Empirical studies on the trade-environment nexus that use panel data face two simultaneous challenges. One is associated with the potential presence of unobserved cross-country heterogeneity, while the other is due to the use of aggregate data. In this paper, we apply both the dynamic fixed effects and empirical iterative Bayes estimators to show that when country heterogeneity is accurately accounted for in the estimation, it is possible to obtain significant impacts of trade variables on the environment, even though we use aggregate data. Furthermore, using both the parameter estimates and indicators of stringency of environmental regulations, we show that at lower levels of stringency, the probability of having pollution-intensive foreign direct investments increases with a decrease in stringency, but the opposite is true for higher levels of regulatory stringency. This implies that pollution havens may exist only if environmental regulations are very lax or nonexistent.