Factor Momentum and the Momentum Factor: Rethinking the Market Momentum Effect
This post presents a visual summary of the paper "Factor Momentum and the Momentum Factor" by Sina Ehsani & Juhani T. Linnainmaa, published in the *Journal of Finance*, 2022, Vol. 77, Issue 3, pp. 1877-1919.
Research Background
- The momentum effect is a pervasive phenomenon in financial markets.
- Traditional views treat momentum as an independent risk factor.
- The existence of momentum challenges the efficient market hypothesis.
- Key question: is momentum truly a standalone factor, or a composite expression of other factors?
- There is no such thing as true "individual stock momentum" — momentum strategies are essentially timing factors.
- Momentum is not an independent risk factor, but an aggregation of the autocorrelation of other factors.
- Individual-stock momentum strategies indirectly perform factor timing.
- Factor momentum can explain all forms of individual-stock momentum.
- Analyzes time-series autocorrelation of 20 common factors.
- Constructs time-series and cross-sectional momentum strategies.
- Compares the explanatory power of factor momentum vs. individual-stock momentum.
- Uses Fama-French five-factor model regressions.
- Factor returns exhibit significant autocorrelation.
- On average, a factor earns 52 basis points per month in the year following a profitable year, but only 2 basis points after a losing year.
- Factor momentum fully subsumes the information in individual-stock momentum.
- Momentum-neutral factors display an even stronger momentum effect.
- Factor momentum concentrates in factors that explain more of the cross section of returns.
- Momentum is not a distinct risk factor; it is an aggregate of factor autocorrelation.
- Investors trading momentum are, in effect, performing factor timing.
- The results have important implications for asset pricing theory and strategy construction.
- Momentum strategy profitability depends on the persistence of factor autocorrelation.
Core Thesis
Methodology
Key finding (time-series strategy): Going long or short factors based on the past year's returns earns an annual return of 4.2% (t = 7.04).
Main Findings
Conclusion and Implications
Key Quotes
> "Our results suggest that momentum is not a distinct risk factor—it times other factors."
> "Momentum in individual stock returns emanates from momentum in factor returns."
*The original post includes charts illustrating the relationship between factor and stock momentum, factor autocorrelation analysis, and a comparison of momentum strategy performance.*