Why do states exit IOs? How often does IO exit happen? And what are the consequences of IO exit for leaving states? Despite recent attention to individual cases and the importance of membership in IOs, little is known about state exit from IOs across states, organizations, and time. Chapter 1 outlines the common logic of IO exit that links withdrawal and suspension: States often use IO exit as a strategy to negotiate institutional change when mechanisms of voice have failed. We summarize our empirical contributions that rely on a new dataset of IO exit across 198 states and 534 IOs from 1913 to 2022. We show that exit is infrequent, intermittent, and often temporary rather than terminal. Factors related to bargaining help predict IO exit, and exit generates negative reputational and cooperative consequences for leaving states. Nonetheless, IO exit is often an imperfect tool in achieving institutional change. Overall, we correct the view of IO exit as recently increasing. We also show that alternative arguments are not correct: IO exit is not widely occurring because of a backlash against globalization, nationalism/populism, IO authority, or legal rules. Moreover, exit is not inconsequential. We end with a roadmap for each chapter.

Chapter 3 outlines and tests our theory of IO exit by applying it to the predictors of IO withdrawal. We argue that many dissatisfied states use the process of withdrawal to broker deals for institutional change in the IO. Many withdrawals are driven by preference divergence from other member states or declining power. Using our IO Exit dataset, we analyze 387 IO withdrawals from 1913 to 2022 across 534 IOs and 198 states. In categorizing the reasons for state withdrawals, we show that two-thirds of IO withdrawals are motivated by the desire to negotiate change rather than by issues that reflect populism, nationalism, or capitulation toward international cooperation. States also use the threat of withdrawal, which supports the notion that exit is a negotiating process with multiple steps rather than a final or singular act. Withdrawal is usually not permanent; half of the time, states return to the IOs they left. States also likely consider costs a priori and avoid withdrawal if the costs are projected to be too high. This prevents many withdrawals from happening in the first place. We do not find consistent support for alternative arguments that backlash against globalization, encroachment from authoritative IOs, nationalism/populism, or legal rules are robust drivers of withdrawal.