Long-term vs. short-term comovements in stock markets: the use of Markov-switching multifractal models

摘要

During financial crises, interest is strong for analysing market comovements. However, a majority of these analyses is based only on correlations. This article uses Markov switching multifractal models to derive new indicators by considering different horizons for dependency among four stock indices (N YSE FTSE DAX CAC) between 1996 and 2008. The detection of crises, extreme volatility comovements or the co-cycle lengths are derived. In this context, September 2008 appears to be an unprecedented example of global crisis, extended to all horizons and markets.