Excess correlation of equity or bond returns. For example, under usual conditions we might observe a certain level of correlation of market returns. A period of contagion would be associated with much higher-than-expected correlation. Some examples are the conjectured contagion in East Asian markets beginning in July 1997 when the Thai currency devalued and the impact across many emerging markets of the Russian default. Contagion is difficult to identify because you need some sort of measure of the expected correlation. It is complicated because correlation's are known to change through time, for example, see Erb, Harvey and Viskanta's article in the 1994 Financial Analysts Journal. In periods of negative returns, correlation's (and volatility) are known to increase, so what might appear to be excessive may not be contagion.