This post is a follow up to a post by Chris Dillow. He makes two claims about macroeconomics. To understand this you need to know about the distinction between mainstream and heterodox economics. It is difficult to define what divides the two groups (papers/chapters are written trying to do so), but we can make some observations. Most economists in academic departments are mainstream, and heterodox economists fall into lots of different schools. Another observation is that mainstream economists almost completely ignore what heterodox economists do. What is not true is that all heterodox economists are left wing and all mainstream economists are not.
Chris’s first claim is that much of heterodox economics is the macro that was fashionable many years ago. The second is about the mainstream. He suggests there was a ‘great forgetting’ in the 1980s, such that apparently new ideas today are just rehashed old ideas that seem to have been forgotten. I think there is some validity to both these points, and I hope to explain here why this has happened and how they are related.
The great forgetting happened because of a methodological revolution in macro, the New Classical Counter Revolution (NCCR). Many at the time focused on one aspect of that revolution - an attack on the mainstream Keynesian macroeconomics of the time. That failed fairly quickly, but what lasted was a methodological revolution, which is often called the microfoundation of macroeconomics. I have written about the NCCR here.
Before the NCCR, macro was almost a different subject to microeconomics. Macro’s main theory, Keynesian economics, could be written in aggregate equations and could be estimated using aggregate data, but it was sometimes hard to relate it to microeconomic theory. Partly for that reason, macro contained different schools of thought, where it was hard to relate one ‘school’ to another.
The NCCR insisted that any macro model should not only be derived from microeconomic theory, but also that the theory was used in a consistent manner. By unifying macro and micro economics, it appealed to young economists in particular. Very quickly academics realised you could also formalise aspects of pre-NCCR Keynesian economics in a microfounded way, and the microfoundations hegemony within mainstream economics became complete. As all macroeconomists within the mainstream talked the same language (microeconomics), to the extent there existed schools of thought, these related to beliefs about parameter values within a common theoretical framework (which in turn those beliefs might reflect ideological preferences).
This reduction in the fragmentation in mainstream macro is an important unifying force. Pre-NCCR you could go to a macro seminar given by someone from a different school of thought and have little idea what they were talking about. Post-NCCR you could do the same and know exactly what they are talking about, and even make comments on their work, but just think it was not very relevant to the real world.
One of the negative aspects of the NCCR was it led to the ‘great forgetting’ Dillow talks about. For some the NCCR was year zero in macroeconomics, with everything that had gone before (that was outside the microfoundations hegemony) of little interest. The example that is close to home for me is described here. I had done a lot of work applying the John Williamson’s FEER model of equilibrium exchange rates, which uses pre-microfoundations techniques. The approach was rediscovered by Obstfeld and Rogoff who used a microfounded model (the ‘New Open Economy’ approach, which applied the microfoundations of imperfect competition to international trade). Which was fine, but as far as I know Obstfeld and Rogoff never acknowledged that earlier literature started by John Williamson. It was a perfect example of a great forgetting.
Having year zeros in any discipline is a bad idea. My view reflects an argument I have often made: microfoundations are fine, but the microfoundations hegemony is not. By hegemony I mean seeing other macro as old fashioned and not.worthy of publication in top journals. Some economists had a stronger reaction and chose to be heterodox economists outside the mainstream. They felt microeconomics didn’t deserve this central role in macroeconomics, or felt it made no sense microfounding macro. I think it’s fair to say that after the NCCR it was more difficult if not impossible for those economists to get on the mainstream train.
The analogue to this is the point Dillow makes, that a lot of heterodox economics is what he remembers learning before the NCCR. That observation isn't a slur on heterodox economists, but rather to note that some of what they do is a continuation of pre-NCCR macro.
A good example of that comes from MMT, where a key idea is that we should use fiscal policy rather than monetary policy for macroeconomic stabilisation at all times (and not just when rates are at the lower bound). That reflects how macro policy was done in the UK in the 1960s and mid 1970s. It is quite possible to argue that case within the mainstream (although most mainstream economists would disagree), but I don’t believe MMT economists would wish to do that.
However the NCCR didn't create heterodox macroeconomics. The heterodox economics I remember from studying at Cambridge in the early 1970s was a rejection of ‘neo-classical economics’. At its simplest, neoclassical economics is how producers who maximise profits interact with consumers who maximise utility in perfect markets (where perfect means there are a large number of producers and consumers for the same product so everyone is a price taker). As Chris notes, the theory implies people are paid their ‘marginal products’.
If you asked most mainstream economists today if what they do assumes the validity of neo-classical economics, they would probably laugh. As I repeatedly say, much of economics today is about studying markets that are far from perfect. (We can debate endlessly whether having perfect markets (and associated optimality conditions) is a useful or misleading reference point.) Economics has also grown well beyond the study of markets and their imperfections, to include game theory and behavioural economics for example. None of this stops right wing economists using ideas from neoclassical economics when it suits them.
Another feature of microeconomics today is that it is dominated by empirical analysis. The days when the heights of the subject were highly mathematical proofs about the nature of general equilibrium are long gone.
However we now need to go back to the NCCR in macro. The microeconomics the proponents of the NCCR took to remake macroeconomics was essentially the neoclassical economics that microeconomists were simultaneously moving beyond. New Keynesian economists brought a form of imperfect competition into DSGE models to enable a formalisation of price or wage rigidity, but it remains the case that the basic DSGE model is based on pretty neoclassical microfoundations. (The clearest exposition of this is a book by Athreya.)
We can now see an additional reason for the barrier between heterodox and mainstream macroeconomists. Macroeconomics has adopted the neoclassical economics that heterodox economists had objected to over decades before the NCCR.
While I personally regard post-NCCR macro as progressive in a Lakatos sense, I believe its hegemony is unwarranted. I learnt a great deal about the world working in pre-NCCR macro, and this has been complemented rather than duplicated by what I learnt from microfounded macro. (I have published plenty of papers in both.) Mainstream macroeconomics now needs to follow micro in becoming more empirical, and one of the ways it can do this is by resurrecting the pre-NCCR style of modelling. This involves both econometric studies of individual relationships or sub-systems, and the use of structural econometric models (what Blanchard calls policy models) that Chris mentions in his post. These models are an eclectic mix of theory and econometric studies of individual relationships or sub-systems.
As I note in this paper, it should be worrying that policy institutions like the IMF and Fed still do a great deal of this kind of analysis, but mainstream macro in the top journals does almost none. The time has passed when we can excuse this fact as reflecting old-fashioned policy institutions, or that such analysis is flawed because of identification problems or the Lucas critique. If mainstream academic macro will not venture into these directions, perhaps some heterodox economists can take their place.
I agree with what you've written in this post, but there's an additional issue: "microfounded" models are not actually microfounded, and it matters a lot. These models all work with representative agents, including those with heterogeneous agents. They don't actually derive macro relationships from the plethora of individual decisions that a true microfounded approach would need to model. Of course, that would be impossible; there are too many decision-makers and too little understanding of how they respond to their changing environments. But worse, what is really distinctive about macroeconomics derives from exactly that complexity. The choice of the "old" macro to abstract from the individual level and work with aggregate patterns at least allows for the effects of that complexity to be studied; microfounded macro doesn't unless it's reverse engineered to have ad hoc features that generate them.
ReplyDeleteThe closest we come at the moment to true microfoundations is agent-based modeling, but this remains highly stylized and not terribly useful for policy work, as far as I know.