Are we in a bubble? Here’s who to ask

How can you tell if we are in a stock market bubble? In my post “No, we are not in a bubble yet,” I listed my four horsemen of the bubble apocalypse. Here are the first two:

First Horseman, Overvaluation: Are current prices at unreasonably high levels according to historical norms and expert opinion?

Second Horseman, Bubble beliefs: Do an unusually large number of market participants say that prices are too high, but likely to rise further?

Today, I want to tackle these two horsemen. In my view, “overvaluation” and “bubble beliefs” are two distinct concepts. Overvaluation means that measured by objective standards and expert analysis, the market is too high. Bubble beliefs are not about objective facts but subjective expectations: do market participants believe the market is overvalued but rising?

Let me give an example to distinguish between overvaluation and bubble beliefs. Japan experienced a giant stock market bubble peaking in early 1990. As of 1989, valuation measures were extremely high in Japan and many experts believed that the market was too high. For example, the working paper version of French and Poterba (1991) was written from the perspective of 1989 and released in March 1990 with the title “Are Japanese Stock Prices Too High?”¹ By the time their paper was published, in October 1991, they had to change the first word of the title so that it read “Were Japanese Stock Prices Too High?” due to the collapse of the bubble.

So the first horseman, overvaluation, was present in the Japanese stock market in 1989. What about bubble beliefs? As of September 1989, bubble beliefs were absent because Japanese investors didn’t think the market was overvalued. A Yale survey of Japanese institutional investors found that 92% thought the stock market would rise in the next year and only 29% thought the market was overvalued.² By March 1990, however, bubble beliefs had arrived: the same survey found that 90% of respondents thought the market would rise but, simultaneously, 70% of respondents thought the market was overvalued. You can tell from these numbers that at least 60% of the respondents believed the market was overvalued but nevertheless would rise; that is my definition of “bubble beliefs.” They were right about the overvaluation but wrong about the market rising; in the 12 months after March 1990, the Nikkei fell 12%, eventually falling 47% between March 1990 and July 1992, and not subsequently reaching its March 1990 level until thirty-three years later in 2023.

So, to summarize, how can you tell if we are in a bubble? First, you should ask disinterested experts whether the market is overvalued. Second, you should ask average investors (such as your friends or your dentist) whether the market is overvalued but likely to rise further.

Overvaluation

Who are the disinterested experts you should ask about overvaluation? Let me start by saying who you should certainly not ask: brokers and investment bankers.

Investment banks are in the business of underwriting securities, and brokerages earn money when you trade. Both have an interest in promoting high stock prices and high trading volume, and I view them as closer to cheerleaders than neutral analysts of stock valuations.

In contrast, academic economists are both expert and disinterested, although also fallible. Let me repeat: fallible. As I detailed previously in “The unholy trinity of bubbles: valuation, volatility, and volume,” in my wayward youth, I myself was an academic economist who thought the stock market was overvalued circa 1996, shortly before it doubled (to be fair, many economists much smarter than me had similar views at the time).  Fallible.

So one approach to deciding whether the market is overvalued is simply to survey a group of distinguished economists and ask them whether the market is overvalued. This approach is less outlandish than it sounds, since the National Bureau of Economic Research (NBER) has a Business Cycle Dating Committee, and in the U.S. a recession is defined as whatever the committee says is a recession.³

The Flagrantly Overvalued Market Committee

So, in an ideal world, we would have a Flagrantly Overvalued Market Committee (FOMC) consisting of distinguished academic economists. I have some FOMC members to nominate, but first let me be clear: I am not claiming these individuals have the infallible ability to perfectly time the market. What I do claim is that many of them have correctly pointed to markets that were overvalued. A proper study would be to formally track their predictions over time to assess their accuracy. I have not performed such as study; all I can say is that I am confident these individuals are worth listening to.

My nominees for the FOMC are Malcolm Baker (Harvard), Markus Brunnermeier (Princeton), John Campbell (Harvard), Aswath Damodaran (NYU), Jeffrey Frankel (Harvard), Kenneth French (Dartmouth), Robin Greenwood (Harvard), Stefan Nagel (Chicago), Jay Ritter (Florida), Antoinette Schoar (MIT), Robert Shiller (Yale), Jeremy Siegel (Wharton), Jeremy Stein (Harvard), Annette Vissing-Jorgensen (FRB), Jeffrey Wurgler (NYU), and Wei Xiong (Princeton).

My FOMC includes a range of perspectives, including a bearish bubble believer (Shiller), a long-time bull (Siegel), an advocate of risk-based rational asset pricing (French), a former member of the real FOMC (Stein), and a current member of the NBER Business Cycle Dating Committee (Frankel).

