Maximum Pseudo Likelihood Estimation Based On Estimated Residuals In Copula Semiparametric Models

Author : Hee Won Lee
Publisher :
ISBN 13 : 9781109782172
Total Pages : 94 pages
Book Rating : 4.7/5 (821 download)

DOWNLOAD NOW!

Download or read book Copula, Simulated Likelihood, and VAR Estimation Based on MCMC written by Hee Won Lee and published by . This book was released on 2010 with total page 94 pages. Available in PDF, EPUB and Kindle. Book excerpt: My dissertation provides new simulation algorithms based on Markov chain Monte Carlo (MCMC) to estimate otherwise intractable models that are applicable in real world empirical problems. My work shows my interest in both theoretical and applied aspects of econometrics. In the first chapter, "Copula Analysis of Correlated Counts", I show that copula modeling enables the analysis of multivariate count data that has previously required imposition of potentially undesirable correlation restrictions or has limited attention to models with only a few outcomes. This article presents a method for analyzing correlated counts that is appealing because it retains well-known marginal distributions for each response while simultaneously allowing for flexible correlations among the outcomes. The proposed framework extends the applicability of the method to settings with high-dimensional outcomes and provides an efficient simulation method to generate the correlation matrix in a single step. Another open problem that is tackled is that of model comparison. In particular, the article presents techniques for estimating marginal likelihoods and Bayes factors in copula models. The methodology is implemented in a study of the joint behavior of four categories of U.S. technology patents. The results reveal that patent counts exhibit high levels of correlation among categories and that joint modeling is crucial for eliciting the interactions among the outcomes. The second chapter, "Review of Simulation Likelihood Estimation", reviews popular simulation methods for calculating the probability of the observed outcome in multivariate discrete choice problems. This probability, which is necessary to construct the likelihood function, involves numerically intractable integration so that simulation methods become indispensable in maximum likelihood estimation and Bayesian model choice. This paper describes two types of approaches. One is developed in classical statistics: the accept-reject, Stern, and GHK simulators. While all of the classical methods produce unbiased estimates of the probability of interest, the performance of the GHK method dominates other classical simulators, which explains why the GHK method has been widely used in empirical work. Another approach is described in the context of Bayesian simulation methods that are studied to derive the normalizing constant of the posterior density: the Chib and Ritter-Tanner methods. An extensive simulation study shows that the Bayesian methods can outperform the classical simulators. The third chapter, "Fluctuations of Land and Stock Prices in the Japanese Economy", examines whether innovations in stock and land prices have contributed to fluctuations in the Japanese economy. From a simple VAR model, the results suggest that land and stock prices are related to movements in real GDP, CPI, and the interest rate. Specifically, I find that shocks to the stock price have more immediate impact on real GDP. Shocks to the land price, which are strongly correlated to the stock market as a risk factor, have significant impacts on CPI even under the zero interest rate commitment. After the bubble burst, both prices were not found to have a significant interaction with interest rate.