What’s wrong with the contemporary economics, Part I: Models are bad

The quantum theory of economics (QTE) seems to have provoked people (because I provoked them lol) . So I need to explain what motivated me to work on QTE, and why in my opinion contemporary microeconomic models are not working.

Well..it’s a long story…because they are wrong in so many ways!

But let me start first with other testimonials. I’m far from being alone in thinking that current economic theories are poorly created. Besides having met several natural scientists who mock economics as a pseudoscience, many economists themselves are skeptical about the predicting power of today's theories. Every year the Swedish academy of Science hands out a Nobel prize in economics, indicating that it should be as much a science as physics or medicine. Yet Noble laureates of economics are never invited together with the other Nobel prize winners to the annually televised discussions about humankind’s future, opened by the king of Sweden, …and I used to wonder why that was so. As a young business student I was rather annoyed about it.

Of course the Nobel prize in economics is not an actual Nobel prize, since it is funded by the National bank of Sweden (and not the Nobel foundation) and it is actually called Sveriges riksbanks pris i ekonomisk vetenskap till Alfred Nobels minne, but on the other hand it is decided by the same academy that makes decisions about the other prizes. So you would expect that it would have the same prestige. Yet many have opposed it, like the famous Swedish economist Gunnar Myrdal or the long time Swedish minister of finance Kjell-Olof Feldt. Even one of the Nobel prize winners, Friedrich Hayek, said in his acceptance speech that if he had been asked whether there should have been a Nobel prize in economics at all, he would have voted against it, because he didn’t think economics was a ”proper science”.

Well distinguished gentlemen...I am sorry to say, but I think you are wrong! There definitely should be an award for economics similar to those for physics or medicine. Economics is a huge part of our lives and defines, just as well as the natural sciences or medicine, how our societies and wellbeing work. I would just like to see them fund their particle colliders or neurosurgical equipment without some economic apparatus working in the society that enabled them to build their gizmos in the first place. So the study of economics is a science...or it should be at least.

But I can understand why they were critical. Both the macroeconomic and microeconomic models are so far from reality that for many they seem completely useless. Like to me for instance. I participate in business life on a daily basis and need to deal with prices, valuations and how they are being calculated in real life. I never use any microeconomic models, couldn’t use them because they model something that isn’t real or useful for me. They deal with ”rational agents” playing ”utility” games in a more or less linear world and the agents usually have no feedback between each other and with the system itself. But in real world where actual people negotiate or try to beat ”the system”, emotional as much as rational agents are interacting with each other and the environment to make individual deals.

Yet all sorts of microeconomic models are used everyday to create convincing predictions about the future valuations of all sorts of things…and not the least about stocks and other investment instruments. But they work really poorly. For example, NY Times organized the famous experiment with a chimpanzee as a stock analyst in the early 90s (I remember reading about it then and being tremendously impressed – it might even have affected my career choice). The chimp was blindfolded and made to "invest money" by choosing a portfolio of stocks by throwing darts at the Wall Street Journal’s list of stocks. Then the reporters asked several tens of financial analytic companies and financial advisors to recommend a portfolio, using the same amount of money. It turned out that the chimp had beaten 40 % of the professional analysts who were using the latest software based on the latest numerical models. I remember thinking: wow! There is something interesting going on here :-)

That was of course almost 20 years ago, but I doubt the models have gotten essentially any better. They are still faulty in the same ways. They model only hypothetical laboratory cases, which do not apply to the real world.. Of course one could suspect that the same should be true for physical models, but I don’t think so. I very much doubt that anyone builds technology with a physical model that is already known in advance to be insufficient or outright faulty! I would guess that rather than go straight ahead building quantitative numerical models and equations, a natural scientist would spend her time tuning the qualitative description of the model until she believes that it is taking all the phenomena, which she knows the system should exhibit, into account satisfactorily. Of course this might not be true every time a natural science theory is created, but as far as I have understood, that’s how it should go in principle. A natural scientist tries to understand the qualities of the system first, then build the model accordingly, and then test it. And if the empirical results don’t match, the description of the model is revised first and only after that the quantitative numerical model fixed.

But in economics it doesn’t seem work like that in practice. Economic models are typically polynomials of the form: Property (for example price) P = A + Bx + Cx2 +.. Then if the model does not work at some point, an economist just changes the values of coefficients A,B,C... or adds higher order terms (eg. Dx3) If the economist has read more mathematics, she might get fancy and use polynomials with differentials of the variable x instead…but it wouldn’t make the process methodologically any better. Adding coefficients just fix the previous erratic model by adding new terms based on experience, that is based on “knowledge” which is acquired fundamentally by induction, without necessarily understanding what is truly happening in the system. She would not really understand how the system should be actually described qualitatively and then using deduction what could happen within the boundary conditions the system has in real life.
What if after time T the system’s equilibrium changes so that the whole variable x gets thrown out of the window…there would be no way of knowing what will happen to property P, and what’s worse, even the whole system change would not be predicted by the model…and that’s exactly what happens practically all the time and everywhere current economic models are being used in practice.

