Friday, July 3, 2015

Behavioural Economics is Rational After All

There are some deep and interesting issues involved in the debate over behavioural economics. Greg Hill posted a comment on my previous blog where he says:
Now, you really have to ask yourself whether the kind of rationality involved in [Thaler's idea of a nudge], where a minimal change in cost results in a significant change of behavior, is same kind of rationality Lucas and Sargent have in mind.
That led me to explain my views a little further.

The most interesting issue [with behavioural economics] is whether we should continue to accept the neoclassical assumption that preferences are fixed. Let's go with that assumption for a moment.

If preferences are fixed, then we face a second question. What form do they take? For a long time, macroeconomists assumed that people maximize the discounted present value of a time and state separable von Neumann Morgenstern expected utility function. The narrow version of behavioural economics asserts that this assumption is wrong; but people are still utility maximizers.

Finance economists seeking to reconcile macroeconomics with finance theory have already taken up that challenge (see the survey here on Exotic Preferences in Macroeconomics). The dominant view in finance theory is that people maximize the present discounted value of a subclass of preferences first formalized by Epstein and Zin. These preferences drop one of the key assumptions of Von-Neumann Morgenstern; that the date at which information is revealed is irrelevant. There are also more radical possibilities. The original version of the Epstein Zin utility function also allows for dropping a more fundamental assumption, called the independence axiom.

My point here, is that neoclassical economics can absorb the criticisms of the behaviourists without a major shift in its underlying assumptions. The 'anomalies' pointed out by psychologists are completely consistent with maximizing behaviour, as long as we do not impose any assumptions on the form of the utility function defined over goods that are dated and indexed by state of nature.

There is a deeper, more fundamental critique. If we assert that the form of the utility function is influenced by 'persuasion', then we lose the intellectual foundation for much of welfare economics. That is a much more interesting project that requires us to rethink what we mean by individualism.

Greg also asks
“can we understand all the failures of classical macroeconomics without giving up both rational choice and the premise that all markets clear?” If anyone can make a persuasive case for the “yes” answer, I believe you can."
My response. Yes we can and should maintain rational choice, rational expectations and market clearing: but that requires a radical change in the way we define equilibrium. As I have done here.


  1. Well, you open the cage and stay in. If maximization is what is wrong, let's change that too.

  2. Roger,

    Thanks for taking the time to respond and to extend our discussion into broader questions.

    You begin with my “nudge” example in which the “default option” for employee contributions to retirement accounts is changed from zero to 5%, after which 20% of the previous “zero savers” begin deferring 5% of their income. I don’t think you explained how this particular change in behavior fits within your maximizing framework. The “reduction in cost” involved in “not having to sign up for deferred income” seems incommensurate with the size of the intertemporal shift of income.

    Of course, you can assume people always act rationally. Perhaps the 20% who decided to defer 5% of their income were “right on the margin” and their “cost saving” in “not having to sign up” was the tipping point. But this seems a bit like explaining the heroin addict’s behavior by saying she has a high discount rate.

    You argue that “the 'anomalies' pointed out by psychologists are completely consistent with maximizing behaviour, as long as we do not impose any assumptions on the form of the utility function defined over goods that are dated and indexed by state of nature.”

    If I understand this argument, you’re claiming that once the utility function is defined over contingent commodities, the behavioral “anomalies” pointed out by Thaler, et al, will fit neatly into a maximizing framework. Could you provide an example? How does this approach eliminate apparently non-rational behavior like “loss aversion,” “mental accounts,” hyperbolic discounting, or making decisions on the basis of “fairness considerations”?

    When you say, “indexed by state of nature,” I believe you're assuming publicly observable states of nature. Since many “psychological states” are not publicly observable, you’re excluding a lot of things Thaler, et al, are interested in.

    I believe the Arrow-Debreu concept of contingent commodities assumes we can envision all possible states as well as their probability of occurrence. Now, suppose the “Grand Futures Market” of Arrow-Debreu opened in 1910. Is it reasonable to assume that all the market participants could envision the state and probability of the first ever world war, or, alternatively, the combination of states of the world that comprise a world war? What about the 2010 price of minerals that were unknown in 1910, or the price of Internet service, or, for that matter, the prices of all the goods and services made possible by all the discoveries and inventions that have taken place since 1910?

    You write, “If we assert that the form of the utility function is influenced by 'persuasion', then we lose the intellectual foundation for much of welfare economics.” Well, I certainly hope my utility function can be influenced by reasons! (See Plato’s Republic or A. K. Sen’s recent work for examples). On the other hand, while I hope my utility function is not very much influenced by advertising, I suspect firms wouldn’t bother with this kind of “persuasion” unless it was effective. Indeed, if you assume profit-maximizing firms that advertise, can you consistently deny any role for "persuasion"?

    Finally, you recommend holding onto the assumptions of “rational choice, rational expectations and market clearing.” I can see these three desiderata coming together in the notion that the expectations of all market participants are consistent with one another. But I cannot see how the contracts arising from the Grand Futures Market of 1910 maintaining this kind of harmonious equilibrium (see examples above).

    I’ll close by sharing with you G.L.S. Shackle’s judgment regarding what he called the “Grand System of Economics.” “In its arresting beauty and completeness, it seemed to need no corroborative evidence from observation.”

    p.s. Thanks for the Epstein-Zin reference; it was interesting, and I can see why it serves your modeling purposes.

  3. Greg
    This semantics. I am using rational in the sense of Von Mises. Human action is rational. Period.

    Now the interesting part begins. How are choices made by human beings in typical situations? Call it maximization if you like. But what is being maximized? Are preferences time independent? Those are the questions currently being asked by game theorists and decision theorists. If rational means, expected utility maximization, neoclassical economics gave up on that a long time ago.

    I see behavioral economics as potentially more subversive. Preferences change and choice sets are incomplete.


Note: Only a member of this blog may post a comment.