Paul Krugman compares the two and calls for a pragmatic methodological pluralism. Most of his discussion is about expectations, but when this topic first came up I had a few other issues in mind:
1. For how long — in today’s America — can an AD-driven recession last? At what point do even the Keynesians toss in the towel and say “By now it is a growth and structural problem, not mainly AD”? After all, the private sector had a chance to create more M2 and it failed. How sharp is the distinction between the short run and long run? At what point do long-run problems twist their neck back and screw up the present day?
2. How much do we heed the concrete results of new Keynesian models on the extent and duration of wage and price stickiness? Many of these models suggest stickiness is not quite the bugaboo it is often made out to be, especially not in the longer run, or not with unemployed workers. It is no longer the Great Depression and it seems at least possible that 2-3 percent of the workforce could lower their reservation wages without setting off a downward deflationary spiral.
3. How much should we integrate Keynesian results with search models of the labor market, a’la Mortensen and Pissarides? That’s actually not easy to do, yet the search models are built on some fairly basic and general microeconomic intuitions. I see little interest in these ideas from the old Keynesians.
4. What to make of the liquidity trap? One of Krugman’s models (with Eggertsson) suggests that a very low rate on T-Bills implies an upward-sloping AD curve, and other counterintuitive results, but most new Keynesian models would not bring you to this conclusion. This is a big difference with important practical policy conclusions.
5. IS-LM has a quite primitive or indeed non-existent treatment of the banking sector; Stephen Williamson stresses this point.
I interpret the old Keynesians as holding an attitude something like: “We know from the Great Depression that an AD problem can be very bad for a very long time. Maybe there are mysteries in how that happened but we need to double down on traditional Keynes, Hicks, and IS-LM.”
The old Keynesian approach has a major presence in the blogosphere but much less influence in current academic macroeconomics. Whether Econ 101 sides with the Old Keynesians I am not sure (it depends who teaches the class), but Econ 2011 in many cases does not.
There are enough AD-denialist arguments running around that the new and old Keynesian perspectives can forge an alliance on some major issues. But as the downturn continues, this intellectual alliance will grow increasingly fragile, mostly over the question of whether long-run or short-run models are relevant.
Not long ago I tweeted this:
Confused by the Right on macro, you’re a New Old Keynesian; confused by the Left, you’re an Old New Keynesian.
I also see old Keynesians as believing that the IS-LM framework follows directly from the quantity theory of money, while new Keynesians are not committed to such a view and may even oppose it. I may write a post devoted to this topic.