As a matter of abstract financial-economic theory, the cost of equity is straightforward. It is the minimum expected return investors require to hold the firm’s equity at the current price. Financial economists may disagree on the best way to estimate the cost of equity or the causal relationships that drive costs of equity, but it is safe to say that we know what we mean by the term when we use it. And, for nearly all equities, we almost always mean an expected return that exceeds the risk-free rate.
But what evidence do we have that the cost of equity for most firms exceeds the risk-free rate? Our ability to directly estimate individual firm costs of equity is limited. Evidence on the existence of an “equity risk premium” comes from the returns to the market as a whole, not from individual firms. Evidence on positive expected returns for stocks comes from tests on large portfolios, not individual stocks.
As for individual firms, the evidence for a positive expected risk premium is now questionable. In pathbreaking work published last year in the Journal of Financial Economics, Hendrik Bessembinder of Arizona State University (2018) reports that the majority of U.S. listed common stocks since 1926 returned (inclusive of dividends) less than the risk-free rate (that is, the one-month Treasury bill) over the lives of their issuers as listed companies, so that the shares of just 4 percent of listed U.S. companies account for all of the gains of the U.S. stock market from 1926 to 2016.
If the cost of equity for most firms was greater than the risk-free rate, then there are several possibilities for explaining these empirical results. One is that investors are risk averse but still willing to hold assets subject to such a highly-skewed return distribution. This seems unlikely (though possible), given that the need to hold the market portfolio increases dramatically when only a handful of firms can pay off significantly over the risk-free rate. It seems unlikely that we can square investor risk aversion and rational expectations with the willingness to hold small subsets of stocks that routinely underperform the market (more on this below).
Two other possibilities are that stocks have been subject to massive mispricing over almost a century or that investors have suffered an equally massive disappointment over that same period. Neither possibility seems plausible. If anything, estimates of the market risk premium suggest that actual stock returns turned out to be better than would be expected in a rational asset pricing framework.
A more plausible explanation is that the cost of equity is around the risk-free rate. I advance this hypothesis in a new research paper. I argue that stock-market investors are risk-neutral or even sometimes risk seeking and not the risk averse investors of our prior asset-pricing models. Yet risk aversion seems to be real and pervasive; our own introspection tells us this as well. Certainly, the assumption that investors are somehow risk averse pervades economics generally, not just finance.
Also real, however, is the cognitive bias of excessive optimism. It is so strong that optimism can survive even outright statements of numerical probabilities to the contrary, perhaps because people believe that unrealistic optimism is always beneficial or are just wired to act optimistically. Excessive optimism acts as a counterweight to risk aversion. Enough optimism and a risk averse individual acts as if risk neutral. More optimism than that and the risk averse individual can act as a risk seeker. Taken together, optimism and risk aversion can produce a set of investors who believe correctly that they are risk averse but act as if they are risk neutral or risk seeking because they are too optimistic.
Common stock is an almost-ideal security to evoke optimism. Compare common stock with debt, both risk-free and risky. With debt, we have a known contractual commitment to pay a certain amount on specific dates. For risky debt, we may have doubt about default and recovery, but even that problem is one-sided: The contractually-promised amounts are upper-bounds on the cash flows the investor can receive. With common stock, there is no contractually-specified amount owed. Limited liability does allow us to bound the value below at zero but we have no upper bound. There are no contractually-defined times that cash must be paid except after a dividend declaration, after which the shareholder has – as to that dividend – a contract claim against the corporation. The ambiguity of outcomes to stock investment can allow more optimism than can exist with respect to contractually-promised cash flows over a set period of time.
But if the cost of equity is around the risk-free rate, why aren’t prices higher than currently justified by discounted cash flow methods? The final piece of the cost of equity puzzle – the low realized return on almost all individual U.S. stocks over their lives – is that common stocks are more like wasting assets than perpetual dividend machines. Expected cash flows are lower than usually modeled because corporate lives are shorter than we assume, and usually much shorter.
The hypothesis could help further explain the failure of active investment management in U.S. equities. If the expected return on most stocks is around the risk-free rate, this problem is even more severe than previously thought.
While the hypothesis that the cost of equity is around the risk-free rate is no doubt controversial, the reality is that the returns to most all U.S. stocks are as consistent with that hypothesis as with a hypothesis of a positive expected risk premium for most individual stocks. That is an empirical reality we must work to understand better.
This post comes to us from J.B. Heaton, the president of legal and financial consultancy J.B. Heaton P.C. It is based on his recent article, “Is the Cost of Equity Greater than the Risk-Free Rate?,” available here.