Financial econometric analysis for investment decision using markov switching model

การวิเคราะห์เศรษฐมิติทางการเงินเพื่อการตัดสินใจลงทุนโดยใช้แบบจำลองมาร์คอฟสวิตชิง

Name: Kongliang Zhu

ThaSH: Econometrics

Classification :.DDC: 330.015195

ThaSH: Finance

ThaSH: Markov processes

Abstract: This thesis concentrates on introducing Markov Switching model for investment decision in financial markets. In this respect, it provides a regime-switching model to analyze the linkages between the movements of stock price, bond yield, market interest rate and exchange rate. The thesis presents Markov Switching approach with high dimensional copula in order to measure the dependency among these variables, and assess the role of stock investment on optimizing multi-asset portfolio returns. It also provides an approach to specify the trading signal for pairs trading strategy taking into consideration the structural change in the pairs return. This thesis consists of three issues. First of all, a review will be made on whether or not the spillover effects exist and take place across exchange rate, interest rate, government bond yield and stock return in different countries’ market. We estimate the correlations between these variables using the MS-VECM model. In addition, the obtained regime probabilities allow us to detect and identify the factor or event affecting the movement of the financial markets. Secondly, this thesis investigates portfolio risk structure for multi-asset allocation issue using a copula-based Markov Switching approach. The high dimensional copula is extended to Markov switching, making the model flexible and able to capture the economic behavior, which changes over time. The Conditional Value at Risk is taken into account in the economic change, thus it will be a more accurate risk measure than the conventional method, which is measured under the one dimension. The empirical results show that this proposed model could capture the financial behavior well since it could detect the two great crises in our samples. Thirdly, this study specifies the trading signal for stock pairs taking into account the structural change in the pairs return using Markov Switching Regression GARCH model. We applied our proposed model to the Stock Exchange of Thailand and the result shows our pairs trading strategy is relatively more effective for financial investment management compared with the single mean return from individual stock method. In addition, to estimate the parameters in the MS-VECM model and Markov Switching copula model, we select a Bayesian estimation technique because the computation in the conventional maximum likelihood function is difficult to reach the global maximum and is not easy to be converged in our case where we have a large number of unknown parameters in the model.

Chiang Mai University Library

Address: Chiang Mai

Email: cmulibref@cmu.ac.th

Name: Hung Nguyen

Role: Chairman

Name: Songsak Sriboonchitta

Role: Member

Name: Napat Harnpornchai

Role: Member

Name: Pathairat Pastpipatkul

Role: Member

Name: Jianxu Liu

Role: Member

Created: 2017

Modified: 2019-12-24

Issued: 2019-12-24

วิทยานิพนธ์/Thesis

application/pdf

CallNumber: Th 330.015195 K82F 2017

eng

DegreeName: Doctor of Philosophy

Level: Doctoral Degree

Descipline: Economics

Grantor: Chiang Mai University

©copyrights Chiang Mai University

ลำดับที่.	ชื่อแฟ้มข้อมูล	ขนาดแฟ้มข้อมูล	จำนวนเข้าถึง	วัน-เวลาเข้าถึงล่าสุด
1	econ81017klzh_full.pdf	11.95 MB	19	2026-05-26 18:16:01

ใช้เวลา

0.053769 วินาที