Title: Behavioral, fundamentals and market determinants of the correlation between asset classes
Authors: Benoit Sevi - Nantes Universite (France)
Yannick Le Pen - Universite Paris Dauphine (France) [presenting]
Abstract: The correlations between stock, bond and commodity returns are of utmost importance for asset allocation but are also informative about the flight-to-quality issue. We estimate these correlations for a sample of US indices with the DCC-MIDAS model. This model distinguished between the long-run and a short-run component in the dynamics of conditional correlation. We extend the initial DCC-MIDAS to include an asymmetry effect of negative shocks on conditional standard deviations and correlations. In the next step we apply the definition of impulse response function in nonlinear models. In particular, we compute correlation impulse response functions (CIRF) in periods of high and low volatility as well as in periods of market downturn and upturn.