Identification and Estimation of Semi-parametric Censored Dynamic Panel Data Models
2013/04/16
研討會日期 : 2013-04-16
時間 : 14:30
主講人 : Professor Ji-Liang Shiu
地點 : Conference Room B110
演講者簡介 : Ji-Liang Shiu received his Ph.D. in Economic from Johns Hopkins University in 2009. He is currently serving as Assistant Professor of Economics at National Chung-Cheng University. His research field is in microeconometrics, applied econometrics, applied microeconomics and labor economics.
演講摘要 : This study presents a semiparametric identification and estimation method for censored dynamic panel data models and their average partial effects using only two-period data. The proposed method transforms the semi-parametric specification of censored dynamic panel data models into a valid semi-parametric family of PDFs of observables without modeling the distribution of the initial condition. Then the censored dynamic panel data models can be identified by a standard maximum likelihood estimation (MLE). The identifying assumptions are related to the completeness of the families of known semiparametric PDFs corresponding to censored dynamic panel data models and observed conditional density functions between the dependent and explanatory variables. This study shows that the families of PDFs corresponding to dynamic tobit models and dynamic lognormal hurdle models satisfy the identification assumptions with two types of data generating process (DGP). This study proposes a sieve maximum likelihood estimator (sieve MLE) and investigates the finite sample properties of these sieve-based estimators through Monte Carlo analysis. This study presents the dynamic behavior of annual individual health expenditures estimated as an empirical illustration using the dynamic tobit model and data from the Medical Expenditure Panel Survey (MEPS).
