Gadgets Versus Fundamentals (Wonkish), by Paul Krugman, in NY Times: Wren-Lewis argues against the many people insisting that the crisis means that we must rebuild macroeconomics from the ground up; he argues, on the contrary, that the crisis doesn’t require a fundamental rethink of macro. And I am very much in agreement there. Basic sensible macro — what we learned from Keynes and Hicks — has actually held up very well in the crisis. To the extent that we have a crisis in macroeconomics as practiced, it comes from the way many economists chose to reject sensible macro. The crisis should (but won’t) kill fresh-water, equilibrium macroeconomics; but IS-LM is looking pretty good.
And let me throw in that I personally have found the crisis remarkably comfortable territory in an intellectual sense, because we’re facing exactly the same sorts of issues I began worrying about in the face of the Asian crisis of the late 1990s. My liquidity trap work (pdf) has, I think, aged pretty well; and even back then I was very much worried about issues of debt overhangs and balance sheet effects.
Now, what’s true is that my old work on balance sheet stuff focused on corporate rather than household debt, and that I was entirely concerned with the balance sheet effects of a movement in the exchange rate as opposed to, say, a drop in housing prices. And I really, really should have connected the dots and seen how a burst housing bubble could produce similar effects — but I didn’t; I thought the end of the bubble would be nasty, but failed to realize how nasty. Mea culpa.
But this was a failure to look at the right variables, not a fundamental flaw in the theory — and once it became clear that we were in a balance-sheet crisis, it was quick work to slot that into our understanding.
So, macro as I understand it hasn’t done badly.
But macro as I understand it isn’t what Wren-Lewis is describing when he talks about
the intertemporal optimising framework that is the heart of modern macro.
Um, it’s not at the heart of MY macro. Or maybe I should say, it’s not at the HEART of my macro. I’m willing to do models with intertemporal optimization, but the optimization is never the core of the story. Instead, it’s just a gadget.
What do I mean by that? Readers who know my work on trade and geography know that a lot of it made use of the Dixit-Stiglitz model of monopolistic competition. This model is, I would say, silly but brilliant — or if you prefer, brilliantly silly. It makes some obviously untrue assumptions about tastes and technology in order to produce a mathematically and conceptually tractable way to think about increasing returns and imperfect competition. And for the past three decades I and many others have made heavy use of this brilliantly silly model to clarify our thoughts about areas, such as trade and geography, where increasing returns and imperfect competition are crucial.
The point is that all this work USES Dixit-Stiglitz, but it’s not ABOUT Dixit-Stiglitz; D-S is a gadget, a tool that helps you work with the fundamental issues, but you don’t ever want to forget that it is no more than that.
I think of intertemporal optimization in macro the same way. The assumptions of hyper-rationality, not to mention representative agents, that underlie many macro models, including some of my own, are just as silly as the symmetric CES preferences of Dixit-Stiglitz (I told you this was wonkish). It’s OK to use those assumptions to make a point, but only as long as you keep your tongue firmly in cheek.
But I guess not everyone even on the sensible side of macro sees it that way. And that is a problem. A gadget is only a gadget, and you should not let it define your field.