« Desirable Effects of Irrational Exuberance? | Main | NBER: TV Appearance and Electoral Success »

Friday, June 01, 2007

How Stubborn Are We?

Here's my second entry in Christopher Hayes discussion of heterodox economists at TPM Cafe:

For the most part, this debate has taken as given that the economics profession is resistant to change and to unorthodox ideas. Is that true?

As a macroeconomist, it's hard to view economics as stationary. When I think about the macroeconomics I learned thirty years ago, which was a traditional Keynesian IS-LM model, and compare it to what we teach today, I see vast changes in both the theoretical models we use to understand the world, and the tools and techniques we use to analyze our models.

Very briefly, from the old style Keynesian model I first learned, to the New Classical model, to the current split between the New Keynesians (markets have frictions, government intervention is helpful), and the Real Business Cycle advocates (markets work, government intervention makes things worse), macroeconomic models have evolved considerably over time. There have also been substantial changes in the tools and techniques that we use. The use of Bayesian estimation techniques, rational expectations solution methods, and the simulation of dynamic stochastic general equilibrium models are but a few examples.

That's a lot of change in thirty years. In fact, one complaint I've heard expressed in this debate is that academic journals will no longer accept papers that employ old style Keynesian economics, i.e. that we have changed too much. But more generally I think the complaint is about the direction the profession has headed more than our resistance to change. We have not gone in the direction heterodox economists would like to see, and that's frustrating when you think you are right.

Are we willing to change when the models conflict with the evidence? I think we are. Are we wedded to the neoclassical model? Macroeconomists investigate frictions, market failures, and information problems. There are disequilibrium models as well as equilibrium models, and we allow for multiple equilibria as a possibility. We model market power to give firms price setting capability, and we impose transactions costs, adjustment costs, and other impediments to continuous market clearing (and I'm sure I forgot some other key frictions). We're not above assuming money illusion, though we try to avoid it.

Though most of this can be embedded within, or as departures from, the standard neoclassical model, a point made elsewhere in this debate, I don't think macroeconomists much care whether the model is neoclassical in the traditional sense or not. There are certainly groups who argue that microfounded, continuous market-clearing, neoclassical style models are the best approach, but find a model that works and you won't be able to keep macroeconomists away from it. If it predicts the future with any certainty, something our current models don't do well, we'll beat down your door trying to get to it no matter what you've assumed. In fact, I sometimes wonder if we aren't too quick to adopt the latest fad, not too slow to change.

I understand that the changes I am talking about assume (for the most part) that people optimize, are rational actors, and so forth, i.e. that the changes in macro have mostly been within the overarching structure of standard economic analysis. The neoclassical framework is a surprisingly flexible structure and the standard framework can handle many of these "heterodox" approaches to macroeconomic modeling so there's no compelling need (in most cases) to develop a new theoretical structure. In fact, over time macroeconomists have moved from using models where the underlying behavior of agents is not optimal, to models where we try to ensure that agents are rational, maximizing actors. Thus, anyone who believes the assumption of optimizing behavior is wrong or that our models miss important social, political, and power relationships, as many heterodox economists do, will understandably be distressed with these developments. That's not the direction they think the field should be headed. But there has been change.

Does the fact that we haven't changed radically imply we won't or can't? I don't think so. Let me give a microeconomic example. When I first entered graduate school and took microeconomic theory, we talked about utility maximization (of course). We were told that economists take preferences as given, psychologists were the ones who worried about how preferences are formed, that was their job not ours. What economists do is take the preferences psychologists give us, and then assume people maximize them. If psychologists tell us that people have a preference for green, triangular shaped goods due to some brain quirk, that's fine, we take that as given and assume they maximize their utility - psychologists can worry about whether it is "rational" in their sense of the word for people to have those preferences.

But things have changed, and we now think a lot about preference formation, what it means to choose rationally and so on - the wall between psychologists and economists is not as high as it once was. For example, later today I am going to a seminar sponsored by the Psychology Department to see Brian Knutson of Stanford present a paper on the neural representations of expected value. This research, which he founded, uses economic experiments with fMRI brain imaging to test alternative theories of choice under uncertainty. That he's a psychologist and neuroscientist, but has written many papers with economists shows how much things have changed. Thirty years ago that would not have happened, not with the regularity it does today. I don't know if this type of work will lead to a revolution in how we model behavior and improve upon the rational actor model used today, something that could affect both micro and macro models, but it does show we are willing to take serious steps in new directions (and it is affecting policy, today neuroeconomists might be tempted to correct the "defect" of preferring green, triangular shaped goods through the use of incentives that "correct" the quirky behavior, i.e. use schemes such as opt-out retirement accounts to maximize enrollment and correct behavior that causes "sub-optimal" choices not to enroll).

Heterodox economists who have argued against the rational actor model may ultimately be proven to be correct, or not, but the question is whether economists are responsive to new evidence, whether they are willing to adopt new models, and whether they are willing to step outside the neoclassical framework if the evidence points in that direction. Others disagree, but my experience says that we are.

    Posted by on Friday, June 1, 2007 at 12:33 AM in Economics, Methodology | Permalink  TrackBack (0)  Comments (23)


    TrackBack URL for this entry:

    Listed below are links to weblogs that reference How Stubborn Are We?:


    Feed You can follow this conversation by subscribing to the comment feed for this post.