Pablo Triana. Professor. IE Business School
27 September 2007
Using financial econometrics to predict the future of markets based on past performances may be very fashionable, but does it make sense?
In his bestseller, Fooled By Randomness, the trader and now famous author Nassim Taleb presents the following ideas: "At first, when I knew hardly anything about econometrics, I wondered why the temporal series of figures showing the activities of retired or deceased people were important when predicting the future. Econometrists never asked this question, which made me think that it was perhaps a silly question on my part... now I am convinced that most of econometrics is not worth the bother and that many financial statisticians know that".
Intriguing thoughts. And controversial. The field of "financial econometrics" appears to be stronger than ever, attracting a great deal of attention from many key academics, and one of its inventors actually received the Nobel Prize only a few years ago. Nevertheless it is difficult not to agree with Taleb. As far as the financial markets are concerned, econometric analysis is possibly less relevant than is usually assumed in theoretical circles.
Econometrics is basically an attempt to predict the future based on past events. As any student of the subject can confirm, this requires the use of extremely complex statistical and mathematical operations. But you do not have to be an expert to understand the fundamentals: explaining tomorrow based on what happened yesterday.
Past market figures are used to predict the future movements of assets using models with particularly picturesque names. However, despite the sophisticated methods and the unquestionable intelligence of their creators, believing is not easy. Simple common sense suggests that past information should not be very useful when guessing what will happened in tomorrow’s financial markets.
Why? Mainly because, as Taleb says, we would be trying to predict what the current financial players are going to do on the basis of what the past players did. On the markets, prices move for one single reason: human activity. Of course, every human being has the intrinsic capacity to take decisions independently. The financial prices of a specific period in the past are the result of the actions taken by the individuals operating in the markets at that precise time. Those prices show the consensual decisions of the players who "were there at the time" in accordance with the relevant circumstances that where applicable at the time.
Econometrists would try to use those prices to predict the prices corresponding to subsequent periods, e.g. today. The problem is that many of the individuals originally involved in generating the prices in the temporal series under analysis are either dead or no longer operate in the markets. Econometrists would therefore be borrowing inactive brains, trying to predict the decision-taking processes of a group of people with a capacity for independent thought based on the decision-taking processes of a different group of people with a capacity for independent thought that are no longer in the game. Why would Peter’s transactions on the stock market 20 years ago be important when predicting Paul’s transactions on the stock market today, especially when Peter has been retired and living in the Bahamas for the last ten years? We could admit the possibility that the data from Peter’s past activities are relevant for predicting Peter’s current activities, but it is indeed dubious to use said data to predict the actions of other human beings.
What financial econometrists are trying to do is similar to predicting the number of goals a football team will score next season based on the goals the team has scored in the past. In the same way that financial prices are the result of specific actions taken by very specific people, the goals are the direct result of the actions taken by the players on the pitch. As any football fan would say, it sounds strange trying to use relevant information about goal figures involving players who no longer play. It is simply a matter of different people and it would seem rather bold to assume that that past temporal series can tell us anything about the future, no matter how complex the techniques used.
The fact that the legendary Ian Rush managed to score 346 goals during his time at Liverpool in the 1980s and 1990s tells is absolutely nothing about the goalscoring abilities of the English team´s new star, Fernando Torres. In the same way, using past figures to predict future financial prices seems to be an exercise in bringing people back from the grave.