Bayesian Markov Chain Monte Carlo for reparameterized Stochastic volatility models using Asian FX rates during...

Area of Science:

• Quantitative Finance
• Econometrics
• Statistical Modeling

Background:

• Stochastic Volatility (SV) models are crucial for financial time series analysis.
• Accurate estimation of SV models is challenging due to parameter autocorrelation.
• The Covid-19 pandemic heightened volatility in Asian FX markets.

Purpose of the Study:

• To improve Markov Chain Monte Carlo (MCMC) estimation for SV models.
• To reduce autocorrelation in SV model parameters.
• To incorporate heavy-tailed distributions for enhanced modeling of financial data.

Main Methods:

• Application of reparameterization techniques to the SV model.
• Integration of the Student-t distribution to induce heavy tails.
• Bayesian computation using MCMC samplers for parameter estimation.

Main Results:

• Successfully reduced autocorrelation in SV model parameters.
• Introduced a heavy-tailed SV model via Bayesian MCMC.
• Demonstrated improved estimation for volatile Asian FX series during Covid-19.

Conclusions:

• The proposed methods enhance MCMC estimation accuracy for SV models.
• Reparameterization and Student-t distributions are effective for volatile financial markets.
• This research provides a robust framework for analyzing FX markets during periods of high uncertainty.