One of the “most well-documented and well-researched” investment factors may not be as powerful as commonly believed, according to researchers at Germany’s University of Mannheim.
That factor is momentum, or the tendency of past winners to continue to outperform past losers in the stock market. First documented in 1993, momentum has been incorporated into numerous asset management strategies over the last few decades. But the market anomaly may be “much weaker than previously thought,” according to new study from University of Mannheim finance professor Erik Theissen and PhD student Can Yilanci.
“A convincing explanation for the existence of profitable momentum strategies is missing. We thus do not know why momentum strategies seemingly allow investors to earn abnormal returns,” the pair wrote in their paper, “Momentum? What Momentum?”
To better understand the momentum effect, Theissen and Yilanci compared the actual returns of momentum portfolios to the expected returns for the individual stocks in the portfolios. This is in contrast to previous momentum studies, which have analyzed the momentum effect at the portfolio level. The Mannheim researchers argued that this portfolio-level risk analysis of momentum returns does not yield clear results, because momentum strategies have high turnover.
“As a consequence, the factor exposure of momentum portfolios varies over time, and it does so in a way that is systematically related to past factor realizations,” they wrote. “The portfolios tend to load positively (negatively) on factors that performed well (poorly).”
This means the average factor exposure “may well be close to zero,” they added.
[II Deep Dive: The Momentum Factor Is Real. Too Bad It Doesn't Work.]
Analyzing for risk at the stock level solves this problem, Theissen and Yilanic argued.
“This procedure, which we denote stock-level risk adjustment, accounts for the turnover in the momentum portfolio because, in each month, the factor exposure is based on the actual composition of the winner and loser portfolios,” they wrote.
Using this method to analyze U.S. stocks from July 1963 to December 2018, Theissen and Yilanci found that the profitability of the momentum factor “largely disappears.” They had the same finding when dividing the portfolios into micro-, small-, and large-cap stocks.
Splitting the overall sample into shorter time periods revealed a momentum premium in the period from 1963 to 1979. But this premium also disappeared when transaction costs were taken into account.
“Our results imply that the momentum effect may actually be much weaker than previously thought, and that the returns to momentum strategies may, to a large extent, be a compensation for risk,” the authors concluded. “This insight has potentially important implications for portfolio managers pursuing momentum-based strategies.”