Small arms kill many more people than any other weapon. They are the weapons of choice in the vast majority of contemporary conflicts and are often used in crime and terrorism. In stark contrast to the devastating consequences of small arms proliferation, the international regulatory framework for the trade in small arms, established in 2001, is a weak non-legally binding instrument. To explain this puzzling discrepancy, I develop a theoretical model of government preferences on international regulation. The model offers several innovations over conventional models of trade policy, including the introduction of negative externalities, rather than protection, as the motivation for trade restrictions, and the incorporation of value-based humanitarian concerns. I operationalize the model with respect to small arms and test it empirically through an original survey of national views on the regulation of the small arms trade. The survey - based on interviewing officials from 118 countries - provides strong support for my theoretical predictions. Governments backing small arms exporters, especially those which are state-owned, favor weak international regulation of small arms. Nondemocratic governments, concerned that international controls might impede their access to arms, prefer weak international regulation as well. By contrast, governments facing high rates of gun violence and governments with humanitarian foreign policy concerns support strong regulation. The survey also reveals cross-regional variation. Africa, Latin America, and Europe support the international regulation of small arms; the Middle East as well as East and South Asia are less supportive of regulation. International control of small arms is, in fact, part of a broader phenomenon of international regulation, which the literature has overlooked. Beyond small arms, the theoretical framework developed in the paper can illuminate the political conflict over other internationally regulated goods, such as drugs and antiquities.