When addressing an externality such as air pollution, regulators can control policy inputs (e.g., pollution taxes and technology standards) or outputs (e.g., emission caps). Economists are familiar with this debate, known broadly as “prices vs. quantities,” but analysts of international environmental agreements have rarely focused sustained attention to such questions. Using an inventory of all international air pollution agreements, we document the historical patterns in instrument choice. Those agreements that require little effort beyond the status quo are usually codified in terms of effort, but agreements that require substantial actions by the parties nearly always deploy a cap on emission quantities as the central regulatory instrument. We suggest that this concentration of experience with emission caps and paucity of serious efforts to coordinate policy inputs may explain why the architects of international environmental agreements appear to believe that emission caps work best. We illustrate what's at stake with the example of international efforts to control the emissions that cause global climate change. We also show that the conventional history of the agreement that is most symbolic of the superiority of emission caps—the Montreal Protocol on Substances that Deplete the Ozone Layer—has wrongly overlooked a little-known provision that operates akin to a “price” instrument.