Identifying contagion effects during periods of financial crisis is known to be complicated by the changing volatility of asset returns during periods of stress. To untangle this we propose a GARCH (generalized autoregressive conditional heteroskedasticity) common features approach, where systemic risk emerges from a common factor source (or indeed multiple factor sources) with contagion evident through possible changes in the factor loadings relating to the common factor(s). Within a portfolio mimicking factor framework this can be identified using moment conditions. We use this framework to identify contagion in three illustrations involving both single and multiple factor specifications: to the Asian currency markets in 1997-1998, to US sectoral equity indices in 2007-2009 and to the CDS (credit default swap) market during the European sovereign debt crisis of 2010-2013. The results reveal the extent to which contagion effects may be masked by not accounting for the sources of changed volatility apparent in simple measures such as correlation.
