Document Type



Doctor of Philosophy (PhD)



First Advisor's Name

Prasad V. Bidarkota

First Advisor's Committee Title

Committee Chair

Second Advisor's Name

Brice V. Dupoyet

Second Advisor's Committee Title

Committee Member

Third Advisor's Name

Sheng Guo

Third Advisor's Committee Title

Committee Member

Fourth Advisor's Name

Kai Huang

Fourth Advisor's Committee Title

Committee Member

Fifth Advisor's Name

Cem Karayalcin

Fifth Advisor's Committee Title

Committee Member


Emerging markets, Volatility spillover, Markov switching, GARCH

Date of Defense



This dissertation investigates the dynamics of mean and volatility spillovers from the U.S. and three large (regional) Asia-Pacific stock markets to ten small (local) ones from June 2008 to May 2013.

After a brief introduction to the main purposes and contributions of my research in Chapter 1, I examine the impact of lagged American and regional returns on the local markets in Chapter 2. By building up a univariate autoregressive model and treating lagged U.S. and regional returns as exogenous variables, I find that the local markets have statistically significant exposure to lagged returns of their own and the U.S. market only. The empirical results suggest that lagged American returns have exerted considerable mean spillover impact upon most of the local markets, whereas the large Asia-Pacific markets involved in this study have few such impacts.

I study the linkage between the U.S. market and each of the regional markets in Chapter 3 by employing two specifications of the bivariate GARCH process—the BEKK and general dynamic covariance (DC) models—to capture common features of equity return data. Based on the results of carefully constructed diagnostic tests, the BEKK model is demonstrated to be more appropriate for the U.S.–China and U.S.–Japan cases, and the dynamic covariance model for the U.S.–Australia case.

In Chapter 4, I discuss time-varying correlation of a local market with the U.S. market and with each regional market by proposing three Markov-switching shock spillover models. A comparison of model performance is drawn based on a series of model selection criteria. In fourteen cases, the local market is found to be more sensitive to regional shocks. Disturbances from two regional markets account for a higher proportion of local variance than those of U.S. origin. I conclude that the regional center, although having little mean spillover effect upon the local markets, has become increasingly influential in volatility transmission. Possible extended studies in the future as well as main findings in the preceding chapters are summarized in Chapter 5.



Li Xu's dissertation_Latex (11067 kB)
Latex package of Li Xu's dissertation



Rights Statement

Rights Statement

In Copyright. URI:
This Item is protected by copyright and/or related rights. You are free to use this Item in any way that is permitted by the copyright and related rights legislation that applies to your use. For other uses you need to obtain permission from the rights-holder(s).