David Hirschleifer conducts a simple thought experiment in a wonderful 2001 article. He imagines an alternative reality in which the foundations of modern finance are a “Deficient Markets Hypothesis” (DMH) and a “Deranged Anticipation and Perception Model” (DAPM), instead of the EMH and Capital Assets Pricing Model (CAPM).² He suggests that had an inefficient market hypothesis been embraced at the dawn of modern finance, it would have been more easily validated and less readily falsifiable than the EMH.

Words have consequences. A risk premium presumes a logical homo economicus who makes investment choices based on sound reasoning and objective reality. At a minimum, a risk premium presumes a collectively rational market, with individuals’ emotions canceling in some psychological variant of the central limit theorem. The “madness of crowds” cannot exist. But suppose in the early days of the EMH, academia had thought like Hirschleifer and acknowledged the human component of capital markets and sought to understand the nature and magnitude of a “fear premium” instead of an ostensibly quantifiable “risk premium.”

A fear premium accounts for human emotion and renders many of the “anomalies” and “factors” that have consumed academic finance in recent decades unsurprising, perhaps even expected. In an efficient market, can a frothy growth stock be priced at a negative risk premium? Under most conditions, of course not.³ Its risk—and its contribution to portfolio risk—is objectively higher than that of the risk-free asset. The market must therefore expect sufficient future growth to justify the premium valuation multiples, no matter how large they may be.

Conversely, with a fear premium, might our frothy growth stock be priced at a negative fear premium? Of course! Whenever the fear of missing out (FOMO) leads investors to view low-risk assets with dread. Charlie Munger often observed, “The world is not driven by greed. It’s driven by envy.” It is difficult to imagine how any investor could view an investment in today’s “Magnificent Seven” as less risky than an investment in cash. But the “Mag-7” could easily sport a negative “fear premium,” priced to offer a lower future return than cash, if investors collectively fear the opportunity cost of idle cash reserves more than they fear the volatility of their beloved growth stocks.⁴

Several myths relating to the equity risk premium (ERP) have persisted for quite a while, so it is important to separate myth from fact to better understand the natural limits of the risk premium. Let’s begin with the unarguable: Stockholders are further down in the capital structure of a business. Bonds have first call on company resources, stocks are secondary. Therefore, for the capital markets to “work,” stocks should be priced at levels that can produce higher expected future returns than bonds. This relationship should hold particularly when comparing stocks to (ostensibly) risk-free government bonds. Unsurprisingly, stockholders have enjoyed outsized returns from their investments. When investors collectively expect a greater return for stocks relative to bonds or cash, we call this expectation the “equity risk premium.”

At its essence, the equity risk premium is the incremental return that we should earn relative to a risk-free asset, as a reward for our willingness to bear risk. But there is no risk-free asset. Cash may have zero volatility, but its purchasing power diminishes over time. Unless we’re planning to spend the money tomorrow, cash is not risk-free. Bonds may have a recurring coupon that creates an illusion of eliminating risk from our ability to fund our long-term spending needs. But most of us have spending needs that vary over time, notably rising with inflation. Long Treasury Inflation-Protected Securities (TIPS), inflation-linked bonds, can be used to fund a real spending plan, well into the future, with minimal risk. In short, the risk-free asset will vary from one investor to another, depending on what future spending our investments are intended to serve.⁵

Accordingly, we need to tightly define our meaning for the term “risk premium,” in order to avoid confusion or disagreements that are rooted in semantics, rather than true real-world differences of opinion. Not only do we need to define what our risk-free asset is (e.g., cash, bonds, or TIPS) and what our risky asset is (e.g., U.S. stocks, global stocks, individual stocks, or some other risky asset). We also need to specify what return difference we are using. A risk premium may be based on single-year expected arithmetic return differences, geometric return differences, log return differences, or compounded multiyear return differences. With such semantic differences, we risk talking past one another in any dialogue on the topic.

Exhibit 2 shows how this can happen. Suppose we are comparing past excess returns, which I define as the log-return difference between stock and bond returns, and you define as the annual arithmetic difference between stock and Treasury bill returns. Over the last 75 years, I would calculate a 5.6% excess return compared to your 8.8%. That’s a big difference, but we’d both be correct.⁶ Any discussion of either past excess returns or the expected equity risk premium must precisely specify what definitions we are using.

Many of modern finance’s foundational theories can be misleading or worse. That doesn’t mean they have no insight to offer. While ample evidence, from Roaring Kitty to Zoomability, may demonstrate that markets in the real world are far from efficient and that homo economicus is as elusive as Bigfoot, by seeking greater clarity and zeroing in on precisely what these concepts mean at the individual level, we can gain a better sense why and how markets and investors behave the way they do. Yes, investors are prone to irrational decision-making, to behavioral biases and cognitive errors. They may also have separate and distinct definitions of the equity risk premium and its components. But that doesn’t mean we should dispense with the ERP or any of the other fundamental principles of modern finance. Our task today should be to think like Hirschleifer did when he developed the equity fear premium and inefficient market hypothesis and seek to bridge the gap between conventional and behavioral finance.