Kurt Mitman | Researcher or hack?


Sorry to be Kurt | Economics might be the “dismal science,” but it’s science nonetheless


kurt_web

Kurt Mitman
Sorry to be Kurt

Photo by Kurt Mitman


Last Thursday night, I gave a TED-style talk as part of TEDxGradfest. I spoke to the hundreds assembled about my research on unemployment, in which I find that most of the persistently high unemployment following the end of the Great Recession can be explained by unemployment benefit extensions.

While answering questions from several attendees at the subsequent reception, I was impressed (though not surprised) by how well people were able to grasp the subtle issues surrounding causality in the research and by the quality of questions that people asked.

I was not impressed (and equally not surprised) by a different strand of questioning that I received after the talk. Several people questioned my political ideology, my motives behind writing the paper and the conclusions that my co-authors and I have found. Unfortunately, as an economist, this is something that I face on a regular basis when talking about my research.

My research in economics falls into two broad categories: consumer credit markets and labor markets. The types of research questions that interest me have both positive and normative components.

Positive implications of a policy say what the effect is on outcomes. For example, in my research on the government bailout guarantees of Fannie Mae and Freddie Mac, we can say the government policy increased the foreclosure rate in the economy. Normative questions ask whether a policy is good or bad in terms of the “welfare” of the people and the economy. We can answer the question: Would people prefer to live in a world with a government guarantee or without one?

While research that only provides positive answers is interesting, part of the reason I was driven to do economic research was to make a difference and to be able to inform policy debate. The only pitfall, however, is that when your research begins to step into the fray of politics, your motives immediately start to come into question.

Of course, a student of logic knows that trying to attack someone’s argument based on personal ideology is an ad hominem fallacy. That doesn’t mean that it doesn’t get used all the time — even by Nobel-prize winning economists.

Paul Krugman, in a blog post from June 10, said that “people who say things like [unemployment benefits raise unemployment] are fundamentally confused about what economic research actually means.” The research with my co-authors, one a tenured professor at Penn, one an economist at the Federal Reserve Bank of New York and another a tenured professor at the University of Oslo find exactly that: The unemployment rate would have been much lower had unemployment benefits not been extended.

Apparently, my co-authors and I are fundamentally confused about what economic research actually means (even though we used measurement techniques highly praised by Krugman in the past to study the effects of minimum wages).

I’ll give Krugman a pass, however, because we hadn’t circulated our paper at that time, so he most likely wasn’t commenting directly on our work, though my co-authors and I have been presenting it all over the world since early this year.

The irony in all of this is that when I first started looking into the design of optimal unemployment insurance with 2012 School of Arts and Sciences graduate and current Amherst College professor Stan Rabinovich, we thought that the best thing to do would be to extend unemployment benefits during recessions.

Stan and I thought that because the availability of jobs was low during times of recession, that it would make sense to extend benefits, because there was no need to try to incentivize workers to search harder. But through the course of our research we stumbled upon another question: What if the availability of jobs — that is, firms’ decisions to hire — was directly affected by the extensions? That idea spawned a follow-up paper, which was the heart of my TED talk last week.

In economics research, we act as scientists — we let the data speak. Neither regression coefficients nor data change with the beliefs or ideology of the authors. You don’t have to trust the author, but numbers don’t lie.

Kurt Mitman is a 7th-year doctoral student from McLean, Va. His email address is kurt.mitman@gmail.com. Follow him @SorryToBeKurt. “Sorry To Be Kurt” appears every Friday.

Discussion

Comments powered by Disqus

Please note All comments are eligible for publication in The Daily Pennsylvanian.