Category: Uncategorized

…without careful arguments on substantive grounds, lagged explanatory variables should never be used for identification
purposes.

This is from a new working paper of mine, coauthored with Cornell’s Tom Pepinsky, an associate professor of government, and Taka Masaki, a PhD student in government, on “lag identification.” This is a very common strategy to avoid problems of endogeneity or reverse causality, especially in political science, although table 1 in our paper shows that this is not exactly uncommon in economics.

We expose the assumption that underlies this strategy, which we term “no dynamics among unobservables.” We also argue that this assumption is almost impossible to defend on substantive grounds, because it requires knowledge about the time series properties of a variable which is unobserved.

We are fairly certain that the germ of the idea for our paper was this post by Phil Arena.

Indeed. Prevalent in poli sci. RT @TomPepinsky: Where is my "lagged IVs because endogeneity"? http://t.co/H9ofBmkhS0 cc @filarena

— Marc F. Bellemare (@marcfbellemare) February 13, 2014

(This post is being simultaneously published today on Tom’s blog.)

Once upon a time In the fall of 2013, I took my paper with Lindsey Novak and Tara Steinmetz on female genital cutting (FGC) on the road. I presented it at invited seminars and at development conferences here and there and received a lot of good feedback before submitting it for publication.

In March of last year, the Journal of Development Economics invited us to revise and resubmit it for publication. One of the best comments we received, however, was that we were only looking at Senegal and The Gambia for the most recent years available, and since there existed comparable data sets for most of West Africa over several years, why not expand the analysis to cover West Africa, thereby gaining in external validity?

And so Lindsey got to work on cleaning 36 additional country-year data sets, i.e., all of the available Demographic and Health Surveys and Multiple Indicators Cluster Surveys for West Africa, for all available years, which included questions about FGC. This occupied us Lindsey for the better part of this last year (and because the data sets are repeated cross-sections which vary a bit in how they were collected both between and within countries, this generated an almost 200-page appendix), but we finally have a revised version of the paper.

An interesting new article (gated) titled “Agricultural Economics and Economics: Influence and Counter-Influence, 1910–1960” by Marc Nerlove (yes, the same Marc Nerlove who won the Clark Medal and made seminal contributions to economics and econometrics) in the latest issue of Applied Economic Perspectives and Policy:

This paper examines the effects which agricultural economists had on developments in economics and on economic policies during the years 1910-1960. These developments are viewed against the backdrop of what was happening in the economy at large and particularly with respect to farmers, and of major books and articles in economics published in the period. Some ideas of agricultural economists and the influence they had are discussed. Agricultural issues and agricultural economists, in the broad sense, had a profound influence on the development of general economics and on economic policies during this period.

The article covers many of the contributions of agricultural economists to economics. Nerlove’s contribution is obviously mentioned, but he also discusses Theodore W. Schultz, who won the Nobel prize for economics in 1979 for his work on agricultural development, among others.

After reading Mastering ‘Metrics last month, however, I am surprised by Nerlove’s omission of the contribution of Philip G. Wright, an agricultural economist who, in collaboration with his son Sewall Wright, derived the instrumental variables estimator in an 1928 book titled The Tariff on Animal and Vegetable Oils.

Similarly, had Nerlove’s article not stopped in 1960, he would have had to mention Yair Mundlak who, for all intents and purposes, derived the fixed effects estimator in a 1961 article in the Journal of Farm Economics–which has since been renamed the American Journal of Agricultural Economics.