54th Congress of the European Regional Science Association: "Regional development & globalisation: Best practices", 26-29 August 2014, St. Petersburg, Russia
This paper presents an original econometric model for estimating a dose-response function through a regression approach when: (i) treatment is continuous, (ii) individuals may react heterogeneously to observable confounders, and (iii) selection-into-treatment may be potentially endogenous. After describing the model, two estimation procedures are suggested: one based on OLS under Conditional Mean Independence (or CMI), and one based on Instrumental-Variables (IV) under selection endogeneity. The paper goes on by presenting ctreatreg, a user-written Stata routine for an easy implementation of such a model, thereby performing a Monte Carlo experiment to test the reliability of the model and of its software implementation. Finally, an application to real data for assessing the effect of public R&D support on companies' R&D expenditure is presented and results briefly commented. The usefulness of such a model for program evaluation is clearly stressed.
treatment effects dose-response function continuous treatment Monte Carlo R&D support