This John Kay rant against modern macroeconomics brought to my mind the Copernican Revolution in astronomy. Not the potted 7th grade story of linear progress, but the tale told in Thomas Kuhn’s somewhat revisionist book.
The way this went was as follows. Ptolemaic astronomy started with the observation that “the planets” (including the sun and the moon) seem to revolve around the earth. It assumed they moved in circular orbits, and made predictions based on that. As people bothered to pay attention, it became clear that this theory gives you the wrong predictions. So people developed the ad hoc concept of “epicycles.” The planets moved in circles-within-circles, with equations developed to account for the actual position of the planets. With more and more observations, the calculations became more and more complicated and a lot of people were unhappy with the increasingly messy picture. Then along comes Copernicus who as a young man had been involved in some neo-Platonist cults featuring sun-worship and a heliocentric worldview. He notes that if you reinterpret the heavens as centered around the sun, you can derive a considerably more parsimonious and theoretically elegant account of positions of various heavenly bodies. All the epicycles are gone! Victory.
Now the problem, as some scolds note, is that while Copernicus’ spheres give considerably better predictions than Ptolemy’s, they actually give much worse predictions than the fully worked out Ptolemaic astronomy with all its epicycles. You’ve gained a huge amount in simplicity and elegance, but actually lost ground in precision and accuracy.
Eventually Johannes Kepler comes around to note that if you keep Copernicus’ heliocentrism but ditch the circles in favor of ellipses, you end up getting the accuracy back. This is a huge win for everyone, because the ellipses have most of the ease-of-calculation benefits of the Copernican system and most of the accurately-forecasts-where-things-are benefits of the Ptolemaic system. The only problem is that everyone’s been plugging along with this idea that motion should be circular, an idea that seemed well-grounded in the theoretical assumptions of the time about divine perfection. Kepler shows you can account for everything much better by using ellipses, but has no idea why ellipses would be relevant to the situation. It’s only some time later that Isaac Newton comes along with an explanation of why orbits would be elliptical.
What does this have to do with macroeconomics? Well, quite a lot.
For one thing, it shows that this putative contrast between inductive and deductive modes of reasoning is a bit faulty. Progress is made on both ends, herky-jerky style. The basic impulse to say that the theory has to make sense and be grounded in a compelling account of how the world works makes major contributions here. It’s why Copernicus ditching geo-centrism for heliocentrism, and it’s how Newton develops the theory of gravity. But the instinct to say that no, the important thing is for the theory to produce actual results is also important. If we’d stuck with Copernicus’s “theoretically compelling” idea about perfect circles, we’d never have noticed that this was actually a totally arbitrary modeling assumption with no basis whatsoever and thus would never have noticed the theory of gravity. What’s more, while in retrospect we see Copernicus as the ancestor of our modern way of thinking, the fact of the matter is that if you were trying to launch a rocket ship somewhere in the early 16th Century you’d be much better off chugging along with the epicycles rather than siding with the guy who knew that the earth orbited the sun.
My view, with both all due respect and all due derision, is that the Robert Lucas types are like the early Copernicans here. There’s something admirable in their insistence that it ought to all work out to an easily modeled system grounded in compelling theoretically considerations. The New Keynesian model is a mess, like late-Ptolemaic astronomy, thrown together to account for observed reality. But you don’t fly to a moon with an elegant model that delivers mistaken predictions about where the moon’s going to be. And what we actually need is a Kepler to give us an elegant model that actually predicts the phenomena, and then a Newton who can explain what that model means.