Betting Against Beta: Why Boring Stocks Win

Key Takeaway

The Capital Asset Pricing Model makes a clear prediction: higher beta should mean higher returns. In practice, this relationship is far flatter than theory predicts -- and sometimes inverts entirely. Frazzini and Pedersen's Betting Against Beta (BAB) factor formalizes this anomaly, showing that a portfolio long leveraged low-beta stocks and short high-beta stocks delivers a Sharpe ratio of approximately 0.75 in U.S. equities. The mechanism is elegant: many investors face leverage constraints that prevent them from amplifying positions in safe assets, so they reach for return by buying riskier, high-beta stocks instead. This demand pressure overprices high-beta assets and underprices low-beta ones, creating a persistent and exploitable premium.

The Flat Security Market Line

The CAPM predicts that the relationship between beta and expected returns should be linear and steep: a stock with twice the market beta should earn twice the market risk premium. Fischer Black recognized as early as 1972 that the empirical security market line (SML) is far flatter than what the CAPM predicts.

This flatness means that high-beta stocks earn less than they should according to theory, while low-beta stocks earn more. In risk-adjusted terms, boring low-beta stocks systematically outperform exciting high-beta stocks. This is not a subtle finding -- the effect has been documented across decades of data.

The key insight is that the SML does not just fail to hold its predicted slope; it fails in a specific, directional way that creates a tradeable opportunity.

How the BAB Factor Works

Frazzini and Pedersen (2014) constructed the BAB factor by ranking all stocks in a market by their estimated beta, then forming two portfolios.

The low-beta portfolio holds stocks in the bottom beta decile, leveraged up to a beta of 1.0. If the average beta of these stocks is 0.6, the portfolio is leveraged by a factor of 1/0.6, roughly 1.67 times.

The high-beta portfolio holds stocks in the top beta decile, de-leveraged down to a beta of 1.0. If the average beta is 1.5, the portfolio is scaled down by 1/1.5, roughly 0.67 times.

The BAB factor is the return difference: leveraged low-beta minus de-leveraged high-beta, producing a market-neutral portfolio with a beta of zero. This construction isolates the pure effect of the beta anomaly, independent of overall market direction.

U.S. Equity Performance

The Sharpe ratio of approximately 0.75 is notably higher than the market portfolio's long-run Sharpe ratio of roughly 0.4 to 0.5. Because BAB has near-zero correlation with the market, it offers substantial diversification benefits when added to a traditional equity portfolio.

Why Does This Anomaly Exist?

The theoretical foundation of the BAB factor rests on one central idea: leverage constraints.

The Leverage Constraint Mechanism

Consider two types of investors. The first -- pension funds, mutual funds, retail investors -- faces restrictions on leverage. Many are prohibited from borrowing, and most cannot use more than modest amounts of leverage. The second type -- hedge funds, proprietary trading desks -- can use leverage more freely but faces funding constraints and margin requirements.

When the first group targets a specific return objective, say 8 percent annually, they cannot achieve it by leveraging a portfolio of safe, low-beta stocks. A portfolio of utilities and consumer staples might return 6 percent unlevered. To reach 8 percent, constrained investors would need 1.3 times leverage -- but they cannot use it.

Instead, they buy high-beta stocks: technology, biotechnology, speculative growth companies. This demand pressure pushes high-beta stock prices up and expected returns down. Simultaneously, the overlooked low-beta stocks remain cheaply priced.

The result is a systematic tilting of the security market line. High-beta stocks become overpriced relative to their risk, and low-beta stocks become underpriced.

Why Arbitrage Does Not Eliminate It

Several frictions prevent arbitrageurs from correcting this mispricing.

Tracking error aversion. Professional asset managers are evaluated relative to benchmarks. A low-beta portfolio will have significant tracking error versus the market index, which creates career risk for the portfolio manager -- even if risk-adjusted returns are superior.

Leverage risk. Exploiting the BAB factor requires leverage on the long side and short selling on the high-beta side. Both introduce risks beyond simple stock selection: margin calls, funding shocks, and short squeezes.

Slow convergence. Beta mispricing can persist for years. A high-beta stock can remain overpriced for an extended period before the market corrects, testing the patience and funding of arbitrageurs.

BAB vs. Low Volatility: What Is the Difference?

The BAB factor and the low-volatility anomaly are related but distinct concepts, and understanding the difference matters for portfolio construction.

In practice, the two are positively correlated -- low-beta stocks tend to also have low volatility. But they are not identical. A stock can have low total volatility but moderate beta if its returns are highly correlated with the market. The BAB factor captures a more specific effect tied to the security market line.

For long-only investors, the practical implications are similar: tilt toward low-beta or low-volatility stocks. For sophisticated investors who can use leverage and short selling, the BAB factor offers a more complete implementation of the underlying anomaly.

Evidence Across Asset Classes

One of the most compelling aspects of Frazzini and Pedersen's research is the breadth of evidence. The BAB effect is not confined to U.S. equities.

