Identification in Binary Response Panel Data Models: Is Point-Identification More Common Than We Thought?

This paper investigates identification in binary response models with panel data. Conditioning on sufficient statistics can sometimes lead to a conditional maximum likelihood approach that can be used to identify and estimate the parameters of interest in such models. Unfortunately it is often difficult or impossible to find such sufficient statistics, and even if it is possible, the approach sometimes leads to conditional likelihoods that do not depend on some interesting parameters. Using a range of different data generating processes, this paper calculates the identified regions for parameters in panel data logit AR(2) and logit VAR(1) models for which it is not known whether the parameters are identified or not. We find that identification might be more common than was previously thought, and that the identified regions for non-identified objects may be small enough to be empirically useful.

[1]  T. Magnac.  Subsidised Training and Youth Employment: Distinguishing Unobserved Heterogeneity from State Dependence in Labour Market Histories , 2000 .

[2]  R. Blundell,et al.  Changes in the Distribution of Male and Female Wages Accounting for Employment Composition Using Bounds , 2004, SSRN Electronic Journal.

[3]  Georg Rasch,et al.  Probabilistic Models for Some Intelligence and Attainment Tests , 1981, The SAGE Encyclopedia of Research Design.

[4]  Peter Schmidt,et al.  Estimation of Models with Jointly Dependent Qualitative Variables: A Simultaneous Logit Approach , 1975 .

[5]  D. Cox.  The Regression Analysis of Binary Sequences , 2017 .

[6]  Ariel Pakes,et al.  Moment Inequalities for Multinomial Choice with Fixed Effects , 2016, Quantitative Economics.

[7]  Ekaterini Kyriazidou,et al.  Estimation of a Panel Data Sample Selection Model , 1997 .

[8]  S. Nickell,et al.  An Investigation into the Incidence and Dynamic Structure of Sickness and Unemployment in Britain, 1965-75. , 1985 .

[9]  Ivan A. Canay,et al.  Practical and theoretical advances in inference for partially identified models , 2016 .

[10]  Charles F. Manski,et al.  SEMIPARAMETRIC ANALYSIS OF RANDOM EFFECTS LINEAR MODELS FROM BINARY PANEL DATA , 1987 .

[11]  Philip A. Haile,et al.  Inference with an Incomplete Model of English Auctions , 2000, Journal of Political Economy.

[12]  Bo E. Honoré,et al.  Panel Data Discrete Choice Models with Lagged Dependent Variables , 2000 .

[13]  Thierry Magnac,et al.  Set Identification, Moment Restrictions and Inference , 2017 .

[14]  Gary Chamberlain,et al.  Longitudinal Analysis of Labor Market Data: Heterogeneity, omitted variable bias, and duration dependence , 1985 .

[15]  Victor Aguirregabiria,et al.  Sufficient Statistics for Unobserved Heterogeneity in Structural Dynamic Logit Models , 2018, Journal of Econometrics.

[16]  Bo E. Honoré,et al.  Trimmed LAD and Least Squares Estimation of Truncated and Censored Regression Models with Fixed Effects , 1992 .

[17]  E. B. Andersen,et al.  Asymptotic Properties of Conditional Maximum‐Likelihood Estimators , 1970 .