Evaluating the predictive power of a procedural model for the European Union legislative process
2003; Taylor & Francis; Volume: 9; Issue: 3 Linguagem: Inglês
10.1080/1357233042000246891
ISSN1743-9337
Autores Tópico(s)European and International Law Studies
ResumoAbstract This article analyses the European Union legislative process with the help of formal models of decision-making. The predictions gained from a procedural model are compared with the power of a much simpler model that is informed by the median voter theorem. The finding of the analysis shows that the simpler model out-performs the more complex one. The article proceeds by deriving the hypothesis that the more salient an issue is for EU policy-makers, the higher the probability becomes that the procedural model out-performs the median model. The test results show that this hypothesis can be confirmed. Acknowledgments The author would like to thank Bernard Steunenberg, Marijn Huijbregts, Christopher Achen, Javier Arregui, Frans Stokman, Robert Thomson, Mika Widgén, Antti Pajala, Vincent Boekhoorn, Adrian Van Deemen, Madeleine Hosli, Thomas König, Sven-Oliver Proksch, Stefanie Bailer, and Gerald Schneider for research assistance, and Mark Rhinard and three anonymous JLS reviewers for helpful comments. This study has been made possible by a grant from the Netherlands Organisation for Scientific Research (NWO). Notes 1. R. Scully, ‘The European Parliament and the Co-decision Procedure: A Reassessment’, Journal of Legislative Studies, 3 (1997), pp.58–73; G. Tsebelis and G. Garrett, ‘Agenda Setting, Vetoes, and the European Union's Co-Decision Procedure’, Journal of Legislative Studies, 3 (1997), pp.74–92; R. Scully, ‘The European Parliament and Co-decision: A Rejoinder to Tsebelis and Garrett’, Journal of Legislative Studies, 3 (1997), pp.93–103. 2. For a summary of empirical studies on EU decision-making and on rational institutionalist theory applied to the EU, see H. Wallace and W. Wallace, Policy Making in the European Union (Oxford: Oxford University Press, 2000), pp.69, 80–81; B. Rosamond, Theories on European Integration (Basingstoke: Palgrave, 2000), pp.130–56; S. Hix, The Political System of the European Union (London: Macmillan, 1999), pp.56–98. 3. The two terms will be used interchangeably throughout this paper. 4. Hix, The Political System of the European Union, pp.88–94; B. Steunenberg and T. Selck, The Insignificance of the Significance: A Comparison of Procedural Models on EU Decision Making (Mimeo: Leiden University Press, 2003). 5. K. Shepsle, ‘Institutional Arrangements and Equilibria in Multidimensional Voting Models’, American Journal of Political Science, 23 (1979), pp.27–59. See also E. Ostrom, ‘An Agenda for the Study of Institutions’, Public Choice, 48 (1986), pp.3–25. 6. B. Steunenberg and T. Selck, Testing Procedural Models of European Union Legislative Decision Making (Mimeo: Leiden University Press, 2003). 7. See C. Achen, Model Comparison (Mimeo: University of Michigan, 2003). Another possible null model for spatial models is discussed by C. Achen, The Compromise Model as a Predictive Baseline (Mimeo: University of Michigan Press, 2003). This model would be based on the mean, or average position within a committee. For the present analysis, the median is preferable over the mean because it is rooted firmly in the spatial modelling literature, and because, just as the procedural approach, it is based on non-cooperative game theory. At the same time, it uses one particular mode of decision-making, in this case simple majority voting, as the decision rule. Therefore, while representing the null hypothesis, it does not assume EU legislative politics to be a completely random process. Rather, just like the more complex procedural model does, this model assumes actor preferences and institutions to be the building blocks for analysing legislative decision-making. See K. Shepsle and M. Bonchek, Analysing Politics (New York: Norton, 1997), p.115. 8. D. Black, The Theory of Committees and Elections (Cambridge: Cambridge University Press, 1958). 9. Shepsle, ‘Institutional Arrangements and Equilibria in Multidimensional Voting Models’. 10. Shepsle and Bonchek, Analysing Politics, pp.82–136; M. Hinich and M. Munger, Analytical Politics (Cambridge: Cambridge University Press, 1997), pp.64–9. 