Bayesian estimation for a semiparametric nonlinear volatility model

Research output: Contribution to journalArticleResearchpeer-review

Abstract

This paper presents a new volatility model which extends the nonstationary nonparametric volatility model of Han and Zhang (2012) by including an ARCH(1) component. This model also allows the errors to be independent and follow an unknown distribution. A Bayesian sampling algorithm is presented to estimate the ARCH coefficient and smoothing parameters. Empirical results show that the proposed model outperforms its competitors under several evaluation criteria.

Original languageEnglish
Pages (from-to)361-370
Number of pages10
JournalEconomic Modelling
Volume98
DOIs
Publication statusPublished - May 2021

Keywords

  • Backtesting
  • Cross-validation
  • Nadaraya-Watson estimator
  • Unknown error distribution
  • Value-at-risk

Cite this