Limited Dependet Panel Data: a Bayesian Approach

Giuseppe Bruno ()

No 161, Computing in Economics and Finance 2005 from Society for Computational Economics

Abstract: Computing power now allows empirical researchers to use intensive computing estimation techniques with nonlinear panel-data models. Maximum Likelihood estimation is often cumbersome, if not analytically intractable, when dealing with such models. Even the simple calculation of the likelihood function can require a joint T-variate multiple integration whose numerical approximation can be poor. Different solutions have been proposed: integral approximation by simulation, the Generalized Method of Moments (GMM), and Markov Chain Monte Carlo (MCMC) algorithms. I examine these techniques using a software application employing Gibbs sampling and Metropolis-Hastings Markov chains. My aims are twofold: first, I assess the numerical reliability of standard econometric packages with nonlinear panel-data models, and second, I develop a posterior simulation for Tobit panel-data models in the presence of serial correlation, where high-dimensional integrals are induced by the serial correlation among censored variables. In this circumstance, although the standard Tobit estimator is consistent it will be inefficient. Building on Wei's work, I implement and test a sampling scheme for the unobserved (censored) data that allows effective Gibbs sampling to be used with the data augmentation algorithm. The Gibbs sampler includes a Metropolis-Hastings step to generate the posterior distribution of the serial correlation coefficient of the model

Keywords: Gibbs sampler; Econometric software; Metropolis-Hastings; Panel Tobit model; random effects. (search for similar items in EconPapers)
JEL-codes: C15 C23 C24 (search for similar items in EconPapers)
Date: 2005-11-11
References: Add references at CitEc
Citations:

There are no downloads for this item, see the EconPapers FAQ for hints about obtaining it.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:sce:scecf5:161

Access Statistics for this paper

More papers in Computing in Economics and Finance 2005 from Society for Computational Economics Contact information at EDIRC.
Bibliographic data for series maintained by Christopher F. Baum ().