« links for 2009-09-23 | Main | "The Crisis of Public Management" »

Thursday, September 24, 2009

"Dark Age in Macroeconomics?"

This is Nick Rowe (it's in response to Paul Krugman and follows up on one of Nick's previous posts):

Dark Age in Macroeconomics? A History of Taught approach, by Nick Rowe: (Or maybe the title should be: "Notes from the Phelps/Lucas Administration"; or "Notes to supplement our fading memories of the late 1970's".)
Is this a Dark Age in macroeconomics? In other words, have we collectively forgotten some (important) stuff that we used to understand?
I want to approach this question by looking at what was taught in the past to economics graduate students, so we can compare what is left out now to what was left out then.
I have a sample of one: my own lecture notes from grad school. I began my MA at UWO in 1977, and continued into the PhD. I took everything in macro/money that was offered. At the time, UWO was arguably the top Canadian department in macro/money (OK, Western grads would argue for; Queens grads would argue against), and would hold up well against anywhere in the world.
Macro 1 (David Laidler). Required course. Review and critique of ISLM (lags, stocks flows and the government budget constraint, are the IS and AD curves really demand curves? [no], the missing AS curve). Crowding out debate. Non-Walrasian macro (Barro and Grossman). Say's Law. Phillips Curve (up to Phelps and Friedman). Consumption function (Friedman/Modigliani). Demand for money. Investment demand.
Macro 2 (Michael Parkin). Required for those continuing to the PhD. (I can't resist quoting from the first page of my notes here: "Economics [is] Understand + Explain Phenomena using Rational models. How could Rational Behaviour [lead to] Disaster? Market Failure."). Review and critique of Neoclassical model of labour market. Lucas and Rapping (from the Phelps volume), and why their model was logically incoherent (Michael Parkin was right on this point). Mortensen's (also from Phelps volume) search theory of unemployment. Theories of implicit wage contracts (sticky wages). Theories of price adjustment (proto New-Keynesian). ISLM plus Phillips Curve (distinction between proto New-Keynesian and New Classical interpretations of Phillips Curve). Adaptive vs. Rational expectations. Policy Irrelevance Proposition ("[deviations of output from y* are] just noise, but obviously false").
Money 1 (Don Patinkin/Peter Howitt). Optional. Hume. Fisher. Lavington. Wicksell. Keynes' Tract, Treatise, and General Theory. Patinkin's Money interest and Prices. Are money and bonds net wealth? Commodity money. Solow/Swan growth model. Money and growth. Optimal quantity of money. Transactions costs. Baumol/Tobin and Miller/Orr models of demand for money.
Money 2 (Joel Fried). Optional. Microfoundations of money, Menger, Ostroy, Jones. Money in general equilibrium theory. Clower constraints. Transactions costs. Financial markets. Tobin. CAPM. Efficient Markets. Modigliani/Miller theorem. Term structure of interest rates. Tobin portfolio choice. Friedman and Monetarism. International finance. Dornbusch overshooting. Exogenous vs endogenous money. Canadian monetary policy.
Advanced Macro (Peter Howitt). Optional. (Lovely quote from the first page: "We are Aristotelian monks, trying to solve anomolies to stop future generations wasting their time doing the same thing.". Non-Walrasian disequilibrium theory (Clower, Leijonhufvud, Barro/Grossman, Malinvaud, Benassy, etc.). Stability. Catastrophe theory(!). Price adjustment under oligopoly. Optimal control theory. Inventories. Phelps/Winter price setting with transient monopoly power (from the Phelps volume, proto New-Keynesian).
(I learned some more money/macro in David Laidler's History of Thought class. But I was the only graduate student in that class, so I'm not going to count it. My colleague Calum Carmichael, who took the same course as an undergraduate, estimates that about one quarter of the Honours economics students took that class.)
I make the follow observations:
1. The Phelps volume was clearly very influential in the late 1970's. This supports Paul Krugman's memory, and my own.
2. The beginnings of the split between New Classical and New Keynesian approaches was already apparent in the late 1970's. I saw several references to the distinction between Fisher and Phelps on the interpretation of the Phillips Curve. (Fisherian market-clearing with misperceptions vs Phelpsian disequilibrium price adjustment). This too supports Paul Krugman's memory.
3. We received a very broad education in short run macroeconomics and monetary theory. Probably much broader than today's students. That tends to support the Dark Age hypothesis.
4. But there is one glaring omission from our education: we did lots of short run business cycle theory but almost no long run growth theory. We briefly covered the Solow growth model, but only as a prelude to money and growth. There was no interest in growth theory per se! If growth theory is important, and it is, that directly contradicts the Dark Age hypothesis. We barely touched on half of macro! The late 1970's were the Dark Age, for growth theory.
Why did we ignore growth theory?
Growth theory wasn't on the agenda. It wasn't that growth was unimportant; just that there seemed to be nothing important to say about it. All the exciting policy debates were about inflation and unemployment, not long run growth. All the exciting theoretical developments were about inflation and unemployment, not long run growth. "Endogenous" growth theories (a stupid misnomer, because growth is endogenous in Solow too, just with an extremely simple functional relationship to the exogenous variables, namely g=n) came later.
Fiscal policy has been off the agenda for much the same reasons, until recently.
(5. We spent surprisingly little time on open economy macroeconomics as well, for a Canadian school.)
OK. Let's compare notes!

