Testing for contagion in international financial markets: which way to go?

This paper tests for the existence of contagion during the 1997/98 Asian crisis. We interpret contagion as a significant change in the way that countryspecific shocks are transmitted across international stock markets. Using the full-information framework of Favero and Giavazzi (2002) we find that the null hypothesis of no contagion is widely rejected. We also uncover evidence of an asymmetric transmission of shocks. Since our results contrast with those obtained by Rigobon (2001, 2002) using a limited-information methodology we present Monte Carlo simulations which show that certain necessary conditions must be satisfied for this method to have power. For parameter values in line with our econometric estimations we conclude that the power of the limitedinformation approach remains relatively low. JEL Classification: C3, F3, F4, G1