One thing I want in an FOMC member is views that vary over time, that is, not being a broken record⁴ or pursuing a broken clock⁵ strategy of always forecasting the same outcome.⁶ We wouldn’t want an FOMC member who says the market is never overvalued (such as an Efficient Markets zealot or a Panglossian techno-bull), nor would we want one who says the market is always overvalued (such as anyone whose nickname includes the word “Doom” such as Dr. Doom, Mr. Doom, or any member of the Doom Patrol).

There are no broken records on my proposed FOMC. For example, as of March 2021, Shiller thought the market was reasonably priced relative to bonds and did not use the word “bubble.”⁷ So Shiller is not a broken record. Similarly, Siegel is not a broken record on the bullish side, writing a spectacularly well-timed piece in March 2000 saying: “But are the high valuations of the tech stocks that drive the Nasdaq index justified? History suggests not.”⁸

Again, let me differentiate between “overvalued” and “bubble.” The individuals I’ve named may have an opinion about whether the U.S. stock market is currently overvalued (or equivalently, low expected return). But not all of them necessarily agree that bubbles exist or are a useful framework for understanding stock markets.

If you don’t like my proposed FOMC, you could instead use the economists selected by the University of Chicago for their excellent survey series. For example, in a 2011 survey, they asked 40 prominent economists whether they agreed with the statement that “Plausible expectations of future dividends, discounted using a plausible risk-adjusted interest rate, explain well the level of stock prices for recently listed internet businesses in 1999.”⁹ They were unable to find any economists who agreed with this statement (28 disagreed).

Of course, this survey asked a retrospective question that suffers from hindsight bias: who wants to defend the absurd prices now that most of the newly listed internet businesses of 1999 have gone bust? At the time, the overvaluation of tech stocks was less obvious. For example, writing in February 2000, Paul Krugman said

Mr. Shiller believes that the whole stock market … is inflated by a speculative bubble. I'm sympathetic but not entirely convinced. The social and psychological hallmarks of a bubble -- like the fact that the TV in my local greasy pizza place is now tuned to CNBC, not ESPN -- are plain to see, but so is the spectacular pace of technological progress. I'm not sure that the current value of the Nasdaq is justified, but I'm not sure that it isn't.¹⁰

Another example of expert disagreement is the stock market in 2021. As mentioned above, Shiller did not feel it was a bubble, but I disagree, based mostly not on valuation but on the other three horsemen (bubble beliefs, issuance, and inflows), and at least one other proposed FOMC member (Frankel) agreed with me.¹¹

Where are valuations today?

First, here’s a March 2024 commentary from Siegel:

I say it’s possible we will get there, but at this point we are not in a bubble.¹²

Second, Damodaran (2024) estimates that the forward-looking expected return on stocks over T-bills is 4.6% as of year-end 2023, not far from its historical average for the period 1960 to 2023. In contrast, he estimates that in 1999, the expected excess return was much lower at 2.05%. So in my terminology, I would describe him as saying the market was overvalued in 1999 but fairly valued in 2024.

So in summary, the two votes that we currently have from the FOMC say that markets are not overvalued currently.

Bubble beliefs

I define bubble beliefs as “market participants say that prices are too high but are likely to rise further.” Why might these beliefs arise?

First, these beliefs are consistent with a “greater-fool theory” speculative motive for trade; investors buy with the intention to quickly sell at a profit. A second explanation for bubble beliefs is investor psychology. Investors have a strong belief in extrapolation: if the market went up last year, they predict it will go up this year. They also have a strong belief in mean reversion: if the market is too high, it will eventually go down. These two beliefs are in harmony when the market is overvalued and falling (say, in 2001), but at war with each other when the market is overvalued and rising (say, in 1999).

Now, you might think it is crazy for investors to simultaneously believe in both extrapolation and mean reversion, but these concepts are also present in virtually all systematic investing approaches; we call them momentum and value. So there is nothing inherently contradictory in believing in reversals at long horizons and continuations at short horizons.

Thus, it is not necessarily crazy to have bubble beliefs. Indeed, consider this statement from George Soros, quoted in Pedersen (2019):

When I see a bubble forming, I rush in to buy, adding fuel to the fire. This is not irrational.

Asking about bubble beliefs is one way to detect bubbles. Presumably, Soros would have agreed in 1999 that the stock market was overvalued but rising. Another approach is to forget about the beliefs and just ask Soros directly about whether there is a bubble, which I discuss next.

Bubbles are not a secret

Bubbles are not a secret. Thus, a simple way to determine if a bubble is occurring is to just ask people if a bubble is occurring. Don’t bother assembling my proposed FOMC, just ask your bartender, your doctor, and your Uber driver.

This is not to say that bubbles are obvious to 100% of the population when the bubble is occurring. During a bubble, you are aware that something is going on, but you are not necessarily sure that is a bubble. Indeed,  hallmarks of bubbles include disagreement about current valuations and appealing narratives that justify current valuations. That is why you need to rely not on your own intuition, but rather on external measures such as expert opinion and surveys of bubble beliefs.

