Wednesday, May 12, 2010

Information Cascades

John Derbyshire has an insightful post on why last week's temporary market meltdown occurred.  I've chosen to reprint it in full because he does a better job of explaining an interesting market phenomenon than I could ever do.  John is author of Prime Obsession.  You can follow John here.


Forget about the “fat fingers” and “black swans.” Last week’s sensational stock-market drop was just an unusually big instance of a normal phenomenon, the information cascade. Basically, here’s what happened: Seeing a lot of people headed over to the starboard side of the boat, a lot more people headed over to that side, too.

There are good reasons to think that this kind of imbalance will become more common. There is, for example, the current political environment.

Consider two market players, 
A and BA sells a stock to B. Why is this happening? Well,A thinks the stock is overvalued, so he wants to sell it. B, contrariwise, thinks the stock is undervalued, so he wants to buy it.

Why do 
A and B have different views, though? They’re both working from the same data. Likely they’re both professionals, with teams of analysts backing them up. (In today’s environment in fact, one of them is probably a computer program; but that’s another story, and doesn’t affect the argument.)

One common reason is that they’re working with different time horizons. 
A might be looking two weeks ahead, B may be looking two years ahead. On this basis, they might both be right: the stock may be overvalued in the short term, but undervalued in the long term.

A lot of trading is like that. Most commonly the trade will be between someone with a higher appetite for risk (short time horizon) and someone with a lower (long time horizon — probably an institutional investor). This long-term/short-term balance is one of the things that keeps the market stable.

Politics can upset this balance by introducing uncertainty into the long-term picture. The bailouts did just that. Investors — and many more long-term investors than short-term — are asking: “How do I know who’s Too Big to Fail? How do I know when there’ll be a bailout and when nature will take its course? 
What are the rules?

In an atmosphere of uncertainty like this, with risk harder to quantify, long-term players become short-term players. Trading strategies will then fail more often. If several fail at once, automated high-frequency trading programs all start doing the same thing (buying or selling) and suddenly everyone’s over on the starboard side of the boat. Information cascade.


......tipping your boat over out on the Openwater.

No comments:

Post a Comment