Stock market cointegration in Europe

Abstract: In this paper we try to uncover long-run dependency structures by testing for cointegration among six major European stock indices during the first eleven years and four months of the 21st century. We find strong support for one cointegrating vector, indicating that the national stock markets follow the same long-run growth path, and that no market is likely to have substantially greater or worse relative performance over time. This can be attributed to the wide-spread economic and financial integration of the markets, e.g. through the EU or EMU. Further, the cointegration relationship means that traditional risk diversification techniques based on covariance lose their effectiveness in the long run. We find the that U.K. is weakly exogeneous and will be the first market to be hit by events that do not affect all countries simultaneously, but that later spread to the others. In addition, we investigate short-run causal relationships, to determine which markets have a significant short-term effect on the others, and determine how the effects of different events will spread in the region.