Start with a flat plain. How could you get a bunch of steep mountains? One way would be to pile rocks up in some areas, but not others. Another approach would be to cut deep valleys, such as those created by the famous four parallel rivers slicing through the East Tibetan plateau. (Well they’re not famous, but they ought to be.)
Pretend the DJIA rose smoothly and steadily for 100 years. Assume the level is theoretically “appropriate.” How could you get bubbles? One approach would be to have stocks occasionally rise far above their theoretically appropriate level. That’s probably how most people envision bubbles. Another would be for stocks to occasionally plunge far below fair market value. Then the “normal” looks like a bubble. I believe that if the EMH is not true, the latter is actually the most likely cause of bubbles. Risk adjusted stock returns have been way too high for the past 100 years, according to standard finance models. This implies stocks are only valued fairly at peaks like 1929 and 2000, and are otherwise greatly under-priced.
So we have have stock “bubbles” for the same reason the Tibet/Sichuan border area has lots of mountains—deep valleys have been cut in the appropriate stock price path.
TravisV recently asked me an interesting question:
At the AEI event with Avent and Pethokoukis, you suggested as an aside that there is an association between volatile NGDP and asset price bubbles. Could you please clarify that reasoning? I don’t see the connection, given that NGDP was very volatile during the 1970″²s and there weren’t any major asset price bubbles.
I see his point, but I think people tend to focus too much on the big rise in prices during a “bubble” and forget that the big fall is equally important. If prices rise and stay high, as with San Francisco or Manhattan housing, then people eventually stop thinking of it as a bubble, and start thinking that it’s “normal” that housing would be expensive in highly desirable areas like San Francisco or Manhattan.
The 1920s and the Great Moderation both saw relatively stable NGDP growth. So why the big bubbles? Because NGDP growth crashed in 1929-30 and 2008-09. In 2008-09 NGDP growth slowed by 9% relative to trend, nothing like that happened in the 1970s. It was the crash that (partly) created the bubble. Without the crashes, we wouldn’t even be talking about the great stock bubble of 1929, or the great housing bubble of 2006. I don’t hear people talking about the Australian housing bubble of 2006.
Keep in mind I am not claiming any sort of deterministic relationship between NGDP instability and bubbles. Merely that asset prices are more likely to be unstable when NGDP is highly unstable. And people see bubbles when asset prices are highly unstable. Certainly there are other factors at work during the 1970s—such as the fact that rising trend rates of inflation can depress real equity prices. This is one reason the 1970s inflation showed up in gold prices rather than stock prices.