Identification and Estimation of Average Partial Effects in “Irregular” Correlated Random Coefficient Panel Data Models

摘要: In this paper we study identification and estimation of a correlated random coefficients (CRC) panel data model. The outcome interest varies linearly with vector endogenous regressors. on these regressors are heterogenous across units may covary them. We consider the average partial effect (APE) small change in regressor (cf. Chamberlain (1984), Wooldridge (2005a)). (1992) calculated semiparametric efficiency bound for APE our model proposed √N-consistent estimator. Nonsingularity APE's information bound, hence appropriateness Chamberlain's estimator, requires (i) time dimension (T) to strictly exceed number (p) (ii) strong conditions series properties vector. demonstrate irregular when T = p more persistent processes. Our approach exploits different identifying content subpopulations stayers—or whose values little periods—and movers—or substantially periods. propose feasible estimator based result characterize its large sample properties. While irregularity precludes from attaining parametric rates convergence, limiting distribution is normal inference straightforward conduct. Standard software be used compute point estimates standard errors. use methods estimate elasticity calorie consumption respect total outlay poor Nicaraguan households.