June 24, 2011

Momentum Crashes

Kent Daniel, Tobias Moskowitz
Print this brief

The Idea:

Hedge investment risk in momentum strategies by anticipating periodic momentum crashes.

The Research

Momentum strategies, commonly used by hedge funds and other large institutional investors, are based on the simple idea that today’s winning stocks will continue to enjoy upward momentum and today’s losing stocks will continue to decline. Yet momentum confounds conventional market wisdom. In theory, stock prices should immediately jump to a level that reflects all information — yet the price trending that characterizes momentum persists for up to a year. Likewise, arbitrageurs should be able to root out inefficiencies and push prices up or down to appropriate, stable levels. But, again, momentum performance suggests otherwise.

To better understand momentum, Professor Kent Daniel examined its performance and its underlying features. First, he established that a standard momentum strategy performs remarkably well over the long term: from 1947 to 2007, winning stocks on the NYSE, NASDAQ, and AMEX returned an average of 16.5 percent annually to investors — 7.7 percent above the average return of Treasury-bills over the same period. Losing stocks returned just one percent annually.

However, there are periods of time when momentum reverses for months or even years. These momentum crashes occur when a bear market rebounds. Typically, the prices of cyclical firms (financial firms, auto manufacturers, or highly levered firms) are pushed down to very low levels over the course of a recession. When the market finally rebounds, these firms dominate. In the months after the start of the most recent such upturn, starting in March 2009, Daniel shows that the past losers were up 156 percent, while past winners were up by only about 6.5 percent. Momentum investors who bet against the past losers faced substantial losses as those losers turned into winners. (Hedge funds, with their high leverage, are particularly sensitive when their short bets reverse in a momentum crash.)

Daniel analyzed the performance of all stocks on the NYSE, NASDAQ, and AMEX from 1927–2010 to identify predictive conditions that consistently signal a forthcoming momentum crash, establishing that a high volatility market that is substantially below previous highs is prone to a momentum crash.

Daniel further investigated how the risk of momentum strategies — measured by volatility and beta (which measures how closely a stock’s returns follow those of the market as a whole) — vary with market conditions: the beta of a momentum strategy is strongly positively related to past market returns. Daniel’s techniques should be helpful in guiding investors both in controlling the risk of momentum strategies, and in anticipating momentum crashes.

Practical Applications

Portfolio managers, fund managers

You can use this research both to help anticipate momentum crashes and to rebalance your portfolio to mitigate any undesired exposure to market movements.

Working Paper, April 2013

Publication type: Working paper

View abstract/citation