Swedroe: Momentum Across Time & Asset Classes

June 12, 2015

The academic study of price momentum has intensified considerably since 1993, the year Narasimhan Jegadeesh and Sheridan Titman’s paper, “Returns to Buying Winners and Selling Losers: Implications for Stock Market Efficiency,” appeared in The Journal of Finance. The authors found that buying winning stocks and selling losers generated significant positive returns over three- to 12-month ownership.


A recent contribution to the body of literature exploring various aspects of price momentum is a May 2015 study from Christopher Geczy and Mikhail Samonov, “215 Years of Global Multi-Asset Momentum: 1800-2014 (Equities, Sectors, Currencies, Bonds, Commodities and Stocks).”


Looking At A Long Timeline

The authors set out to create and examine the longest possible history of the global asset momentum effect to more thoroughly understand its time-series properties and to better explain general characteristics of the phenomenon already discovered in more recent history.


Using databases from Global Financial Data and additional information available through Bloomberg, the authors created an expanded data set that goes back to 1800 and includes 47 country equity indexes, 48 currencies (including the euro), 43 government bond indexes, 76 commodities, 301 global sectors and 34,795 U.S. stocks.


Following is a summary of Geczy and Samonov’s findings:

  • The 215-year history of global multi-asset class price momentum generated for the study showed that returns to momentum were consistently significant for each of the six asset classes examined (country equities, currencies, country government bonds, commodities, global sectors and U.S. stocks).
  • Between 1800 and 2014, country equity momentum (price-only) had the largest long/short spread of the six asset classes. The premium was 0.88 percent per month (t-stat 10.6). Using total returns, the premium was smaller, but still clocks in at 0.57 percent per month (t-stat 6.8). The second-largest momentum premium appeared in currencies, with a spread of 0.51 percent per month (t-stat 9.6). U.S. stock momentum generated a premium of 0.51 percent per month (t-stat 6.0), global sector momentum generated a premium of 0.36 percent per month (t-stat 6.6) and global country bond momentum generated a premium that averaged 0.13 percent per month (t-stat 2.3). A cross-asset class momentum strategy, consisting of four asset classes, generated a premium of 0.45 percent per month (t-stat 10.2).
  • Inverse commodity momentum generated a premium of 0.45 percent per month (t-stat 5.5). This finding on inverse momentum is interesting and it contradicts other studies. However, Geczy and Samonov used spot prices where other researchers used futures prices. They employed spot prices because futures prices have only become available in recent decades. However, spot prices, unlike futures prices, are not investable.
  • The long/short momentum portfolios exhibited significant variation of beta compared with the average return of the corresponding asset class from which the portfolios were formed. In other words, the longer an up- or downmarket state persists, the larger the absolute value of the momentum portfolio’s beta, creating a dynamic risk profile for momentum over a given market cycle. Overall, for every additional month a given market state continued, the combined momentum long/short portfolio beta increased by 0.01 percent (t-stat 19.0) in the direction of the market return.
  • There is consistent evidence of long-run reversion and short-run continuation, with the exception of U.S. stocks. U.S. stocks experience a short-run reversal, which is why the most recent month is excluded by most practitioners.
  • Over the long run, bond momentum leads country equity returns.
  • Momentum outperforms its “cousin” effect—trend—although both remain highly significant.
  • About 60 percent of the premium in equity momentum comes from the long side. For currency momentum, the figure is 35 percent. For global sector momentum, the figure is 44 percent. For cross-asset-class momentum, half the premium comes from the long side. These are important figures, because long-only momentum avoids the crash risk of long/short momentum strategies.
  • Momentum profits are much higher following upmarkets than following downmarkets.
  • While the average five-year rolling correlation among the seven long/short momentum portfolios over the full period was quite low—9 percent—with long cycles of variability, correlations have been trending upward since reaching a low point in the 1950s. As of May 2014, the five-year cross-momentum correlation reached an all-time high of 41 percent. The upward trend in correlations strongly indicates that the diversification benefit of momentum investing across global asset classes has been significantly diminished. It seems likely the increase is due to continued global market integration as well as increased allocation of investment capital to momentum strategies. This has risk management implications for momentum-tilted strategies.



Find your next ETF