11. B. Steunenberg, ‘Decision Making under Different Institutional Arrangements: Legislation by the European Community’, Journal of Institutional and Theoretical Economics, 150 (1994), pp.642–69; G. Tsebelis, ‘The Power of the European Parliament as a Conditional Agenda Setter’, American Political Science Review, 88 (1994), pp.128–42; C. Crombes, ‘Legislative Procedures in the European Community’, British Journal of Political Science, 26 (1996), pp.199–228. 12. L. Gormley, P.J.G. Kapteyn and P. VerLoren Van Themaat, Introduction to the Law of the European Communities (London: Kluwer, 1998), p.192; P. Craig and G. De Búrca, EU Law: Texts, Cases, and Materials (Oxford: Oxford University Press, 1998), pp.131–2, 142–3. 13. E. Kirchner, Decision-Making in the European Community: The Council Presidency and European Integration (Manchester: Manchester University Press, 1992). 14. K. Kollman, ‘The Rotating Presidency of the European Council as a Search for Good Policies’, European Union Politics, 4 (2003), pp.51–74. 15. J. Tallberg, ‘The Agenda-Shaping Powers of the EU Council Presidency’, Journal of European Public Policy (forthcoming 2003). 16. B. Steunenberg and A. Dimitrova, Interests, Legitimacy, and Constitutional Choice: The Extension of the Codecision Procedure in Amsterdam (Mimeo: University of Twente, 1999). 17. The model hereby follows Crombes, ‘Legislative Procedures in the European Community’, and diverges from the account given by Steunenberg, ‘Decision Making under Different Institutional Arrangements: Legislation by the European Community’. For a discussion on whether the Commission can be requested by the Council or by the EP to submit a proposal, see Craig and De Búrca, EU Law: Texts, Cases, and Materials, pp.141–2. 18. The model considers the Council to be a multi-member committee. For an analysis which is based on the unitary actor assumption, see A. Laruelle, ‘The EU Decision Making Procedures: Some Insight from Non Cooperative Game Theory’, in M. Hosli, A. van Deemen and M. Widgrén (eds.), Institutional Challenges in the European Union (London: Routledge, 2002), pp.89–112. 19. D. Kreps, A Course in Microeconomic Theory (Princeton, NJ: Princeton University Press, 1990), p.7. Note, however, that in the field of game theory there also exists ‘pure theory’, that is, models which only constitute a basis for further theorising without immediate applications to real-world phenomena. See Rebecca Morton, Methods and Models (Cambridge: Cambridge University Press, 1999), pp.60–61. 20. Such a model refers to the hypothesis that a certain model factor, in this case the decision-making procedures of the European Union, have no statistically significant effect on the outcome. See B. Tabachnick and L. Fidell, Using Multivariate Statistics (New York: Harper Collins, 1996), pp.34–6. 21. B. Bueno de Mesquita and F. Stokman (eds.), European Community Decision Making: Models, Applications, and Comparisons (New Haven, CT/London: Yale University Press, 1994). 22. Achen, The Compromise Model as a Predictive Baseline. 23. See R. Thomson and F. Stokman, Research Design for the DEU Project (Mimeo: University of Groningen, 2003). For an introduction to the data-gathering method, see Bueno de Mesquita and Stokman (eds.), European Community Decision Making: Models, Applications, and Comparisons, and B. Bueno de Mesquita, Predicting Politics (Ohio: Ohio University Press, 2002). For the DEU data set, in order to measure whether an issue is politically contentious, the European Union daily news bulletin Agence Europe was consulted. Estimates on actor positions were normalised to fit the comparative analysis better. The DEU data set will be made publicly available for replication at the Netherlands national social science data archive Steinmets Archive (www.niwi.knaw.nl/us/dd_star/dd_star.htm). 24. Steunenberg and Selck, Testing Procedural Models of European Union Legislative Decision Making. 25. Values for the status quo variable are missing on many issues because of the fact that some of the corresponding observations in the data set represent new policies on the level of the European Union. This means that it is not possible to give a numerical estimate for the situation which would apply in case the policy-makers would not be able to agree on some new European Union legal measure. 