ABSTRACT

Furthermore, the idiosyncratic shock of the “ground-zero” country is allowed to potentially infl uence the return of the second country over and above that captured by the common shock during periods of market turbulence. is describes the fact that pure contagion occurs when a country-specifi c shock becomes a global factor during a crisis. It is captured by augmenting the return equation of country 2 with the idiosyncratic shock of country 1 during the crisis period. is results in the following set of equations for the regime in which the idiosyncratic shock of the groundzero country experiences high volatility:

r z z

r z z z

= μ + σ + σ = μ + σ + σ + δσ (23.3)

A fi nal assumption of normality of the structural shocks enables us to estimate the full model via maximum likelihood along the lines of the methodology for Markov switching models (see Hamilton, 1989).*

23.2.2 Testing for Shift Contagion

Shi contagion occurs when the transmission of common shocks changes between regimes. To empirically test the null hypothesis of “no shift

contagion,” we conduct a likelihood ratio test specifying the null and alternative as follows:

H H σ σ σ σ

= ≠ σ σσ σ

(23.4)

e test statistic has a χ2 distribution with one degree of freedom corresponding to the restriction of equality of the ratio of coeffi cients between the two regimes.