Economics Graduate Students Should Learn about Unemployment
To be interesting and helpful to a broad group of graduate students, first-year macroeconomics courses should teach unemployment. More technical material could be pushed to the second year.
As discussed in a previous post, first-year macroeconomic courses have become optional in some graduate programs, and are at risk of being cancelled altogether elsewhere. This post tries to explain why these courses are facing so much pushback, and what how they could be revamped.
The defense of first-year macro courses
In defense of first-year macro courses, many macroeconomists argue that the courses teach students useful mathematical tools—such as dynamic programming, optimal control, and differential equations—which are not taught elsewhere. These courses are therefore needed to equip students with important tools.
While all this is true, the material can be taught in a week or two, for instance during a summer math camp. I taught this material in that format when I was at the LSE, and it worked. For anybody interested, the course material—lecture notes and homework—is available online. I think it’s hard to justify the existence of a one-year course sequence solely with dynamic programming.
What is actually taught in first-year macro?
The defense of first-year macro courses based on math tools is symptomatic of the problem with these courses: they do not teach macroeconomic content that’s interesting to non-macro students. An exception is the section on economic growth, which everybody usually enjoys. This is why most non-macro people (student and faculty) want to get rid of first-year macro courses in the first place.
It isn’t hard to see why the content covered in first-year macro is uninteresting for non-macro folks. A good fraction of the course is devoted to the Real Business-Cycle (RBC) model, in wich:
The economy is always efficient so business-cycles are costless.
Money is neutral so monetary policy plays no role.
In fact, since the economy is always efficient, stabilization policies (fiscal or monetary) have no reason to exist.
There is no unemployment, so even in bad times everybody who wants to work is able to find a job.
Because money is neutral, inflation is a non-issue: people don’t care about it at all.
It isn’t surprising that non-macro students are reluctant to learn all of this material, since it only has a flimsy connection to reality and will therefore not be useful to their own research.
Of course, it’s possible to add “frictions” of all sorts to the RBC model to remedy some of these limitations. But these make the model much more complicated, and they do not do justice to the phenomena that they aim to describe. Accordingly, the model with frictions is not very useful to non-macro folks either.
The most common “friction” added to the RBC model is a constraint that prevents firms from changing prices as often as they wish. The resulting New Keynesian (NK) model is taught in many graduate programs. While the model features somewhat-costly business cycles and monetary nonneutrality, it is widely reviled by non-macro graduate students, for good reasons:
There is no good justification for the constraint on price changes, which is cheekily referred to as “Calvo fairy” by students. (The fairy can be replaced by a “menu cost”, but that isn’t more realistic.)
The NK model is very complicated, which makes it hard to learn and renders it useless for other fields. (It usually takes an entire one-hour section just to derive the price index in the model.)
So non-macro students naturally wonder what is the point of all of this.
When should the RBC and NK models then be taught?
There is of course value in teaching RBC models and NK models, both because they are part of the history of macroeconomic thought and because they describe interesting mechanisms. Any student who aims to work for a central bank also needs to know the NK model inside-out, since it is the workhorse model there. But these models are probably better covered in second-year macro field courses, where all the students intend to do research in macroeconomics.
What can be taught in first-year macro instead?
If the RBC and NK models are moved to second-year courses, there will be a big void in the first-year macro curriculum. To fill this void, we should develop material that covers important macro questions and that could interest and be useful to a broad range of students—irrespective of their field.
In my view one quarter of the first-year macro sequence could be devoted to studying the labor market and unemployment. Although unemployment is the foremost problem associated with business cycles for workers, politicians, and policymakers, the topic is often never covered in first-year macro. This creates a huge blindspot for econ graduate students. The models and techniques covered in a course on unemployment can also be helpful across non-macro fields, from development to labor to public economics.
Last year I taught one quarter of the first-year macro sequence at Brown. The content of the course is freely available online: lecture videos, lecture notes, readings, and homework. And for those who only want to follow the lectures, there is a YouTube playlist that covers the entire course. The students were engaged and I think liked the course—so it might be possible to teach macro content that is valuable to all graduate students.
Such material could adequately fill part of first-year macro courses. It fulfills the two missions of first-year material:
The models and techniques covered in the course can be broadly applied, not only in macro but also in non-macro fields, from development to labor to public economics, as well as in economic theory and industrial organization.
The material might generate interesting research questions across fields—since unemployment affects so many facets of the economy and society.
Brief overview of the course and how it might benefit non-macro students
Here is a brief overview of what the course covered and how it could benefit non-macro students:
The course starts by reviewing labor market facts. This section might be interesting to any student with applied interests, since the labor market is an integral part of any economy.
Next, the course develops a matching model of the labor market. In the matching model, unlike in the neoclassical model, all trades are mediated by a matching function, which gives rise to unemployment. Matching models are helpful to describe markets in macro, labor, finance, or industrial organization.
Then we discuss wages. This part covers the labor market institutions that determine wages—unions, minimum wage, corporate policies. It also discusses various wage functions that can be used to describe wages in worker-firm relationships—such as fixed wages, rigid wages, and bargained wages. This section might be of particular interest to labor and theory students.
The next section covers unemployment fluctuations over the business cycle. This is more of a macro topic, although the analysis based on elasticities could be helpful to students in labor and public economics.
The following part decomposes unemployment into frictional and rationing components. This section shows how to build a labor market with a lack of jobs and simultaneously job search. Such a model has numerous applications, not only in macro but also in labor, public, urban, and development economics.
Next, the course introduces the concepts of efficient unemployment and unemployment gap. These concepts are particularly helpful in macro, labor, and public economics. And insofar as this section studies efficiency in matching markets, it might also be helpful to theorists.
The last two parts of the course cover labor-demand policies (including minimum wage and public employment) and unemployment insurance. These policies are heavily studied in labor and public economics. Such policies are also implemented in developing countries, so the material might be of interest to development economists too.
In teaching this course I became convinced that it is possible to teach macro content that non-macro students find interesting, and that might even be useful to them. This material departs from canonical macro material—especially the RBC and NK models. But it might be better to teach this material than having first-year macro courses be cancelled altogether.