Betting against beta (BAB) is an investment strategy, and the name of the return factor built from it, that holds assets with low beta and sells short assets with high beta. The low-beta holdings are scaled up with borrowing and the high-beta shorts are scaled down, so that each side has a beta of one and the net position has no exposure to the overall market.
The term comes from a paper of the same name by Andrea Frazzini and Lasse Heje Pedersen, published in the Journal of Financial Economics in 2014. They reported that such portfolios earned positive average returns across many asset classes and offered an explanation based on limits to borrowing. BAB has since become a standard factor in academic research and in factor investing.
Background
The capital asset pricing model predicts that an asset's expected return above the risk-free rate rises in proportion to its beta. Tests from the early 1970s onward, notably by Black, Jensen and Scholes, found that the actual relationship was flatter. Low-beta shares earned more than the model predicted and high-beta shares earned less.
Fischer Black showed in 1972 that a flatter line follows if investors cannot borrow freely. Frazzini and Pedersen built on this idea. Many investors, such as pension funds and mutual funds, cannot or will not use leverage. To reach a higher expected return they overweight risky, high-beta assets instead. That demand pushes up the prices of high-beta assets and lowers their future returns, while low-beta assets are neglected and offer better returns per unit of risk.
Formula and example
The return on the BAB factor is:
- and are the returns on the low-beta and high-beta portfolios.
- and are the betas of those portfolios.
- is the risk-free rate.
Dividing by beta is what scales each side to a beta of one. Suppose the low-beta portfolio has a beta of 0.7 and earns 6 percent above the risk-free rate, and the high-beta portfolio has a beta of 1.4 and earns 8 percent above it. The long side contributes 6 ÷ 0.7 = 8.57 percent. The short side costs 8 ÷ 1.4 = 5.71 percent. The BAB return is 8.57 − 5.71 = 2.86 percent. The high-beta portfolio earned more in absolute terms, but less per unit of beta.
For each $1 of capital, the example holds about $1.43 of low-beta assets and is short about $0.71 of high-beta assets. The construction is market-neutral in beta terms but not in dollar terms.
Evidence
The original study covered US shares, shares in twenty international markets, government bonds, corporate bonds and futures. In each, portfolios built this way had positive average returns and a high Sharpe ratio over the sample. In bonds the pattern appears as short-maturity bonds, which have low beta, earning more per unit of risk than long-maturity bonds.
The theory makes further predictions that the authors tested. When funding conditions tighten, leveraged investors must cut positions, and the factor should lose money at those times. Investors who can borrow, such as leveraged buyout funds, should hold low-beta assets, and investors who cannot should hold high-beta assets. The paper reported support for both.
Criticism
Later research has questioned how much of the return an investor could actually capture. A study by Novy-Marx and Velikov argued that the factor's performance owes much to its construction. The original method weights stocks by beta rank instead of by market value, which gives heavy weight to very small, illiquid companies. Trading those shares is expensive, and after realistic transaction costs the excess return shrinks considerably.
Other researchers find that much of the effect overlaps with known factors. Low-beta shares tend to be profitable, conservative companies, and controlling for those traits in a factor model reduces the unexplained return, or alpha. Some argue the anomaly is better described by volatility than by beta, or that it reflects a taste for lottery-like stocks more than borrowing limits. Estimated betas are also noisy and change over time.
Risks in practice
A BAB portfolio is leveraged and short, so it depends on financing. It tends to lose when credit is scarce and lenders demand more collateral, which is when losses are hardest to bear. Because the long and short sides differ in dollar size, the position is not protected against every market move, and a sharp rally led by high-beta shares can cause large losses. Short selling adds borrowing fees and recall risk.
Investors seeking the effect without leverage or shorting often buy low-volatility or minimum-variance funds, which simply overweight calmer stocks. These keep full systematic risk exposure to a falling market and are a related but different strategy. Past factor returns, including those of BAB, are no assurance of future results.