Greg Mankiw brings our attention to this piece by Gilles Raveaud, published in that paragon of thoughtful journalism, Adbusters, which critiques Mankiw's introductory textbook from the viewpoint of heterodox (or "post-austistic" as it is known in Europe) economics. Read the article for some idea of what the heterodox economists' concerns are. Here is more. In response, Mankiw does not directly address any of the criticisms, which I believe to be a mistake. This is a common theme for Mankiw. He'll link to pieces attacking him and argue that since he is criticized equally from the left and from the right, he must be saying the correct thing. (Dani Rodrik is also guilty of this fallacy.) This is plainly wrong since it only pushes aside the substantive issues at hand. Perhaps he does not have time to respond in full and so prefers to remain pithy, but in this case it seems negligent to do so, especially since I'm sure he has excellent responses.
My general rejoinder to the post-autistic economists is that their criticisms were far more valid 30 years ago than today. As someone (whose name I cannot remember) observed, the history of economics can be modeled by an hourglass. Before Samuelson's neo-Keynesian synthesis in 1945, economists addressed many different issues, but in what we might call "literary" fashion, that is, without quantitative tools. These tools were constructed during the 1950s, 1960s, and 1970s, but in doing so, the scope of economics was severely restricted. Since the late 1980s, however, the scope of economics has broadened each year. It has become a far more empirical field--in many subdisciplines it is no longer kosher to publish papers that would better be classified as pure mathematics. Freakonomics would not have been possible in 1980. In my view, nearly all of the post-autistic economists' proposals, short of economics becoming second-rate sociology, are already integrated into the current state of the field. To give an example, I went to a graduate student's talk at MIT a few weeks ago at which one of the professors present said "Well, in your model, agents are forward-looking and rational, but there's no a priori reason to assume that they aren't myopic." If that phrase can be spoken without the slightest hint of derision or irony in one of the top economics departments in the world, then the post-autistic economists have won before they have even begun. I see a post-post-autistic economics, one in which their particular critique has already been answered.
Thus, I am confused about why the post-autistic economists do not see things the same way. Do they not read leading journals? In the current issue of the American Economic Review, for example, there are at least three articles on behavioral economics, which seeks to integrate psychology and economics, an article that explains war through political biases and another that presents a theory in which trade agreements are driven by domestic lobbies, and an article on "the benefits of social networks in prisoner-of-war camps." Do they not attend seminars like the one I mentioned above? There seems to be a large identity effect. If you call yourself a heterodox economist, then you cannot participate in the mainstream discussion, even though you are talking about the same things. This distinction seems arbitrary, though it would seem too easy to shout "call yourself neoclassical!". It is possible that post-autistic economists were so digusted with the first-year graduate courses that they never bothered to go to, say, any labor economics seminars and see that you can write papers on basically anything you want (yes, even at the University of Chicago). Another explanation is that they do not wish to adhere to the standards of the field (producing testable predictions from mathematical theories, testing those implications with econometrics). If either of these are the case, I have no sympathy for the heterodox.
Enough ranting. I am happy to agree with the heterodox on a weaker point. Undergraduate education in economics is badly in need of reform. This is the bulk of what the Raveaud article discusses. Undergraduate education and, in particular, the introductory course no longer reflect the methodology and diverse substance of more advanced levels of mainstream economics. This is certainly the focus of Raveaud's criticism of Mankiw's textbook. One line in the article tells this story well:
Unemployment is an example of the market being imperfect. For Mankiw, if unemployment exists, it is only because of human inventions such as unemployment benefits, trade unions and minimum wages.
Ironically, Mankiw is the author of a theory arguing that menu costs--the small costs that firms incur when they change prices--are responsible for sticky prices and thus recessions and unemployment. But this does not appear in his own introductory textbook.
