A Bayesian semi-parametric estimation of the binary response model using Markov Chain Monte Carlo algorithms is proposed. The performances of the parametric and semi-parametric models are presented. The mean squared errors, receiver operating characteristic curve, and the marginal effect are used as the model selection criteria. The graphic processing computing is implemented to estimate the optimal bandwidth within the kernel density estimation for the semi-parametric model. Simulated data and Monte Carlo experiments show that unless the binary data is extremely unbalanced the semi-parametric and parametric models perform equally well. However, if the data is extremely unbalanced the maximum likelihood estimation does not converge whereas the Bayesian algorithms do. Finally, an application to evaluated the unemployment rate based on the PSID data is presented.

This dissertation consists of three essays evaluating model selection criteria in both sampling theory and Bayesian analysis. In chapter one, I compare the Bayesian model selection criteria (DIC, PDIC and MSEF) and the conditional Kolmogorov test for the spot asset pricing models (Vasicek and CIR models); MCMC and block Bootstrap methods are applied. In chapter two, I compare parametric and semiparametric methods for the binary response models. The comparison is made by model specifications, ROC area, and marginal effects. Monte Carlo simulation, quasi-maximum likelihood and kernel density methods are applied. In chapter three, I compare two bandwidths of the kernel density: the standard bandwidth and computationally optimized bandwidth. The computationally optimized bandwidth is obtained by using the graphic processing unit (GPU) that shortens the computational time.

This important collection brings together leading econometricians to discuss advances in the areas of the econometrics of panel data. The papers in this collection can be grouped into two categories. The first, which includes chapters by Amemiya, Baltagi, Arellano, Bover and Labeaga, primarily deal with different aspects of limited dependent variables and sample selectivity. The second group of papers, including those by Nerlove, Schmidt and Ahn, Kiviet, Davies and Lahiri, consider issues that arise in the estimation of dyanamic (possibly) heterogeneous panel data models. Overall, the contributors focus on the issues of simplifying complex real-world phenomena into easily generalisable inferences from individual outcomes. As the contributions of G. S. Maddala in the fields of limited dependent variables and panel data were particularly influential, it is a fitting tribute that this volume is dedicated to him.

Volume 40B of Advances in Econometrics examines innovations in stochastic frontier analysis, nonparametric and semiparametric modeling and estimation, A/B experiments, big-data analysis, and quantile regression.