Tuesday, October 9, 2012

Mathematics, Economics, & the Nobel Prize

With the announcement of this year's Nobel Prize in Economic Science less than a week away, here's a recent working paper that you'll surely enjoy: "The use of mathematics in economics and its effect on a scholar's academic career", by Miguel Espinosa, Carlos Rondon, and Mauricio Romero. (Be sure that you download the latest version - dated September 2012.)

Here's the abstract:
"There has been so much debate on the increasing use of formal mathematical methods in Economics. Although there are some studies tackling these issues, those use either a little amount of papers, a small amount of scholars or cover a short period of time. We try to overcome these challenges constructing a database characterizing the main socio-demographic and academic output of a survey of 438 scholars divided into three groups: Economics Nobel Prize winners; scholars awarded with at least one of six prestigious recognitions in Economics; and academic faculty randomly selected from the top twenty Economics departments worldwide. Our results provide concrete measures of mathematization in Economics by giving statistical evidence on the increasing trend of number of equations and econometric outputs per article. We also show that for each of these variables there have been four structural breaks and three of them have been increasing ones. Furthermore, we found that the training and use of mathematics has a positive correlation with the probability of winning a Nobel Prize in certain cases. It also appears that being an empirical researcher as measured by the average number of econometrics outputs per paper has a negative correlation with someone's academic career success." (Emphasis added; DG)
The first of the highlighted conclusions doesn't surprise me. I'm not sure that I like the second one, though!

© 2012, David E. Giles


  1. Well, the way I see it, it's relatively easy to run a regression, change the specification a bit, run another regression, ad infinitum. And more than a few papers, with tables and tables of regression coefficients, seem to follow that recipe..

    But it takes sharper, more critical thinking to figure out which regressions really mean anything, are valid beyond the sample used in estimation, or tell us something important we didn't already know or are at all relevant to anything other than a small group of academics in one obscure sub-discipline.

    So, consider it a folk theorem: the more regressions in a paper, the less important the paper really is.

    In that light, the negative correlation isn't as surprising.

  2. Thanks! But I suspect that most authors (and readers) of papers in (say) "Journal of Appled Econometrics" would disagree with you.

  3. Someone's work says a Nobel prize winner is more likely to come from Switzerland. Why? They consume more chocolate.
    Have fun:

  4. thanks for sharing..