Equity Risk Premium Forum: The Deficient Market Hypothesis
“If we thought of the equity premium as a fear premium — if we had the luxury of going back 60 years and labeling it a fear premium — a lot of the so-called anomalies that we’ve talked about would not be anomalies at all. They would be totally reasonable and expected.” — Rob Arnott
In the second excerpt from the Equity Risk Premium Forum discussion, Laurence B. Siegel and fellow participants Rob Arnott, Cliff Asness, Elroy Dimson, Roger G. Ibbotson, Martin Leibowitz, Rajnish Mehra, and Jeremy Siegel delve deeper into the nature of the concept.
Arnott suggests at the outset that the description “risk premium” might be something of a misnomer that obscures more than it reveals. From there, the participants broaden their range of inquiries, exploring, for example, what a Roman centurion who safely invested a drachma at 4% interest might have generated over the centuries and why that almost infinite wealth was never created. That leads them to consider why there aren’t more billionaires.
This installment of the conversation concludes with an analysis of why the equity risk premium is so large. Generally, calculations suggest that to make the move “from riskless to risky,” as Leibowitz puts it, people need a premium in the range of 4% to 6%, which leads Mehra to wonder whether risk accounts for that entire premium or whether other factors may contribute.
Rob Arnott: For at least 20 years, I’ve been an advocate of the notion that we shouldn’t call it a risk premium. We should call it a fear premium. Many of you may remember David Hirshleifer’s famous thought experiment in 1999, in which he said: Suppose a school in Chicago had come up with the deficient (or deranged) market hypothesis and Bill Blunt (i.e., not Bill Sharpe) at Stanford had come up with DAPM, the disorderly asset pricing model; they would be declared the most validated and proved set of hypotheses in the social sciences.
He was joking, but he meant that if your starting point was market inefficiency, you could find ample proof of that, just as many efficient-market types say it’s well documented that the market is efficient. If it had been called a fear premium from the beginning, the value effect would be expected — not as a risk factor but because buying loathed and feared companies is scary. The size effect would be expected but relatively weak, because buying small companies that are not widely understood engenders a little more fear than buying well-established companies.
Roger’s liquidity factor would be expected. Long-horizon mean reversion would be expected. Even momentum would be expected, based on fear of missing out. If we thought of the equity premium as a fear premium — if we had the luxury of going back 60 years and labeling it a fear premium — a lot of the so-called anomalies that we’ve talked about would not be anomalies at all. They would be totally reasonable and expected.
Roger G. Ibbotson: I think that the fear premium is an interesting concept, and I’ll give it some thought. I’ve used the word “popularity,” which includes all kinds of premiums, whether they are risk or non-risk. And I think that risk has become too dominant in the discussion of asset pricing because the key idea is preferences.
We started out with the capital asset pricing model, where you are afraid of only one thing, one kind of risk. Ultimately, we generalize it to include many dimensions of risk, but we want to generalize it even further, to non-risk characteristics. For example, I don’t think of liquidity (actually the lack of it) as a risk, even though the literature talks about liquidity risk. You can conceive of a liquidity factor, but that factor does not make liquidity a measure of risk. Illiquidity may be a source of fear. However, there are a lot of preferences that go beyond fear.
But I agree with you, Rob, that fear encapsulates a broader notion than risk as we measure it. It’s an interesting concept, but it might not be general enough.
Jeremy Siegel: I’d like to address Raj’s original article, which asks, “Why is the equity risk premium so big?” Everyone has twisted and turned, used the Von Neumann-Morgenstern utility function, and done various other things to get an answer. Does anyone here have an explanation that they feel satisfied with for why the equity risk premium is so large and persistent and universal?
Rajnish Mehra: I’ve tried to give some answers. I think the borrowing constraint stuff that I did with George Constantinides and John Donaldson is one answer. If younger people can’t borrow to buy enough equities to hedge their future income uncertainty and older workers have mostly resolved their income uncertainty, then (as we wrote):
“[F]luctuations in [the] consumption [of older workers] occur from fluctuations in equity income. At this stage of the life cycle, equity income is highly correlated with consumption. Consumption is high when equity income is high, and equity is no longer a hedge against fluctuations in consumption; hence, for this group, it requires a higher rate of return.”
And this middle-aged group is the dominant, price-setting group in the equity market. So, this market segmentation story is, I think, a reasonable explanation for equity prices that are low enough to provide, on average, a high rate of return.
Laurence B. Siegel: Some decades back, I wrote that the equity market is much riskier than it looks from the Ibbotson chart because nobody gets those returns. The evidence that nobody gets those returns is that we’re not all rich. From time to time, almost everyone has cash flow needs, emergencies, times when you need to withdraw from the market or at least can’t contribute to it. As Jeremy has said, you spend the “income,” but income is a legal concept denoting whatever is subject to the income tax. More likely you spend your market “profits” in whatever way your mental accounting defines “profit.” So, the vagaries of human life make it impossible to realize a 5%, 6%, 7% equity premium.
