We develop a game theoretic model of the conditions under which the European Court of Justice can be expected to take “adverse judgments” against European Union member governments and when the governments are likely to abide by these decisions. The model generates three hypotheses. First, the greater the clarity of EU case law precedent, the lesser the likelihood that the Court will tailor its decisions to the anticipated reactions of member governments. Second, the greater the domestic costs of an ECJ ruling to a litigant government, the lesser the likelihood that the litigant government will abide by it (and hence the lesser the likelihood that the Court will make such a ruling). Third, the greater the activism of the ECJ and the larger the number of member governments adversely affected by it, the greater the likelihood that responses by litigant governments will move from individual noncompliance to coordinated retaliation through new legislation or treaty revisions. These hypotheses are tested against three broad lines of case law central to ECJ jurisprudence: bans on agricultural imports, application of principles of equal treatment of the sexes to occupational pensions, and state liability for violation of EU law. The empirical analysis supports our view that though influenced by legal precedent, the ECJ also takes into account the anticipated reactions of member governments.