Last week, I finally had the opportunity to read Lisa Martin’s recent piece on compliance entitled Against Compliance. Prof. Martin meticulously evaluates the literature on compliance and concludes that compliance is the wrong dependent variable to uncover the causal effect of international law on state behavior. Martin’s review of compliance studies is comprehensive and her analysis is sound. She correctly identifies the conceptual and empirical problems that continue to trouble compliance research. Against Compliance prompted me to rethink long and hard about compliance, and stars aligned to make me consider Martin’s arguments all week along. Lisa Martin makes an important contribution, and I do agree with some of her points. But I am not ready to give up on compliance.
Martin argues that compliance tells us nothing about treaty implementation or effectiveness. Soon after I read this criticism, I got hit with the early syllabi posting requirement. Faculty are required to post representative syllabi on the first day of the registration period. I complied. But what does compliance mean? My impression is that have a good number of folks put TBAs on their early syllabi because they have yet to figure out the readings and assignments. Incomplete syllabi indicate poor implementation. Students at best get a vague idea about the course. I am also skeptical about the effectiveness of this rule. Which plays a larger role in students’ decision-making? Early syllabi or online professor reviews? And don’t many students just take the courses their advisors recommend. I fully agree with Martin. We should not confuse compliance with implementation and effectiveness, and we need to be careful about what we causally attribute to international law. Yet I am still for compliance. I think throwing away compliance could do more harm than good. Compliance does indicate a particular kind of causal effect. In a nontrivial number of cases, compliance is the pivot in the behavioral chain. If our goal is to examine conformity of state behavior to what is prescribed, following Oran Young’s (1979:104) definition, compliance is the first step in assessing the effect of international law on state behavior. Outliers aside, doesn’t implementation usually rest on compliance? And if compliance is not followed by good implementation, don’t we want to know when and why this happens?
Prof. Martin diligently discusses the known challenges that plague the empirical study of compliance. How can we infer the causal effect of international law on state behavior if states create laws that reflect their interests? What does compliance tell us if parties to a treaty have no intention of violating it? The universe clearly wanted me reflect upon these questions in the busiest time of the semester. I got hit with the university’s yearly compliance training. As any “good” assistant professor would do, I conscientiously completed my training and agreed anew to abide by many university policies. But so what? Compliance with these rules is my default position. For instance, it is in my interest to comply with tenure and promotion policies. In the same vein, I would never abuse human subjects, embezzle money or harass anyone whether university policies instruct me or not. I agree with Lisa Martin that the study of compliance could benefit from “the direct measurement of the policies states implement and from counterfactual analysis.” I also recognize her concerns about the limitations of statistical solutions. Yet I am still for compliance. Though imperfect, statistical specifications do address some of the problems we face in causal inference. I am not entirely convinced that the techniques that help us identify whether treaties “constrain or screen” are always ineffective (von Stein, 2005). And other creative solutions also exist. Grieco, Gelpi and Warren (2009), for example, explore what happens to compliance when government preferences change, operationalizing change by partisan shifts. Some who study compliance with EU law have long examined inconvenient compliance to get valid inferences (e.g. Boerzel). My own work experimentally manipulates compliance costs and assesses politicians’ support for compliance when compliance requires significant deviation in behavior. Similarly, Michael Findley, Michael Tomz and Dustin Tingley use experiments to tackle the inference problem. Scholars at USCD-ILAR also rigorously tackle selection effects and endogeneity problems.
There are works that do convincingly demonstrate the causal effect of international treaties and conventions. Instead of moving away from compliance, I think it is important to broaden our understanding of causal effects and theorize different kinds of effects. Compliance, when properly conceptualized, measured, and analyzed, does indicate a particular kind of behavioral effect. To understand the different effects international law and institutions can have on state behavior, I think it is important to keep compliance alive.
I cannot possibly do justice here to Martin’s and others’ contributions. Neither can I capture the essence of the Opinio Juris symposium on law and compliance. And I now have to comply with the faculty profile update requirement to get my name off the blacklist. I have been defecting on this for months. I am still for compliance. How about you?