November 15, 2007
Historical Probabilities and Markets II Posted by John Steele Gordon at 10:15 AM EST I certainly agree with Fredric Smoler that the price of Confederate bonds in Amsterdam was a measure of the speculators’ perception of the probability of Confederate victory, not a measure of the actual probability. Speculators in Europe, at least ten days away from the North American battlefields, could not have had full knowledge of what was going on. And, as Mr. Smoler notes, they were not competent to evaluate the military situation in any event. Then, given the fact that wars can be lost “for want of a nail,” or because of highly improbable chance events that nonetheless occur, the speculation in Confederate bonds was sheer gambling, not reasoned investment. I can think of two other examples off the top of my head of military actions that rational people would not have predicted. I don’t think Macedonian bonds would have sold well in Amsterdam when Alexander the not-yet-Great set out to conquer the Persian Empire. And would speculators have bet on Nimitz instead of Yamamoto before the Battle of Midway, even if they were in on the secret that Nimitz was reading the Japanese naval code? Markets are notorious for over- and underestimating the probability of future events due to “the madness of crowds.” From tulipomania in the early seventeenth century to the dot-com bubble that burst in 2000, the history of markets is filled with what appears, at least in retrospect, to be remarkably irrational behavior. This irrationality is not always to the up side. It’s a commonplace that “the stock market has predicted ten of the last three recessions.” I believe that British government bonds held their value even in Britain’s darkest days in 1940. Of course, there’s a difference here. Everyone knew that Confederate bonds would be worthless in the event of a Confederate defeat, as indeed, they were. But would Britain, even defeated, have repudiated its debt? I doubt it, and, obviously, so did investors. Hitler, after all, wanted peace with Britain and would have cut a deal far short of unconditional surrender. Just by the way, British consols, which are government bonds that never mature, have been the best measure we have of the price of money over the last 250 years, as there is no yield to maturity to factor into the current price. The only other factor is the likelihood of Britain defaulting on the interest. Even in 1940, not many thought that a serious risk. While crowds are no better than individuals in predicting future events that are contingent on many factors, they are very good at one kind of prediction. If you fill a very large jar with, say, jellybeans, and invite people to guess how many beans are in the jar, the individual guesses will be all over the map. But the average of all the guesses will always be spot on. There are now markets on the Internet in which people can bet on the outcome of political races. I believe they do a better job of predicting outcomes than the polls. The reason, perhaps, is that people must put their money where their mouth is in a market, whereas they can tell a pollster anything they feel like. Also, of course, polls are supposed to be of a carefully selected sample of the total electorate, many of whom couldn’t care less. People in a political marketplace betting on outcomes are all interested in that outcome and trying to analyze what will happen. So the jellybean effect is in operation. One poll that has an excellent record of predicting outcomes in presidential races is the polls of the Weekly Reader, which polls school kids in very large numbers. It is not a “scientific poll” in any sense, but it has a very good track record of being right. The reason, perhaps, is that while grownups often tell pollsters what they think they ought to say, the kids just echo what they hear at home. From the mouths of babes . . .
|
