A Peculiar Science

As economists are routinely (and sometimes casually) ridiculed in this space, a recent article in the Financial Times by science writer Philip Ball was found to be of particular interest. Titled Baroque fantasies of a peculiar science, the commentary covers many of the criticisms those in the hard sciences often times lob toward the practitioners of the dismal science.

What was even more interesting than the commentary by Philip Ball was the response at Mark Thoma's blog, Economist's View.

First, Philip Ball's Baroque fantasies:

It is easy to mock economic theory. Any fool can see that the world of neoclassical economics, which dominates the academic field today, is a gross caricature in which every trader or company acts in the same self-interested way – rational, cool, omniscient. The theory has not foreseen a single stock market crash and has evidently failed to make the world any fairer or more pleasant.

The usual defence is that you have to start somewhere. But mainstream economists no longer consider their core theory to be a “start”. The tenets are so firmly embedded that economists who think it is time to move beyond them are cold-shouldered. It is a rigid dogma. To challenge these ideas is to invite blank stares of incomprehension – you might as well be telling a physicist that gravity does not exist.

That is disturbing because these things matter. Neoclassical idiocies persuaded many economists that market forces would create a robust post-Soviet economy in Russia (corrupt gangster economies do not exist in neoclassical theory).
...
Economics arose in the 18th century in a climate of Newtonian mechanistic science, with its belief in forces in balance. And the foundations of neoclassical theory were laid when scientists were exploring the notion of thermodynamic equilibrium. Economics borrowed wrong ideas from physics, and is now reluctant to give them up.

Recall that Isaac Newton, the famed 18th century physicist who first described universal gravitation and who later moved Pound Sterling to the gold standard as Master of the Mint in London, also lost a fortune in the South Sea Bubble of 1820 1720.

In an odd initial connection between physics and economics, the man regarded by many as the greatest figure in the history of science once commented, "I can calculate the movement of the stars, but NOT the madness of men."

The criticisms by Philip Ball are many.

One can go through economic theory systematically demolishing all the cherished principles that students learn: the Phillips curve relating unemployment and inflation, the efficient market hypothesis, even the classic X-shaped intersections of supply and demand curves. Paul Ormerod, author of The Death of Economics, argues that one of the most limiting assumptions of neoclassical theory is that agent behaviour is fixed: people in markets pursue a single goal regardless of what others do. The only way one person can influence another’s choices is via the indirect effect of trading on prices. Yet it is abundantly clear that herding – irrational, copycat buying and selling – provokes market fluctuations.

There are ways of dealing with the variety and irrationality of real agents in economic theory. But not in mainstream economics journals, because the models defy neoclassical assumptions.

There is no other “science” in such a peculiar state. A demonstrably false conceptual core is sustained by inertia alone. This core, “the Citadel”, remains impregnable while its adherents fashion an increasingly baroque fantasy. As Alan Kirman, a progressive economist, said: “No amount of attention to the walls will prevent the Citadel from being empty.”

Needless to say, these views did not sit particularly well with the author or readers of one of the more popular economics blogs. Here's a sample of the 70+ comments that the posting of the Ball commentary inspired.

Mark Thoma: I don't mind the criticism, it's good for us and there are truths in what is said. But I always resent the arrogance of scientists from other fields thinking they can mosey on over to economics for a few minutes, diagnose our ills, and solve all our problems. I'd suggest they solve the problems in their own discipline first, or show a bit more humility when giving advice to others, especially when, as above, they are clearly unaware of vast swaths of literature such as the published work on corruption. And along those lines, and for the record, we're well aware of and have models for the list of things he mentions in his "abundantly clear" theory of market fluctuations. If he actually tried to build these models rather than simply casting aspersions at the existing paradigm, he'd find it isn't as clear as he thinks.

realpc: Mark I absolutely agree with you! Physicists in glass houses should not throw stones. What absurd lunacy -- expecting economics, or any science, to be perfect.

Phycisists only have to make sense of lifeless matter, and they have either failed or stalled at their endeavor. Economists, and other social scientists, have to make sense of things far more complicated.

There is nothing wrong with people straying outside of their fields, that is not the problem. But what arrogance this guy has.