This cross-asset evidence strongly supports the leverage-constraint hypothesis. The same mechanism -- constrained investors reaching for return through beta rather than leverage -- operates across all major financial markets.

Practical Implementation

Long-Only Approach

The simplest implementation is a long-only tilt toward low-beta stocks. This can be achieved through low-volatility or minimum-variance ETFs and funds. This approach sacrifices some of the theoretical alpha (since you miss the short-side contribution) but avoids leverage and shorting costs.

Long-Short Implementation

A full BAB implementation requires leveraging the low-beta leg and shorting the high-beta leg. This is typically available only to institutional investors and hedge funds. Key considerations include borrowing costs for the leverage, short-selling costs and availability, and rebalancing frequency for beta estimation.

Beta Estimation

The quality of beta estimates matters significantly. Frazzini and Pedersen use a shrinkage estimator that blends each stock's estimated beta with the cross-sectional average. This reduces estimation error and turnover. Using rolling one-year daily returns to estimate beta, with shrinkage toward 1.0, is a reasonable starting point.

Combining with Other Factors

BAB combines well with value and momentum. The three factors have low pairwise correlations, so a multifactor portfolio that includes BAB captures distinct return premia. However, investors should be aware that BAB can underperform significantly during sharp market rallies led by speculative, high-beta stocks.

Risks and Drawdowns

The BAB factor is not a free lunch. It carries specific risks that investors must understand.

Leverage risk. The levered low-beta portfolio amplifies losses during low-beta selloffs. Even boring stocks can decline sharply in broad market stress.

Short squeeze risk. High-beta short positions can experience violent squeezes during speculative rallies, as seen in early 2021 with certain heavily shorted stocks.

Regime dependence. BAB performs best in calm, moderately rising markets. In aggressive risk-on rallies driven by speculative enthusiasm, high-beta stocks can outperform dramatically, causing the BAB factor to suffer.

Crowding. As more investors have adopted low-volatility and BAB strategies, there is evidence that the premium has narrowed somewhat, particularly in the most liquid markets.

Independent Backtest: BAB Factor by Decade

Methodology: Long leveraged low-beta stocks (bottom 30% by estimated beta, scaled to beta of 1.0) minus short de-leveraged high-beta stocks (top 30%, scaled to beta of 1.0). Monthly rebalancing using 1-year rolling beta with shrinkage toward 1.0. U.S. equities, January 1927 through December 2025. Returns before transaction costs and financing charges.

The BAB factor delivered strong risk-adjusted returns across most decades, with a full-sample Sharpe ratio of 0.48 -- substantially above the market's long-run Sharpe of roughly 0.40. The 2010s and early 2020s mark a significant weakening, coinciding with the dominance of high-beta mega-cap technology stocks and the widespread adoption of low-volatility strategies. The leveraged construction creates larger drawdowns than most factor portfolios, with the 38.5% maximum drawdown reflecting the amplification inherent in the strategy.

Cross-Market Evidence

Frazzini and Pedersen (2014) documented the BAB effect across all major asset classes, finding consistent evidence that the security market line is too flat everywhere -- not just in U.S. equities. Asness, Frazzini, and Pedersen (2012) extended this to show that the leverage constraint explanation unifies the low-risk anomaly across asset classes. The breadth of cross-asset evidence distinguishes BAB from factors that are primarily equity phenomena and supports the leverage-constraint mechanism as a fundamental feature of capital markets.

The Leverage Constraint Frontier

The Betting Against Beta factor occupies a distinctive position among investment factors: it is grounded in a clear theoretical mechanism rather than being a purely empirical anomaly.

Black (1972) first identified the flatness of the empirical security market line, and Frazzini and Pedersen (2014) formalized the leverage-constraint explanation into a tradeable factor. The theoretical elegance is matched by extraordinary empirical breadth -- few other factors demonstrate consistent premia across equities, bonds, credit, currencies, and commodities simultaneously.

However, the practical challenges of harvesting the BAB premium are substantial. The leveraged construction requires continuous financing and rebalancing, creating costs that erode gross returns. Beta estimation noise introduces turnover and potential misclassification. The short side faces borrowing costs, short-sale constraints, and periodic squeeze risk. And the 2010s demonstrated that the premium can vanish for extended periods when leverage-unconstrained capital flows into low-beta strategies.

For practitioners, the evidence suggests deploying BAB exposure as one component within a diversified multi-factor framework rather than as a standalone allocation. The low correlation with value (Fama and French 1993), momentum (Jegadeesh and Titman 1993), and quality (Novy-Marx 2013) makes BAB a valuable portfolio diversifier. Long-only implementations through minimum-variance or low-volatility strategies capture a meaningful fraction of the premium while avoiding the leverage and shorting costs that make full BAB implementation prohibitive for most investors. The persistence of the anomaly likely reflects the structural nature of leverage constraints -- pension funds and mutual funds are unlikely to be permitted unlimited leverage regardless of how well-documented the BAB premium becomes.