Our starting point is a post on Unlearning Economics, dated March 5, which argues that the flaws of mainstream economics contribute to lousy policy on several fronts: downplaying the role of monopoly, cheerleading for the shareholder value imperative in the corporate world, knee-jerk support for trade agreements under the banner of comparative advantage, and regressive macroeconomic policy, among others. A particularly pointed paragraph brought up the Reinhart-Rogoff 90% affair and accused the economics profession of dereliction of duty by not taking action to rebuke the wrongdoers:
Where was the formal, institutional denunciation of such a glaring error from the economics profession, and of the politicians who used it to justify their regressive policies?UE’s conclusion is that mainstream economics needs to be taken down several notches, which would open more space for alternative approaches to economics and, indeed, alternative approaches to policy that place more weight on human outcomes, broadly understood, than the formalistic criteria of efficiency, etc.
Simon Wren-Lewis responded by arguing that UE has it exactly backwards. Restricting himself to UE’s critique of macroeconomics, SWL says, yes, reactionary politicians have invoked “economics” to support austerity, but “real” economists for the most part have not gone along. True, there were a few, like Reinhart and Rogoff and those in the employ of the British financial sector (“City economists”) who took a public stand against sensible Keynesian policies in the wake of the financial crisis, but they were a minority, and, in any case, what would you want to do about them? Economists, like professionals in any field, will disagree sometimes, and having a centralized agency to enforce a false consensus would ultimately work against progressives and dissenters, not for them. Let’s put the blame where it really belongs, says SWL—on the politicians and pundits who have brushed aside decades of theoretical and empirical work to promulgate a reactionary, fact-free discourse on economic policy.
Yes—but, adds Brad DeLong. He largely agrees with SWL, but delves more deeply into the Reinhart-Rogoff affair. He shows that, even without the famed Excel glitch, a cursory look would reveal that R-R were trumpeting nonexistent results:
• The 90% debt cliff was an artifact of the way R-R set up their bins. Replace binning with a continuous relationship between growth and debt and the cliff disappears.
• The correlation between growth and debt supported no particular causal interpretation, and R-R provided no other evidence to support their particular causal argument.
• The correlation itself was so weak that the practical implication of R-R’s claim was nil. Fiscal stimulus that could make or break a recovery was being rejected on the basis of future economic growth effects that would be too small to measure.
So the R-R claim that fiscal consolidation was necessary and urgent was unfounded from the get-go, and these two were both respected mainstream economists, so what can we infer? DeLong’s takeaway is that economists do need to recognize that they operate in a political environment (the sewers of Romulus) in which their work will be seized upon by interested groups, with real practical outcomes. In this situation, the profession as a whole has a responsibility to assess high profile but dubious work. Although he isn’t explicit, my reading is that DeLong wants some sort of professional quality control, but not institutionalized in the way UE seems to call for.
I have a few points to make in addition to those already in the public stream (or sewer). First, I think that, apart from the specific value of particular economic theories, there is a giant problem in mainstream economics resulting from a false certainty that the overarching theory, the economy as an arena in which agents are constantly maximizing subject to constraint, with their interactions entirely mediated by markets, is correct. Believing they are operating within a correct framework, most economists grossly narrow the scope of the questions they allow themselves to ask. This shows up in virtually every applied area, from the economics of the family, to industrial organization, labor markets, environmental problems, etc. Just to take one example, consider the immense labor economists have put into estimating the social cost of carbon, as their primary instrument for “optimal” climate policy. It’s true that one needs such a magnitude if the goal is optimality, but why is that the goal? In the end, it’s because there’s no other model of decision-making: all agents are assumed to be constrained maximizers, and the public objective is simply to correct for the non-maximizing effects of an externality like greenhouse gas accumulation. You can’t even pose a question like, what are the risks that a climate catastrophe will destroy the foundations of human civilization? That’s not an assumption, that climate change will do this—it’s a question, which one could investigate. Start with the social cost of carbon framework, and you will be sure not to ask it, or dozens of other questions about the meaning of climate risk for fundamental human values. (Where in your integrated assessment models will you find the cost of cultural memory resulting from sea level rise and coastal inundation?)
The false certainty about core theory has in turn given rise to a pernicious tendency in econometrics to calibrate rather than actually test models. This is true almost by definition in most structural econometrics: a set of equations is derived from theory, and their parameters are estimated from an available dataset. This procedure makes sense if you know the structure is right, since you aren’t actually testing it. Incredibly, the bar is lowered still further, since many theories remain in circulation even when their structural estimations fail out of sample or are inconsistent with one another. There doesn’t exist in mainstream economics a culture of radical self-testing, since there is no professional cost to having one’s results disconfirmed by a subsequent study. Hey, we’re all just playing with our models, which is OK since they follow the proper rules, with maximizing agents and everything. The world of DSGE modeling is rife with this, but you’ll see the same thing in the micro world; it just hasn’t been called out as vociferously up to now.
