Petri Jylha, Imperial College

Matti Suominen, Aalto University

Tuomas Tomunen, Columbia University

Betting against beta

All MBAs and Finance students learn in their basic finance courses the Capital Asset Pricing Model (CAPM), a celebrated theory largely attributable to the Nobel price winner William Sharpe. This theory states that riskier assets in equilibrium should earn higher returns, and that the relevant measure for a stock’s risk should be its “beta,” a measure of the stock’s systematic risk. Technically a stock’s beta equals its correlation with the stock market index, scaled by the ratio of its volatility to the market index volatility.

All well in theory, but in practice the CAPM has failed miserably. In the real life the stocks with the higher risk measures, i.e., the high beta stocks, have over the recent decades systematically earned lower returns than the low beta stocks. In fact, investing in low beta stocks has become a highly popular investment strategy in the financial market, one that is today aggressively marketed to all major institutional investors.

Be careful: Betas do not measure what you think they measure!

[drizzle]In a recent working paper “Beta Bubbles,” we suggest a potential reason why the logically well motivated CAPM fails to work in practice. Most importantly, we show that the stock’s beta in reality seems to measure not so much the stock’s level of risk, but rather how frequently it is being traded.

Studying US stock markets starting from 1970s, we find that the low beta stocks are commonly held by few passive long-term investors. These stocks have low average turnover, in fact they do not trade at all on 25% of the days and on nearly 70% of the days their trading volume is less than 0.1% of the stock’s market capitalization. Intuitively, these stocks are so rarely traded that they rarely co-move with the market. It does not necessarily mean that the low beta stocks are less risky, just that the traditional risk measure beta fails to measure their risk. These low beta stocks are more prone to jumps, i.e. large market revaluations of their value.

High beta stocks, in turn, are owned by active short-horizon investors that continuously monitor the market and trade. These short-horizon investors’ entry and exit from the stock market seems also to occur in tandem with the returns of the entire market. For both reasons, stocks owned by the short-horizon investors co-move highly with the market. As the high beta stocks are also more widely held, their risk is more evenly distributed amongst investors and the investors require less return from holding them. Hence the poor future returns to the high beta stocks.

Importantly, the stocks betas change over time as the stocks’ popularity changes. For instance, 20% of the stocks in the lowest beta decile of stocks (the 10% of the stocks with the lowest beta) had an above median beta in the previous year. When a stock goes out of fashion and institutions sell the stock, we find that its beta declines. When a stock becomes popular among the active institutional investors its beta rises.

Low beta bets make sense – but prudence required

The low beta stocks are thinly owned and thinly traded. As they are unpopular among investors today they in principle make up for great investments. However as the low beta investing has now become a popular investment theme, there is large risk that an investor investing in low beta stocks today is in for a big surprise. The investor may find that the prices of low beta stocks run up as he tries to take positions in these stocks. After all, these are illiquid and thinly traded stocks. Secondly, he may find that these stocks’ betas increase, as according to our working paper the betas are a function of the investor population and the betas increase as the number of investors increase. Thirdly, the investors investing today in low beta stocks should be expecting that these stocks’ prices drop vastly when all the investors following the low beta investment theme today eventually try to get rid of their former low (now higher) beta stocks.

So indeed, there is reason to be careful with your low beta bets!

Low Beta Bets

Low Beta Bets

Low Beta Bets

See full study below.