Betting Against Beta

The CAPM makes a sharp promise: expected return should rise steeply with Beta (β). Take on twice the market's risk and you should be paid roughly twice the market's risk premium. In the data, that line is far too flat. Low-beta stocks earn more than the CAPM says they should, and high-beta stocks earn less. Betting Against Beta (BAB), from Frazzini and Pedersen, is the strategy built to harvest exactly that gap — and its explanation is one of the cleanest arguments in asset pricing.

BAB is the sharp, beta-specific version of the broader The Low-Volatility Anomaly: where the low-vol story is about total risk, BAB isolates market beta and turns the flat line directly into a trade.

Why the line is too flat

The mechanism does not require anyone to be irrational — only constrained. Lots of investors (pension funds, mutual funds, most individuals) want more than the market's return but are not allowed to borrow, or are unwilling to. If you cannot use leverage but crave higher returns, your only lever is to overweight high-beta stocks — they give you built-in leverage without a margin loan. That crowd bids high-beta prices up, which pushes their future returns down. The security market line gets flattened and lifted.

The BAB portfolio

The unconstrained arbitrageur does the opposite of the crowd: buy low-beta stocks and lever them up to a beta of 1; short high-beta stocks and scale them down to a beta of 1. Because both legs are set to the same beta, the position is market-neutral by construction:

$$r^{\text{BAB}}_t = \frac{1}{\beta_L}\big(r_{L,t}-r_f\big) - \frac{1}{\beta_H}\big(r_{H,t}-r_f\big) .$$

Here $r_{L,t}$ and $r_{H,t}$ are the returns of the low-beta and high-beta baskets, $\beta_L$ and $\beta_H$ their betas, and $r_f$ the risk-free rate. Dividing the low-beta leg by its small beta levers it up; dividing the high-beta leg by its large beta scales the short down. The two betas cancel, leaving a bet purely on the flatness of the line.

Worked example

Say the low-beta basket has $\beta_L = 0.7$ and returned $8\%$, the high-beta basket has $\beta_H = 1.4$ and returned $9\%$, and the risk-free rate is $2\%$. On raw returns the high-beta basket won — a naive investor concludes "more risk paid off." Now run BAB:

$$r^{\text{BAB}} = \frac{1}{0.7}(8\% - 2\%) - \frac{1}{1.4}(9\% - 2\%) = 1.43\times 6\% - 0.71\times 7\% = 8.6\% - 5.0\% = 3.6\% .$$

The strategy makes $3.6\%$ market-neutral, even though high-beta had the higher headline return. Check the neutrality: the long leg carries beta $\tfrac{1}{0.7}\times 0.7 = 1$, the short leg $\tfrac{1}{1.4}\times 1.4 = 1$, so net beta $= 1 - 1 = 0$. All the return came from the mispricing of risk, none from market direction.

Where it stumbles

• Funding risk is the whole catch. BAB requires leverage to lever the low-beta leg. So the very constraint it exploits — the cost and scarcity of borrowing — is also its Achilles heel. When funding tightens in a crisis, the low-beta longs must be delevered and the high-beta shorts squeeze, so BAB itself crashes. It is effectively short liquidity.
• Beta estimation error. The whole trade hinges on measuring $\beta_L$ and $\beta_H$ correctly; noisy or stale betas leave residual market exposure the strategy pretends it does not have.
• Overlap with other bets. Low-beta baskets lean toward bond-proxy sectors (utilities, staples), so a naive version smuggles in a duration bet that suffers when rates rise. It also overlaps heavily with the low-vol, quality, and size factors.
• Crowding and decay. Once "min-vol" and BAB products became popular, low-beta valuations richened, compressing the forward premium.

In interviews

State the anomaly against the CAPM crisply: the realized security market line is too flat, so low-beta stocks have higher risk-adjusted returns than high-beta stocks. Then give the mechanism — leverage-constrained investors reach for return by overweighting high-beta, bidding it up — and the fix: a beta-neutral book, long levered low-beta and short delevered high-beta. The sharp closer is that BAB is short funding liquidity, so it is not free money: it crashes in the exact deleveraging events when you most want it to hold up. See The Low-Volatility Anomaly for the total-risk version of the same story.