Testing for contagion in US industry portfolios - a four-factor pricing approach

We conduct an empirical investigation into the financial contagion hypothesis in the context of 12 US industry portfolios. Using a four-factor asset pricing model we measure contagion as the excess co-movement between idiosyncratic portfolio shocks, and test for an increase in the frequency of contagion during the 2007-2009 crisis sub-sample. We find evidence of 22 instances of financial contagion during the noncrisis sample period, and 21 such occurrences during the 2007-2009 crisis period, at the 5% level. It appears that the frequency of contagion remained steady or declined during the crisis for the industries that had a relatively high frequency of contagion prior to the crisis, but increased for those industries that had relatively few such incidences. Interestingly, the financial sector exhibited the least number of contagion instances across both crisis and noncrisis periods.