Jul 26, 2010

Why economics is not a science

If you're not familiar with the Prisoner's Dilemma, here's a nice Wikipedia article about it for you. For the purposes of this discussion, all you need to know about it is that it's a simple experiment that has shown, time and time again and all over the world, that people almost never make the choices they'd be expected to make if they were purely rational beings.

Rational choice theory (oh, alright - but do your own Googling next time!) is the theory that says that even though that is patnetly the case, we should still govern our lives (and set policy, and make economic predictions) as if they did. Because it's the right, or more correctly "rational" thing to do. I very much suspect that rational theorists' definition of "rational" is "that which will make our mathematical models work".

Imagine if medicine was run that way: "yes, sometimes people don't respond as they should to treatment, but that's just because they're irrational! Everyone will have penicillin shots and peanut butter for breakfast, because that's what's best for them. Even when it isn't."

Modern economics - mostly microeconomics, but macroeconomics takes some of the former's assumptions and runs away with them, too - just can't be made to work without the assumptions of rational choice theory. And rational choice theory is flat out wrong, in the scientific sense of being incogruent with observed reality.

Which is why we should be very, very wary of anyone who dismisses our objections to, e.g., severe public spending cuts on the basis of economic forecasts and economic theory. Not because it's always and necessarily wrong, but because it is an ideology dressed up as science. So we should evaluate the claims of the current government as ideological rather than evidence-based: not "is it necessary, based on economic forecasts, to privatise GP surgeries and cut housing benefit?" but "is it the right thing to do to expose surgeries to potential bankruptcy and people to potential homelessness?".


  1. Perhaps this is now a bit old to still comment, but I do feel the desire to defend my discipline a little...

    I think rationality is a superb concept to analyze people's decisions. People in general are very good at rationally deciding what's best for them. (see, e.g. the abortion debate...assuming women know what is best for themselves should actually seal the discussion. But we all know why it doesn’t...)

    If people’s decisions disagree with what a model predicts, it tells us to look for additional or different motivations people have for reaching the decisions they do. If irrationality is assumed, we have nowhere to go: seeing a disconnect between predictions and observed behavior cannot guide understanding – it’s just irrational and unpredictable.

    It is true that it is a social science, and not a natural science. Which makes it inherently more inexact, since two people generally vary more widely in their intentions, desires, needs, and endowments than two, say, atoms.
    This inexactness is of course compounded by the fact that exposing humans to ‘experiments’ is more strictly regulated than exposing atoms, bacteria, or fungus to experiments – with good reason.

    This being said, I agree with your distrust of the mentioned predictions. However, it is the government’s agenda I do not trust, not economics. Government riding on the coattails of a science in order to justify and seemingly give some air of objectivity to their own agenda is the real scandal.

    Economics distinguishes clearly between ‘how do we create the biggest surplus’ and ‘how do we want to distribute this’. Politicians of any stripe will generally gloss over the fact that these two questions need to be answered separately, and that disagreeing with their answer to the second does not make one ‘uninformed’, or ‘stupid’ or ‘against progress’ or ‘against efficiency’ or any of the other bullshit they say to create the illusion of “it is not really my fault that some people will suffer, it’s SCIENCE!” (jazzhands)


  2. Interesting comment; and you're right, I did not distinguish between "social science" & "natural science" in my piece. Mostly because the argument about why the social sciences want to call themselves social sciences is one that even I am not brave enough to wade into! :)

    I have to dispute the assertion that without assuming rationality we can't do any statistical modelling because all the predictability ofagents falls out the bottom, though. We know that people act irrationally in many prdictable ways: they smoke, they eat things that are bad for them, they speed through red lights even though it's potentially deadly, they give money to charity even though they get no reward for it.

    The Prisoner's Dilemma experiment challenges rational choice theory, but it gives results that are predictable, replicable, and have a predictive power of their own. So economists do have a basis in experimental data to build their theories on, it's just that too mny of them are married to the simplicity of rational choice theory and the mathematical tools it gives them, without much concern for whether or not the math describes the world as it is.