Beware of the Bias (Digest Summary)

Recently published work exposes a bias in equity risk premium (ERP) research that is creating the misleading view that equities are the best investment over the long run.

The author reports that there are two important reasons for practitioners to be aware of a bias in ERP research. First, practitioners should be aware that ERP research primarily reflects the U.S. experience. The author cites results from a study that seem to indicate that equities do not always outperform bonds in countries other than the United States. Also, an investment in global bonds over 1980–2012 would have delivered a better return than an investment in global equities over the same period.

Second, practitioners should be aware that the current long-term outlook for real equity returns is 3–3.5% a year. The low outlook for equity returns has significant implications for some institutional investors. For example, charities’ spending power would steadily decline because the average spending rate of 4% of their portfolios would outpace the returns generated. Another example is U.S. corporate pension funds, which would require an abnormally large real equity return of 10% to compensate for the current low bond yields and maintain their target nominal return of 7.6% of the portfolio.

The author’s spin on the biases in ERP research are somewhat overstated. As a matter of economics, equities should have higher expected returns than bonds because of their higher risk. Findings to the contrary should be viewed skeptically because these results may be attributable to data mining or other biases. Moreover, the study the author references shows that, for the most part, equity returns outperform bond returns. Even in the case of an investment starting in 1980, the study indicated the amount by which global bonds outperformed equity returns is de minimis because both asset classes have an annualized real return of 6.4%.