This is very similar to my own experience, we also did very little growth theory (nothing beyond Solow-Swan, also as a prelude to looking at whether money was "superneutral"), and I didn't take any international at all - it wasn't part of the macro sequence (the international economy was not considered very important for understanding business cycle fluctuations). The emphasis was on short-run stabilization policy, monetary policy in particular. However, my experience was a bit different in that by the time I got to graduate school in the early 1980s, the split between saltwater and freshwater economists was well underway.

Paul Krugman says:

But by 1980 or 1981 it was basically clear to everyone that the Lucas project – the attempt to explain the evidently Keynesian behavior of the economy in terms of nothing but imperfect information – had failed. So what were macroeconomic theorists supposed to do?

The answer was that they split. One faction said, in effect, “OK: we can’t explain what we think we see in terms of full maximization. So we have to assume that there are some limits to maximization – costs of changing prices, bounded rationality, whatever.” That faction became New Keynesian, saltwater economics.

The other faction said, in effect, “OK: we can’t explain what we think we see in terms of full maximization. So we must be interpreting the data wrong – things like changes in the money supply must not be driving recessions, because theory says they can’t.” That faction became real business cycle, freshwater economics.

Here's what I said about this just under two and a half years ago (edited slightly). As you can see, even though this was written well before Krugman's statement, it basically agrees with his assertion that everyone knew the New Classical model was in trouble by 1980 or 1981 (the Mishkin paper noted below was published, I believe, in 1982, but given the long publication lags the results were well known long before then). It also agrees with his comments that one faction, the New Keynesians, built upon the old Keynesian structure by giving it rational agents and microfoundations who operated in an environment beset with rigidities of one type or another (these rigidities prevent agents from fully neutralizing nominal shocks such as changes in the money supply), and the other faction reemerged as the real business cycle school:

I entered graduate school in 1980. Though it started with a pretty traditional IS-LM framework with some AD-AS thrown in, most of our time was spent learning the New Classical model. Much of the research effort at that time, at least the effort I was made aware of, was to try and punch holes in the result that comes out of the New Classical framework that only unanticipated money can affect real variables like output and employment.

This assault came on both theoretical and empirical fronts. Mishkin, for example, had published an empirical paper in the early 1980s that challenged work by Barro and others from the later 1970s supporting the New Classical model and its implication that only surprise money matters. On the theoretical front, the old Keynesian model -- which had been criticized for, among other things, lacking microeconomic foundations and lacking rational expectations -- was being reconstructed into the New Keynesian model. This model would eventually overcome theoretical objections that plagued the older Keynesian model, and it would also do a better job than the New Classical model of explaining the magnitude and persistence of business cycles and other features of the macroeconomic data. We learned some about Real Business Cycle models - but for the most part that work went on elsewhere and would surface later with more force as an alternative to the New Keynesian framework. But we were certainly made aware of the real business cycle model, e.g. arguments about reverse causality to explain statistical money income correlations. I'd say the same about growth theory - we did the Solow-Swan basics, but very little beyond that. Stabilization policy was the main issue we worried about at the time.

Does money matter? I thought so, that's what my dissertation was all about, it gave theoretical and empirical reasons to doubt the New Classical result that expected money does not affect output, but the issue of whether money matters was not settled until later. We now accept, for the most part, that the Fed can affect real interest rates and also affect the real economy, but at that time there was a very strong split within the profession on this issue. It wasn't until later that a general belief that anticipated monetary policy was a potentially useful stabilization tool surfaced in the profession. It's sometimes surprising to me today how complete the conversion on that issue has been, though it's certainly not 100%.

So, it wasn't generally agreed that money mattered, i.e. that money was a useful policy tool for stabilizing the real economy. But the Keynesian economics I learned at the time, which was in the implicit and explicit labor contracting framework for the most part, did say that money mattered. In fact, since the point was to challenge the New Classical result that money did not matter, the focus was mostly on monetary policy. As for fiscal policy, the Keynesian model we talked about - beyond the simple IS-LM version we learned at first - paid very little attention to fiscal policy, though papers such as Barro's "Are Bonds Net Wealth" were part of the conversation. Thus, when I went to graduate school - and this was partly due to who was teaching the courses - the primary focus was on whether and how changes in monetary policy affected the real economy.

In any case, even though it was a few years later than Nick's experience, we also spent considerable time on the ideas that Krugman notes have since been lost as we entered our recent "Dark Ages."

    Posted by on Thursday, September 24, 2009 at 12:09 AM in Economics, Macroeconomics, Methodology | Permalink  Comments (79)


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