Felix Salmon has a fascinating article on the Gaussian Copula. Obviously I had noticed that financial market participants had made some mistakes in the recent past, but I had no idea how crazy they had been.
I’d say that one implication of Salmon’s article is that the invention of the CDS was very unfortunate event. Oddly I recall him arguing the opposite position. I don’t see how he can after writing
For five years, Li’s formula, known as a Gaussian copula function, looked like an unambiguously positive breakthrough,
Then the model fell apart. Cracks started appearing early on, when financial markets began behaving in ways that users of Li’s formula hadn’t expected. The cracks became full-fledged canyons in 2008—when ruptures in the financial system’s foundation swallowed up trillions of dollars and put the survival of the global banking system in serious peril.
When the price of a credit default swap goes up, that indicates that default risk has risen. Li’s breakthrough was that instead of waiting to assemble enough historical data about actual defaults, which are rare in the real world, he used historical prices from the CDS market.
Foolish reliance on Li’s model lead to disaster and it was made possible by CDS markets which convinced participants that they had many observations on the probability of default. They were convinced that prices revealed these probabilities because they had an insane mystical faith in the strong form efficient markets hypothesis and a schizophrenic simultaneous belief that they could beat the market.
If the only problem with CDSs is that they supplied those nut cases with the prices which they missused in calculations, that would, I think, be enough to show that the world would have been better off if CDSs had been banned.
I rant more after the jump.
here and there on the web, I typed that I think that high trading volume leads to volatility, that liquid markets are a bad thing and, in particular, that the fact that the market for CDSs has higher trading volume than the market for the underlying bonds is a very bad thing. I forget where someone asks me if I reject the “price revelation” argument in general. I do.
Financial market participants managed to convince themselves that they could beat efficient markets (already a logical contradition) because the market price of a CDS was an accurate enough measure of the probability of defaults that correlations in the changes in the prices of 2 CDSs could be used to estimate the (assumed to be constant) correlation between two (assumed to be normal) latent variables making it possible to estimate the probability that two bonds would both default.
That is they assumed that market prices contained information which no one even claimed to be able to obtain any other way. So how did the information get into the prices. One hypothesis is that the Zeitgeist exists and has rational expectations. I can’t think of another explanation.
Now given what market participants did with a whole lot of prices which changed very often, I think it is a bad thing to give them a whole lot of prices which change often. That is one reason why it would have been better if the CDS had never been invented.
My comment over at “Wired”
I recall a aol instant messenger debate we had about CDSs. You position was that they were useful since the CDS market is more liquid than the underlying bond markets. It seems to me that this article is demolishes your position. Liquid markets generate a huge number of numbers. They are used by traders to price assets and control risk. According to this article such use has recently destroyed the financial system. One of the many many problems is that Li assumed that the CDS market was efficient. This means that without a CDS market, people recognised that they couldn’t calculate the probability of default exactly, but once there were CDS they thought the market price contained that information. Where did it come from ? Why the magic of the strong form efficient markets hypothesis.
If no one can learn something, then it can’t be reflected in the market price. Not a subtle point. To believe that CDS prices are probabilities one has to abandone methodological individualism, that is believe in the rationality of the Zeitgeist or something. I can believe that I don’t know something, but other people do and they trade on markets so the market price will tell me something about their secret information. I can’t believe that no one knows something but it is still equal to the price at which markets populated by the ignorant people trade.
Traders were not only mystical, they were schizophrenic. They assume that markets are efficient and then try to beat the market. If the market is efficient, the best trading strategy is to buy and hold the market. If it is inefficient, then CDS prices are not probabilities. The trading strategies which brought down the system are only rational if some prices but not others are known to be rational. How can anyone have claimed with a straight face to believe such a thing.
The assumption that correlations are constant is another quite separate gross error as is the assumption that the distribution of something is normal, because it would be nice if it was. Both mean that the number “99%” which you type while noting that it is not 100% was not the true probability.
In any case, liquid markets were very damaging, because without them, people could not have made the same mistakes. Price revelation had huge negative social value because beliefs based on the prices were further from the truth than those people would have had without the prices. Traders might have been equally wrong about the expected value, but, if they had known they were ignorant, their subjective probabilities would have been close to true probabilities and we would all be much richer.