His motivation is obviously entirely political -- promoting anti-capitalism. Notice that he doesn't have any better ideas, which is typical of the far left.

Ian Matiland: On the subject of hard scientists "moseying over," I recall from somewhere that Max Planck began his academic or professional career as an economist but switched to physics because (I paraphrase) "economics was too difficult". Great discussion.

T.R. Elliot: I'm certainly not qualified to defend economics, but much of what we consider to be neoclassical economics seems to be founded on certain axioms regarding the behavior of humans--rationality, independence (non-herding behavior), etc--and then much can be derived from this. As such, it seems a reasonable approach as a simplification, as long as it is understood that these axioms are not true, at least not to the degree that we can trust what pops out of the analysis when the crank is turned.

There is good evidence that humans aren't rational, that humans do herd, that prices are not rational nor do they move in ways economic theory predicts.

My primary heartburn, unlike a previous commentor, is not with those on the left. My problems is with those on the libertarian or conservative side of the spectrum who seem to use economics as a cudgel to silence debate on what are often as much moral issues as they are economic issues--i.e. tax policy, trade, incentives, and the relationship of economics and humans in general with the ecosystem.

That said, economics provides a necessary language in which to discuss these issues.

All the same, I'm not opposed to the errant intellectual, from within or without, jumping into the fray, creating a little healthy debate, and perhaps knocking a few idols off there pedestals. It's not clear to me that Philip Ball has done as such. He seems to be ranting in a way more appropriate for a blogger.

jac: Evagrius and T.R. Elliot make excellent points. Although the last thing I want to do is to resurrect the "science debates" of the 1990s, it does seem to me that the problem is the elevated epistemological status of science, and the aspirations of neo-classical economics to these lofty heights. While I think it is true that in describing human behavior, rather than, say, the behavior of an electron, economists have a task that is far more difficult to describe "objectively," ideally we can take the insights of economists with the same acknowledgement of their limitations as we do those of physicists. Neither will ever have the whole truth, if there is such a thing. A little humility is in order for all scientists, whether natural or social.

joan: It seems to me that Neoclassical Theory is analogous to Newton's theory of gravity. But physics went on from there, where economists seem to never resolve any disputes, but just have the same arguments over and over. Part of the reason is that economics is harder, in the sense that experimental evidence is more difficult to obtain. But many economist have agendas and so fill the literature with contrived studies to produce results that support their positions. Such studies are praised or condemed based on the agenda of the reviewer, not on the validity of the study. There seems to be no mechinism in economics that rewards truth seeking and punishes agenda promoting. This makes finding and agreeing on anything beyond basic theory very difficult and real progress almost impposible. An example of the difference in approach is the debate on global warming. After 20 years fo research and debate, physical scientist are reaching a consensus on the problem, but economist who oppose action are claiming that the scientist are biased, and cling to their original positions, even if they must ignor the evidence. When confronted with evidence physicists change their minds, many economists do not.

Sean Matthews: Note though, that if you ask a physicist why there is gravity, he (or she) *will* be able to tell you.

Cyrille: Yes, even the Nash equilibrium is over-hyped. In fact, my brother's thesis was based on demonstrating an example of a game that spirals away from the Nash equilibrium. In defense of economics though, one must admit that, while physics is an incredibly difficult subject, it does not have to deal with particles thinking "mmm, I'm expected to keep my movement in the absence of forces, then let's disintegrate into neutrons and alpha rays instead to beat the market and make a profit". If you make accurate prediction in physics, they stay accurate. If you are known as extremely accurate and make public prediction in economics, then they should automatically be negated by market anticipations...

Bruce Webb: I was having a little discussion right here with an apparently trained economist who insisted that Walmart's policy of grinding down prices from their suppliers could not possibly harm workers. Which seemed odd to me, clearly the money saved would have to come from somewhere whether that was the employee holiday party fund or the boss's annual bonus. He airily explained himself by saying that real economists simply assumed that companies worked at some kind of efficiency frontier (you people will probably know the term.) Well no, in a typical company there is one or another measure of slack, and they even have specialists who will come in to "turn the company around". Oddly enough they are called "turn-around specialists".

