Wednesday, February 19, 2014

Did Shiller Predict the Burst of the Tech Bubble?

by Charlie Clarke

I've noticed that it's pretty much unquestioned that Shiller predicted the tech bubble.  But I don't think it's correct, at least not in the way most people mean when they say it.  Predicting the bursting of a bubble means (to most people) predicting prices were high and then fell or collapsed even.  What I believe the record shows and what I think used to be well known was that Shiller was too early on the tech bubble.

Here is the earliest article I could find with Shiller and the tech bubble.  The date is November 25th, 1996. David Wessel writes, "Yale University economist Robert Shiller says his measures put the ratio of stock prices to past earnings at historic highs. "When prices get high relative to fundamentals," he says, "either the fundamentals go up or the price goes down." At best, he expects the market to be stagnant for the next decade."  

Shiller would speak about his bubble prediction to Greenspan the next month who then coined the term irrational exuberance.

So how did it go:

Shiller gave a ten year horizon over, which he thought at best stocks would flat.  Instead stocks were up 90%.  That is 6.6 percent per year compounded return.  He seems to have missed the bubble in its entirety.  

But the bubble was really in the tech sector, so maybe looking at the NASDAQ will make Shiller look much better.  

No, it actually looks a little worse.  The NASDAQ was up 95% over the period and if a bubble occurred it is much later than Shiller stared predicting it.

Did Shiller do anything special?  Did he just shout bubble until stocks finally went down and then claim credit?  Maybe there is a more charitable interpretation, but I don't think it has anything to do with what people mean when they say, "he predicted the bubble."  The whole dispute between EMH proponents and Behavioral proponents is if someone can reliably say "we are in the midst of a bubble" and for prices to be lower than that date.

Let's look back at Shiller's quote, "When prices get high relative to fundamentals," he says, "either the fundamentals go up or the price goes down."  Actually, the statement is wrong.  The correct statement is, "When prices get high relative to fundamentals, either the fundamentals go up or the prices will rise more slowly."  That's a statement the whole finance profession can get behind.  When prices are high, expected returns tend to be lower.  That's exactly what happened, stocks grew at 6.6% instead of their long-term average of 9.55%.  That's not to say that over ten years that will always happen.  There is a lot of volatility in stocks, and they can have low returns for a long time.  But this episode happened to come out pretty close to what the standard efficient markets type analysis would have given.

These anecdotes are so powerful that it's important to get them right. 


  1. Are there any examples of a real honest-to-god bubble prediction that holds up? I mean, even if EMH is right surely someone would have once luckily bet on the price of widgets falling right before they fell.

    I ask because it seems like almost every famous bubble predictor, once people really dig into his record, didn't. At least, not with the sort of specificity that would impress. Is Taleb legit?

  2. Taleb is not a good example, because he doesn't really offer up predictions. His big idea is just that the market makes large moves more than you think. In that form it isn't really a testable hypothesis, and even transforming it to something testable, it won't be a moment in time prediction.

    John Paulson, star of Michael Lewis's, The Big Short definitely called the housing collapse. He made boatloads of money. Given any historical boom or crash, it's pretty easy to find a trader that called it. The problem is that traders are always betting their beliefs, and they are often wrong. After Paulson's big bet against housing, he made several bad bets against Treasuries. Thus, it's harder for the media to find a trader that doesn't let them down pretty quickly. Even someone with a great trading record will make a lousy oracle.

    I think Shiller did better on housing. I plan to look into that next week. I guess I would say, it's pretty easy to find someone with one good bubble prediction, but surprisingly hard to find someone with two. Then at some point we switch to the overall record and anoint the Warren Buffets, Peter Lynches, and George Soroses of the world.