Abstract
dc:descriptionThis thesis consists of three main essays in Bayesian Econometrics. The first essay is about Bayesian Model Averaging with non conjugate priors with an application to growth regressions. Bayesian Model Averaging is a novel technique for model selection and model averaging with a lot of virtues in terms of predictiveness. Despite the vastness of the literature, most papers deal with the natural conjugate case for the parameters of the regressors, mainly because one can get analytical expression for the posterior moments and the marginal likelihood. In this essay I extend the existing literature taking into account non conjugate priors, by considering two limiting cases of a (non conjugate) multivariate student-t distribution. For the estimation of the posterior probabilities, moments, marginal likelihood and predictive density, I proceed to Laplace Method, as it is proposed by Tierney and Kadane (1986) and Tierney, Kass and Kadane (1989) and a multivariate version of Theorem 3, of the latter paper, is derived. The approximations are nicely, incorporated to the MC(3) algorithm. An application to growth regressions shows that the departure from the natural conjugate setting results to some striking differences regarding the model selection procedure. The predictive results are strong and support the choice of the non conjugate setting. The second essay deals with the issue of objective Bayesian analysis in dynamic panel models with arbitrary cross sectional and intertemporal dependence. The objectiveness of the analysis refers to the estimation of Jeffreys prior of the model parameters. I extend the work of Phillips (1991) in a dynamic panel setting and I show, as Phillips did in a time series context, that flat priors are not suitable in dynamic models to express ignorance. The arbitrariness of the structure of stochastic term deals with the fact that no functional assumptions are made for the error process. I allow for general covariance matrices Ω and Σ, denoting the correlation among the time periods and cross sectional units respectively. I estimate the marginal posterior distribution of the autoregressive parameter for several models and I proceed to a comparison with the flat prior case. My results are similar to those of Phillips (1991). In the the third essay, I propose a Bayesian most stringent test for serial correlation. I compare a number of tests for autocorrelation that exist in the literature with the Bayesian one, in terms of power and stringency for several sample sizes and number of regressors. I show that the minimization of the Bayesian Shortcoming with respect to the prior parameter(s) is equivalent to the maximization of Power with respect to the same parameter(s) and that under particular parametric settings, the resulting test is the Most Stringent. My findings show that for certain beta priors the Bayesian test attains very good overall performance in the alternative hypothesis parameter space. I also estimate the envelope power and under these certain prior distributions the proposed test is the most stringent.
Degree
thesis:*- Grantor dc:publisher
- Athens University Economics and Business (AUEB)
- Year dc:date
- 2016
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Tasiopoulos, Anastasios
- Τασιόπουλος, Αναστάσιος
Subjects
dc:subject × 24- Μέθοδοι επιλογής υποδείγματος
- Μη συζυγής εκ των προτέρων κατανομή
- Παλινδρομήσεις υποδειγμάτων μεγέθυνσης
- Δυναμικά υποδείγματα Πάνελ
- Διαστρωματική και διαχρονική εξάρτηση συνδιακύμανης
- Εκ των προτέρων κατανομή κατα Jeffreys
- Συνάρτηση ισχύος και Περιβάλλουσα
- Αυστηροί, Σημειακά βέλτιστοι και Τοπικά πιο ισχυροί στατιστικοί έλεγχοι
- Αυτοσυσχέτιση
- Bayesian model selection
- Non conjugate priors
- Growth regressions
- Dynamic panel models
- Cross sectional and Intertemporal dependence
- Jeffreys prior
- Power function and Envelope
- Stringent, Point Optimal and Locally most powerful tests
- Autocorrelation
- Οικονομικά και Επιχειρήσεις
- Κοινωνικές Επιστήμες
- Οικονομετρία
- Economics and Business
- Social Sciences
- Econometrics
Rights
dc:rights- Statement dc:rights
-
- BY
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/39223