Financial vs. Non-financial Stocks: Time-varying Correlations and Risks

We analyze the time-varying co-movements of both financial and non-financial stock returns across countries to analyze the conditional correlation exhibited by cross-country pairs during the recent financial crisis. Using an asymmetric bivariate GARCH model, the analysis is conducted for a number of developed and developing countries. Given the origins of this current crisis, we expect increased correlation between financial sectors. However, recent correlations are not excessively large when compared to those earlier in this decade. Principal components analysis reveals one common driver of these pairwise correlations which may be related to U.S. returns and market liquidity