The capital asset pricing model has suffered a few indignities of late. The most recent comes from a series of academic articles questioning whether beta—or the one moving part of the model that determines the expected return of risky assets—is doing a good job of predicting subsequent returns. These studies find that high-beta stocks don't perform as well as they should, and that low-beta stocks do better than the model would predict. In other words, the slope of the security market line, or the tradeoff between risk and return, isn't a straight line.
This is a problem for efficient-market types because it suggests that high-beta securities are overpriced and low-beta securities are too cheap. If cheap and expensive securities exist within a free and open market, there should be an opportunity for sophisticated investors to earn riskless profits by selling expensive securities to buy the cheap ones.
It's tempting to conclude that CAPM and beta may be of limited use if the data don't match the theoretical relationship between risk and return according to the model. However, Andrea Frazzini and Cliff Asness of AQR Capital Management and Harrison Hong and David Sraer of Princeton University believe that the problem is not with the theory, but with the limitations of capital markets. These limitations explain why high-beta stocks become overpriced and why their prices don't fall, even when sophisticated investors are aware that a mispricing exists.
In an article just published in the Financial Analysts Journal, Frazzini, Asness and Lasse Pedersen of New York University attribute the beta anomaly to investors' unwillingness to leverage their investment portfolios. Assets may be seen as existing on a continuum of risk based on covariance with the market for risky assets (beta). The market portfolio, which according to the authors consists of 42% stocks, 48% bonds and a few other assets, provides a beta of one.
According to modern portfolio theory, this value-weighted market portfolio provides the highest return per unit of investment risk, or Sharpe ratio. Except it doesn't.
Bonds consistently have higher Sharpe ratios than stocks. This fact isn't of much use to most investment advisors because their clients want a higher expected return than they can get from bonds alone. So an advisor selects a portfolio that consists of a mix of stocks and bonds that provides a higher expected return with more risk. But the authors point out that you don't have to increase your stock allocation to get a higher expected return. You can just borrow money and invest more in bonds. As you increase your leverage, your risk goes up but so does your expected return.
Theoretically, you can create a portfolio of bonds that has the same risk as a balanced equity/bond portfolio through leverage. The problem is that many investors can't easily create a leveraged portfolio.
This "leverage aversion" increases demand for riskier assets among investors who are willing to accept greater risk for greater return but are not willing or able to do so through a leveraged portfolio. Mutual funds and pension managers may also be restricted from increasing leverage to generate greater returns. So investors demand more risky assets, which drive down their risk-adjusted performance. Safer assets become less popular, resulting in a higher Sharpe ratio. Aversion to leverage or institutional barriers to creating leveraged portfolios among investors creates the price pressures that allow less risky investments to outperform in an otherwise efficient market.
Frazzini and his co-authors provide convincing evidence that the outperformance of low-beta investments isn't just the result of mining data from a brief period of U.S. asset returns. Between 1926 and 2010, stocks had a Sharpe ratio of 0.35 while bonds achieved a much higher 0.47. A portfolio that maximized Sharpe ratio consisted of 88% bonds and 12% stocks—much more heavily weighted toward bonds than the theoretically optimal value-weighted market portfolio from modern portfolio theory.
While stocks indeed outperform bonds historically by beating the risk-free rate by 6.71% versus 1.56% for bonds, the authors remind us that this is theoretically irrelevant. They insist that an investor focus not on asset allocation but on risk parity; that it to say, compare portfolios that have the same amount of risk in order to judge their performance. By leveraging the most efficient 88% bond portfolio to match the risk of a traditional 60/40 portfolio, it was possible to achieve an annual return that is 3.34% higher.
Thus, one could create a portfolio that is as risky as a 60/40 portfolio by using a leveraged portfolio heavily weighted toward bonds and achieve a 334 basis point performance improvement. This surprisingly robust performance by a leveraged low-risk portfolio also outpaced a sample of global markets and within different time periods.
Investor Impulses
