clipped from www.economist.com

It takes a model to beat a model

The benchmark macroeconomic model, though not junk, suffers from some obvious flaws, such as the assumption of complete markets or frictionless finance. Indeed, because these flaws are obvious, economists are well aware of them. Critics like Mr Buiter are not telling them anything new. Economists can and do depart from the benchmark. That, indeed, is how they get published. Thus a growing number of cutting-edge models incorporate one or two financial frictions. And economists like Mr Brunnermeier are trying to fit their small, “blackboard” models of the crisis into a larger macroeconomic frame.

But the benchmark still matters. It formalises economists’ gut instincts about where the best analytical cuts lie. It is the starting point to which the theorist returns after every ingenious excursion. Few economists really believe all its assumptions, but few would rather start anywhere else.