Essays in international market correlations

Modern finance theory is developed based on the concept of diversification among assets and asset classes as measured by correlations, and the concept of efficiency of financial markets in the sense of optimal risk-return trade-offs across assets and across time. In the area of international finance, the integration and efficiency of world markets have long been issues of debate. By focusing on the sources of international market correlations and the informational contents of correlations over different horizons, this dissertation sheds fight on these issues. As financial markets contain large amount of noise and sample correlations are highly unreliable, the structural model and the term structure analysis allow us to remove the noise and better estimate correlations among international markets.; The first chapter studies the geocultural, economic, and financial reasons for world market correlations. The explanatory powers of those factors are significant in the 1972–1997 period and have been increasing over time. While realized correlations fluctuate widely over time, the model predicted correlations have been quite stable and have been steadily increasing in the past 26 years. This indicates that world equity markets have become more correlated and more integrated over time. The model does better in predicting future correlations than historical sample correlations. The second chapter studies the impact of return autocorrelations on world equity market correlations over different holding periods. The term structure of correlations demonstrates that the degree of international diversification depends on investment horizons and the relative efficiencies of individual equity markets. Correlations tend to increase with holding periods on the short end and decrease with holding periods over multiple years. The third chapter studies the sensitivity of quanto option prices to correlations between foreign equities and currencies and demonstrates that quanto options are more valuable when the correlations are low. The fast approximation of implied correlations can be used for real time trading and empirical testing of the informational contents of implied correlations.