The Returns to Carry and Momentum Strategies

Abstract

We find that global time series carry strategies (across bonds, commodities, currencies, equities and metals) can be explained by a set of lagged macroeconomic variables. The payoffs to carry strategies disappear once futures returns are adjusted for their predictability based on these macroeconomic variables. On the other hand, momentum strategies are only weakly affected by lagged macroeconomic variables but are significantly related to measures of hedge fund capital flow. When studying these two findings together and over time we find that while momentum strategies were highly co-moving with carry strategies and therefore business cycle predictors between 1994 and 2002, when Hedge Fund AUM was low, correlation has since decreased. The decrease in correlation has coincided with significant increases in hedge fund AUM, and limits to arbitrage have become more relevant in explaining momentum returns. We embed these findings within a broad empirical investigation of time series carry and momentum strategies across 55 futures contracts spanning the asset classes bonds, currencies, commodities, equities and metals. Our results provide a possible avenue for identification strategies to disentangle the role of limited arbitrage effects on futures returns and systematic risks that are associated with time-varying expected returns in explaining momentum returns.

Keywords: carry, momentum, predictability, business cycles, limits to arbitrage, hedge funds

JEL Classification: C53, G11, G12, G13, F44, D84