Template-Type: ReDIF-Paper 1.0
Series: Tinbergen Institute Discussion Papers
Creation-Date: 1999-11-02
Number: 99-082/4
Author-Name: Luc Bauwens
Author-Workplace-Name: CORE, Université Catholique de Louvain
Author-Name: Charles S. Bos
Author-Email: cbos@few.eur.nl
Author-Workplace-Name: Erasmus University Rotterdam
Author-Name: Herman K. van Dijk
Author-Email: hkvandijk@few.eur.nl
Author-Workplace-Name: Econometric Institute, Erasmus University Rotterdam
Title: Adaptive Polar Sampling with an Application to a Bayes Measure of Value-at-Risk
Abstract: Adaptive Polar Sampling (APS) is proposed as a Markov chain Monte Carlomethod for Bayesian analysis of models with ill-behaved posteriordistributions. In order to sample efficiently from such a distribution,a location-scale transformation and a transformation to polarcoordinates are used. After the transformation to polar coordinates, aMetropolis-Hastings algorithm is applied to sample directions and,conditionally on these, distances are generated by inverting the CDF.A sequential procedure is applied to update the location and scale.Tested on a set of canonical models that feature nearnon-identifiability, strong correlation, and bimodality, APS comparesfavourably with the standard Metropolis-Hastings sampler in terms ofparsimony and robustness. APS is applied within a Bayesian analysisof a GARCH-mixture model which is used for the evaluation of theValue-at-Risk of the return of the Dow Jones stock index.
Classification-JEL: C11; C15; C63
Keywords: Markov chain Monte Carlo; simulation; polar coordinates; GARCH; ill-behaved posterior; Value-at-Risk
File-Url: https://papers.tinbergen.nl/99082.pdf
File-Format: application/pdf
File-Size: 693248 bytes
Handle: RePEc:tin:wpaper:19990082