This semester, I audited a class called “Political economy and its future,” taught jointly by Dani Rodrik (Kennedy School) and Roberto Unger (law school). The course examined interrelated contemporary issues – the global financial crisis, stalled productivity growth, authoritarianism and nationalism, increasing inequality, and (unexpectedly) the current pandemic – and asked what an alternative set of institutional arrangements might look like, and how we might begin to get there.
The exchanges between Rodrik, a more-or-less neoclassical (but not neoliberal!) economist, and Unger, a social theorist from a leftist tradition, were enlightening. The foreground theme of the course was alternatives to the market economy, but I found the background theme, a reorientation of the social sciences, even more interesting.
Unger’s critique of economics
Unger thinks the problems of economics are a variant on the fundamental problem with the contemporary social sciences: the “suppression of structural imagination,” which leads them to present as an explanation what is rather a “retrospective rationalization of established arrangements.”
He argues that the familiar criticisms of economics – its unrealistic assumptions about perfect competitition and rationality – miss the point. The real problem, he says, is that economics is not a causal science, as it sometimes purports to be, and is instead a form of logic. This kind of analytic apparatus with no need for facts cannot have a causal theory aside from that derived from “means-ends rationality.” It’s unfalsifiable, he argues – particular models can be tested and rejected, but there is nothing in the elementary schema that can be impugned. As I understand, this is a variant of the following critique: economic theory posits that what we see in the world is always the result of individuals doing the best they can subject to constraints, which is impossible to falsify – just change their preferences or change their constraints and you can explain anything.
But hasn’t economics become increasingly empirical? Sure, but that “copious flesh hangs awkwardly from a frail skeleton,” Unger says, by which I think he means that the causal engine is still the result of an unrealistic maximization problem: either with standard rationality assumptions, or imported from a different discipline (behavioral or neuroeconomics), or developed on the spot, thus highly context-dependent.
He argues there are three types of economics:
Pure economics, which is empty of institutional assumptions or commitments. This is exemplified by general equilibrium theory.
Fundamentalist economics, a la Hayek, which mistakenly associates the abstract idea of a market economy with a contingent, historically specific set of institutional arrangements – namely contract and property law.
Equivocating economics, exemplified by macroeconomics, which claims to study “law-like” relationships (like the Phillips curve) among large-scale aggregate variables against a murky institutional background. This economics recognizes that changes in the background matter (this is a nod to the Lucas critique, I think), but the background is static in the formal analysis, so institutional details are neatly sidestepped.
Which leads him to conclude:
To the extent this economics is rigorous, it is empty of institutional insight; to the extent it has institutional implications, it has them by virtue of equivocation and confusion.
He goes on to discuss Keynes as a “lesson in the dangers of partial intellectual rebellion.” He argues that Keynes intentionally focused on the politically acceptable aspect to his doctrine, boosting aggregate demand; his follower Hicks then developed the IS-LM scheme as a formulaic reduction of his doctrine; his American followers, led by Paul Samuelson, “downsized” the theory into a theory of countercyclical management, which was labeled “macroeconomics” and placed alongside existing economic theory, redubbed “microeconomics”. And then economists managed to rederive everything in the new theory by microfounding it, abandoning whatever was innovative in Keynes’s original thought.
Rodrik’s response was that this type of economics claims to do too much, that the field has moved from positing programmatic visions (Smith, Marx) to a discipline – a deductive, analytical tool, with which it is possible to distinguish positive and normative claims. He gives the examples of Chicago school economists versus “Krugman-type” liberals who combine the same tools of economics with very different normative judgements about inequality and views on the efficacy of government. He doesn’t want economics to attempt to come up with a universal framework for thinking about structural alternatives; that is rightly the job of democratic consensus.
Unger sees this as economists abdicating responsibility under the guise of intellectual modesty.
A postscript: I’ve never heard economics criticized for being insufficiently mathematical. Unger argues that math may or may not be useful to economic analysis, but not the current “toy mathematics that is simply the adornment of deductive reasoning,” which he says is “sterile as a methodological basis for thinking about transformation.” Indeed, optimization theory is not especially deep, and I agree with Unger that it’s not essential – it’s largely an aid to formal deduction. Economists would probably say that working through the math sometimes produces counterintuitive results; I just don’t think Unger is particularly interested in social science that explains counterintuitive empirical facts as the result of an optimization problem that takes institutions as given.