This website uses cookies.

By clicking the "Accept" button or continuing to browse our site, you agree to first-party and session-only cookies being stored on your device to enhance site navigation and analyze site performance and traffic. For more information on our use of cookies, please see our Privacy Policy.

  1. Home
  2. ASSA Annual Meeting
  3. 2022
  4. Preliminary Program
  5. Nonlinear Panel Data Models
« Back to Results

Nonlinear Panel Data Models

Paper Session

Friday, Jan. 7, 2022 3:45 PM - 5:45 PM (EST)

Hosted By: Econometric Society
  • Chair: Xavier D'Haultfoeuille, Center for Research in Economics and Statistics

Dynamic Heterogenous Distribution Regression Panel Models, with an Application to Labor Income Processes

Ivan Fernandez-Val
,
Boston University
Wayne Gao
,
University of Pennsylvania
Yuan Liao
,
Rutgers University
Francis Vella
,
Georgetown University

Abstract

We consider dynamic distribution regression models for panel data with heterogeneous coefficients across individuals. The objects of interest are functionals of the coefficients including linear projections on individual covariates, cross sectional stationary distributions of the outcome variable of interest, and counterfactual distributions and effects of the outcome variable resulting from manipulating the initial conditions or covariates. The coefficients and their functional are estimated via fixed effect methods, which are debiased to deal with the incidental parameter problem. We propose a multiplier bootstrap scheme to carry out uniform inference on function-valued functionals. This scheme is computationally attractive because it avoids coefficient re-estimation, and is proven to be consistent for a large class of data generating processes, including the reference point where there is coefficient homogeneity. An empirical application using labor income data from the PSID shows that our model predicts significantly smaller effects of tax policies than traditional models based on homogeneous autoregressive processes. A counterfactual education experiment consisting on giving high school degrees to all individuals with less than 12 years of schooling results in an increase of the long run stationary distribution of labor income between 1 and 4\% at the bottom tail, but has no effect at the upper tail.

Time-Varying Linear Transformation Models with Fixed Effects and Endogeneity for Short Panels

Irene Botosaru
,
McMaster University
Chris Muris
,
McMaster University
Senay Sokullu
,
University of Bristol

Abstract

This paper considers a class of fixed-T nonlinear panel models with time-varying link function, fixed effects, and endogenous regressors. We establish sufficient conditions for the identification of the regression coefficients, the time-varying link function, the distribution of the counterfactual outcomes, and certain (time-varying) average partial effects. We propose estimators for these objects and study their asymptotic properties. We show the relevance of our model by estimating the effect of teaching practices on student attainment as measured by test scores on standardized tests in mathematics and science. We use data from the Trends in International Mathematics and Science Study, and show that both traditional and modern teaching practices have positive effects of similar magnitudes.

Identification and Estimation of Average Partial Effects in Semiparametric Binary Response Models

Laura Liu
,
Indiana University
Alexandre Poirier
,
Georgetown University
Ji-Liang Shiu
,
Jinan University

Abstract

Average partial effects (APEs) are generally not point-identified in binary response panel models with unrestricted unobserved heterogeneity. We show their point-identification under a index sufficiency assumption on the unobserved heterogeneity, even when the error distribution is unspecified. This assumption does not impose parametric restrictions on the unobserved heterogeneity. We then construct a three-step semiparametric estimator for the APE. In the first step, we estimate the common parameters using either conditional logit or smoothed maximum score. In the second step, we estimate the conditional distribution of the outcomes using the local polynomial regression, given regressors that depend on first-step estimates. In the third step, we average this conditional distribution over a subset of conditioning variables to obtain a partial mean which estimates the APE. We show that this proposed three-step APE estimator is consistent and asymptotically normal. We then evaluate its finite-sample properties in Monte Carlo simulations, and illustrate our estimator in a study of determinants' of married women's labor supply.

Discussant(s)
Koen Jochmans
,
University of Cambridge
Laura Liu
,
Indiana University
Ivan Fernandez-Val
,
Boston University
JEL Classifications
  • C2 - Single Equation Models; Single Variables