In my experience over the last few years I have encountered any number of economists more than willing to deploy theoretical constructs as the equivalent of empirically tested concepts. Whether it is the setting of CEO pay or the usefulness of setting policy for next year based on an economic projection over "the Infinite Future Horizon", some people are letting their Pareto/Ricardo/George overrun their common sense and that thing some of us call "reality".

I like y'all and everything. But methinks you protest Mr. Elliott too much.

steve: Your comment about string theory is misinformed and a cheap shot. The well-established framework of quantum field theory + classical gravity is capable of explaining *all* current experimental data in physics. If there is a malaise in physics it is because our fundamental theories are *too* good. String theory is meant to describe quantum gravity, whose effects are negligible except in the first 10{-43} s of the big bang, or in the final 10^{-43} s of evaporation of a black hole. That's it. All the noise over string theory is in one tiny corner of physics (although a lot of the smartest people are working in that corner), and the controversy exists specifically because we don't have the *experimental capability* yet for testing theoretical predictions in that regime. This doesn't affect tour de force applications of physical theory, like building quantum computers or predicting with fine accuracy the number of neutrinos produced in the sun.

Looked at from a broader perspective, physical theories are remarkably good, and cannot be compared in terms of intellectual depth or rigorous testing to any other theoretical frameworks in science. The level to which, e.g., evolution (let alone trade or growth theory) has been tested is a joke compared to quantum electrodynamics or special relativity.

steve: BTW, let me say that I'm not negative about economics at all. The problems are, due to their complexity, much harder in a sense than the ones in physics. In physics, very high value is placed on knowing where, in a messy system, it is possible to extract clean or robust results. Because we have some very successful examples, it is easy to judge when you might be wasting your time. This is much harder for economists, as successful examples are much more rare.

It seems to me that economists sometimes delude themselves about the value of their models. The slavish worship of mathematics often lends only a spurious air of rigor or technicality to the subject. When a physicist looks at the unrealistic assumptions made in the formulation of the model, it is often obvious that the subsequent highly technical analysis is a waste of time.

anne: No; I do not find an argument with Steve Hsu, and I did not think Philip Ball was compelling or even understanding in the complaints about economics. My complaints are wholly separate. Steve's complaint, if I understand properly, was over the quality of descriptive or predictive results that are even possible in economics no matter the technical application. What is possible in physics, even if in probability terms is immensely more definite than in economics; and, economics is in this respect astonishingly complex.

Mark Thoma: One final comment.

I think we (economists and non-economists) may be speaking past each other a bit here.

Here's how I use terms:
Classical model - model with labor theory of value
Neoclassical model - model with a utility theory of value

The neoclassical model is used to describe a macro "model" that was in vogue prior to Keynes, but it is not the terminology we use today to describe the strains of macro models (RBC, New Keynesian, older is New Classical to specifically distinguish it from the term "neoclassical" and so on).

When I hear that the neoclassical model doesn't work, I assume they mean they are rejecting the utility maximization framework used to derive demand curves, etc.

Saying that macro models do not work has little to do with the rejection of this framework (so why even mention the stock market or actual macro variables of interest unless you think it's a macro model).

Now, I understand the psychologists have doubts about our rationality assumptions, and I've tried to keep up with neuroeconomics developments, but saying that the utility maximization framework (and it's direct relative, profit maximization) is wrong is a harder case to make than pointing to some macro outcome our models missed.

I don't believe the case against the utility maximization framework has been made (and hence, macro people are trying to derive their models from "first principles" to see if that will help), and there is certainly nothing better to replace the existing paradigm. I find a lot of the criticism about economics muddled on this point, i.e. using the failure of macro models to predict this or that to reject the utility maximization framework.

What to make of all this?

Would Alan Greenspan have been better at identifying bubbles if he had discarded an economic model or two and spent an afternoon in the office of a sub-prime lender in 2004 or with a day-trader in the late 1990s? Probably.

Could contemporary economics be made into a harder science if the participants were removed and replaced with actors who weren't increasingly conditioned to react to one financial bubble after another? Definitely.

Indulging a voyeuristic bent by reading all 70+ comments confirms the soundness of the decision to switch from a business major back to engineering after that first freshman microeconomics class over twenty years ago.

Stopping after only four years seems a wise decision as well.