Macroeconomics can be more open without losing rigour
The Oxford project to rebuild macroeconomic theory that we have covered this week leaves little doubt that mainstream macroeconomics is in deep need of reform. The question is how, and whether the standard approach, DSGE modelling, can be sufficiently improved or should be jettisoned altogether.
The main challenge is that in its conventional form, DSGE macroeconomics does not really allow for the large-scale financial panic we saw in 2008, nor for some of the main contending explanations for the slow recovery and a level of economic activity that remains far below the pre-crisis trend. It is hard not to see these misses as an existential challenge to the field.
Several contributions to the project, however, push back quite hard at any nihilistic conclusions. Paul Krugman, while admitting the blind spot of conventional DSGE, argues that existing macroeconomic theory — perhaps used eclectically (and with eclecticism applied the wisdom urged by Dani Rodrik) — is “good enough for government work” and policy advice. Ricardo Reis points out that the work of fixing DSGE is well under way, listing the huge efforts made (especially by younger researchers) to incorporate the insights that earlier versions of the approach left out. And Jesper Lindé points out that while DSGE did not predict the crisis, nor did its main rivals — and thinks DSGE will continue to be judged at least by central banks as the best tool they have got. (But it should be said that some more heterodox thinking has been better at emphasising the possibility of a big financial crisis.)
These points are valid as far as they go. But they do not go far enough. Krugman’s “models for different seasons” approach leaves unfulfilled one of the main purposes of economic modelling listed by Olivier Blanchard in his contribution. That is to provide an intellectual core to the field — a common framework within which economic researchers can formulate their different views with a hope of reaching agreement on which is better. Having such a framework is crucial for improving our knowledge of the macroeconomy. The question is whether something like DSGE can fulfil it or it is too hopelessly constraining.
The advances celebrated by Reis are necessary — they bring the economic relationships from which the DSGE model is built closer to reality — but seem insufficient. The framework must be flexible enough to allow for explanations that depart further from the mainstream than just tweaks to the standard account to be at least admitted to the contest between rival theories. In particular, there should be room for formalised theories in which the economy is subject to multiple equilibria — what could be called paleo-Keynesian views — which show how both large crises and protracted periods of stagnation are possible. Not because they are necessarily true (though they probably are more true standard macro), but because it is only if such theories can be formulated within the common framework can they be properly assessed against more conventional rivals.
That entails a double responsibility: of the overarching theoretical framework to have room for such theories, and of the proponents of the multiple-equilibrium view of the economy to formalise them within the broad “core” model. There is work to be done on both sides. An effort at the latter is presented by Wendy Carlin and David Soskice, who very didactically also compare how their approach differs from another recent multiple-equilibrium macroeconomic model. The detailed intuitive explanations that go with the mathematics make this an important paper to read also for non-specialists.
Carlin and Soskice make clear where their “microfoundations” depart from convention. One important feature is to allow participants in the economy to have divergent expectations of the future. Another “key assumption is that nominal wage inflation is bounded by a floor around zero, which captures the ability of employees to coordinate on opposition to nominal wage cuts”. These are not macroeconomists’ traditional kind of microfoundations. But it is an abuse of language to say they are not microfoundations at all. They are plainly efforts at building a macro model up from assumptions about micro behaviour. The test should be which assumptions make for better explanations of the phenomena we observe.
The point here is that microfoundations, which help interpret the causes of macro patterns (such as economy-wide consumption, investment and so on), should not be confused with specific microfoundations, let alone the highly artificial ones of hyper-rational expectations and costlessly maximising behaviour. Once that distinction is clear, the overall DSGE approach could be liberated enough to allow for a large range of different assumptions about microeconomic behaviour and the macro relationships that they give rise to. That pluralism will inevitably give rise to much disagreement about which micro assumptions are ad hoc and which are well-founded. But, as Samuel Bowles and Carlin’s recent FT op-ed points out in the context of economics education, pluralism integrated within a common framework is preferable to “pluralism by juxtaposition”, where fundamentally different methodological approaches do not attempt to engage with one another at all.
The fundamental difficulty of microfoundations is that we simply do not have a comprehensive and convincing theory of economic behaviour at the micro level. That is simply the state of our knowledge, and macroeconomics methodology must proceed from this state of imperfection. So there is no option but to try many different modelling choices — many more than have tended to be admitted to the good company of academic economics. But it is surely better for those modelling choices to be comparable inside an overall structure where their plausibility, predictive ability, match with the data and causal interpretation are all commensurate. A more expansive and liberal form of DSGE still looks to be the best arena in which this contest can play out.
- China’s mildly improving growth figures do not reflect the true strength of the economy’s pick-up — because fake statistics from previous years failed to represent the depth of the slowdown then.
- The UK’s National Audit Office has established that the true taxpayer cost of private-public partnerships is much higher than plain government borrowing, at least when they take the form of Britain’s private finance initiative.
In January 2018, a group of economists convened by Oxford published their reply to Queen Elizabeth’s question of why nobody in the profession saw the 2008 crisis coming. Free Lunch examined the “Rebuilding macroeconomic theory” project, assessing the recommendations on how macroeconomic theory should evolve to do a better job.
You can read the whole series below, as well as other articles by Martin Sandbu on this subject.