“KEN, THERE IS NO ECONOMIST WITH NUMERICAL EXPERTISE THAT MEETS ECONOMETRICA’S REQUIREMENTS TO BE AN ASSOCIATE EDITOR”

The Editor of Econometrica declared this to be the view of the Co-Editors of Econometrica regarding the community of economists that are trying to bring numerical methods to economic analysis.

This declaration came after over a year of e-mail exchanges between the Editors of Econometrica and me regarding the manner in which Econometrica treats computational issues. These exchanges focused on the mathematical errors made in published papers as well as the content of rejection decisions. In those discussions, not a single Co-Editor challenged any mathematical criticism I made. It was also true that not a single Editor had any interest in engaging in any discussion of the substantive questions I raised.

Others (not me) who were aware of these exchanges suggested to Econometrica that it appoint Associate Editors who had expertise in these computational methods. The Editor told me that he and his colleagues discussed this at their August, 2007, meeting but decided against it. When asked why, his response was that no one “met Econometrica’s requirements”.

When pressed, he said that he was worried about the number of submissions in that area. I said that perhaps Econometrica could take a leadership role in computational economics and appoint some appropriate Associate Editors even though there were not a large number of submissions in computational economics. The Editor said that Econometrica would not do that.

While the number of submissions in computational methods is surely part of any reasonable discussion of needed Associate Editors, this does not alter the fundamentally insulting nature of the declaration. I have been Co-Editor of a couple of journals. When we made associate editor appointments, we examined the range of expertise an individual had, not just rather there were enough submissions for him in one specific field. If Econometrica wanted to appoint an Associate Editor with expertise in numerical and computational methods, they could appoint someone who could also handle papers in some other areas of economics so that his total workload met Econometrica requirements.

Therefore, it is clear that, in their opinion, there is no economist who has numerical expertise who also has the necessary expertise to handle Econometrica submissions in macro or IO or general equilibrium theory or game theory or econometrics or any other type of economics paper suitable for Econometrica.

The claim that there are not enough submissions is also questionable. Take, for example, the lead article of 2007. It is based on the premise that standard numerical methods are impractical for large constrained optimization problems. However, the “problems” that were listed had been addressed successfully in the last 40 years in the math programming literature. It is also clear that many economists were aware of these developments in the past, and even referred to them in Econometrica papers in the late 1980’s and early 1990’s. Apparently no current Co-Editor is aware of this recent work. While the lead article of 2007 may have been primarily an econometrics paper, an editor might also have sent it to an Associate Editor with numerical expertise to check out the paper’s claims. Also, a paper later in 2007 had a numerical section of questionable validity. Again, while it was primarily a game theory paper, it would have been useful to have it examined by someone with numerical expertise. Right there we have two papers out of the less than fifty papers published in 2007 that might have benefited from being examined by someone with numerical expertise.

These comments are not meant to advocate the appointment of one associate editor with computational expertise. One such appointment would smack of tokenism unless they could find someone who knows numerical methods for optimization and complementarity and variational inequalities and approximation and quadrature and differential equations and convex analysis and algebraic geometry and nonlinear equations and regular perturbation and singular perturbation and nonlinear functional analysis, and is familiar with how they can be used to solve problems in static game theory, dynamic game theory, general equilibrium, macroeconomics, public finance, and econometrics. Econometrica might have a difficult time finding such a person, but the lack of such people is at least in part due to the fact that Econometrica Editors, in their editorial roles and in their roles as professors, have seen to it that there is little incentive for anyone to gain expertise in even a small fraction of these topics.

The basic problem is systemic, not solved by minor gestures. In the same conversation, I gave the Editor the names of four individuals who were quite capable of evaluating much of the numerical work being done in economics today. Since that time I have heard of two cases where Econometrica expressed disdain for the idea that computational work should be taken seriously; it is highly unlikely that any one of those four people would have been involved in the editorial process for those two submissions, just as it is unlikely that they would have been involved in any of the other cases I have discussed. In that conversation, I also argued that, whatever the number of submissions were, Econometrica could use AE appointments as a way to take a leadership role in solving these problems. The Editor said they were not interested in doing this.

Unfortunately, this kind of dismissal of the computational economics community is, in my experience, typical of how Econometrica treats computational issues and the people who want to advance computational techniques in economics. Other pages in this website document this by reporting my communications with Econometrica on various matters. While these comments are critical, no where do I argue that Econometrica made a bad decision in terms of the final outcome. I do not know the opportunity cost of Econometrica pages nor the value of my submission relative to others, and how it compares to other work in terms of advancing Econometrica’s agenda. That is private information known only to the editorial board of Econometrica. My comments focus on mathematical and computational matters, and where the editors and referees of Econometrica appear to be unaware of the basic literature on numerical methods, even those closely related to economics. Even when one believes that his submission has little chance of accepted, he still hopes to benefit from constructive and informed criticism that helps him improve his work. I have not benefited from any such feedback from Econometrica in many years.

One can conclude from Editor’s declaration that there currently are no Associate Editors (or Co-Editors) who have expertise in numerical and computational methods. I also infer that the same can be said for the editorial boards of the two new journals, Theoretical Economics and Quantitative Economics, since they will have the same standards as Econometrica. Some of the work I am currently finishing is closely related to work done by individuals who are Editors at the Econometric Society journals; the fact that these individuals would likely handle this work makes it irrational for me to submit that work to any of those journals.

This is not a call for a general boycott of any sort. I will continue to do referee reports for Econometrica; I just finished two this week. In general, it would be irrational for me, given my professional situation and alternative uses of time, to submit any of my sole-authored work to any of the Econometric Society journals. As I have in the past, I will usually defer to junior collaborators about where our joint work is sent because they are in different professional situations and face different trade-offs. Otherwise, I have better things to do with my time than deal with the attitude Econometrica has towards computational and numerical work.