Abstract
In this paper, we study the effect of a small ceteris paribus change in the marginal distribution of a binary covariate on some feature of the unconditional distribution of an outcome variable of interest. We show that the RIF regression techniques recently proposed by Firpo, Fortin, and Lemieux (2009) do not estimate this quantity. Moreover, we show that such parameters are in general only partially identified, and derive straightforward expressions for the identified set. The results are implemented in the context of an empirical application that studies the effect of union membership rates on the distribution of wages.
Keywords
unconditional partial effect; partial identification; unconditional quantile regression;
JEL codes
- C14: Semiparametric and Nonparametric Methods: General
- C31: Cross-Sectional Models • Spatial Models • Treatment Effect Models • Quantile Regressions • Social Interaction Models
Replaced by
Christoph Rothe, “Partial Distributional Policy Effects”, Econometrica, vol. 80, n. 5, September 2012, pp. 2269–2301.
Reference
Christoph Rothe, “Unconditional Partial Effects of Binary Covariates”, TSE Working Paper, n. 09-79, September 2009.
See also
Published in
TSE Working Paper, n. 09-79, September 2009