Unconditional Quantile Regression for Panel Data with Exogenous or Endogenous Regressors
Download eBook for Free
|PDF file||0.2 MB||
Use Adobe Acrobat Reader version 7.0 or higher for the best experience.
Download Support Files
Stata Code for the QRPD and IV-QRPD Estimators
Included in this zip archive is a .DO file that contains code intended to be used with the Stata statistics software, and a PDF file containing instructions for its use. This version of the Stata code and instructions was last updated on November 19, 2013.
|zip file||0.1 MB||
The file(s) provided above are ZIP-formatted archives, which most modern systems can natively unpack. If your computer does not unpack the archive when you double-click it, you may need to use a separate decompression program such as UnZip.
Quantile regression allows the impact of the explanatory variables to vary based on a nonseparable disturbance term. Panel data are frequently used in applied research because fixed effects control for unobserved heterogeneity and aid identification. The inclusion of fixed effects or use of a location-shift model in a quantile framework changes the interpretation of the estimates by separating the disturbance and including an additive fixed effect term. This paper introduces a quantile estimator for panel data which uses within-group variation for identification but allows the parameters of interest to be interpreted in the same manner as cross-sectional quantile estimates. The estimator maintains the nonseparable disturbance term. The fixed effects are never estimated or even specified and the estimator is consistent for small T. An IV version is also introduced. The estimation technique is straightforward to implement in standard statistical software. As an application of the estimator, the impact of anticipated income shocks on the distribution of consumption is estimated.