Identification of Nonlinear Dynamic Panels under Partial Stationarity

2023-12-30

AI Summary

This paper addresses identification in nonlinear dynamic panel models—such as binary choice and ordered response models—with restricted dependent variables, under partial stationarity assumptions that accommodate multiple lags of the dependent variable and various endogenous covariates. To overcome the strong distributional and temporal independence assumptions inherent in conventional approaches, the authors propose weakened partial stationarity conditions and develop a moment-based identification framework unifying partial and point identification. Monte Carlo simulations and empirical analysis using an ordered choice model demonstrate the method’s robustness in finite samples. An application to income-level classification further validates its practical relevance and estimation reliability. The core contribution lies in relaxing classical strict stationarity and parametric error-distribution assumptions, thereby establishing a more robust, flexible, and implementable identification foundation for dynamic discrete choice models.

Technology Category

• Reasoning under Uncertainty: Relational Probabilistic Models
• Search and Optimization: Mixed Discrete/Continuous Search
• Intelligent Robots: State Estimation

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems
• Economics, Online Markets and Human Computation: Sustainability of Web economics

Abstract

This paper studies identification for a wide range of nonlinear panel data models, including binary choice, ordered response, and other types of limited dependent variable models. Our approach accommodates dynamic models with any number of lagged dependent variables as well as other types of (potentially contemporary) endogeneity. Our identification strategy relies on a partial stationarity condition, which not only allows for an unknown distribution of errors but also for temporal dependencies in errors. We derive partial identification results under flexible model specifications and provide additional support conditions for point identification. We demonstrate the robust finite-sample performance of our approach using Monte Carlo simulations, and apply the approach to analyze the empirical application of income categories using various ordered choice models.

Problem

• Identifies nonlinear panel data models with partial stationarity
• Handles dynamic models with lagged variables and endogeneity
• Applies to binary choice and ordered response models

Innovation

• General identification for nonlinear panel models
• Partial stationarity condition for flexible specifications
• Robust finite-sample performance via Monte Carlo