Expected return and volatility spillover among the stock markets of the Philippines, ASEAN, China, Japan, EU and US: Multivariate GARCH-BEKK approach and its use in dynamic portfolio optimization

Date of Publication


Document Type

Master's Thesis

Degree Name

Master of Science in Financial Engineering


Ramon V. Del Rosario College of Business


Financial Management Department

Thesis Adviser

Ruperto Majuca


The study aims to examine the interactions of stock return movements and integration among the stock markets of the Philippines, ASEAN, China, Japan, European Union and the United States by determining and measuring the direction and magnitude of transmission of expected return and volatility of daily returns of their stock market indices from 2007 to 2012 using a six-variable generalized autoregressive conditional heteroscedasticity Baba, Engle, Kraft and Kroner (GARCH-BEKK) model. Based on the results, there is persistence of both ARCH and GARCH effects within the same stock markets. Expected return spillover transmits evidently from relatively bigger markets to smaller markets. The bidirectionality of the spillover of returns between China and Japan, ASEAN and Japan, Philippines and ASEAn suggests significant level of stock market integration in East Asian region. Cross-market volatility spillover is more prevalent compared to cross-market expected return spillover as shown by the number of significant spillover effects. The resulting variance-covariance matrix from the GARCH-BEKK model is applied to engineer Markowitz optimized portfolios theoretically composed of the stocks of the six markets. Comparing the risk of portfolios generated using conventional and GARCH-BEKK variance-covariance matrix while holding the expected return constant, the portfolios generated using GARCH-BEKK variance-covariance matrices generated lower risk for both 2012 and 2013. For 2012 and 2013, the optimal portfolio generating 0.0479% return has a standard deviation of 1.1744% and 1.2118% respectively using the conventional variance-covariance matrix while the optimal portfolio using GARCH-BEKK variance-covariance matrix has a standard deviation of 1.1721% and 1.0905%. This is a reduction of 0.0023% risk and 0.1213% risk for 2012 and 2013. A back-testing of the reward-to-risk performance of the portfolio generated using GARCH-BEKK variance-covariance matrix shows a higher ex-post Sharpe ratio compared to the portfolio generated using the traditional variance-covariance matrix (2.68 against 2.55).

Abstract Format






Accession Number


Shelf Location

Archives, The Learning Commons, 12F Henry Sy Sr. Hall

Physical Description

1 computer optical disc ; 4 3/4 in.

This document is currently not available here.