Did Power Politics Cause European Integration? Realist Theory Meets Qualitative Methods

Click to increase image sizeClick to decrease image size Acknowledgments Andrew Moravcsik is professor of politics and director of the European Union Program at Princeton University. The author is grateful to Michael Desch for repeatedly encouraging him to critique Europe United, to Alexander Lanoszka for insightful comments, suggestions, assistance, and even some felicitous wording, and to Christina Davis, Robert Keohane, Anne-Marie Slaughter, and Marc Trachtenberg for useful advice. Notes Sebastian Rosato, Europe United: Power Politics and the Making of the European Community (Ithaca, ny: Cornell University Press, 2011). Rosato, Europe United, 2. The book is an exemplar not just of poor qualitative methodology, but also of poor social scientific methodology. The analysis does not present explicit hypotheses, make appropriate causal process observations, choose cases that have variation on the independent variables (to have a quasi-experimental treatment group comparable to a control group), properly take into account confounding factors that precisely estimate the effect of the favored independent variable, or measure phenomena in an unbiased fashion. It also repeatedly fails to test theoretical claims using data different from those employed to develop those same claims. The fundamental basis of such standards is common to all forms of social scientific analysis, qualitative and quantitative. Gary King, Robert Keohane, and Sidney Verba, Designing Social Inquiry: Scientific Inference in Qualitative Research (Princeton, nj: Princeton University Press, 1994). Rosato, Europe United, 10. Ibid. , 218. Nowhere does Europe United conclude that other causes played a significant role, except to assert occasionally that ideas mattered but were endogenous to geopolitics—an exception that proves the rule. To be sure, not all the evidence presented in the book actually supports a geopolitical interpretation. The book includes some caveats, some of which involve entire decades of eu history being explained using economic factors. Yet none of this evidentiary ambivalence is reflected in Europe United's conclusions. See the discussion under “Internal Validity” below. Desmond Dinan, Ever Closer Union: An Introduction to European Union, 3rd ed. (Boulder, co: Lynne Rienner, 2005), 17–25. Dinan, Ever Closer Union, 27–29. Among more recent analyses are Raymond Poidevin, “Introduction to the Debate, ” in Power in Europe? Great Britain, France, Germany, Italy and the Origins of the eec, 1952–1957, ed. Ennio Di Nolfo (Berlin: de Gruyter, 1992), 546–50 passim; Raymond Poidevin, Robert Schuman (Paris: Beauchesne, 1997) ; Hanns Jürgen Küsters, Die Gründung der Europäischen Wirtschaftsgemeinschaft, (Baden-Baden, Nomos Verlagsgesellschaft: 1982) ; Hans von der Groeben, Aufbaujahre der Europaäischen Gemeinschaft: das Ringen um den Gemeinsamen Markt und die Politische Union (1958–1966) (Baden-Baden: Nomos Verlagsgesellschaft, 1982) ; Hans-Peter Schwarz, Die Zentralmacht Europas: Deutschlands Ruückkehr auf die Weltbuhne (Berlin: Siedler, 1994) ; Timothy Garton Ash, In Europe's Name: Germany and the Divided Continent (New York: Random House, 1993). Though they disagreed on most things, political scientists Stanley Hoffmann and even Ernst Haas (a neo-functionalist indelibly associated with economic welfarism and technocracy) both stressed geopolitics in precisely this context. Stanley Hoffmann, “Obstinate or Obsolete? The Fate of the Nation-State and the Case of Western Europe, ” Daedalus 95, no. 3 (Summer 1966): 862–915; Ernst Haas, The Uniting of Europe: Political, Social and Economic Forces, 1950–1957 (Notre Dame, in: University of Notre Dame Press, 1958), 155. Alan Milward, The Rise and Fall of a National Strategy: The uk and the European Community, vol. 1 (London: Routledge, 2002). Andrew Moravcsik, The Choice for Europe: Social Purpose and State Power from Messina to Maastricht (Ithaca, ny: Cornell University Press, 1998), 137, 474, also, 90–95, 102–103, 135. For recent literature interpreting the 1950s and 1960s in an increasingly commercial light, see N. Piers Ludlow, “The Green Heart of Europe? The Rise and Fall of the cap as the Community's Central Policy, 1958–1985, ” in Fertile Ground for Europe? The History of European Integration and the Common Agricultural Policy Since 1945, ed. Kiran Klaus Patel (Baden-Baden: Nomos Verlagsgesellschaft, 2009), 79–98; Laurent Warzoulet, Le choix de la cee par le France: L’Europe économique en débat de Mendès France à de Gaulle (1955–1969) (Paris: Comité pour L’Histoire Économique et Financière de la France, 2010). For an overview of recent trends, see Andrew Moravcsik, “De Gaulle and European Integration: The New Revisionism, ” Journal of Cold War History 14, no. 1 (Winter 2012): 53–77. Craig Parsons also adopts a synthetic approach, arguing that ideological causes were important in the negotiation of various eu treaties, but economic interest locked them in. See Craig Parsons, A Certain Idea of Europe (Ithaca, ny: Cornell University Press, 2006). Rosato maintains (correctly) that in case study analysis the provision of circumstantial evidence, that is, evidence of structural conditions that give actors “motive, means, and