Market beta measures a stock’s sensitivity to the overall market return. It is the slope coefficient from regressing a security’s excess return on the market’s excess return:
In the CAPM, beta is the sole determinant of expected returns. Higher-beta stocks should earn proportionally higher risk premia because they carry more undiversifiable (systematic) risk.
Empirical status
Fama and French (1992) delivered a direct challenge: in the 1963-1990 period, the cross-sectional relation between beta and average returns is flat. The slope from Fama-MacBeth regressions of returns on beta alone is 0.15%/month with a t-statistic of only 0.46. This result holds even in the longer 1941-1990 sample.
Beta’s explanatory power is subsumed by size and book-to-market equity. Once these variables are included, beta adds nothing to the cross-section of expected returns.
Role in factor models
Despite its failure as a sole pricing variable, the market excess return (RM-RF) remains the first factor in all major multi-factor models: fama-french-three-factor, carhart-four-factor, fama-french-five-factor, and the QMJ framework. RM-RF averages about 0.43%/month (t=1.76) in the 1963-1991 sample.
In commercial risk models
Bloomberg’s MAC3 includes market beta as a style factor but also adds residual volatility as a separate factor. Commercial risk models treat beta as a risk descriptor rather than a return predictor.
Absorption of the low-beta anomaly by FF5
Fama and French (2016) show that the five-factor model absorbs the low-beta anomaly through its RMW and CMA loadings. Low-beta stocks have strongly positive RMW and CMA slopes, meaning their returns covary with those of profitable, conservatively investing firms. High-beta stocks have strongly negative RMW and CMA slopes, resembling unprofitable aggressive investors.
The five-factor model reduces average absolute intercepts on Size-beta portfolios from 0.106%/mo (three-factor) to 0.072%/mo. The strong positive CAPM intercepts for low-beta portfolios disappear, and the negative CAPM intercept for megacap high-beta stocks (-0.36%/mo, t=-2.29) becomes inconsequential (-0.06%/mo, t=-0.45). There is a positive relation between multivariate beta and average returns that conforms well to the five-factor model, even though the univariate CAPM beta-return relation is flat.
Betting against beta
Frazzini and Pedersen (2014) show that the flat SML is explained by leverage constraints: investors who cannot lever up the tangency portfolio instead tilt toward high-beta assets, bidding up their prices and compressing their returns. A BAB factor (long leveraged low-beta, short high-beta) earns significant positive alpha across equities, bonds, and futures. This provides an economic theory for the low-beta anomaly beyond the statistical observation in Ang et al. (2006).
Key debate
The failure of beta does not necessarily invalidate risk-based pricing. It may indicate that a single market factor is insufficient and that additional risk dimensions (size, value, profitability, investment) are needed to capture the full structure of expected returns.
Sources
- Capital Asset Prices: A Theory of Market Equilibrium under Conditions of Risk (File, DOI)
- The Cross-Section of Expected Stock Returns (File, DOI)
- Common risk factors in the returns on stocks and bonds (File, DOI)
- Dissecting Anomalies with a Five-Factor Model (File, DOI)
- Risk, Return, and Equilibrium: Empirical Tests (File, DOI)
- Betting Against Beta (File, DOI)