Degree Type

Creative Component

Semester of Graduation

Fall 2020

Department

Statistics

First Major Professor

Cindy Long Yu

Degree(s)

Master of Science (MS)

Major(s)

Statistics

Abstract

Stochastic volatility (SV) is known to be advantageous to capture important stylized features in real financial markets. Based on a time-discretized return dynamics model involving SV, we develop a Markov Chain Monte Carlo (MCMC) method for the estimation of the structure parameters and SV as the latent variable in the model. Simulation studies are conducted to test the performance of the MCMC method in terms of the convergence to the true parameters and volatility. Applying the method to the time series data of two distinct assets in the real markets, we empirically assess the inference that provides evidence that the mixture of the two assets helps to reduce volatility without compromising long-term returns. According to the model diagnostics, we find out that the inclusion of a jump in the model is likely to enhance the estimation.

Copyright Owner

Jang, Minsung

File Format

PDF

Embargo Period (admin only)

11-12-2020

1

Share

COinS