Job Market Paper

Explaining Momentum within an Existing Risk Factor Model


Abstract I propose that abnormal returns generated by price momentum can be explained within the framework of an existing risk factor model such as the Fama-French three-factor model. Two features of a systematic factor, weakly positive autocorrelation and the leverage effect, generate a small positive alpha in the factor portfolio scaled by its own past returns. The momentum portfolio magnifies this alpha by taking long positions in stocks with highly positive (negative) betas and short positions in stocks with highly negative betas given a positive (negative) realized factor return. Time-varying stock betas enhance the degree of magnification significantly. I demonstrate that a simulated market in which asset returns obey the CAPM or the three-factor model can produce realistic momentum dynamics and substantial abnormal profits. In empirical tests, I show that a replicating portfolio with time-varying betas accounts for 50% of the Fama-French alpha of the canonical momentum portfolio and 75% of the value-weighted momentum portfolio. Among firms larger than the NYSE median, the momentum strategy is no longer profitable after taking into account the dynamic replicating portfolio. The residual alpha can be attributed to small firms suffering recent losses and can be completely explained away with the addition of a financial distress factor.

Latest Version: PDF
Last updated: Nov. 2012


Main Research Area:

Empirical asset pricing

Topics of Interest:

  • Pricing anomalies, particularly momentum
  • Market participation, behavioral biases and limits to arbitrage
  • Firm characteristics and asset prices
  • Effect of idiosyncratic risks on asset pricing
  • Disaster risks

Thesis Committee

Prof. Jiang Wang (chair)
Prof. Leonid Kogan
Prof. Hui Chen