opportunity” consistent with a theory, only establishes “correlation” or “initial plausibility, ” not “causality. ” Demonstration of causality requires direct evidence of real-world “causal mechanisms” linking concrete causes and mechanisms to outcomes (in this case specific decisions on European integration) working in ways that support one theory above others. The analogy is to criminal law, where “motive, means, and opportunity” are sufficient to indict but not to convict a particular suspect. See, for example, Rosato, Europe United, 16–19, 42–43, 50, 175–176. See also fn. 13. Rosato, Europe United, 50. Circumstantial evidence (or “data-set” observations) comprises facts and texts about the endpoints of a process—cause and consequence (independent and dependent variable). Direct evidence (or “causal-process” observations) comprises detailed “process-tracing” observations of the “causal mechanism” linking cause and effect. This includes explicit hypotheses about actor intentions, the nature of domestic coalitions, the timing of policy change, and policy consistency. In a small-n research design, circumstantial evidence is invariably easy to find for any hypothesis plausible enough to test, and thus represented in the secondary literature. It is often employed to select and specify theories and hypotheses. Claims about process multiply the predicted observations many times and tend to be more specific, concrete, and varied, and involve collecting data of which the analyst is unaware when selecting hypotheses. For this reason, it is prudent to use circumstantial evidence to select and specify plausible theories, but rely on direct process-tracing evidence to test them, using explicit hypotheses. Collier, Brady, and Seawright, “Sources of Leverage. ” For an example in this area, see Moravcsik, Choice for Europe, 28, 55, 69. Rosato, Europe United, 83–96, 153–63, 209–18. Ibid. , 151–53. Ian S. Lustick, “History, Historiography, and Political Science: Multiple Historical Records and the Problem of Selection Bias, ” American Political Science Review 90, no. 3 (September 1996): 614–16. Lustick points out that we need to sample secondary sources the same way we do other data. However, he overlooks one implication of this, namely that no secondary-source based analysis can be truly revisionist on an issue already treated by historians. The only way to be a revisionist, as Rosato aims to do, is to reinterpret the underlying primary evidence. By comparison, in the corresponding chapter of my Choice for Europe, 56 percent of empirical space is devoted to economic factors and 44 percent to geopolitical evidence. In each case, the strongest arguments and evidence in favor and against the theories are weighed. This method was selected even though Choice for Europe, unlike Europe United, introduced an interpretation that was at the time genuinely novel, so perhaps greater attention to new documentation supporting the economic case might have been justified. See footnotes 12 and 13 above. Explication added. Alan Milward, The Reconstruction of Western Europe (London: Methuen, 1984), 395, on the Schuman Plan generally, 362–420. Though twenty-five years old now, Reconstruction postdates much of the historiography Europe United cites. Europe United seeks to dismiss my claim that geopolitical arguments mattered by stating that “the little supportive evidence Moravcsik does find for noneconomic arguments validates ‘liberal constructivist’ variants of the argument, rather than ‘Realist power-balancing’ ones. ” Rosato, Europe United, 13. This misreads my argument. True, over fifty years of integration, most of which Europe United does not analyze, most noneconomic factors I observed at work were ideological. But in “early European Community history, ” the only period where the two studies overlap, I argue that realist balancing was the dominant noneconomic factor. Moravcsik, Choice for Europe, 478, also 474–79. Specifically in the case of Germany in the eec negotiations: “Some argue backwards from Adenauer's support for integration to … federalist idealism … But Adenauer does not appear to have believed strongly in European political union … Adenauer did, however, consistently exploit the domestic and international legitimacy of integration … to promote geopolitical objectives. From 1948 onward, Adenauer's central geopolitical objective appears to have been to assure Germany's defense, control over West Berlin, and opposition to the Soviet Union … One means of committing a wide range of allied governments to German goals was to embed Germany in … various schemes for European integration … Without geopolitical concerns, Germany would surely have supported a British-style fta … ” Moravcsik, Choice for Europe, 91, 102, 477, also generally 99–103. These facts are documented in Moravcsik, Choice for Europe, 90–103, 137–39. Rosato, Europe United, 218–19. Four errors are evident here: the content of the document is misread to mean the reverse of what it actually says; Milward's gloss on the document, which clarifies this, is ignored; the document is misdated; and it is cited to the wrong page of Milward's book (page 200) ; cf. Alan Milward, The European Rescue of the Nation-State (London: Routledge, 1992), 175. The underlying source of the confusion appears to lie