International Asset Pricing, Diversification and Links between National Stock Markets

Abstract: This thesis consists of three self-contained empirical studies on international financial economics. The common economic theme is that all three studies deal with international stock market return and risk. The common econometric theme is that all estimations are performed using simulated annealing, a stochastic optimization algorithm in the Markov Chain Monte Carlo family, in an attempt to solve the well-known problems associated with multivariate regime-shift and GARCH models when using conventional optimization algorithms. The first chapter contains a short summary of the thesis together with a relatively non-technical description of simulated annealing. The second chapter uses a multivariate regime-switching model to investigate and endogenously date changes in return characteristics of the four largest Nordic stock markets. We find that the deregulated time-period, specifically after 1982, is associated with higher expected return, higher volatility, stronger links with international stock markets and higher correlation between the Nordic stock markets. We calculate reward to risk ratios, i.e. expected return per unit of volatility risk, for each Nordic stock market, and find that these ratios have not decreased after liberalization. We also argue that because of the much higher country-specific volatility in the deregulated period, the possibility to diversify both to other Nordic countries or internationally provides an attractive opportunity to lower portfolio risk without sacrificing return. The third chapter extends previous tests of the conditional CAPM using different asymmetric and non-diagonal multivariate GARCH-M specifications for eight large national markets and the world market simultaneously. We find that a model with double asymmetric effects and a time-varying price of world covariance risk supports all tested asset-pricing restrictions and that the previously often employed symmetric diagonal specification is overwhelmingly rejected. We find very limited support for the view that increased correlations have created less gain from world market diversification. Instead, investors from all investigated countries could expect statistically significant benefits from world market diversification. However, the evidence suggests that investors from all countries could expect statistically significant benefits from international diversification but that gains are considerable larger for investors with smaller home markets than for US and Japanese investors. The fourth chapter estimates a conditional version of the APT model simultaneously for 16 OECD national stock markets and 14 observable macroeconomic variables to investigate the contributions to risk-premiums from macroeconomic risk. We find that a non-macroeconomic factor interpreted as an international stock market factor and a macroeconomic factor interpreted as an US-dollar exchange rate factor contributes substantially to the risk-premiums for all stock markets. A macroeconomic factor interpreted as an international inflation factor is an important determinant of risk-premiums in many countries and a macroeconomic factor related to the slope of the US yield curve factor systematically contributes negatively to the risk-premiums. This factor seems to be closely related exactly to the periods of a downward sloping US yield curve. Our evidence also suggest that observable macroeconomic variables considered only explain a small portion of excess return variance, 13.7% on average while the non-macroeconomic factors together capture 38.7% of return variance.