To analyse the role that the World Trade Organization (WTO) plays in enforcing international trade agreements, this paper first explores what countries can achieve alone by characterizing optimal private trigger strategies ( PTS ) under which each country triggers a punishment phase by imposing an explicit tariff based on privately observed imperfect signals of the other country's concealed trade barriers. It identifies the condition under which countries can restrain the use of concealed barriers based on PTS and establishes that countries will not reduce the cooperative protection level to its minimum attainable level under the optimal PTS . This paper then considers third-party trigger strategies ( TTS ) under which the WTO allows each country to initiate a punishment phase based on the WTO's judgement about potential violations. By comparing the optimal PTS and optimal TTS , it demonstrates that the WTO facilitates a better cooperative equilibrium by changing the nature of punishment-triggering signals from private to public , which in turn enables countries to use a more efficient punishment, such as an asymmetric and a minimum punishment.

This paper uses recursive methods to characterize the payoff frontier of self-enforcing trade agreements between countries of asymmetric size. We show that at points on the frontier where only one country's incentive constraint binds, the efficient agreement will be a non-stationary one that starts with a positive trade distortion but eventually reaches free trade. Our analysis illustrates how (i) relative country size, (ii) consumption smoothing incentives, and (iii) sunk investments affect the form of efficient trade agreements. In contrast to previous work on gradualism, our results are obtained from a model in which the economic environment is stationary.

Antidumping (AD) trade protection policies allow government agencies to recalculate AD duties based on foreign firms' most recent pricing behavior. We examine the resulting dynamic pricing problem of a foreign firm facing such policy. We show that the expected pattern of AD duty recalculations over time crucially depends on the foreign firm's ex ante expectations of possible outcomes of AD policy enforcement. Our empirical analysis then confirms the role of ex ante expectations in explaining observed patterns of AD recalculations. Many of our model's results are applicable to other situations where enforcement of policy is tied to the subject's behavior.