A wavelet based approach to measure and manage contagion at different time scales

• We apply wavelet analysis to study tail dependence between US stocks.
• Based on decomposed time series we build portfolios that minimize short-run volatility.
• Stronger dependence between US stocks is not only present after the outbreak of financial crisis (2008) but also in the long run.
• Portfolios that minimize the short-run volatility outperform portfolio compositions that are based on raw return series.

We decompose financial return series of US stocks into different time scales with respect to different market regimes.First, we examine dependence structure of decomposed financial return series and analyze the impact of the current financial crisis on contagion and changing interdependencies as well as upper and lower tail dependence for different time scales.Second, we demonstrate to which extent the information of different time scales can be used in the context of portfolio management. As a result, minimizing the variance of short-run noise outperforms a portfolio that minimizes the variance of the return series.