In this paper I offer an alternative identification assumption that allows one to test for changing patterns regarding the international propagation of shocks when endogenous variables, omitted variables, and heteroskedasticity are present in the data. Using this methodology, I demonstrate that...