As a new and rapidly evolving polity, with clear formalized rules and a great deal of publicly available data, the European Union presents a fascinating object of study for political scientists. Until recently, it was often treated as sui generis, the preserve of dedicated Euro-watchers. Most attention was focused on grand events—treaties, judicial cases, and individual cases of specific interest. But increasingly, the general tools of political science are being brought to bear on the EU and its activities, including humdrum ones.

This book contributes to that literature by studying the outcomes of 66 European Commission legislative proposals for directives or regulations, which are detailed, technical, and very specific. It seeks to test different general formal or rational choice theories of legislative decision making derived from political science to see which offers the greatest predictive power. In particular, it pits procedural models that use decision-making rules and formal procedures against bargaining models that emphasize informal negotiations before formal legislative procedures begin (although such bargaining takes place under the shadow of formal rules for decision making) and a third group of mixed models that combine both bargaining and procedural games.

The research design is simple but powerful. Seven chapters use the same data set of the 66 proposals (themselves divided into issues, leading to a data set of a maximum of 162 issues). Information about matters such as actors' positions, outcomes, and votes was gathered from different sources, including a panel of experts. Thereafter, each chapter compares the predictions of its model, or two or three models, against the legislative outcomes. Most chapters also seek to offer one or two case study illustrations, looking at particular directives. The final chapters compare the predictive power of the different models both among themselves and with two baseline models using the mean and the median of the different actors' positions.

The central conclusion is that in general, bargaining models offer the strongest predictions, whereas procedural and mixed ones fare less well. Thus, confining analysis to formal procedures misses major elements of decision making. But, an additional and disquieting conclusion is that these models perform only slightly better than the baseline model based on the mean of the actors' positions, an analysis that requires little development and is not based on any theory.

The chapters are often admirably honest in their evaluation of their particular model. Thus, for instance, Stefanie Bailer and Gerald Schneider accept that the differences between a baseline Nash bargaining game model and more sophisticated versions are small, Thomas König and Sven-Oliver Proksch find that the procedural exchange model they favor does no better than competing models, while Mika Widgren and Antti Pajala argue that detailed modeling of the voting procedures actually reduces the accuracy of predictions. The evaluating chapter by Christopher Achen is also clear and transparent in its comparison of different models and its somewhat devastating conclusion of the limited gains of models compared to the baselines.

The book illustrates both the positive features of the new generation of EU studies and their limitations. It offers valuable cumulation by using existing work on legislative studies. Equally, it achieves a remarkable degree of coherence through its analysis of a common data set and the ability to appraise different explanations, a rare feature in political science. The chapters have a clear objective linked to theory, namely, showing the predictive power of theoretically derived models. The data set is impressive and is representative of EU legislative decision making. There is great rigor, with limitations of the data set and models explicitly acknowledged. Although quantitative analysis is dominant, case studies of individual pieces of legislation are set out, and Achen explicitly argues that case studies remain valuable in identifying explanatory factors and formulating tentative generalizations, and also as illustrations of theory.

But the book also faces important difficulties. The overall conclusion that many models perform less well than baseline nontheoretically derived ones suggests that much of the theorizing has actually produced less accurate predictions than nontheoretical work. Although the evaluating chapter by Achen seeks to defend this in terms of other domains where for many years a similar position held, it does call into question the resources and effort poured into developing so many sophisticated models of EU legislative decision making—and whether they would have been better applied to other domains in which surprising outcomes and explanations could have been found. König and Proksch end their analysis by suggesting rather plaintively that the failure of its models to offer better predictions “seem[s] to confirm the scepticism about empirical research in general and comparative testing of models in particular” (p. 235). In addition, the models are dense—the overview offered by the penultimate chapter would have been better placed at the start in order to allow readers to quickly understand groups of models. Equally, the case studies work only partially: It is often unclear what insight each model would bring to the particular case. Finally, and perhaps unfairly, the work is very dry: Many of the illustrations underline just how technical EU decision making is—cases such as the scrap-build penalty for fisheries or a directive on ballasts in electric lighting seem designed to bear down on even the most ardent Euro-scholar.

All the same, The European Union Decides is a highly rigourous book that has an impressive research design and execution. It will be suitable for advanced scholars of the EU and of legislative politics in general. It serves as an excellent example of the advantages and difficulties of modeling and data analysis.