Thursday, February 23, 2006

Economic Prediction

From time to time, especially in the twilight afternoon hours, I put my sense of humor to bed and take up with a serious issue. Today, I am involved in a debate inside my own head, and with a couple of professionals I know, about the nature of economics as a science. Are those who require scientific methodological rigor in economics committing "scientism" or not?

[Thanks to for the image.]

I personally believe in the scientific method and in its application to economics. After all, my father was none other than Edward C. Harwood, founder of the American Institute for Economic Research, coauthor of "Useful Procedures of Inquiry" and of "A Current Appraisal of the Behavioral Sciences" (available here), and a true scientific-method philosopher if there ever was one. But at the same time, I confess to be able to identify with Mises and Hayek's argumentation about the limitations of this methodology in economics.

I also have observed that the "pure-science" econometricians, computer modelers and monetarists don't always produce irreproachable or relevant work. I would conjecture this is mostly due to their lack of understanding of the challenges inherent in "forecasting the skill of the skill forecast," which lack is afforded by our inexperience with what must be considered rudimentary tools, at least to date.

The preceding quote is from meteorologist Henk Tennekes and concerns climatology. He has written a truly fascinating piece over at Tennekes paraphrases Karl Popper as follows:

"Popper wrote, this demand [that scientists be held accountable for the accuracy of their predictions] leads to “infinite regress”: computations of forecast skill are much harder than the forecasts themselves, and the next level, forecasting the skill of the skill forecast, is insurmountable when a complex system such as the climate is involved. Popper concluded that the positivist claims of science are in general unwarranted. "

Again about climatology, Tennekes also says:

"The constraints imposed by the planetary ecosystem require continuous adjustment and permanent adaptation. Predictive skills are of secondary importance."

(The whole article is great, by the way. I recommend it highly.)

Although I may not wish to go so far as to equate economics with climatology and state that "predictive skills are of secondary importance" in economics, I will state my belief that economic prediction is limited (at the moment at least) by economics' dependence upon a social "ecosystem" (context) that also requires "continuous adjustment and permanent adaptation," due (1) partly to our lack of understanding of human behavior (although perhaps to a lesser degree than climatologists' untenable grasp on climate's "living" ecosystem, with its random causality sources like sunspots and solar radiation, etc.) Secondly, their lack of understanding is also partly due (2) to the evolution of technology, politics, international dynamics, social evolution and other contributory factors that modify human behavior unpredictably (for the moment), and partly (3) to the nonlinearity of it all, or at best to the complexity of its multiple linearities.

A perspective overview of the tendencies of market and public choice behavior has become increasingly discernible to the economic scientist, and this may be enough to allow us to understand most of what we need to know to make some pretty valid predictions. On the other hand, judging from the plethora of conflicting theory that still exists in the field today, it is also possible (but not certain) that human action, studied in its own right by another conflicted social science, psychology, may need to be a lot better understood, and thus "predictable," before we can come to grips with our subject. Human behavior, or at least certain facets of it, is still complex (although perhaps closer to our grasp than sunspots) and will most likely require more than these last 200 years of scientific inquiry to get it right.

On the other hand, perhaps we can go for another goal, which is what Tennekes calls the "vulnerability approach." This sounds much more attainable. (See Comment 26 in the article.) In any event, I will bet my rationality on the ultimate success of the scientific method to the extent that success is measured by the achievement of some measure of the capacity to predict; but in the meantime, my heart strings long for answers now and become easily bewitched by Mises's and Hayek's premise that economics has its scientific limitations by definition.


Post a Comment

Links to this post:

Create a Link

<< Home