Home » Research Blog » The low-risk effect: evidence and reason

The low-risk effect: evidence and reason

The low-risk effect refers to the empirical finding that within an asset classes higher-beta securities fail to outperform lower-beta securities. As a result, “betting against beta”, i.e. leveraged portfolios of longs in low-risk securities versus shorts in high-risk securities, have been profitable in the past. The empirical evidence for the low-risk effect indeed is reported as strong and consistent across asset classes and time. The effect is explained by structural inefficiencies in financial markets, such as leverage constraints for many investors, focus on the performance of portfolios against benchmarks, institutional incentives to enhance beta and – for some investors – a preference for lottery-like securities with high upside risks.

Baltussen, Guido, David Blitz and Pim van Vliet (2019), “The Volatility Effect Revisited”.

The below summary consists of quotes from the paper. Headings and text in brackets have been added.
The post ties in with the SRSV summary on implicit subsidies.

What is the low-risk effect?

“The low-risk effect [refers to] the empirical finding that higher risk [as approximated by market beta] is not rewarded with a higher return in global stock markets, nor within other asset classes…High-risk stocks do not have higher returns than low-risk stocks in all major stock markets.”

“The low-risk effect is not about a slight failure of the CAPM [capital asset pricing model], but about the total absence of a positive relation between risk and return. Many studies even find the relation to be clearly inverted.”

“The main driver of the low-risk effect appears to be volatility, which implies that it is essentially a low-volatility effect…Volatility and beta are closely related metrics, since the beta of a stock to the market index is equal to its volatility times its correlation with the market index, divided by the volatility of the market.”

“If the relation between risk and return is flat, then a long low-risk and short high-risk hedge portfolio will show an average return of zero and a strong negative CAPM beta…[One can] construct a so-called Betting-Against-Beta factor which is designed to turn this into a positive premium with zero beta, by dynamically levering the long low-risk portfolio up, to a beta of 1, and de-levering the short high-risk portfolio down, also to a beta of 1.”

For a previous explanation of the low-risk effect and betting-against-beta strategies view post here.

What is the empirical evidence for the low-risk effect?

“The empirical evidence for the low-risk effect…goes back all the way to the very first empirical asset pricing studies in the nineteen seventies.”

“[Based on] historical return series for US stock portfolios sorted on 60-month market beta, with data going back to July 1963…we plot the performance of ten decile portfolios sorted on beta, and…the performance of the 5×5 size/beta portfolios…The graphs clearly show a flat, or even slightly negative relation between risk and return.”

“Portfolios are constructed by sorting the 1,000 largest US stocks on their past 36-month volatility, with data starting in January 1929. [The first graph below] shows that the full-sample relation between risk and return is clearly flat instead of upward sloping, and even becomes inverted in the highest-risk spectrum. [The second graph below] shows that this result is robust over time.”

“The low-risk effect is remarkably robust from a geographic perspective (present in all major developed and emerging markets), from an industry perspective (present within and across industries), and from a time perspective (consistent over time).”

“Compared to the Fama-French factors, the volatility premium is not only stable through time, but also large in magnitude. Over the full sample period, the average premium is 5.8% per annum with a volatility of 9.0%. This translates into an annualized Sharpe ratio of 0.65 and an accompanying t-statistic of 5.7, well above all common thresholds for statistical significance.”

“Many anomalies are known to be concentrated in small-cap stocks and therefore difficult to exploit in reality…but… the low-risk anomaly is strongly present among the largest, most liquid U.S. stocks…[Also] that the low-risk effect exists within industries and countries, and also across industries and countries.”

“Altogether, there appears to be a low-risk effect within every asset class. The relation between risk and return only seems to be positive across entire asset classes, since stocks have higher returns than bonds, and corporate bond returns are higher than government bond returns, in the long run.”

“The low-risk effect cannot be explained by factors such as value, profitability, or exposure to interest rate changes.”

What explains the low-risk effect?

“The most popular explanations for the low-risk effect [are]…

  • The low-risk effect has been linked to the limits to arbitrage that arise from various practical constraints, in particular shorting and leverage constraints. Heterogeneous beliefs cause the investors with the most optimistic expectations to drive up the price of risky assets…For many investors the possibility to sell short or use leverage is restricted by mandate or means…In the presence of leverage constraints, however, investors looking to increase their return are forced to tilt their portfolios towards high-beta securities…When leverage constraints are tighter, the low-risk anomaly tends to be stronger…
  • A second explanation for the low-risk effect is a focus on performance relative to others instead of absolute performance. The CAPM assumes that investors only care about absolute returns, but in reality many investors focus on beating the market average…Low-risk stocks are unattractive for benchmark-relative investors, because they involve high tracking error and lower expected return…If investors only care about performance relative to others, then in equilibrium the relation between risk and return is flat…
  • Profit-maximizing asset managers have a strong incentive to create high-beta products due to the highly asymmetric nature of the flow-performance relationship…Most flows are attracted by funds with the best performance in asset classes that also had a good performance…Mutual fund managers, who have an incentive to attract investor flows, have a preference for stocks with higher idiosyncratic volatility…[There is also] evidence that sell-side analysts prefer high-volatility stocks…Analysts implicitly or explicitly have option-like reward structures, which incentivizes them to focus on high-risk assets…
  • High-risk stocks are attractive to lottery-seeking investors, because they offer limited downside risk combined with unlimited upside potential. A preference for skewness can even support the existence of an inverse risk-return relationship, i.e. that instead of requiring a compensation for taking on risk, investors may actually be willing to pay a premium for it…”

Related articles