Mcmc application to stochastic volatility model

Abstract

The main objective of my Masters thesis is to study the Dynamic structure that return prices or exchange rates exhibits. Intensive studies have been conducted for inferring the parameters and missing information. [1] paper studies this for ASEAN (Association of South East Asian Nations) markets. To study this I have chosen SV model as it mimics most of the stylized facts that exchange rates show. Also I have applied a Bayesian computation approach to infer the parameters associated with the SV model. I have chosen MCMC technique as the main approach since it solves some rigorous calculation issues which other techniques cannot overcome. Our data series includes exchange rates of Indian Rupees (INR) with United States Dollars (USD), and the period covers the crises of the ASEAN markets in 1997. Most of the part of thesis includes understanding of the basic and advanced concepts involved while applying MCMC technique to SV model. I have also studied MCMC application to some other models, which I have not mentioned in this thesis, like Geometric Brownian model etc, but I have concentrated my studies on SV model for the application of MCMC. Finally I have estimated the parameters involved in SV model by producing results from my own written MATLAB codes. The results produced are quite expected because of high level of persistence involved in the data. Also, the properties of MCMC can be quite easily visible from the graphs shown in the results section. It would be interesting for further research to come up with more time-effective codes.