Researchers have long been interested in whether environmental regulations discourage investment, reduce labour demand, or alter patterns of international trade. But estimating those consequences of regulations requires devising a means of measuring their stringency empirically. While creating such a measure is often portrayed as a data-collection problem, we identify four fundamental conceptual obstacles, which we label multidimensionality, simultaneity, industrial composition, and capital vintage. We then describe the long history of attempts to measure environmental regulatory stringency, and assess their relative success in light of those obstacles. Finally, we propose a new measure of stringency that would be based on emissions data and could be constructed separately for different pollutants.