The Strategic Nature of Compliance: An Empirical Evaluation of Law Implementation in the Central Monitoring System of the European Union Thomas K ¨ onig University of Mannheim Lars M ¨ ader University of Mannheim This compliance study models correct and timely implementation of policies in a multilevel system as a strategic game between a central monitoring agency and multiple implementers and evaluates statistically the empirical implications of this model. We test whether compliance is determined by the anticipated enforcement decision of the monitoring agency and whether this agency is responsive to the probability of enforcement success and the potential sanctioning costs produced by noncomplying implementers. Compared to other monitoring systems, the centralized monitoring system of the European Union (EU) is praised for exemplary effectiveness, but our findings reveal that the monitoring agency refrains from enforcing compliance when the probability of success is low, and the sanctioning costs are high. This results in a compliance deficit, even though the selective enforcement activities of the monitoring agency are almost always successful before court. T he enforcement of compliance by a central agency, which monitors lawmaking activities for imple- menting agreed policy goals and eventually lit- igates noncompliance by starting an infringement pro- ceeding, is an effective ex post control mechanism for combating violations in multilevel systems, thus “reduc- ing non-compliance to a temporal phenomenon” (Tall- berg 2002, 609). In addition to international agreements and treaties, which come into force through ex post ratifi- cation by the signatories (Calvert, McCubbins, and Wein- gast 1989; Chayes and Chayes 1993, 1995; Downs, Rocke, and Barsoom 1996; Putnam 1988; Simmons 1998, 2009), the goals of federal policies are implemented by lower- level local or regional lawmakers (Elazar 1987; Filippov, Ordeshook, and Shvetsova 2004; Lijphart 1999; Scharpf 1988). Multilevel systems are prone to collective action problems as the number of collaborating actors increases Thomas K¨ onig is Professor of Political Science II, University of Mannheim, PO Box 103462, D-68131 Mannheim, Germany (koenig@uni- mannheim.de). Lars M¨ ader is a Post-Doc at the Mannheim Centre for European Social Research (MZES), University of Mannheim, D-68131 Mannheim, Germany (lars.maeder@uni-mannheim.de). We are especially grateful to the AJPS editors and three anonymous reviewers. Replication data can be found at http://dvn.iq.harvard .edu/dvn/dv/ajps as well as at http://www.sowi.uni-mannheim.de/lspol2/englisch/08downloads01.html. 1 Even in restrictive standing systems as with the World Trade Organization (WTO), in which only governments can bring cases, a court ruling will allow a less informed public to observe noncompliance and in the event of unsatisfactory government behavior will impose sanctioning costs, which may increase compliance (Carrubba 2009; Dai 2005). (Scharpf 1988, 1997), which raises disputes on the opti- mal design of their material competencies (e.g., Alesina and Spolaore 1997; Boadway and Shah 2009; Frey and Eichenberger 2004; Weingast 1995). Because “free riding is the dominant strategy for large groups in the absence of a leviathan or of countervailing norms that can induce actors to monitor and punish defection” (Hooghe and Marks 2003, 239), compliance is a central challenge for the functioning of multilevel systems and may also impact the decision on their optimal design for lawmaking. One way to cope with this challenge is to establish a court which can punish defections and facilitate com- pliance with the agreed policy goals at the federal and international level (Carrubba 2005, 2009; Staton 2006; Vanberg 2004). This literature considers courts to act as a fire alarm by providing a venue where the actors can bring cases and signal possible violations. 1 For the American Journal of Political Science, Vol. 58, No. 1, January 2014, Pp. 246–263 C  2013, Midwest Political Science Association DOI: 10.1111/ajps.12038 246