Professor Frank Partnoy should be applauded for presciently considering many of the topics that the recent financial crisis has brought to the forefront. In one of his articles, “Why Markets Crash and What Law Can Do About It,” Partnoy implicitly assumes that there are only two plausible reasons for a market to crash: either a market can crash as the result of a rational response to new information, or a market can crash as the result of a market failure. The analysis of the October, 1987 crash in my previous posts reveals that Partnoy’s presumption about what can cause a market crash, though widely accepted, is incorrect.
Partnoy begins his analysis with a review of the Efficient Market Hypothesis (EMH), which posits that stock prices generally reflect available information. As Partnoy correctly observes, when markets crash, the price changes are often so substantial that they cannot be explained by the sudden availability of new information (or by minor deviations from the EMH). The October, 1987 crash provides a nice case in point. The stock market dropped 22% in one day with relatively little new information. Perhaps the market did fall that day for reasons that can be explained by behavioral psychology and cognitive errors, as Partnoy concludes. But there is a third possibility as well, which Partnoy does not consider.
The problematic step in Partnoy’s analysis comes when he equates the EMH with the claim that “changes in prices should be essentially random responses to the revelation of new information” (p. 748). The mistaken implicit assumption here is that prices are likely to be relatively stable in the absence of either new information or a market failure. But the use of portfolio insurance in 1987 suggests a third way in which prices can change dramatically. If the demand for equities is upward sloping in some price ranges, then two quite distinct equilibrium points might exist even without the introduction of new information. In fact, there is good reason to believe that it was a move from one viable price equilibrium (with holders of portfolio insurance fully invested) to another viable price equilibrium (with holders of portfolio insurance largely divested) that triggered the market crash on October 19, 1987. Just because share prices dropped by 22% on October 19 without dramatic new information (or rose 11% today!) does not mean that the EMH was violated.
The challenge presented here to Partnoy’s presumption that market crashes must be caused by either new information or a market failure is not simply a matter of semantics. Recognizing that markets can crash, and yet be both informationally efficient and properly functioning, is crucial to crafting an appropriate regulatory response.
In challenging one aspect of Partnoy’s argument I do not want to diminish the many excellent points Partnoy makes in his article. First, Partnoy is one of the few authors to acknowledge and directly address the significance of the October, 1987 crash (p. 747). Partnoy also does a nice job of addressing and dismissing, as irrelevant for practical purposes, the rational bubble literature (p. 750). Finally, while a market crash caused by the presence of multiple equilibriums would not be the result of a market failure, there is certainly something odd about the preferences revealed by the widespread use of portfolio insurance.
TrackBack URL for this entry:
http://www.typepad.com/services/trackback/6a00d8345157d569e20105357fbbad970b
Links to weblogs that reference One Complaint with Partnoy:
1. Posted by fedgovernor on October 14, 2008 @ 4:21 | Permalink
"The stock market dropped 22% in one day with relatively little new information."
This sentence is self-evidently false.
Once the market had dropped 10%, there was new information. The information was this: "The stock market is falling precipitiously."
You might not believe that's very useful information, but it's information nonetheless. That information causes people to act, and to act rationally to sell their stock on the theory that the market is going to decline further.
That this causes the further decline is irrelevant to the person who just sold his stock and only matters to the person who did not sell his stock.
2. Posted by on October 14, 2008 @ 13:20 | Permalink
Under the EMH, information is anything that affects prices and is not known in the present, right? When participants are using portfolio insurance, or more generally when wealth affects the desire to hold particular assets, price changes are information because they affect the demand curve for the assets.
What is the supply curve in your model? One could argue that the supply curve for stocks is fixed absent new issues, but that wouldn't give multiple equilibriums. Therefore I assume that the supply curve is the total number of shares holders would be induced to sell at a given price. Since portfolio insurers sell on the way down, you would have a potentially downward-sloping supply curve too, correct?
3. Posted by Mike Guttentag on October 14, 2008 @ 15:31 | Permalink
Responding to comment 2 above, there are, as you are probably aware, many different forms of the EMH. I think in most all of these variations the EMH is a claim about how information about the firm’s future cash flows is incorporated into current stock market prices. It is certainly possible to describe an EMH that includes information about the firm’s current stock price, but I think to do so would be unusual.
As for your second point, my assumption is a fixed supply curve (of equity securities in the short term), but I don’t see why that would prevent multiple equilibrium. I wish I were capable of blogging a graph, but in simple terms you could imagine a demand curve snaking around a fixed, vertical supply curve. Although I think the intuitions are straightforward, if you want to see a formal treatment that supports my claim look to, for example, Sanford Grossman & Zhongquan Zhou, Equilibrium Analysis of Portfolio Insurance, Journal of Finance Vol. 51, Issue 4, page 1379 (September, 1996).
| Sun | Mon | Tue | Wed | Thu | Fri | Sat |
|---|---|---|---|---|---|---|
| 1 | 2 | 3 | 4 | |||
| 5 | 6 | 7 | 8 | 9 | 10 | 11 |
| 12 | 13 | 14 | 15 | 16 | 17 | 18 |
| 19 | 20 | 21 | 22 | 23 | 24 | 25 |
| 26 | 27 | 28 | 29 |
