Essays on International Stock and Bond Returns

This thesis consists of three chapters on the dynamics of asset returns, with a focus on global stocks and bonds. The first chapter investigates the contagion effect between the European stock and bond markets, and between the Greek bond market and other European bond markets. The perspectives of nonlinear contagion effects and the predictability of contagion are also investigated in the first chapter. The main findings are as follows. Firstly, the European sovereign debt crisis generally leads to contagion effects between domestic stock and bond markets, and this is more likely in relatively smaller countries. The financial crisis had generally led to a higher level of flight-to-quality, whilst this has also been found over the tranquil period, especially in the relatively larger countries. Secondly, the contagion effect between the Greek and other European bond markets started appearing at least four months earlier than the beginning of the European debt crisis. Thirdly, strongly significant copula estimation results reinforce the findings of the existence of nonlinear contagion effect in the Eurozone area. In addition, the information asymmetry carried by the counterpart of the GJR model significantly increases the ability of the Student-t copula to detect changes of dependence structure. Finally, conditional volatility as an explanatory variable is found to be statistically significant in explaining and predicting the contagion across at least five countries, and the level of exchange rate shows its predictive power in contagion for at least four countries. The interest rate (the level of risk free rate for the Eurozone area) is found to have the weakest predictive power amongst all the explanatory variables considered. The second chapter examines the bi-directional relationships between stock returns and trading volume, and between trading volume and volatility. By using the nonlinear Granger causality test, we find the existence of both bi-directional relations between stock returns and trading volume, and between trading volume and volatility. Further to this, from limiting the sample period to the widely known tranquil period (1994 to 2006), an interesting result is found. In comparison to the full sample test, statistically significant nonlinear results are also observed from the tranquil period. However, the nonlinear feedback from stock returns to trading volume, and the nonlinear feedback from volatility to trading volume are shown to be much stronger during the tranquil sample period than the other way round. The third chapter evaluates the effects of fundamental factors on international stock returns. Dividend, earnings and interest rate are considered as fundamental factors. The results from the international stock markets are mixed: some markets see dividends playing a more significant role in explaining the variation of stock returns, and some markets see earnings playing a more significant role. However, neither dividend nor earnings can predict the returns changes in a few markets. In order to investigate this problem, we take one step further through estimating the effects of changes of interest rates upon dividend and earnings discount models. However, our analysis only finds a slight influence there. This suggests that other unexamined factors are more important, consequently, further research is required for clarification.