by Massimo Guidolin and Manuela Pedio
We use impulse response functions computed from linear and nonlinear, Markov switching models to investigate the strength of four alternative contagion channels. These are the flight-to-quality, flight-to-liquidity, risk premium, and correlated information channels. We study the differences among estimates and impulse response functions across linear and nonlinear models to identify and measure cross-asset contagion. An application to weekly Eurozone data for a 2007-2014 sample, reveals that a two-state Markov switching model shows accurately estimated but economically weak contagion effects in a crisis regime. These results are mainly explained by a flight-to-quality channel. Furthermore, we extend our analysis the analysis to investigate whether European market may be subject to contagion when exposed to external shocks, such as those originated from the US subprime crisis.
Keywords: Contagion channels, Markov switching models, vector autoregressions, impulse response function, flight-to-quality, flight-to-liquidity, risk premium. JEL codes: G12, E43, C32.
IGIER - Università Bocconi