Implementing Momentum: What Have We Learned?
Adrienne Ross (AQR Capital Management), et al.
An abundance of academic evidence and theory exists on the efficacy and intuition behind momentum investing, yet a limited number of studies discuss the feasibility of running momentum portfolios in practice. And no study to date has directly analyzed implementation costs for a live momentum portfolio. As a result, many are still quick to dismiss momentum as difficult or costly to implement because of its high turnover. In this paper, we use seven years of live data to evaluate the implementability of momentum investing. We show that live momentum portfolios are capable of capturing the momentum premium, even after accounting for expenses, estimated trading costs, taxes, and other frictions associated with real-life portfolios.
Andreas Neuhierl (U. of Notre Dame) and Michael Weber (U. of Chicago)
We document a large return drift around monetary policy announcements by the Federal Open Market Committee. Stock returns start drifting up 25 days before expansionary monetary policy surprises, whereas they decrease before contractionary surprises. The cumulative return difference across expansionary and contractionary policy decisions amounts to 2.5% until the day of the policy move and continues to increase to more than 4.5% 15 days after the meeting. The return drift is a market-wide phenomenon, holds for all industries, and many international equity markets. In the cross section of stocks, size, value, profitability, and investment do not exhibit differential return drifts. Momentum is an exception, because past losers plummet around contractionary monetary policy surprises. A simple trading strategy exploiting the drift around FOMC meetings increases Sharpe ratios relative to a buy-and-hold investment by a factor of 4.
Michal Czerwonko (Concordia University)
November 25, 2017
High idiosyncratic volatility (IVOL) stocks exhibit extended low returns followed by persistently higher returns — which results in a tradable anomaly. Return patterns due to IVOL explain long-term reversals and, to a substantial degree, momentum and investment anomalies; moreover, they affect the economic significance of profitability anomaly. The pattern of initial underreaction is linked to a rational but not efficient asset pricing story: uninformed investors hold on to underperforming high IVOL stocks in anticipation of recovering losses in year-end trading.
Time-Series Momentum, Carry and Hedging Premium
Marat Molyboga (Efficient Capital Management), et al.
November 21, 2017
This paper examines the performance of time-series momentum across 65 futures markets from all major asset classes, including equity indices, fixed income, currencies and commodities, for the period between January 1975 and December 2016. We find that the basis between spot and futures contracts explains approximately 36% of the performance of time-series momentum indicating that time-series momentum and carry are related. Conditioning trading signals on the sign of the basis improves the Sharpe ratio of time-series momentum by approximately 0.17 and is robust across sub-periods, choice of position-sizing in the implementation of time-series momentum and the lookback period used in the calculation of the basis. The improvement in performance is particularly strong during the early stages of recessions that tend to exhibit very poor stock market performance. Therefore, our strategy can substantially improve investors’ welfare. We investigate whether time-series momentum and carry are related to hedging premium by examining the positions of hedgers in the Commitment of Traders (COT) reports. We find strong evidence that indicates that time-series momentum is capturing hedging premium whereas the carry trade is only weakly related to hedging premium. Thus, time-series momentum and carry are related because both strategies benefit from the time-series and cross-sectional variability in basis, and yet they are distinct because time-series momentum alone is linked to hedging premium.
Risk Adjusted Momentum Strategies: A Comparison between Constant and Dynamic Volatility Scaling Approaches
Minyou Fan (Queen’s University Belfast)
November 28, 2017
We compare the performance of two volatility scaling methods in momentum strategies: (i) the constant volatility scaling approach of Barroso and Santa-Clara (2015), and (ii) the dynamic volatility scaling method of Daniel and Moskowitz (2016). We perform momentum strategies based on these two approaches in an asset pool consisting of 55 global liquid futures contracts, and further compare these results to the time series momentum and buy-and-hold strategies. We find that the momentum strategy based on the constant volatility scaling method is the most efficient approach with an annual return of 15.3%.
Deep Learning and the Cross-Section of Expected Returns
Marcial Messmer (University of St. Gallen)
December 2, 2017
Deep learning is an active area of research in machine learning. I train deep feedforward neural networks (DFN) based on a set of 68 firm characteristics (FC) to predict the US cross-section of stock returns. After applying a network optimization strategy, I find that DFN long-short portfolios can generate attractive risk-adjusted returns compared to a linear benchmark. These findings underscore the importance of non-linear relationships among FC and expected returns. The results are robust to size, weighting schemes and portfolio cutoff points. Moreover, I show that price related FC, namely, short-term reversal and the twelve-months momentum, are among the main drivers of the return predictions. The majority of FC play a minor role in the variation of these predictions.