Abstract EN:

The first part of the dissertation deals with the issue of portfolio diversification in an international context and the integration of international financial markets. In chapter i. We show that the variance of a portfolio can be substantially reduced through diversification, which results in a decrease of non systematic risk. In chapter 2. It is empirically proved that international diversification is more efficient than domestic diversification. Because correlation coefficients between markets are relatively low. The impact of exchange risk on the total risk of an international portfolio is measured in chapter 3. Different ways of eliminating exchange risk are explored, and the concept of global risk is introduced. We conclude that the reduction of global risk through international portfolio diversification cannot be measured due to the difficulty to define an international systematic risk. The international systematic risk can be measured only in an integrated financial market. In chapter 4, the concept of integration is discussed and several methodologies to test it are presented. We reach the conclusion that the integration hypothesis can be neither accepted nor rejected. In the second part, we attempt to test the international market integration hypothesis by studying the structure of relationships between world markets. After testing normality and independance through time of daily and monthly returns of 8 stock market indexes. The stability of correlation coefficients is tested: the results show that stability is a function of the length of the interval used to compute the returns. This can be explained by the existence of an underlying structure of markets, which is stable on the long term, but deviates from equilibrium on short periods due to random shocks.