However, there is certainly some folly in the premise of Raveaud's article, which is that an
introductory textbook should present a discipline in all of its complexity and marginalia. Imagine a post-austistic physics student: "But, sir! Why are we learning Newtonian mechanics? This is ridiculous! Why is there no mention of Einstein?" The standard introductory textbook in economics is far more distant from the current practice of economics than it need be and should be updated. But an introductory textbook for a sequential discipline, in which one course builds upon another, must necessarily present what is easy in that field, with mere glimpses of the complications to come.
The flaw in my analogy with physics is that economists are typically presumed to have a responsibility to educate their students about the
real world, whereas physicists are not. This claim may seem backwards at first, but it is not. The belief is that students should leave an introductory economics classroom armed with tools to inform their interactions with the world (and many teachers of economics perpetrate this belief). A beginning physics student, on the other hand, is really only expected to know how to solve basic problems in physics: "What is the acceleration of such and such ideal object under such and such force?" not "How do you build an airplane?". Yet an economics student with the same amount of training might be asked "At what level should the government set the minimum wage?"
This presumption about economics education should be dropped, because economics is much more like physics in this respect than, say, political science or literature. That is, novices in physics are not expected to build airplanes and their counterparts in economics should not be given the idea that they can set macroeconomic policy. The post-autistic economists are half-wrong that economics is not a unified discipline: it
is unified with respect to its methodology (mathematical agent-based modeling and econometric testing of theories) but highly variegated with respect to the explanations of economic phenomena that it arrives at.
Raveaud makes this point for me:
Mankiw knows that the vast majority of his students are not going to become economics majors. He is not interested in training economists – his textbook is too simplistic to prepare a student for advanced study in economics. As he explicitly tells his teaching fellows, Mankiw’s interest is in shaping the minds of thousands of citizens and future leaders around the world.
I could imagine several different uses for an introductory course. One would be to uncritically introduce the student to the discipline in the style of a natural science course. But the current course does not do that. No one uses the simplistic theories found in Econ 101, though they may be useful heuristics. And most economics is conducted on the level of statistics now--even theorists must pay attention to the statistical relationships that their theories imply--yet there is usually no mention of statistics or econometrics in Econ 101. Another good use of the introductory course would be to present a wide variety of results from economics (necessarily simplified, of course) in the style of a first course in psychology, perhaps. But the current course has focused on theoretical results that have been around, in most cases, for at least a century.
(Another kind of course may examine the foundations of economics, like Stephen Marglin's course at Harvard, which is mentioned in Raveaud's article.
Here is the syllabus for the heterodox class introductory course at Harvard. The course does look very good and I think it's fantastic that such a course is taught. But, it is interesting to note that the course still uses Mankiw's textbook. Here is what Marglin has to say about that:
There is a choice of textbook ... How should you choose between the texts? Mankiw is a “conservative” text ... Baumol-Blinder is a “liberal” text ... I would say that Baumol-Blinder is more nuanced and careful in its defense of the market, but what is striking is the similarity between the texts, not the differences. This is not, after all, so surprising: the unity of the economics profession on the principles of economics transcends political differences that separate Republicans and Democrats.)
I'd personally like to see all three sorts of courses implemented. The current Mankiw-style course serves none of these purposes. It does not induct students into the discipline, it does not present recent and interesting results or debates, and it does not examine the foundations of economics. What, then, does it do? I don't quite agree with Raveaud that it merely serves as indoctrination into a conservo-libertarian neoliberal world order. But it is certainly harmful. Free trade, for example, is usually discussed in terms of comparative advantage. But there are equity effects that are large and well-known in the profession (the Stolper-Samuelson theorem, for example, proves that free trade negatively effects the less relatively abundant factor in a country). The standard theory about the minimum wage is that it increases unemployment, but there's a great deal of debate about that proposition in mainstream empirical labor economics. And the introductory treatment of macroeconomics is usually so simplistic that it cannot be used for any reasonable analysis. I do believe that the canon must be learned
before it can be critiqued. The introductory course does neither.