Hendrickson responded recently to criticisms of Econ 101 made by Noah Smith and Mark Thoma. Mark Thoma thinks that Econ 101 has a conservative bias, presumably because Econ 101 teaches students that markets equilibrate supply and demand and allocate resources to their highest valued use and that sort of thing. If markets are so wonderful, then shouldn’t we keep hands off the market and let things take care of themselves? Noah Smith is especially upset that Econ 101, slighting the ambiguous evidence that minimum-wage laws actually do increase unemployment, is too focused on theory and pays too little attention to empirical techniques.
I sympathize with Josh defense of Econ 101, and I think he makes a good point that there is nothing in Econ 101 that quantifies the effect on unemployment of minimum-wage legislation, so that the disconnect between theory and evidence isn’t as stark as Noah suggests. Josh also emphasizes, properly, that whatever the effect of an increase in the minimum wage implied by economic theory, that implication by itself can’t tell us whether the minimum wage should be raised. An ought statement can’t be derived from an is statement. Philosophers are not as uniformly in agreement about the positive-normative distinction as they used to be, but I am old-fashioned enough to think that it’s still valid. If there is a conservative bias in Econ 101, the problem is not Econ 101; the problem is bad teaching.
Having said all that, however, I don’t think that Josh’s defense addresses the real problems with Econ 101. Noah Smith’s complaints about the implied opposition of Econ 101 to minimum-wage legislation and Mark Thoma’s about the conservative bias of Econ 101 are symptoms of a deeper problem with Econ 101, a problem inherent in the current state of economic theory, and unlikely to go away any time soon.
The deeper problem that I think underlies much of the criticism of Econ 101 is the fragility of its essential propositions. These propositions, what Paul Samuelson misguidedly called “meaningful theorems” are deducible from the basic postulates of utility maximization and wealth maximization by applying the method of comparative statics. Not only are the propositions based on questionable psychological assumptions, the comparative-statics method imposes further restrictive assumptions designed to isolate a single purely theoretical relationship. The assumptions aren’t just the kind of simplifications necessary for the theoretical models of any empirical science to be applicable to the real world, they subvert the powerful logic used to derive those implications. It’s not just that the assumptions may not be fully consistent with the conditions actually observed, but the implications of the model are themselves highly sensitive to those assumptions. The meaningful theorems themselves are very sensitive to the assumptions of the model.
The bread and butter of Econ 101 is the microeconomic theory of market adjustment in which price and quantity adjust to equilibrate what consumers demand with what suppliers produce. This is the partial-equilibrium analysis derived from Alfred Marshall, and gradually perfected in the 1920s and 1930s after Marshall’s death with the development of the theories of the firm, and perfect and imperfect competition. As I have pointed out before in a number of posts just as macroeconomics depends on microfoundations, microeconomics depends on macrofoundations (e.g. here and here). All partial-equilibrium analysis relies on the – usually implicit — assumption that all markets but the single market under analysis are in equilibrium. Without that assumption, it is logically impossible to derive any of Samuelson’s meaningful theorems, and the logical necessity of microeconomics is severely compromised.
The underlying idea is very simple. Samuelson’s meaningful theorems are meant to isolate the effect of a change in a single parameter on a particular endogenous variable in an economic system. The only way to isolate the effect of the parameter on the variable is to start from an equilibrium state in which the system is, as it were, at rest. A small (aka infinitesimal) change in the parameter induces an adjustment in the equilibrium, and a comparison of the small change in the variable of interest between the new equilibrium and the old equilibrium relative to the parameter change identifies the underlying relationship between the variable and the parameter, all else being held constant. If the analysis did not start from equilibrium, then the effect of the parameter change on the variable could not be isolated, because the variable would be changing for reasons having nothing to do with the parameter change, making it impossible to isolate the pure effect of the parameter change on the variable of interest.
Not only must the exercise start from an equilibrium state, the equilibrium must be at least locally stable, so that the posited small parameter change doesn’t cause the system to gravitate towards another equilibrium — the usual assumption of a unique equilibrium being an assumption to ensure tractability rather than a deduction from any plausible assumptions – or simply veer off on some explosive or indeterminate path.
Even aside from all these restrictive assumptions, the standard partial-equilibrium analysis is restricted to markets that can be assumed to be very small relative to the entire system. For small markets, it is safe to assume that the small changes in the single market under analysis will have sufficiently small effects on all the other markets in the economy that the induced effects on all the other markets from the change in the market of interest have a negligible feedback effect on the market of interest.
But the partial-equilibrium method surely breaks down when the market under analysis is a market that is large relative to the entire economy, like, shall we say, the market for labor. The feedback effects are simply too strong for the small-market assumptions underlying the partial-equilibrium analysis to be satisfied by the labor market. But even aside from the size issue, the essence of the partial-equilibrium method is the assumption that all markets other than the market under analysis are in equilibrium. But the very assumption that the labor market is not in equilibrium renders the partial-equilibrium assumption that all other markets are in equilibrium untenable. I would suggest that the proper way to think about what Keynes was trying, not necessarily successfully, to do in the General Theory when discussing nominal wage cuts as a way to reduce unemployment is to view that discussion as a critique of using the partial-equilibrium method to analyze a state of general unemployment, as opposed to a situation in which unemployment is confined to a particular occupation or a particular geographic area.
So the question naturally arises: If the logical basis of Econ 101 is as flimsy as I have been suggesting, should we stop teaching Econ 101? My answer is an emphatic, but qualified, no. Econ 101 is the distillation of almost a century and a half of rigorous thought about how to analyze human behavior. What we have come up with so far is very imperfect, but it is still the most effective tool we have for systematically thinking about human conduct and its consequences, especially its unintended consequences. But we should be more forthright about its limitations and the nature of the assumptions that underlie the analysis. We should also be more aware of the logical gaps between the theory – Samuelson’s meaningful theorems — and the applications of the theory.
In fact, many meaningful theorems are consistently corroborated by statistical tests, presumably because observations by and large occur when the economy operates in the neighborhood of a general equililbrium and feedback effect are small, so that the extraneous forces – other than those derived from theory – impinge on actual observations more or less randomly, and thus don’t significantly distort the predicted relationship. And undoubtedly there are also cases in which the random effects overwhelm the theoretically identified relationships, preventing the relationships from being identified statistically, at least when the number of observations is relatively small as is usually the case with economic data. But we should also acknowledge that the theoretically predicted relationships may simply not hold in the real world, because the extreme conditions required for the predicted partial-equilibrium relationships to hold – near-equilibrium conditions and the absence of feedback effects – may often not be satisfied.