I believe that Robert Shiller is gradually losing confidence in his own research. He still has confidence in the findings. The findings are rock-solid and there is no reason to doubt them. But what really matters are the how-to implications of the research — what should investors be doing differently because of Shiller’s finding that valuations affect long-term returns? Shiller has made several statements in recent months indicating that he is pulling back from an expansive understanding of the significance of his work.
Q4 2020 hedge fund letters, conferences and more
Timing The Market With The CAPE metric
The high-water mark for Shiller’s belief that his work “revolutionizes” (this word appears in the subtitle of Shiller’s book) the investment advice field was in July 1996. At that time, Shiller published a paper predicting (based on what the historical data teaches us about how the stock market works) that investors who stuck with their high stock allocations despite the sky-high valuation levels of the day would come to regret it within 10 years. That was the clearest statement that Shiller has made that the CAPE metric can be used to time the market effectively.
That prediction did not prove out, of course. We did see a horrifying crash in late 2008. And the paper in which Shiller made the prediction contained the proper caveats that the historical data cannot be employed to make precise predictions. So I do not think it is fair to say that Shiller was absolutely wrong in his warning. My personal belief is that he did more good than harm in pointing us to the dangers of the high stock valuations that eventually led to the 2008 economic crisis.
But still….
Most investors who heeded Shiller’s warning and lowered their stock allocations in 1996 are kicking themselves today for doing so. We are 25 years down the line and today’s CAPE value is much higher than the one that applied in 1996. CAPE values have been remarkably high for that entire stretch of time, with the exception of a few months immediately following the 2008 crash. All of us who believe that valuations matter have been feeling like the boy who cried “wolf!” We keep seeing a price crash only a bit over the horizon but it never actually arrives.
My sense is that Shiller has been feeling a desire to hedge his bets a bit.
A few years back he offered a brief statement in an interview indicating that he once believed that the CAPE metric could be used for market timing but that he no longer believed that that was the case. I found that statement shocking. If irrational exuberance is a real thing, then the riskiness of stocks is not a constant but a variable. Investors who want to keep their risk profile constant over time MUST engage in market timing to do so. To my mind, market timing is the entire point of Shiller’s work. So it was disconcerting to hear him cast doubt on whether it can be used for that purpose.
Shiller’s Anti-Timing Statement Was Unclear
However, the full reality is that Shiller’s anti-timing statement was exceedingly brief and unclear. He did not offer any analytical reason why he had lost confidence in the idea. My guess is that he was just frustrated that prices had remained so high for so long and embarrassed that his 1996 prediction had failed in such a spectacular manner. My belief until recently was that Shiller still believes that market timing is a good idea when stock prices are at truly crazy levels.
Recent comments are causing me to question whether this remains the case. Shiller recently published an article arguing that today’s stock prices are not as crazy as them seem when you consider how low interest rates have gone. All of the points made in the article were legitimate. There was nothing flat-out wrong about it. But today seems to me to be an odd time to be offering reassuring words about stock prices. The CAPE level that applies today is higher than the CAPE level that applied just prior to the onset of the 1929 crash and the Great Depression that followed from it.
My view is that stocks would be more dangerous today than they were in 1996 even if the CAPE level were lower, which it is most decidedly not. We can never say with precision when a price crash will arrive. That’s a point made in Shiller’s article that is of course a solid statement. But Shiller’s theory -- that it is shifts in investor emotion rather than economic developments that cause stock price changes -- suggests that, the longer a bull market continues, the more danger it is in of crashing. It is investors’ loss of belief that irrational exuberance can continue indefinitely that causes the pretend money to disappear. Keeping confidence up obviously becomes more difficult as more and more years pass and more and more is riding on the bull market prices being real. So I see this as a particularly dangerous time.
Keeping Risk Profile Constant
For me, the key to making sense of things is getting away from the practice of trying to identify bubbles and when their bursting points and thinking of market timing as a means of keeping one’s risk profile constant over time. The investor who lowers his stock allocation with the aim of keeping his risk profile constant wins the bet on the day he makes the change because his risk is reduced regardless of what happens to prices.
Stocks are risky today. Investors should be going with lower stock allocations than what they would go with if the CAPE level were near its mean value of 16. Shiller did not dismiss the idea that stocks are risky today in his article. But he downplayed it. He offered reassurances that things are not as bad as the CAPE value considered by itself signifies. Could it be that having a “perma bear” like Shiler give up the fight is a sign that the bull has gone as far as it can go?
Rob’s bio is here.
