Economist Noah Smith, assistant professor for Finance at Stony Brook University, recently wrote a post titled “Most of What You Learned in Econ 101 is Wrong” on Bloomberg view. He basically argues that standard introductory Economics textbooks—citing Mankiw’s “Principles of Economics”—solely refer to (outdated) basic models and do not incorporate advances in the profession. Although he did not choose the catchy title of the Bloomberg post himself, he claims that “… Mankiw’s book, like every introductory econ textbook I know of, has a big problem. Most of what’s in it is probably wrong.”
In reply, David Henderson, research fellow at the Hoover institution and associate professor of Economics at the Naval Post-Graduate School in Monterrey, defending the Econ 101 textbook points out: “Here’s what’s striking. In an article that purports to show that Mankiw is wrong on many issues, he [SMITH] doesn’t point out how he [MANKIW] is wrong on ANY issues.” To be fair, it should be mentioned here that Smith moderates in his reply: “… is that “right” and “wrong” are not very descriptive, helpful adjectives in this situation.”
In my opinion, Noah Smith is correct that “right” and “wrong” are problematic adjectives to evaluate basic theory in Economics. On the other hand, I agree with Henderson that Smith dropped some strong claims in the original Bloomberg view post but only supports them half-heartedly. The truth—if there is any—is, as so often, probably more moderate. In what follows I explain why and to which extent I agree with Smith’s opinion on introductory Econ classes and why not:
Why I disagree with Noah Smith
Known to every economist and valid as ever before, Joan Robinson once famously said “A model which took account of all the variegation of reality would be of no more use than a map at the scale of one to one.” I have frequently experienced students untrained in Economics to come to class and strive for the most complicated model there is to take into account and depict every eventuality of a real world situation. Every economist, who at least once attempted to write down a model, is perfectly aware of the fact that such an approach is frankly quite useless. Econ 101 ideally teaches students the ability to abstract from an incomprehensible environment to focus on questions of matter.
As an example for a stylized model N.S. mentions the basic minimum wage model, which predicts that the introduction of a lower price floor—the lowest wage employers may legally pay in a given market environment—leads to an increase in unemployment. Naturally, this (and similar stylized) model(s) do(es) and cannot apply to every real-world labor market. The validity of this model necessitates well-behaved preferences, rational behavior and an also otherwise friction-less environment. Moreover, it neglects various indirect benefits of minimum wages. While I deeply hope that decision makers do not base their actions solely on such a model, it is, in my opinion, still a valid and tremendous starting point for a comprehensive analysis. In addition, it serves an enormous educational purpose as a vehicle to teach students to abstract. This holds true for both theoretical and empirical analysis likewise unless we are eager to revert to ad-hoc estimation techniques.
Why I agree with Noah Smith
I disagree with N.S.’s claim “There’s no reason a college econ student shouldn’t learn how to run regressions in 101.” While teaching introductory regression to students both in Austria and at Northwestern I have experienced an eagerness of students to apply their newly acquired skills exorbitantly over-interpreting results which are far from robust. Thus, I do not believe that theory is necessarily to blame when Economics graduates in professional jobs over-interpret their stock of knowledge to make far-reaching decisions. Results from a basic regression analysis are equally likely as stylized models to be over-emphasized. An issue about which I entirely agree with Smith, if I interpret his contributions correctly, is the missing link between theory and empirics in basic Economics education. As of now, we are teaching stylized models in introductory classes and basic regression analysis parallel, often in the same quarter, not succeeding in highlighting their connection and necessary interplay for well-founded decision making.
I was fortunate enough to attend a class by Dale Mortensen on Labor Economics a couple of years ago. He picked five central theories of Labor Economics and devoted two classes to each of them. The first introduced a model rigorously on mathematical grounds whereas the follow-up class discussed stylized facts, available data, empirical contributions and the limitations of theory. Although I did not exactly end up being a Labor Economist, I have always admired the design of Dale’s class and plan to replicate his approach—when feasible—in the future. The course I am referring to, however, was a PhD elective and I am uncertain to which extent this interplay is replicable in freshman or sophomore classes. But I agree with Noah Smith’s discussion of theory and empirics in Econ 101 in so far as I believe that the missing link is the link itself.