Bayesian methods for economic survey data and its applications

点击翻译按钮获取中文摘要

In this research, I considered the Bayesian estimation method for economic survey data using Markov chain Monte Carlo (MCMC). First, I proposed the approach to Lorenz curve and Gini index estimation with contaminated data using Bayesian nonparametric analysis. The Bayesian nonparametric approach uses a Dirichlet process prior. I also presented a method of removing the contaminated observations.Next, I proposed the Bayesian mixture model with a finite number of normal distributions for estimation of Lorenz curves and Gini index from grouped data. I used Tomlinson's smoothing method and Stephens' birth-death process.Furthermore, I considered a problem of estimating two types of poverty, that is, chronic poverty and transient poverty, which were defined by Ravallion. I presented the Bayesian mixture model with errors-in-variable.I wrote three articles that were submitted to international journals. One of them was published in Journal of Econometrics. I also presented the articles at international conferences held in Australia and India.