Relative vs. Absolute Rationality

I’m reading Tim Harford’s "The Logic of Life" – it’s the first book I bought on my Kindle.  He uses a definition of rationality which I hadn’t seen before, which is simply that people respond to incentives.  I think this model of people as relatively rational has much more support than the idea that they are absolutely rational – that they choose optimal strategies to reach their goals, that they behave in an unbiased fashion.

And I think this is a good way of squaring the ideas that there is lots of evidence for human rationality and lots of evidence for human irrationality.  Before, I’d been thinking of the resolution as just that sometimes people are rational and sometimes they are irrational, depending on how complex the decisions and which heuristic modules are invoked.  But it feels much more correct to say that people rarely get the answer exactly right, but that they generally respond in the right direction when things change.

This definition rescues the implications of rationality-assuming economic analysis from the "But people aren’t rational!" attack.  Sure, people aren’t (absolutely) rational, but since they are (relatively) rational, policy makers[1] can influence behavior by assuming that people will respond in the correct direction to changes in incentives.  And they had better be wary of creating incentives without considering the consequences on behavior.

[1] Or anyone else engaged in mechanism design for humans.

GD Star Rating
loading...
Tagged as:
Trackback URL:
  • http://apperceptual.wordpress.com/ Peter Turney

    This sounds like Herbert Simon’s “satisficing”:

    http://en.wikipedia.org/wiki/Satisficing

  • http://profile.typekey.com/andrewgelman/ Andrew

    It is possible to rational just because you feel like it, not because of any incentives. For example, in solving the daily crossword in the newspaper. Responding to incentives is fine, but I really see this as two dimensions:

    (1) You can be doing something (a) as a response to an incentive or (b) without an incentive.

    (2) You can act (a) rationally or (b) irrationally.

    You can have all four possibilities, I think: 1a/2a, 1a/2b, 1b/2a, 1b/2b.

  • Aaron M.

    I do not see how a person can act, either consciously or unconsciously, without an incentive. Maybe it’s better to say that all action has a goal, whether or not the actor realizes it.

  • http://rolfnelson.blogspot.com Rolf Nelson

    Suppose you can buy a good bicycle for $500. I want to start a business selling bicycles of comparable quality.

    Because I believe people are somewhat (“absolutely”-but-fallibly, rather than “relatively”, in your parlance) rational, I can predict my bicycle will probably sell well at $450 but not at $550. If I make get lots of cool ads, maybe I can get people to buy it at $600, but the ads had better be awfully cool. People will probably not buy it for $50,000.

    Would you be able to duplicate this basic finding in your notion of “relative rationality”? (That is, not just the notion that more people will buy at $5 than would buy at $50,000, but also that there will be a sharp discontinuity around $500 or so?) If so, what’s an example where absolute rationality and relative rationality would produce different predictions?

  • http://profile.typekey.com/andrewgelman/ Andrew

    Aaron,

    Yes, “incentive” is different from “goal.” I can have a goal of finishing the crossword puzzle with no incentive. Or, to put it another way, some of the things I do have incentives and some do not. Or, to put it even another way, some of the things I do have bigger incentives than others. My effort, and even the extent to which I behave rationally, is not always related to the incentive (if any).

    I agree that incentives often help; I just don’t think they’re always there.

  • Joseph Knecht

    Andrew, isn’t there a reason you finish the crossword puzzle? For example, because you enjoy finishing them, or you like completion. If you didn’t have some reason to finish it, you wouldn’t finish it, and would instead do something else.

    I’m assuming something like the following as the meaning of incentive: In economics, an incentive is any factor (financial or non-financial) that provides a motive for a particular course of action, or counts as a reason for preferring one choice to the alternatives.

    In which case, you certainly have an incentive. Maybe you mean something else by incentive, but you should state what you mean if that’s the case.

    When you say, “It is possible to rational just because you feel like it, not because of any incentives.”, I’d ask, why do you feel like being rational? There’s your incentive.

  • http://profile.typekey.com/Unholysmoke/ Ben Jones

    I can have a goal of finishing the crossword puzzle with no incentive.

    …you certainly have an incentive. Maybe you mean something else by incentive, but you should state what you mean if that’s the case.

    Andrew and Joseph, what experiential differences do you expect to see based on your disagreement?

    Cut out the word ‘incentive’ and tell us (without any synonyms like ‘motive’) what you mean by it, and then see if you still disagree.

    Do I win a prize?

  • Sam

    In teaching economics I always bring up the problem of the Rational Actor. Any model of behavior in economics depends on the principle. As an extreme case I introduce a problem: There is a woman who abandons her very young children for a few days while she parties with friends. Most students cannot accept that her actions are rational. They confuse “reasonable” with “rational.” (“Reasonable” always has a moral dimension to it — thus it can be used in law.) But her actions are rational in that she has weighed ( implicitly ) her returns from being a responsible mother against her returns from being a party girl and found higher returns in her partying. This can, of course, be seen in terms of incentives: her basic incentive structure is to seek returns/pleasures; over the short-term ( and probably the limit of her horizons anyway) her incentive for partying is greater than her more complicated and time-bound incentive as nurturer and mother. This framework can be extended to narcotics abusers, the Manson family, etc.

  • http://profile.typekey.com/andrewgelman/ Andrew

    Ben,

    One area in which I’ve done research is on the effects of incentives for increasing participation in surveys. We estimate the effects to be a few percentage points in participation. This is a classical example of an incentive: you pay people and they’re more likely to do something.

    Joseph and Sam,

    To me that just sounds like circular reasoning. All behavior is then rational and incentive-driven, in which case the concepts of rationality and incentives have been drained of all meaning.

    Umm, here’s the definition of incentive from dictionary.com: “something that incites or tends to incite to action or greater effort, as a reward offered for increased productivity.” I think there has to be some “something,” not just a warm feeling. For example, when people talk about giving incentives to kids to get better grades, they’re talking about money, or gas in their cars, or whatever, not just that warm feeling.

    To get back to the original blog entry, I can’t stop you from saying that rationality is “simply that people respond to incentives.” To me, though, that simply removes much of what is distinctive about rational thinking, and much of what is distinctive about incentives, and makes these into empty universal concepts that describe all behavior.