in part in Rosato's lack of familiarity with contemporary usage of the term “partial union”—which appears in all these sources. In Europe United Rosato consistently interprets it to mean unions among a small number of countries (e. g. among the six, like the eec, rather than larger groupings, like the fta). But in the 1950s the term was used specifically to designate Monnet's sectoral schemes, such as the coal and steel community, transport cooperation, or an atomic energy union. The eec was in fact seen as an alternative to these. For example, see the discussion of the François-Poncet quotation in Rosato, Europe United, 219. Rosato, Europe United, 219. Documents diplomatiques français (1955, tome 1), docs. 297, 684, 685. The misreadings cited here do not appear isolated or exceptional. I checked several dozen citations. Nearly half, well beyond those reported above and the misread secondary authorities cited later in this essay, were misrepresented seriously. For some more detailed examples, see “External Validity” below. To be sure, I selected citations that seemed suspicious, but I also focused only on citations central to the core theoretical claims in Europe United. For example, Rosato similarly misreads a statement about sovereignty in the only other major source he cites to support the very point discussed above, from historian Werner Bührer, to state that German business supported only the “oeec Organisation for European Economic Cooperation type of integration, ” not the eec. Rosato, Europe United, 218–19. In fact, Bührer states unambiguously that business “commented favourably on the common market negotiations” and a page later concludes, “to sum up, German industry supported the common market. ” Werner Bührer, “German Industry and European Integration in the 1950s, ” in Western Europe and Germany: The Beginnings of European Integration, 1945–1960, ed. Clemens Wurm (Oxford: Berg, 1995), 104–6. This leaves Europe United with no credible evidence on this vital point of contention with economic interpretations. Another serious issue of causal inference is that some of the critical claims in the book, as we have seen, are ad hoc. Europe United concedes that conventional realism does not predict, for example, that governments will support the ecsc and not the edc; it predicts the opposite. It offers an addendum, not derived from anything in the initial theoretical chapter, to explain this, with no further empirical confirmation from independent data—then counts it as full confirmation for realist theory. One might question, on this basis, whether any observation could have disconfirmed realism. But I remain here focused on the central concern, namely the reporting of evidence. Rosato, Europe United, 183–97, 226. Europe United seeks to clarify matters with two pages of selected quotations from leaders purportedly relating the final decision in 1958 to underlying geopolitical motivations. Selection bias calls any such effort into question. Yet even specially selected evidence proves ambivalent. The first, from Antoine Pinay, states that the policy goal is to save “Europe's economy and its freedom. ” It seems odd to interpret this as entirely supportive of realism, as Europe United does, rather than the conventional multicausal economic account. Crisper process-level hypotheses designed to distinguish between explanations might have helped here. Rosato, Europe United, 185–86. Even more succinct in Rosato's more policy-oriented International Security article: “The answer is simple: Europe kept moving toward monetary union despite the end of the Cold War because of prosperous economic conditions. ” Sebastian Rosato, “Europe's Troubles: Power Politics and the State of the European Project, ” International Security 35, no. 4 (Spring 2011): 73. Rosato, Europe United, 183, 226. Indeed, there is a consensus in the newer historiography, which Europe United does not acknowledge, that concerns about French competitiveness by economic actors were important on their own terms until 1958, when de Gaulle devalued the currency and reformed the budget. Warzoulet, Le choix de la cee par le France; Moravcsik, Choice for Europe. Rosato, Europe United, 9. Rosato's recent article in International Security is framed as a reflection on Europe's current state. Rosato, “Europe's Troubles. ” Any scholar who has studied policymaking processes closely, or any who has experienced it firsthand while serving in government, will surely need little convincing that implementation, interpretation, adjudication, adaptation, legal process, constitutional adjudication, and political bargaining often intervene between broad legal provisions and outcomes. This is even more true of international regimes, which, according to our best theories, are meaningful precisely because the intervening process is highly uncertain, much as is the case with constitutional delegation. Robert Keohane, After Hegemony: Cooperation and Discord in the World Political Economy (Princeton, nj: Princeton University Press, 1984). This subsequent half-century witnessed “ … no further moves toward political or military unification. The various economic agreements worked out in these years simply strengthened and extended the structure put in place in the 1950s. Even the decision to negotiate a