My job market paper is A Panel Data Model with Time-variant Heterogeneity: A Bayesian Treatment with an Application to the Translog Distance Function, in which a panel data model with unit-specific time effects is considered. In this paper, Bayesian inference and MCMC techniques are applied to implement the models. The finite-sample performance of this model is compared to that of alternative models in the literature dealing with time varying effects. An empirical application in analyzing U.S. largest banks efficiency levels are presented.
My job market paper is built on the co-worked paper with With Dr.Sickles and Dr. Tsionas: A Bayesian Treatments to Panel Data Models with an Application to Models of Productivity
A co-worked paper with Dr. Hulusi Inanoglu, Dr. Michael Jacobs, Jr. and Dr. Robin Sickles is Analyzing Bank Efficiency: Are “too-big-to-fail” Banks Efficient? In this paper, We use synthesized parametric and semi-parametric stochastic frontier models and Quantile Regression Method for panel data to estimate the efficiencies of top 50 banks in the U.S.
I have also engaged in the PEMEX project with Dr. Sickles and Dr. Ngyuen: A Study of Productivity and Efficiency in the Mexican Energy Industry: The Case of PEMEX. In this paper, we study the optimizing behavior of PEMEX by estimating the cost shares and undertake the estimation using duality between the cost and production function, which facilitates our specification. This approach allows us to find the cost shares under different levels of returns to scale. Our results indicate the presence of substantial distortions in cost shares. The suggestion is thus to increase the capital use and decrease the labor use to remove such distortions.
There is a working paper in progress with Dr. Sickles and Dr. Qian is Semi-nonparametric Methods in the Stochastic Frontier Model. In this paper, we assume the density function of the efficiency term vi is spanned by Laguerre polynomials. We use a finite order polynomial to approximate the efficiency term and apply maximum likelihood methods to estimate parameters and standard errors in the Stochastic Frontier model.