A bubble is a social, not solitary, event. A bubble is like a romantic relationship: it involves two or more people engaged in observable activities with initially positive feedback loops. If you’re in a bubble, you can tell something is going on, even if you are not sure what to call it. Bubbles, like romance, are events involving strong emotions, decision-making under uncertainty, and most of all, other people. You may feel joy, you may feel pain, you may be confused, but you will certainly be aware that something is happening and it’s not just happening to you. In the previously mentioned Krugman example, Krugman was not oblivious; he was just unsure.

If you are not sure whether you are currently in a romantic relationship, the best solution is to ask the other parties involved (hint: “Let’s not put a label on it” probably means “no”). The same goes for bubbles: if you are not sure, just ask.

If you are unsure whether a bubble is occurring, you could ask a random sample of individuals “is a bubble occurring?” An even better question would be “do many people think a bubble is occurring?” In February 2000, Krugman was unsure about the first question but would surely have answered “yes” to the second question.

Asking directly about bubbles has the virtue of simplicity, but has the downside that different people use the word “bubble” to mean different things. In any case, there are only occasional surveys asking this question. Here are some:

• In a retrospective study, Dhar and Goetzmann (2006) find that nearly half of the investors who bought tech stocks in 1999-2000 agreed that it had “definitely” been a bubble.
• In December 2017, a CNBC survey found 80% of institutional investors agreed that “bitcoin valuation is a bubble.”¹³
• In March 2024, a Bank of America survey found 40% of fund managers said stock markets are in an “AI bubble.”¹⁴

Sadly, these survey results have limited usefulness because they are snapshots at one moment in time, as opposed to a historical time series over many years.  What does it mean that 40% of fund managers think that stock markets are in an AI bubble? I have no idea.

For example, suppose I told you that 47% of individual investors thought that the market was overvalued. Sounds pretty scary, right? That’s even worse than the 40% of fund managers who currently say we are in a bubble. Well, it turns out (as I will explain in the next section) that on average 43% of investors think the market is overvalued, so that 47% is about average and nothing to worry about.

While we don’t have a long historical time-series of asking investors, “is there a bubble?” we do have long historical time series of asking about bubble beliefs. I turn to this topic next.

Too high but rising?

There are many surveys of institutional and individual investors in the U.S. and other countries. See for example Barsky (2009) and Shiller, Kon-Ya, and Tsutsui (1996) for Japan. As already mentioned, we saw bubble beliefs in Japan in early 1990, with investors saying the market was overpriced but rising. We saw similar results in the U.S. in 1999/2000 according to Fisher and Statman (2002), and Vissing-Jorgensen (2003).

My favorite source for bubble beliefs is the series of surveys conducted by the Yale School of Management. These surveys, originally conducted by Robert Shiller (see Shiller (2000)), are an often-overlooked resource that in my view are an invaluable market timing signal (and are freely available and updated on the web).¹⁵

Specifically, the survey collects responses for (among other questions) the following two questions from U.S. individual investors from 1989 to today:

1. One-Year Confidence Index: Percent respondents saying the Dow Jones Industrial Average will go up over the next year.
2. Valuation Confidence Index: Percent respondents saying, “Stock prices in the United States, when compared with measures of true fundamental value or sensible investment value, are too high.”

The table below shows values for selected dates as well as the average from 1998 to 2023 (after 2001 the numbers they report are trailing 6-month averages):

I invite you to examine the graphs at the website yourself, but let me describe what I see as the main points.

For the first question (will the market go up) there is a strong element of extrapolation in the responses: when the market has risen over the past year, many respondents predict the market will continue to go up next year.  For example, in May 2023, the market had recently fallen, and thus a relatively low number of respondents predicted rising prices. You could just relabel the “One-Year Confidence Index” as “trend-following” or “momentum” or “extrapolative expectations.”

For the second question (is the market too high), we see a strong mean-reverting element. The times when investors say that the market is overvalued are exactly the times when the market actually is overvalued: The two highest values in history are April 2000 and June 2021. Similarly, very few individual investors said the market was overvalued in May 2009, exactly when the market was at its lowest.

In these two questions, we see the war raging inside the heads of investors: momentum versus value, the eternal struggle. The first question reflects the madness of crowds: foolishly expecting the recent past to predict the future, like a reckless youth day-trading bitcoin. The second question reflects the wisdom of crowds: sagely expecting a return to normalcy, like Warren Buffett patiently waiting for the market to revert to fundamentals.

Putting the two questions together, we can see bubble beliefs. Near the peak of the 2000 and 2021 bubbles, more than 70% of individual investors said the market was overpriced, but 70% or more also thought the market would continue to go up.

Where are bubble beliefs today?

As of the latest date of Yale’s U.S. Stock Market Confidence Indices, December 2023, we are nowhere near a bubble. About half (47%) of the respondents think the market is overvalued, which is close to the historical average and not near the highs of 70% observed in 2000 and 2021. No bubble.

References