that CTR are vital components of modern financial models, these results have far-reaching consequences because, in the presence of LRD, traditional asset pricing methods/tests are no longer valid. (b) We also show that portfolios focusing on only the most persistent carry trades (highest Hurst coefficients among the ones above 0.5) can significantly outperform their all-encompassing counterparts. This implies that Hurst coefficients contain important investment information which can be used to enhance the performance of practical carry trade strategies.