Many variables of interest to comparative economists are difficult to measure or are unavailable for empirical research. To test for the effects of such variables, we offer a methodology based on examining the bias that is introduced when such variables are omitted from a model's estimation. Further, our methodology is able to use data that are available only for some countries, using a procedure to estimate missing observations. We apply the methodology to test whether the high level of environmental damage in East Europe can be explained by a tendency to export commodities intensive in pollution. We find no support for this hypothesis.