Momentum has received a lot of attention in the asset-pricing literature over the past several decades, and for good reason. Trending behavior is a staple in markets. In contrast with other pricing “anomalies”, short-term return persistence—positive and negative—is a robust factor across asset classes. The fact that momentum is deployed far and wide in the money management industry and hasn’t been arbitraged away suggests that the persistence factor is persistent. The question is whether momentum as traditionally defined can be enhanced? Yes, according to a small but growing corner of research that looks at price trends through an “acceleration” lens.

Momentum is generally defined as the directional bias for asset returns to persist, particularly over a 6- to 12-month period. The modern age of momentum research begins with Jegadeesh and Titman’s 1993 study “Returns to Buying Winners and Selling Losers: Implications for Stock Market Efficiency.” Fast forward to the present and you’ll find a small library of research that extends the analysis in a variety of directions, including the recent focus on the so-called acceleration factor.

There are several ways to define acceleration, but the general concept is simply a methodology for measuring changes in momentum—“the first difference of successive returns,” as a recent paper explained (“The Acceleration Effect and Gamma Factor in Asset Pricing”). What’s the value of monitoring and measuring acceleration? This study finds that it provides “better performance and higher explanatory power than momentum.” As such, “momentum can be considered an imperfect proxy for acceleration.”

That’s an intriguing comment since momentum is already viewed as a solid framework as a risk factor and as the raw material for profitable trading strategies. But can we squeeze even more from this realm of asset-pricing analytics in the search for robust signals? Perhaps.

Another line of research along these lines comes to us by way of Morningstar, which recently published an academic study that found that acceleration is quite useful for anticipating severe market losses. “The Economic Value of Forecasting Left-Tail Risk” reports that the geometric return for the most recent six-month period less its equivalent over the preceding six months, along with trailing 1-year return, are powerful factors for predicting negative skewness in returns. The results suggest, according to the authors, “that it is possible to reduce tail risk without giving up returns.”

There are a number of variations one could devise in trying to mine acceleration as a risk metric. David Varadi has explored several possibilities, including what he labels the volatility of acceleration (VOA). Noting that this indicator has interesting properties for estimating volatility and adjusting asset weights, he writes that “the VOA framework is one step in the direction of looking at alternative and possibly better measures of volatility.”

The research on acceleration and its applications is still in its infancy, but the early efforts certainly look intriguing. It’s premature to abandon momentum in favor of acceleration. But there’s a compelling case for expanding the definition of price persistence.

jlivermoreThe Morningstar paper uses the same data to calculate the regression coefficients and perform the tests. Tests should be carried out only with coefficients that would have been known at the time (i.e. walk forward).

Pingback: Quantocracy's Daily Wrap for 11/25/2015 | Quantocracy

Pingback: 11/30/15 – Monday’s Interest-ing Reads | Compound Interest-ing!