Doing Economics

Comments. As I mentioned above, focus on what will have the highest
possible returns at the margin. This means you should avoid discussing the
quality of the writing, or details that are either cosmetic or stylistic. Do not
mention typos (almost every working paper in existence has them, and
more than a few published papers do as well).
Ask yourself: Is there another available identification strategy which
would allow the authors to examine the same research question and assess
the robustness of their results? Do you know of another data set that would
allow answering the same research question in order to improve the external

---

validity of the authors’ results? Can the authors get to the same result by
relaxing an assumption, thereby making their work more general? Can you
think of better motivations—meaning either more general ones or simply
more convincing ones—for their work than what they currently have in the
paper? Can you think of ways to simplify the authors’ argument, their
theoretical framework, or their empirical framework? All of these higher-
level, higher-return comments are fair game when you are a discussant.
One way to think about serving a discussant is the following: if you had
to write a referee report about this paper, what would your major comments
be—those things you would want the authors to do before you could
recommend the authors be given a chance at revising and resubmitting or
accepted?
Some advice about discussing a paper it to pick your best three
comments about the paper and stopping there (see, for instance, Blattman
2010), because anything after your top three comments will not be as good.
While I understand that reasoning, serving as a discussant is an act of
service, and if you have a number of constructive comments for the authors,
you should not hide your light under a bushel, and you should aim to
mention all those comments in one sitting. Relatedly, serving as a
discussant is not “gotcha” journalism. Your job is not to find a fatal flaw in
what the author has done, but rather to help improve the work. If there is a
fatal flaw, the best way to let the author know is in private conversation
after your discussion of their paper.
In no case should serving as a discussant be about you. If you can
formulate a set of brilliant, game-changing remarks on a paper, you should
certainly do it. Though there is no shortage of brilliance in the economics
profession, there is often a shortage of generosity. If you really wish to
stand out, be generous and constructive—whether you are serving as
discussant or performing some other professional task.