David Levine on Behavioral Economics: The Good, the Bad and the Middle Ground

Cite this Article
David K. Levine, David Levine on Behavioral Economics: The Good, the Bad and the Middle Ground, Truth on the Market (December 06, 2010), https://truthonthemarket.com/2010/12/06/david-levine-on-behavioral-economics-the-good-the-bad-and-the-middle-ground/

This article is a part of the Free to Choose Symposium symposium.

Behavioral economics: love it or hate it – there seems to be no middle ground. Lovers take the obvious fact people are not frictionless maximizing machines together with the false premise that economists assume that they are to conclude that all of economics must be wrong. The haters take the equally obvious fact that laboratories are not the real world to dismiss all laboratory evidence that conflicts with their pet theories as irrelevant. In the end they seem primarily to talk past each other.

To do economics – behavioral economics or any other kind of economics – it is wise to start with correct knowledge of the facts. Here are some facts that are not as well known as they should be:

  • Modern economics is primarily concerned with the difficulties in making decisions in uncertain and changing circumstances. Learning economics from undergraduate textbooks is dangerous in this respect: most undergraduate textbooks are decades out of date, while modern game-theoretic/information-theoretic/mechanism-design methods have been the focus of economic research both theoretical and applied during those missing decades.
  • The most widely used economic theories do an excellent job of predicting behavior in the laboratory. In competitive auctions, in strategic voter participation and in repeated prisoners’ dilemma games the outcome computed directly from the theory matches with a high degree of accuracy the behavior observed in the laboratory.
  • The issue of the relevance of laboratory studies to non-laboratory behavior has been extensively examined: more money and bigger stakes induce people to behave more carefully, but not in a qualitatively different way than they behave in the laboratory.
  • Modern economic theory and laboratory methods have been successfully combined – that this is a success will probably surprise both the pro- and anti-behavioral economics camps – in the (highly lucrative) field of applied mechanism design.

It must be emphasized that not only does modern economic theory work well in the laboratory, it works well in practice as well. Popular opinion holds that since economists failed to predict the crisis there must be something wrong with economics, to which, no doubt, the solution is behavioral economics. But what modern economic theory tells us about crises is that we cannot predict them. Suppose that we could tell you reliably “next week the stock market will crash.” If this was a reliable prediction then everyone would believe it and so the stock market would not crash next week – it would crash now. Crashes by their nature must be unpredictable. It is odd that nobody criticizes physicists for the lack of predictive power of their models – yet the fact is that it is as central to modern physics that it is impossible to predict which slit a photon will pass through as it is for us to predict when a crises will occur. Perhaps though, the impossibility of predicting crises means that economics is useless? However, we can accurately predict the consequences, for example, of bank bailouts – and we did: you are now living with the consequences of the fact that you ignored us.To have any hope of understanding either what economics does or how it can be improved, it is necessary to understand the theory properly. As always, the devil is in the details. For example, here is a cautionary example for economists who reject behavioral theory as irrelevant: The fact that the economic theory of sophisticated agents works well at the macroeconomic level should not lead us to conclude that people are in fact sophisticated (although of course there is plentiful other evidence that they are). As Becker pointed out years ago even simple naïve machines will wind up at the equilibrium of a simple competitive auction if they are budget constrained.

Any practical application of a theory necessarily ignores many details – for example, wind resistance is typically not studied when designing single story buildings, although of course it is essential when designing large skyscrapers. The same is true in economics, and many of the forces discussed in behavioral economics are quantitatively quite small. Take, for example, altruism – or the topic now sometimes called in the behavioral literature “social preferences” – meaning that people are not completely selfish. Of course economists have been studying “social preferences” for decades before behavioral economists argued that their presence somehow invalidates existing theory. No matter: there is grist for all mills. On the one hand people surely do make charitable contributions – so we can hardly take as a justification for models of selfish behavior that “this is what people are like.” On the other hand people do not give a very large proportion of their income to charity – so it is right also to argue that this is not a very significant force. Here is where a proper understanding of theory becomes crucial – and where the power of equilibrium theory – the central notion in economics and one completely missing in other social science disciplines – becomes central.

The key fact is that a small change in incentives for many people, or a large change in incentives for a few may – or equally may well not – have a large impact on the equilibrium behavior of the entire group. Equally true is that economic theory can tell us which is which. In competitive auctions and in strategic models of voter participation small changes have little impact – is it then surprising that this is where the theory performs well in the laboratory? In other settings is makes an enormous difference – the classical example being in repeated games. Consider strictly selfish players who play to a known and fixed endpoint. They have no reason to behave altruistically when the game will end in one period, and simple reasoning shows us that once they realize this they will have no reason to behave altruistically when the game will end in two periods…three periods…and so forth. In games such as the prisoners’ dilemma where there is a conflict between private and social interests, this results in the worst of all possible worlds. On the other hand – in a game played many periods people do not need to be that altruistic to behave cooperatively in just one period. And this small change in preferences is enough to radically change the strategic nature of the game – and gives rise to equilibria – observed all around us – in which people motivated primarily by selfish motives in fact behave altruistically.

Given that in some circumstances small differences in individual behavior makes a large difference in the aggregate – it well behooves us to understand what those small deviations from “standard” economic theory are. And if behavioral economics cannot revolutionize a subject that is already very successful, it certainly offers the promise of refining our theories in the areas where our predictive power is currently weak.