Cross-sectional and Time-series Momentum in Equity and Futures Markets: Trading Strategies and the Role of Correlation Risk

Abstract

The purpose of the thesis is to investigate momentum trading strategies in equity and futures markets and to explore the links between momentum profitability and the equity market correlation of the economy. The first topic focuses on cross-sectional equity momentum patterns by modeling a stock’s price path as the interaction between a long-term growth component and a number of fluctuating price components that oscillate around the long-term trend at various distinct frequencies. Based on this specification, the dependence of momentum profitability on the asset price response to oscillations at various frequencies is explored. The evidence is consistent with a behavioural overreaction-to-private-information and underreaction-to-public-information explanation of the momentum patterns. Cross-sectional momentum profitability is found to be robust to realistic transaction costs and is shown to be optimized in terms of minimising the effects of transaction costs for a 6-month holding horizon. Simple stop-loss rules are shown to improve the performance of strategies with long-term holding horizon by discarding big and growth stocks, which achieve higher levels of price efficiency and therefore realise their momentum potential faster than small and value stocks. The second topic focuses on the source of profitability for cross-sectional momentum portfolios and other commonly used long-short zero-cost factor-mimicking portfolios and investigates whether these abnormal premia are justified as compensation for bearing correlation risk. Using a novel dataset on correlation swaps and building on the fact that large equity market declines are accompanied by increases in stock correlations, it is shown that correlation risk is priced in the cross-section of stock returns even after including conventional risk factors. Moreover, it is documented that a significant part of long-short portfolios’ return premia is explained by exposure to correlation risk. Interestingly, the inflow of capital into long-short hedge fund strategies coincides with increases in the realized equity market correlation, and consequently with decreases in the price of insurance against unexpected correlation surprises. Finally, the profitability and the mechanics of time-series momentum strategies in futures markets are explored. A time-series momentum strategy involves the volatility-adjusted aggregation of univariate strategies and therefore relies heavily on the efficiency of the volatility estimator and on the quality of the momentum trading signal. The evidence shows that trading signals generated by fitting a linear trend on the asset price path maximise the out-of-sample performance while minimising the portfolio turnover. The momentum patterns are found to be strong at the monthly frequency of rebalancing, relatively strong at the weekly frequency and relatively weak at the daily frequency. In fact, significant reversal effects are documented at the very short-term horizon. Regarding the volatility-adjusted aggregation of univariate strategies, the Yang-Zhang range estimator constitutes the optimal choice for volatility estimation in terms of maximising efficiency and minimising the bias and the ex-post portfolio turnover.