Physicists and technologists usually are not caught that easily I think. That’s because their method is to try to build knowledge about their system by understanding everything that happens qualitatively in the system. Then through deduction and by understanding boundary conditions, make the model. Only after the model is made would they then make decisions about which details are essential and which details can be left out to simplify the numerical model, especially if it turns out to be too complicated to solve.

I’m not sure if I am able to convey my idea or if I am too idealistic about how physical modeling is used today in practice…but I’m convinced that the people creating technology would never approve the same inaccuracy and lack of realism in their physical models when they are designing their gadgets and gizmos, in the same way that financial analysts approve known unrealistic assumptions in their economic models. I have referred previously to Nicholas Nassim Taleb and his works….I recommend him again. He makes the same point much more eloquently than I ever could. I think Robert Lucas was also an early critic of the same fallacy of creating “knowledge” and economic models through induction (but he was talking about macroeconomics only) [1].

I also think that a "quantitative model driven" way of creating theory leads to a funny phenomena, that quantitative models create qualitative concepts! Since many of the economic models are created by first building polynomials based on inductive experience, the coefficients N of the new terms Nxn that are added in order to “improve” the model, are given attributes like “utility” or “indifference” or “intrinsic economic value”. Of course the attributes need to be credible sounding enough so that the users of economic models could justify their fallacy ;-) But more often than not, I am unable to relate them at all to real life business practices. And at the end the day it is in these practices that prices and economic valuations are formed in real life, isn't it?

So why do we do it?! How can we be so mistaken and let ourselves create fantasy (imaginary properties in microeconomics) from models rather than create models that would try to imitate reality? Taleb or Lucas were not at all the first ones to notice this error. For example philosopher/mathematician and the other writer of Principia Mathematica, A.N. Whitehead noticed:

It is very arguable that the science of political economy, as studied in its first period after the death of Adam Smith (1790) did more harm than good. It destroyed many economic fallacies, and taught how to think about the economic revolution then in progress. But it riveted on men a certain set of abstractions which were disastrous in their effect on the modern mentality. It dehumanized industry. There is only one good example of a general danger inherent in modern science. Its methodological procedure is exclusive and intolerant, and rightly so. It fixes attention on a definite group of abstractions, neglects everything else, and elicits every scrap of information and theory which is relevant to what has remained. The method is triumphant provided that the abstractions are judicious. But however triumphant, the triumph is within limits. The neglect of these limits leads to disastrous oversights… The methodology of reasoning requires the limitations involved in the abstract. Accordingly, the true rationalism must always transcend itself by recurrence to the concrete in search of inspiration. A self satisfied rationalism is in effect a form of anti-rationalism. It means an arbitrary halt at a particular set of abstractions. [2]

Well…that’s quite a mouthful =). I got this quote actually from Seneca Quandry and he kindly popularized it a bit:

We commit the ‘fallacy of misplaced concreteness’ when we treat our models as more real than the phenomena they are intended to represent. When natural resources don’t appear in economic models (sometimes they don’t, sometimes they do) economists may be tempted to ignore their significance. When economists do this, they are committing Whitehead’s fallacy. [3]

--- So that’s why I came up with QTE. It would an attempt to build abstractions to economy from phenomenology and empiria rather than fall too much in love with our computer models.
If I get around to it, I would also like to study the concept of “utility” in more detail. For some reason I like it even less than many other dubious concepts in economic theories. I think the invention of utility is often accredited to philosopher John Stuart Mills, but I think it is probably mathematician Daniel Bernoulli, who used it first in order to find a solution to the St Petersburg paradox. I think he failed at it though, and so has the whole concept of utility failed, at least in economic theory….but that’s another story...to be titled "What’s wrong with the contemporary economics, Part II" ;-)

Cheers for now
Quintessential

References:
[1] internet page March 30th 2009, http://en.wikipedia.org/wiki/Lucas_critique and the references therein
[2] Whitehead, A.N. 1929. Process and Reality. New York: Harper Brothers.
[3] Quandry, S. March 26th 2009, private communication

PS As before I publish this a way too prematurely, as an almost unreadable draft….but again I’m too eager to wait the 10 proof reads that I need to do always! So sorry for that.Q

PPS Edits done on April 2nd. Million thanks to Elia Scribe for help!Q