The Evolution of Economics

A common criticism of neoclassical economics is its assumption of human rationality, an assumption which is clearly at odds with the behavior we see around us. The usual defence is that irrational behavior may average out, as people mostly make good decisions, and that we have a theory of rational behavior but not irrational. While these are excellent defenses of rationality as a simplifying assumption, it is important to remember that the result of reasoning with such assumptions is merely an approximation, and does not perfectly reflect reality.

While vast quantities of economic analysis have been done with the rationality assumption, neither of its defenses still hold. Fields like behavioral economics have thoroughly demonstrated that humans are consistently irrational, and evolutionary psychology has given us at least the outlines of a theory about why.

Those who have seen only a few studies of human irrationality may doubt the word "thoroughly", but it is quite true. The fabulous but out-of-print book Irrationality, for example, contains chapter after chapter of different, consistent types of irrationality, backed up by studies. Our brains just aren't very good at certain types of common, simple reasoning. Nor are these errors irrelevant to our daily lives. For example, Smart Money Mistakes is a discussion of how these irrationalities lead people to make poor financial decisions. One need only contrast the evidence on index fund performance with the vast quantity of money still invested in conventional mutual funds, with their higher ongoing fees, higher charges, higher tax burdens, and lower performance, to see human overconfidence made clear.

And while the field of Evolutionary Psychology tends a bit towards "Just So" stories, fitting explanations to its data, there is still something to it as an explanation for why we do what we do. Sure, some of our strange behaviors are due to inefficient heuristics by brains not suited to explicit calculation - these are particularly common in areas of logic and probability. But in spheres like sex and romance, it seems abundantly clear that our behavior really is partly governed by strange impulses meant to benefit our genes.

These ideas are certainly not unknown to academic economics. As David Friedman writes in a piece on econ and EvPsych:

Economics is built on a simple assumption—that individual behavior can best be predicted by assuming that each individual will take the actions that best achieve his objectives. The justification for that assumption, somewhat misleadingly labeled “rationality,” is that we have no good theory of mistakes, no way of predicting what particular irrational action an individual will take. That leaves the rational element as the best—although imperfect—way of predicting behavior.[1]
Evolutionary psychology offers, among other things, a theory of mistakes—an alternative to the rationality assumption.

Given this state of events, with a science rich in work whose assumptions are rightly under question, it seems clear that economics needs to grow. If it is to survive as an empirical science, its beautiful framework of elegant mathematical proofs must be extended by the clinging vines of addendums and qualifications: "For type I and III irrationalities, the result still holds. For types I and IV, it still holds if the deviation is below 0.73. For type V, small deviations from rationality will produce small deviations in the final equilibrium." Messy though such clauses are, they are necessary if our theories are to apply to the messy world in which we reside.

So what would a paper in 21st-century economics look like? Fortunately, we need not merely speculate. Scott Beaulier and Bryan Caplan have provided us with a fine example in Behavioral Economics and Perverse Effects of the Welfare State [DOC]. The basic idea is simple: when agents are irrational, you cannot harm them by offering additional choices. Yet people have theorized that welfare harms the poor by encourage them to remain dependent, rather than investing in education or career, which would be better in the long-term. This appears to be a contradiction.

The resolution for this, of course, is to point out that people are not rational. More specifically, they tend to make decisions for short-term pleasure which they later regret. And there is some evidence that poor people have this problem even more than the rest of us (ie consider the rate of teen pregnancy and single motherhood in that population). Hence offering them present value (a handout) which leads them to sacrifice future value (the career potential from starting to work now) may lead them to make an irrational choice.

The paper describes this in more detail, explores what the idea implies for public policy, and how it might affects laissez-faire societies. While the general concept is a relatively straightforward application of behavioral economics, such work is unfortunately unusual, and this one does an excellent job of descending into the jungle of the real world and emerging with the treasure of specific conclusions in hand.

As many others have written, the economics profession is currently heavily oriented towards mathematics. One wonders if this has influenced the focus on rationality, which allows for the clean proofs that mathematicians so adore. Not that the intuitive approach is necessarily better - the arguments of libertarian popularizers like Murray Rothbard and Henry Hazlitt seem to consist to a large degree of lovely intuitive proofs of market optimality which depend utterly on assuming rationality.

I wonder if it is the computer scientists, and agent-based modelers who will inherit the future, with their ability to build models and simulation populations of agents with varying types of irrationality. Irrationality can be produced on the computer with a few lines of Java - so easily, in fact, that it even happens accidentally at times! While who knows how much tortuous multidimensional calculus such models would require for closed-form solutions.

Either way, it seems to me that the paradise of rationality is over. The shape of irrationality is becoming clear, and the smooth first order approximation, the line of rationality, must give way to more complicated and close-fitting curves. The science must evolve, or perish.

Share this