Martin Leibowitz: On that score, I’m reminded by an event that took place when Sidney Homer and I were writing Inside the Yield Book. It goes back to the 1960s and early 1970s. After we had written the book, Sidney asked me a question. He said, “Suppose a Roman centurion had invested one drachma at 4% and this compounded in a totally safe way over the years?” He asked me to calculate what that total amount would be today.
This turned out to be a very difficult problem because standard calculators couldn’t do the math. Even using a computer didn’t work. I had to use logarithms, and when I got the answer, it turned out to be far more than the total wealth of the world at that point in time.
L. Siegel: I’m calculating it while you speak . . .
Leibowitz: The next question was an even better question. Sidney asked, “What happened to it all?”
L. Siegel: Where did all the money go? Of course, I would say that all that wealth was never created in the first place. The idea of investing a drachma at 4% for 2,000 years is a thought experiment that has never been put into practice.
J. Siegel: People consumed the dividend. The growth-of-a-dollar, or drachma, calculation assumes that we invest the dividend.
L. Siegel: Consumption!
J. Siegel: You consume the dividend.
Leibowitz: Consumption, wars, pandemics.
J. Siegel: No. Just consume the dividend. You don’t need any of that other stuff.
L. Siegel: It’s 2.6 x 1034 drachmas.
Arnott: I did that as a thought exercise in one of my Journal of Portfolio Management papers. In a footnote, I hypothesized one ounce of gold, which at the time was $350 an ounce. So, 1/350th of an ounce of gold back at the birth of Christ growing at 5% and the outcome was a sphere of gold as large as the earth’s orbit around the sun.
L. Siegel: And if you add a few more millennia and go back to the days of the pyramids, the sphere of gold might be larger than the universe.
Elroy Dimson: If you look at Victor Haghani’s website, you see where he asks, “Where Are All the Billionaires?” He used the long-term data that we’ve been discussing to work out how many billionaires there ought to be if it weren’t for all those who are siphoning it all off.
Haghani was one of the LTCM partners who started up another firm to look after the modest amount of wealth that he still had. He’s done that in a TEDx talk as well. It’s very amusing. But the problem is that what he was modeling was somebody who never spends any of it. If people behaved like that, there would be loads and loads of billionaires, but they would be worse off than somebody who doesn’t have any money at all. They both end up having spent nothing, but the Victor Haghani clients would have spent their time also worrying about how things are going.
L. Siegel: The billionaires wouldn’t really be worse off because they would have a non-expiring option to stop being misers and live a little, but the point you’ve made is indeed very funny.
J. Siegel: Larry, I want to go back to your point that the market is actually riskier than we perceive. Raj’s original model is a model of consumption maximization under uncertainty, with risk and all the rest, and it can’t derive the premium. There are some variations of his model where you have a minimum amount of consumption, and so on. But the standard models that have been derived to try to explain the equity risk premium have already taken into account your point about the market being riskier than what we see.
Leibowitz: What’s the problem with just looking at the issue of moving from a riskless asset into a risky asset and asking the question: What level of premium does it take to achieve a sufficiently satisfactory level of success, of beating that base level over a typical relevant investment period like 5 years or 10 years?
Ibbotson: It’s not too high.
Leibowitz: When you do that, you get numbers of 4% to 6%, which is in the range of the numbers we’ve been talking about. So, that is not unreasonable in terms of how people would think about making the move from riskless to risky.
Mehra: So, Marty, let me set the stage a little bit. What’s happening is that we’re observing a premium, 6.5%. That’s an observation. Now, you try to come up with a model that is consistent with other observations in the insurance literature, other macro models, other possible estimates of risk aversion, and so forth. That model, which is consistent with other observations and with macro, generates a risk premium of only about 1% or 1.5%.
The question is: Why such a big difference between the observation and the model answer? There’s no dispute about the size of the realized premium. But how much of it is a risk premium, and how much of it is due to other factors? That is something that I wanted to bring up today in a serious way. How much of this 6.5% is a premium for bearing risk itself?
Once the existence of a premium is known — once it is in the information set — it must persist if it is a genuine risk premium because the risk continues to be there. If it’s a factor premium, it does not have to persist. All factors come into and go out of fashion. People will say, “value is working.” So, at that stage, there may be a value premium; or “size is working,” or “momentum is working,” or “accruals are working.” So, I wouldn’t say that those are risk premiums; those are factor premiums.
The question is: Is this premium that we observe for equities a risk premium? We have several theories that address the question, and some of them would say that not all of the 6% is a risk premium. They say part of it is a risk premium and the rest is a premium for other things.
If you liked this post, don’t forget to subscribe to the Enterprising Investor.
All posts are the opinion of the author. As such, they should not be construed as investment advice, nor do the opinions expressed necessarily reflect the views of CFA Institute or the author’s employer.
Image credit: ©Getty Images/Halfpoint
Professional Learning for CFA Institute Members
CFA Institute members are empowered to self-determine and self-report professional learning (PL) credits earned, including content on Enterprising Investor. Members can record credits easily using their online PL tracker.