26. For cases with more than three dimensions, the salience estimates of all actors for each issue are aggregated and only the three most salient issues are used. The Council amendment option for consultation is computed differently than prescribed by the models. Here, to amend a Commission proposal, the Council unanimity winset, and not the unanimity set, is computed. The second-order agenda-setter, that is, the Council Presidency, then amends the proposal, whereas in the formal model the Commission anticipates this move and submits a ‘non-amendable’ proposal. See Steunenberg, ‘Decision Making under Different Institutional Arrangements: Legislation by the European Community’, and Crombes, ‘Legislative Procedures in the European Community’. For a definition of the different sets, see G. Tsebelis, Veto Players: How Political Institutions Work (Princeton, NJ: Princeton University Press, 2002). 27. Applying the Wilcoxon signed-rank test shows that the populations are statistically not clearly distinguishable from each other. This, however, might have to do with the limited number of cases that were used for this analysis. 28. Hinich and Munger, Analytical Politics, pp.64–7. 29. Shepsle, ‘Institutional Arrangements and Equilibria in Multidimensional Voting Models’. 30. The equation of the utility function U for some actors i over two issues, x and y, can be given as: 31. Hinich and Munger, Analytical Politics, pp.52–9. 32. Tsebelis, ‘The European Parliament as a Conditional Agenda Setter’. 33. T. König and M. Pöter, ‘Examining the EU Legislative Process: The Relative Importance of Agenda and Veto Power’, European Union Politics, 2 (2001), pp.329–51. 34. C. Hubschmid and P. Moser, ‘The Co-operation Procedure in the EU: Why Was the European Parliament Influential in the Decision on Car Emission Standards?’ Journal of Common Market Studies, 35 (1997), pp.225–42. 35. For an overview of the theoretical literature, see Hix, The Political System of the European Union, pp.56–98, and Steunenberg and Selck, The Insignificance of the Significance: A Comparison of Procedural Models on EU Decision Making. 36. J. Coleman, ‘Systems of Social Exchange’, Journal of Mathematical Sociology, 2 (1972), pp.145–63. 37. J.B. Kadane, ‘On Division of the Question’, Public Choice, 13 (1972), pp.47–54. 38. G. Kramer, ‘Sophisticated Voting over Multidimensional Choice Spaces’, Journal of Mathematical Sociology, 2 (1972), pp.165–80. 39. J. Buchanan and G. Tullock, The Calculus of Consent (Ann Arbor, MI: University of Michigan Press, 1962). 40. T. Selck, The Effects of Issue Salience on Political Decision-Making (Mimeo: Leiden University, 2003). 41. Bueno de Mesquita and Stokman (eds.), European Community Decision-making: Models, Applications, and Comparisons. 42. The DEU data set provides a quantitative measure on how important a certain issue is for the political actors that are involved in a decision. For a discussion on the measurement of issue salience, see J. Van den Bos, ‘The Policy Issues Analysed’, in Bueno de Mesquita and Stokman (eds.), European Community Decision Making: Models, Applications, and Comparisons, p.55. 43. This sort of information is also an input factor for the Wilcoxon signed-rank test, which was employed for the overall model comparison in Table 2. 44. To test the stability of the estimates, relative model performance has further been regressed on the MCS scores with the help of logistic regression. The results confirm the analysis for simple logit (p > χ2 = 0.0079) on whether any of the two models performs better, and also on multinomial logit (p > χ2 = 0.0207) and ordered logit (p > χ2 = 0.0164) for all three conditions, that is, for the case that the procedural model predicts worse than median, that both models predict the same, and that the procedural model predicts better than the median. 45. Wallace and Wallace, Policy Making in the European Union, p.80. 46. Hix, The Political System of the European Union, p.94. 47. J. Van den Bos, ‘The Policy Issues Analysed’, in Bueno de Mesquita and Stokman (eds.), European Community Decision Making: Models, Applications, and Comparisons, pp.33–65; D. Payne, R. Mokken and F. Stokman, ‘European Union Power and Regional Involvement – A Case Study of the Political Implications of the Reform of the Structural Funds for Ireland’, Aussenwirtsschaft, 52 (1997), pp.119–49; B. Rittberger, ‘Impatient Legislators and New Issue Dimensions: A Critique of Garrett and Tsebelis’ ‘Standard Version of Legislative Politics'’, Journal of European Public Policy, 7 (2000), pp.554–75. Additional informationNotes on contributorsTorsten Selck Torsten Selck is an Assistant Professor of Political Economy in the Department of International Relations at the University of Groningen.
Referência(s)