Have international pollution protocols made a difference?

Evaluating the effectiveness of international agreements is inherently difficult due to problems such as self-selection, spillovers, anticipation effects, and aggregate-level data. In this paper, I provide new and arguably more credible estimates on the effects of three major pollution protocols on SO2, NOx, and VOC emissions. I do so by combining a newly available global dataset on emissions dating back to 1970 with a generalized version of the synthetic control method. By constructing “synthetic” controls that mimic the pre-treatment development of each affected country, I mitigate bias caused by self-selection and non-parallel emission trends. The broader data coverage - both geographically and over time - allows me to examine the importance of spillovers and anticipation effects. Results from the estimation show that all three protocols induced emissions reductions well beyond a (synthetic) counterfactual development.