single currency did not signal a seismic shift of the existing system, only a modification of it. ” Rosato, “Untied States of Europe, ” Stephen M. Walt's Blog at ForeignPolicy. com, June 23, 2010, http: //walt. foreignpolicy. com/posts/2010/06/23/guestₚostₜheᵤntiedₛtatesₒfₑurope. It is unclear whether this is a valid claim: standard qualitative methods do not generally recommend selecting on the dependent variable (where integration decisions occur), on a single value of the independent variable without controls (where the Soviet threat is high), but especially not on both at once, as appears to be the case here. Most would recommend selecting on the independent variable alone, comparing the effect on decision making of the treatment (high Soviet threat in the 1950s) with its absence (low Soviet threat in the 1990s), with process-tracing to increase observational evidence throughout. Even if one believed that later attempts to promote integration were substantively unsuccessful or insignificant, one should still seek to explain these “nondecisions. ” Europe United itself concedes that such comparisons would not favor realism, since the eu is far more ambitious and successful for most, if not all, of the post-Cold War period, but they are set aside because decisions subsequent to 1958 do not matter. Two analysts sympathetic to selecting on the dependent variable for process-tracing purposes, Andrew Bennett and Alexander George, point out that “the most damaging consequences of case selection bias arise from selecting only cases whose independent and dependent variables vary as the favored hypothesis suggests, ignoring cases that appear to contradict the theory, and overgeneralizing from these cases to wider populations. ” Bennett and George, Case Studies and Theory Development in the Social Sciences (Cambridge, ma: mit Press, 2005), 24, emphasis in the original. Lustick, “History, Historiography, and Political Science. ” Rosato, Europe United, 6. Joseph Weiler, “The Transformation of Europe, ” Yale Law Journal 100, no. 8 (June 1991): 2413. Ibid. , 2410 (emphasis added). After 1963, Weiler explains, “the European Court of Justice in a series of landmark decisions established four doctrines that fixed the relationship between Community law and Member State law and rendered that relationship indistinguishable from analogous legal relationships in constitutional federal states. ” Weiler, “Transformation of Europe, ” 2413. Further citations to articles by Jeffry Frieden and William Bernhard, Lawrence Broz, and William Clark similarly do not do justice to the basic literature in international political economy. None of the cited authors believe, as Europe United attributes to them, that the choice of an international monetary regime is insignificant; indeed, all have committed their careers to the opposite proposition. Rosato, “Europe's Troubles, ” 67; Horst Ungerer, A Concise History of European Monetary Integration: From epu to emu (Westport, ct: Quorum Books, 1997), 164. Rosato, Europe United, 9. Moravcsik, Choice for Europe, 1. Hence I start my empirical analysis in Choice for Europe with the sentence: “The history of the European Economic Community begins with a failure. ” Moravcsik, Choice for Europe, 86. This consensus includes Jean Monnet's sympathetic collaborator and preeminent biographer, François Duchêne. Of course some might argue, with Duchêne, that the impact of the ecsc was symbolic, that is, despite its substantive unimportance, it created an enduring institutional template. Yet an argument about symbolic or transaction-cost reducing institutionalist framing is neither advanced in Europe United, nor consistent with realism. François Duchêne, Jean Monnet: The First Statesman of Interdependence (New York: Norton, 1980), 393, 396–99. The European Atomic Energy Community (euratom), the other major initiative of the 1950s—unaccountably ignored in Europe United, since one would expect it to be of security interest—was also a failure. This is not a controversial claim. See, for example, Warzoulet, Le choix de la cee par le France. Indeed, economic explanations can even account and the of European integration in the but that beyond the of this Andrew Moravcsik, “The European World Economy no. 1 A more and thus more realist theory would an and to those institutionalist and theories that or to into their and Moravcsik, a is but I of no case of a work in eu studies or international my for repeatedly to cited sources the opposite of their One way to make such errors and thus would be to fifty to one (or of supporting source (in an is to the by a See Andrew Moravcsik, A for Qualitative Political Science and Politics no. 1 Moravcsik, and Qualitative Political Qualitative and Research no. 1 (Spring 2012): For corresponding in the American Political Science see Making data in this way is also the best way to to about sources. For example, in to some of errors in my Choice for Europe, full source for in a on de Gaulle in Choice for Europe is to be later this in at a new Qualitative based at which I with of articles by other that basis, I a further to those at the of the Journal of Cold War I will also an of this article on my at with citations from Uniting Europe and its sources. Rosato any of sources in this is I him to using the same based