Time-varying systematic risk in the stock exchange of Thailand : Evidence from multivariate garch and kalman filter estimates
Issued Date
2015
Issued Date (B.E.)
2558
Available Date
Copyright Date
Resource Type
Series
Edition
Language
eng
File Type
application/pdf
No. of Pages/File Size
98 leaves.
ISBN
ISSN
eISSN
Other identifier(s)
ba189093
Identifier(s)
Access Rights
Access Status
Rights
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Rights Holder(s)
Physical Location
National Institute of Development Administration. Library and Information Center
Bibliographic Citation
Citation
Muttalath Kridsadarat (2015). Time-varying systematic risk in the stock exchange of Thailand : Evidence from multivariate garch and kalman filter estimates. Retrieved from: http://repository.nida.ac.th/handle/662723737/3698.
Title
Time-varying systematic risk in the stock exchange of Thailand : Evidence from multivariate garch and kalman filter estimates
Alternative Title(s)
Author(s)
Advisor(s)
Editor(s)
item.page.dc.contrubutor.advisor
Advisor's email
Contributor(s)
Contributor(s)
Abstract
The purpose of this study was to use multivariate GARCH and the Kalman
filter to estimate the time-varying systematic risk or beta. Much research has found
that estimating systematic risk with a market model using the traditional regression
approach violated classical assumptions regarding both the stationary assumption and
independent identically distributed of the innovations. This study focuses on using
various models of multivariate GARCH and the Kalman filter to improve this beta
estimation. As the GARCH model is a popular model used in volatility clustering
data. The model allows for the forecasting of the variance of return to vary
systematically along periods. Further, the Kalman filter approach recursively
estimates the beta series from the time update function, which can create a series of
conditional betas that vary through time from the market model. Therefore,
systematic risk of the market can be estimated more precisely by using these two
models. The model also allows for conditional variances and conditional covariance
between individual portfolio returns, which in this study uses equity sector indexes
and the market portfolio returns to respond asymmetrically to past innovations,
depending on their sign as well.
The data used in this study were the daily return of the Thailand Stock Exchange Industries Group Indexes fromJanuary 2007 to June 2014. There are eight groups of equity sector indexes: agriculture and food industry, consumer products industry, financial sector, industrial sector, property and construction sector, service sector, and the technology sector.First, this study estimates the beta using ordinary least squaresin order to ascertain the characteristics of each sector in the Thai stock market. The study also found the ARCH effect and autocorrelation in traditional regression. Next, the study used the multivariate GARCH model in the VECH model and BEKK model specification to estimate time-varying beta. The results showed that all of the sector indexes revealed a time-varying variance. Moreover, the pattern of asymmetries in the covariance of returns was also found, which is evidence that covariance will be higher during a market decline. After that, three models of the Kalman filter, the random walk model, the random coefficient model, and the autoregressive model, were used to estimate the time-varying beta. The results showed that most of the equity sector indexes revealed a time-varying pattern with the Kalman filter model except for the consumer product industry. Moreover, the study also compared the forecasting accuracy among the models. In terms of in-sample forecasting, the multivariate GARCH VECH model performed the best among the models. In terms of out-sample forecasting, the results also confirmed that the multivariate GARCH VECH model and the Kalman filter AR(1) model were superior to rolling OLS according to lower MAE and RMSE. However, the evidence indicating which methodology is the best estimator between these two models is not clear.
This study contributes to financial participants a more precise estimation of systematic risk, which is one of the most important risks in the financial market, by using more proper methodologies, multivariate GARCH, and the Kalman filter framework. The results of the study can provide greater understanding of timevarying systematic risk that is useful information for investors and all financial market participants as well.
The data used in this study were the daily return of the Thailand Stock Exchange Industries Group Indexes fromJanuary 2007 to June 2014. There are eight groups of equity sector indexes: agriculture and food industry, consumer products industry, financial sector, industrial sector, property and construction sector, service sector, and the technology sector.First, this study estimates the beta using ordinary least squaresin order to ascertain the characteristics of each sector in the Thai stock market. The study also found the ARCH effect and autocorrelation in traditional regression. Next, the study used the multivariate GARCH model in the VECH model and BEKK model specification to estimate time-varying beta. The results showed that all of the sector indexes revealed a time-varying variance. Moreover, the pattern of asymmetries in the covariance of returns was also found, which is evidence that covariance will be higher during a market decline. After that, three models of the Kalman filter, the random walk model, the random coefficient model, and the autoregressive model, were used to estimate the time-varying beta. The results showed that most of the equity sector indexes revealed a time-varying pattern with the Kalman filter model except for the consumer product industry. Moreover, the study also compared the forecasting accuracy among the models. In terms of in-sample forecasting, the multivariate GARCH VECH model performed the best among the models. In terms of out-sample forecasting, the results also confirmed that the multivariate GARCH VECH model and the Kalman filter AR(1) model were superior to rolling OLS according to lower MAE and RMSE. However, the evidence indicating which methodology is the best estimator between these two models is not clear.
This study contributes to financial participants a more precise estimation of systematic risk, which is one of the most important risks in the financial market, by using more proper methodologies, multivariate GARCH, and the Kalman filter framework. The results of the study can provide greater understanding of timevarying systematic risk that is useful information for investors and all financial market participants as well.
Table of contents
Description
Dissertation (Ph.D. (Economics))--National Institute of Development Administration, 2015