So where do these biases come from? What would need to be changed in order for mainstream economics to be a reasonably reliable, self-correcting body of knowledge? I don’t have all the answers, but here’s a hypothesis: the simplistic, formulaic introductory economics course—the infamous Econ 101—has brainwashed not only generations of students for whom this is their only exposure to the field, but also generations of professional economists themselves. The so-called “economic way of thinking” dished out to neophytes serves equally as the way sophisticates understand their own field of study. The empirical basis for this claim is my own anecdotal experience working with economist colleagues in a number of educational and research institutions over the years. I might have this wrong, but hear me out.
The first causal mechanism is selection. After all, Econ 101 is where the light begins to glimmer in a few heads: hmmmm, maybe I should think about becoming an economist. A professor (or grad student) is going on and on about rational this and optimal that, illustrated with simple geometry, and the beauty of it seizes upon some small proportion of the students scribbling notes. Someone in this multitude is thinking, yeah, this is the real deal. And that someone goes on to become an economist.
The second is the power of discourse. Econ 101 introduces the framework and language economists use to think through the problems they’ll face further down the road. Yes, they will complexify and qualify it as their knowledge deepens, but the fundamental terms are set in place. Many of the framing effects are subtle: thinking of economic relationships solely as encapsulated in their moments of exchange, adopting a particular conception of rationality, etc. This is not to say that there exists some mentally liberated existence beyond discourse and framing—hardly. But mainstream economics is a discourse, and it has no self-awareness of it. This is a problem because, for many of the practical areas economists get involved in, there are other, competing discourses (management, political theory, psychology, cultural and historical theories) that economists can barely perceive, much less understand. These leads to the hermetic quality of economic practice that many have noted.
The third is the role of default assumptions. Economics is a giant, multifarious discipline with lots of detailed, arcane specialties. Some people spend their lives studying health care markets. Others the economic impacts of immigration. Or petroleum markets. Or the effects of cash transfer programs in low income countries. And so on. In their own fields of specialization economists often develop complex understandings of the forces at play, and cross-disciplinary applied work with colleagues from other backgrounds is becoming more common. (This contradicts what I just said about hermetic economics; it’s an evolving contradiction.) One of the stock arguments of those who deny the existence of any mainstream at all is that if you look at these subfields you will see well-recognized scholars wielding all sorts of hybrid models, and there is truth to this. But the process that gives us intensive specialization also gives us widespread ignorance of the specialization of others. The economist who burrows deeply into markets for health services has to rely on a hazy sense of the rest of the discipline and how her work connects with it. Where does that hazy sense come from? To some extent (yeah, this is quite a hedge) it comes from Econ 101 or at most the field surveys undergraduates take in addition to their core requirements. Consider a world in which there is a discipline like economics and four equally populated subdisciplines A, B, C and D. All economists take Econ 101, after which they specialize in their own subdiscipline. In 101 they learn that markets mostly work, agents are mostly rational, and economic policy is mainly about marginal tweaks to keep the machine humming. Then they dive into their subdisciplines. In A they learn that behavior is complicated, institutions matter, and markets are embedded, with contradictory impacts. It’s all messy and fascinating, and it keeps them busy. But the denizens of A believe, due to their initial training in Econ 101, that B, C and D, together the bulk of the discipline, are all basically variations on 101-ness, so they see their specific problems as exceptions to the general account of how the economy works. And the same is true of the B, C and D tribes. Each of them comes to understand how different their subdiscipline is from the norm laid forth in Econ 101, but also how these differences should be seen as exceptional. Thus, taken together, the economics profession would simultaneously be creative and heterodox in its day-to-day work and rigid and orthodox in its general view of the entire field and the principles that ought to govern it.
So this has gone on rather long, hasn’t it? My point is that, while SWL is right that there is a lot of work in economics that can be used as a resource for thinking clearly about making the world a better place, UE is right that mainstream economics is also a big part of the problem. I agree with DeLong that the profession needs to take more seriously the problem that there are few if any professional incentives that lead economists to scrutinize their own work lest it be subsequently disconfirmed. A false sense of having the correct overarching model, hammered out in Econ 101, pervades the entire field and undermines what ought to be the effects of the ongoing turn toward empiricism. From the header: mostly rot, some treatment, could be more.