Nearly a year ago (7/12/08) I posted here on "
Falling from the Period of Financial Distress into Panic and Crash". In that I noted my own work on this concept, which Charles Kindleberger claimed in his Manias, Panics, and Crashes has been the most common pattern of speculative bubbles and crashes (37 out of 47 bubbles listed in Appendix B of his 2000 4th edition). What is involved is for there to be a gradual decline in prices initially after the peak of a bubble, with the crash coming sometime later. The paper I cited then on this by Mauro Gallegati, Antonio Palestrini, and me ("The period of financial distress in speculative markets: interacting heterogeneous agents and financial constraints," available on my website), has now been accepted for publication and is forthcoming in Macroeconomic Dynamics.
The three patterns that Kindleberger, drawing on the work of Hyman Minsky, argued we have generally seen are ones with such a period of distress as described above, ones that go up to a peak and then crash hard (which are what most theoretical models of crashes predict), and ones that go up to a peak and then decline gradually without a crash, but usually a bit faster than they went up. In the last few years I would argue we have seen all three patterns. The peak-followed-by-crash pattern looks like the oil market last year, which hit $147 per barrel last July only to fall hard to $32 per barrel by November. The more or less symmetric up-then-down-without a crash pattern looks like the US housing market, which, according to the Case-Shiller index, began rising in 1998, peaked in mid-2006, and has been going down since about the way it went up, with quite a ways to go.
Last year I had it in my mind that the global financial derivatives market smelled like a period of financial distress pattern, and now I think that it was indeed. The peak was in August 2007, when the problems in those markets first began to appear. The crash was the dramatic "Minsky moment" in mid-September 2008.
Which brings me to a fourth pattern that is somewhat mysterious, a variation on the pattern that does not have a crash. Whereas most such bubbles go down more rapidly than they went up, and some go down at about the same rate, there is one that has gone down at a much slower rate, indeed may still be in its decline. I am referring to housing in Japan. A graph of the pattern up to 2005 can be seen at http://en.wikipedia.org/wiki/File:EconomistHomePrices20050615.jpg. Around 2006 there was a brief cessation of the decline, but it has since resumed. In any case, that figure shows that the index rose in three years from 150 to its peak around 200, but took ten years thereafter to get back to 150, and it took 15 years from the peak in 1991 to get back down to the level it was in 1986, a clear asymmetry in the direction of going up much faster than it has gone down.
Now, according to private communication from Kindlebeger to me, this is the only major bubble in world history to exhibit such a pattern, and why it has done so remains a mystery. I saw a paper (still unpublished) some years ago that argued that it was Japanese banks manipulating the real estate market to keep the value of their most important collateral from declining too rapidly in the face of broader financial pressure that have been behind this pattern, but I have not seen that confirmed. That would suggest that the pattern has had deep implications for the broader Japanese economy. In any case, this curious decline in Japan remains a mystery (and some say that US housing prices could go down for a much longer time than many think, pointing to this strange case), but it may ultimately have to do with the Japanese wishing to preserve their broader economic system in the face of pressures to more deeply transform it.