In this chapter of my series I hope to address a subject that arises in much of the economic literature and one mistake often made even by professional economists, and that is the belief that empirical studies can trump theory, and that they can prove theory wrong.

Now, on the face of it this sounds paradoxical, as theory exists to explain empirical evidence, but economics is slightly different than other sciences. You, in social sciences there are several differences from empirical physical sciences. First, social sciences are not strictly repeatable, unlike the physical sciences, as we are dealing with volitional beings who do not offer mechanistic responses. Second, social sciences offer a form of proof unavailable to physical sciences, introspection. You cannot think of how you, and logically all beings, would respond in order to understand electrons, but an introspective understanding of human action does allow one to develop economic and other theories through pure introspection.

I could go on at length, but I will simply refer readers to my posts  "The Limits of "Scientific" Management", "The Limits of Econometrics" and "The Limits of Technocracy". In those I discus a related topic, the failure of efforts to both "scientifically" manage the economy and to offer precise measurements, and that is what I hope to cover here.

Let us start by looking at the example that inspired this post. While reading the WSJ editorials I saw a professor who wrote that studies failed to show gains from bank mergers, and tried to explain it away. Now, this sounds reasonable, until one imagines how such studies worked. They looked at the banks before and after, picked certain measures, such as profitability, or staffing, or some other numeric measures, and compared them before and after. But that is the problem. First, any study has a fixed time limit, while in business, time horizons are actually infinite. Perhaps in 2 years there was no significant gain, but mergers had taken place with a five year visions, or ten. So the studies failed to miss the advantage. Or perhaps they did suffer reduced profits, but in merging they set themselves up in an advantageous position which would bear fruit in a decade. Unless research takes place over many decades, it is quite possible it misses many gains.

The second problem is that many such studies focus on entirely monetary measures. Now some do not, but many, especially government studies, confuse monetary measures with reality. As I discussed in "The Rubber Yardstick", these measures can be distorted by many factors which cause monetary values to deviate from real values. However, that is not the only possible mistake these can cause. For example, during the era of massive mergers and acquisitions we often heard studies about how such mergers were economically "unjustified", how little money was put into recapitalizing, or how profits remained the same or declined, and so on. What such studies fail to realize is there are other reasons to take over companies, such as preserving one's dominant market position, or preventing competition with a failing firm from reducing profits farther, or any of a million reasons, all of which might not show higher profits, but only because they simply prevented a future loss of profits. Only if we assume profits would stay the same or increase do flat profits show "no benefit".

Of course there was the other sort of study during the M&A era. That one tried the even more futile approach of arguing that M&A activity did not "help the economy", that it was carried out just to "plunder companies" and tried to show that somehow productivity was reduced by M&A. However, that is absurd. If companies were taken over and sold off, then someone bought them, so how was their productive capacity diminished? And if the resources were moved from area to another, is that not likely a sign profits were better in another area of business and hence the move improved efficiency, not reduced it? Unless they allege greed drove individuals to buy a company and then plunder it to sell for the lowest price they could get rather than the highest.

But that last shows the real problem with economic studies. Whether carried out with an ideological axe to grind (like the M&A ones) or not (like the banking studies), they still carry with them inherent assumptions. They can range from assumptions about what is valuable and what is not, whether the economy is growing, shrinking or stagnant, to  how long it should take for profits to be realized, as well as how one should measure improvements, what economic indicators are viable, and a host of other assumptions. Sometimes these assumptions are valid, sometimes they are not, but in all cases there are assumptions and thanks to those assumptions, every study carries with it some degree of approximation.

And that is why I say empirical studies cannot prove or disprove theory. Theory is based on incontrovertible truths about human nature. That humans value scarce resources, that they have hierarchical preferences, that they prefer things now to the future, that uncertainty requires a premium and so on. Those cannot be debunked by empirical measurements of inexact monetary figures, freighted with huge assumptions and approximations, taking place over a limited time horizon.

And so, though it is a mistake which even drags in professional economists, the economic fascination with econometrics is an example of bad economics.

POSTSCRIPT

Here are the earlier installments in this series:

Bad Economics Part 1 - A discussion of how prices disprove theories of resource depletion
Bad Economics Part 2 - A debunking of the many theories based on "defective" or "damaging" competition
Bad Economics Part 3 - An examination of the many absurd claims about deregulation

Please visit the blog regularly, as it appears this series is actually going to be updated regularly, unlike many of my other series ideas.