Combining Value and Momentum
This is a recap of a research paper by Gregg S. Fisher, Ronnie Shah (Gerstein Fisher research department) and Sheridan Titman (Professor of Finance, University of Texas, Austin) published in the Journal of Investment Management, Vol. 14, No. 2, (2016), pp.33-48.
This paper considers several popular portfolio implementation techniques that maximize exposure to value and/or momentum stocks. Our analysis illustrates how a strategy that simultaneously incorporates both value and momentum outperforms a combination of pure‐play, independently-formed value and momentum portfolios. Our analysis also addresses the optimal way to integrate the two factors.
Both value and momentum are characteristics of companies that tend to deliver excess equity returns. Value stocks trade at a depressed level, so, stated inversely, their book value is typically high relative to their market price (i.e., greater than one). A number of studies find evidence that stocks with a high book-to-market ratio tend to have greater average returns than low book-to-market stocks. Momentum stocks are those that have delivered high returns in the recent past. A momentum effect has also been documented: Stocks with high returns over the previous 6-12 months tend to outperform stocks with low past returns. We studied the possibility of combining these two factors in a single portfolio.
Generally, value (book/market ratio) is a relatively slow-moving characteristic, but momentum is fast moving and typically causes greater turnover in the portfolio. As a result, momentum is associated with relatively high transaction costs. A pure-play value portfolio outperforms the overall market even after transaction costs are taken into account, but a pure-play momentum portfolio may underperform. However, there is still a role for momentum: When optimally introduced into a pure-play value portfolio, a momentum factor can improve the risk-adjusted return (Sharpe ratio).
We considered two alternative implementation strategies that utilize value and momentum characteristics jointly. The first strategy ranks firms by momentum and value separately, and then computes the average rank to calculate the stock’s score. The second strategy buys stocks only when both characteristics are favorable and sells them only when both characteristics are unfavorable. In addition, because the momentum signal decays much more quickly, the second strategy puts greater weight on value. With this strategy, momentum does not trigger trades but influences the portfolio by delaying or avoiding them.
The two different approaches use information about value and momentum differently. The first approach may buy a stock that has negative momentum, as long as the value signal is strong enough. The second approach may buy a stock if both signals are positive, even if neither is extreme. The first approach results is much higher trading costs, as changes in momentum scores have a large influence on trades. The second approach does not fully take into account either signal, but incurs far less trading.
Both portfolio designs are beneficial, and it is not clear-cut from our analysis which one is preferred. Over our sample period, both approaches generate similar increases in Sharpe ratios relative to combinations of pure-play portfolios. Our results suggest the first approach generates slightly higher Sharpe ratios when trading costs are low and the second approach generates slightly higher Sharpe ratios when trading costs are high.
Academic literature has identified a number of stock characteristics that are associated with significant excess returns. In this paper, we explore how two of these characteristics—momentum and value—should be optimally combined to form portfolios. Our analysis indicates that incorporating information from multiple characteristics in a single portfolio results in better performance both before and after transaction costs when compared to combinations of single-characteristic strategies. In this paper, we limit our analysis to combining value and momentum signals. Addressing how value and momentum strategies can be enhanced by incorporating other firm characteristics such as profitability, asset growth or analysts’ revisions would be a straightforward extension of this analysis.