Paper Comments: "Unobservable Selection and Coefficient Stability: Theory and Validation" (Emily Oster)

ABSTRACT: "A common heuristic for evaluating the problem of omitted variable bias in economics is to look at coefficient movements after inclusion of controls. The theory under which this is informative is one in which the selection on observables is proportional to selection on unobservables, an idea which is formalized in Altonji, Elder and Taber (2005). However, this connection is rarely made explicit and the underlying assumption is rarely tested. In this paper I first show how, under proportional selection, coefficient movements, along with movements in R-squared values, can be used to calculate a measure of omitted variable bias. I then undertake two validation exercises. First, I relate maternal behavior on child birth weight and IQ. Simple controlled regressions give misleading estimates; estimates adjusted with a proportional selection adjustment fit significantly better. Second, I match observational and randomized trial data for 29 relationships in public health. I show that on average bias-adjusted coefficients perform much better than simple controlled coefficients and I suggest that a simple form of this adjustment could dramatically improve inference in many public health contexts." (Read More)

COMMENTS: A quite simple, yet very useful, paper that deals with the omitted variable bias. The author suggests an alternative solution against the "artificial control" of coefficients. The results are even more useful because the author also suggests an estimator to evaluate if coefficients already estimated in previous papers by other authors are stable enough. While the applications tried by the author may be less appealing, further applications on more "aggressive" microeconometric o macroeconometric studies are quite possible and, thus, more intriguing.


Popular posts from this blog

Endogeneidad y Exogeneidad en los Modelos Econométricos [MARTIN, JM.]

"De los modelos ARDL - Primera Parte" (Traducción)

¿Qué es un Proceso Generador de Datos ARIMA(1,1)? [J.M. MARTIN]