Back to search

Athens University Economics and Business (AUEB)

Topics in bayesian econometrics

Abstract

dc:description

This thesis consists of four essays on the application of Bayesian econometrics to selected topics in economics. The first essay, proposes a new method of inference for Dynamic Stochastic General Equilibrium models, using a Bayesian semi-parametric approach. The method considers a set of moment conditions involving latent and observed variables and applies entropic latent variable integration via simulation using a least favorable distribution for the vector of the unobservables. A Bayesian blockwise empirical likelihood approach based on the Bayesian expansive block empirical likelihood posterior is used to provide inference for the parameters of the model. In the second essay, an approach of Bayesian inference for the Stochastic Ray Production Frontier and for technical inefficiency is proposed. The method combines the stochastic ray production function with equations derived from the first-order conditions of the related profit maximizing framework and suggests a Bayesian procedure for their joint estimation as a system. The third essay introduces technical inefficiency into the production function of a Dynamic Stochastic General Equilibrium model. Τechnical inefficiency is incorporated via a composed error term in which the first error term is a two-sided error component representing measurement noise and the second is a non-negative error expressing technical inefficiency. Inference for model parameters and technical inefficiency is performed via the Bayesian expansive block empirical likelihood posterior. In the fourth essay, the statistical and computational efficiency of a series of alternative sampling methods in simulating posterior densities derived from Dynamic Stochastic General Equilibrium models are compared. Five different sampling methods consisting of two different formulations of the random walk Metropolis-Hastings algorithm, the Hamiltonian Monte Carlo algorithm, the Metropolis adjusted Langevin algorithm and the simplified manifold Metropolis adjusted Langevin algorithm, are considered and compared against a series of performance criteria for models of different dimensionality.

Degree

thesis:*
Grantor dc:publisher
Athens University Economics and Business (AUEB)
Year dc:date
2020

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Paravalos, Evaggelos
  • Παράβαλος, Ευάγγελος

Subjects

dc:subject × 18

Rights

Language dc:language
eng

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/47699

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Paravalos, Evaggelos; Παράβαλος, Ευάγγελος. Topics in bayesian econometrics. Athens University Economics and Business (AUEB), 2020. http://hdl.handle.net/10442/hedi/47699