Published online by Cambridge University Press: 13 June 2011
Recent cross-national studies have returned their attention to the structural determinants of political regimes, highlighting in particular the factor of “landholding inequality” as a decisive barrier to democratization. This article provides thefirstsystematic test of such hypotheses at the microlevel and proposes a new account of authoritarianism's durability by examining the crucial case of pre-World War I Prussia. The article analyzes the results of a roll-call vote on a watershed piece of legislation that was defeated on the eve of World War I-legislation that would have democratized suffrage rules in Germany's largest state. When examined systematically, this historically and theoretically important vote reveals two surprising lessons: first, landholding inequality undercuts the prospects of democratization even holding income inequality constant. Second, the nature of elite competition and electoral considerations, shaped by the institutional configuration of nondemocratic regimes, can also thwart democratization, even when socioeconomic conditions may appear to make a society ripe for regime change.
1 Iversen, Torben, “Democracy and Capitalism,” in Wittman, Donald and Weingast, Barry, eds., Oxford Handbook of Political Economy (Oxford: Oxford University Press, 2006), 618Google Scholar.
2 Gerschenkron, Alexander, Bread and Democracy in Germany (1946; Ithaca, N.Y.: Cornell University Press, 1989)Google Scholar.
3 Acemoglu, Daron and Robinson, James, “Why Did the West Extend the Franchise? Growth, Inequality and Democracy in Historical Perspective,” Quarterly Journal of Economics 115, no. 4 (2000)CrossRefGoogle Scholar; idem, Economic Origins of Dictatorship and Democracy (Cambridge: Cambridge University Press, 2006)Google Scholar; Boix, Carles, Democracy and Redistribution (Cambridge: Cambridge University Press, 2003)CrossRefGoogle Scholar; Ben Ansell and David Samuels, “Inequality and Democratization” (Paper presented at the annual meeting of the Midwestern Political Science Association, Chicago, Ill., April 3, 2008).
4 Clark, Christopher, The Iron Kingdom: The Rise and Downfall of Prussia, 1600—1947 (Cambridge: Harvard University Press, 2006), 670-81Google Scholar.
5 Weber, Max, “Das preussische Wahlrecht ,” in Max Weber Gesamtausgabe, vol. 15 (Tübingen: J. C. B Mohr, 1984), 233Google Scholar. In pre-1914 Germany suffrage rules for state legislatures varied from state to state despite a precociously universal male, direct, and secret ballot for the national parliament. Prussia's three-class system stood out because Prussia was Germany's largest state and its suffrage system was one of the most regressive, combining universal male suffrage with a highly inegalitarian weighting of votes based on tax contributions that gave disproportionate influence to the wealthy (described more fully below). The system was remarkably robust, withstanding at least seventeen efforts at reform between its creation in 1849 and its demise in 1918, a time when suffrage was being democratically reformed in most other German states. For a European-wide view on suffrage reform in this period, see Mattmüller, Markus, “Die Durchsetzung des allgemeinen Wahlrechts als gesamteuropaeischer Vbrgang,” in Junker, Beate, Gilg, Peter, and Reich, Richard, eds., Geschichte und Politische Wissenschaft (Bern: Francke Verlag, 1975)Google Scholar.
6 See, most notably, Moore, Barrington, Social Origins of Dictatorship and Democracy (Boston: Beacon Press, 1966)Google Scholar. For a discussion of how Moore's argument and mode of analysis has in turn shaped several generations of scholarship on democratization, including, for example, Luebbert, Gregory, Liberalism, Fascism, or Social Democracy (Oxford: Oxford University Press, 1991)Google Scholar; see Mahoney, James, “Knowledge Accumulation in Comparative Historical Analysis: The Case of Democracy and Authoritarianism,” in Mahoney, James and Rueschemeyer, Dietrich, eds., Comparative Historical Analysis in the Social Sciences (Cambridge: Cambridge University Press, 2003)CrossRefGoogle Scholar.
9 Schattschneider, E. E., Party Government (New York: Holt, Rinehart, and Winston, 1942), 48Google Scholar.
10 For paradigmatic examples of each approach, see Rueschemeyer, Dietrich, Stephens, Evelyne Huber, and Stephens, John, Capitalist Development and Democracy (Chicago: University of Chicago Press, 1992)Google Scholar; and O'Donnell, Guillermo and Schmitter, Philippe, eds., Transitions from Authoritarian Rule: Tentative Conclusions about Uncertain Democracies (Baltimore: Johns Hopkins University Press, 1986)Google Scholar.
11 Moore (fn. 6); Rueschemeyer, Stephens, and Stephens (fn. 10).
12 See, for example, Paige, Jeffrey, Coffee and Power: Revolution and the Rise of Democracy in Central America (Cambridge: Harvard University Press, 1997)Google Scholar; Kurtz, Marcus, Free Market Democracy and the Chilean and Mexican Countryside (Cambridge: Cambridge University Press, 2004)CrossRefGoogle Scholar; Varshney, Ashutosh, Democracy, Development, and the Countryside: Urban-Rural Struggles in India (Cambridge: Cambridge University Press, 1998)Google Scholar; Riedinger, Jeffrey, Agrarian Reform in the Philippines: Democratic Transition and Redistributive Reform (Stanford, Calif.: Stanford University Press, 1995)Google Scholar; and Kuhonta, Erik, “The Political Foundations of Development: State and Party Formation in Malaysia and Thailand” (Ph.D. diss., Princeton University, 2003)Google Scholar.
13 Gerschenkron (fn. 2).
14 Moore (fn. 6).
15 Rueschemeyer, Stephens, and Stephens (fn. 10), 80, 83, 84, 91–92.
16 See, e.g., Paige (fn. 12); Varshney (fn. 12); Kurtz (fn. 12); Jong-Sung You, “Explaining Corruption in South Korea, in Comparison with Taiwan and Philippines: The Role of Economic Inequality, Growth, and Democratization” (Paper presented at the annual meeting of the American Political Science Association, Washington, D.C., September 1,2005); Bowman, Shearer Davis, Masters and Lords: Mid 19th Century U.S. Planters and Prussian Junkers (Oxford: Oxford University Press, 1993)Google Scholar.
17 Acemoglu and Robinson (fn. 3,2000, 2006).
18 Boix (fn. 3)
20 In Boix's phrase, “The absence of landlordism constitutes a necessary precondition for the triumph of democracy”; see Boix (fn. 3), 40.
21 Acemoglu and Robinson (fn. 3, 2006).
22 Vanhahen, Taru, Prospects of Democracy: A Study of 172 Countries (London: Routledge, 1997)Google Scholar.
23 Ansell and Samuels (fn. 3), 4.
24 Schattschneider (fn. 9), 45.
26 Collier, Ruth B., Paths toward Democracy: The Working Class and Elites in Western Europe and South America (New York: Cambridge University Press, 1999)CrossRefGoogle Scholar; Llavador, Humberto and Oxoby, Robert J., “Partisan Competition, Growth, and the Franchise,” Quarterly Journal of Economics 120 (August 2005)Google Scholar.
27 The debate over the German “special path” (Sonderweg) is enormous. Some of the major works include Wehler, Hans-Ulrich, Das deutsche Kaiserreicb, 1871—1918 (Göttingen: Vandenhoeck and Ruprecht, 1983)Google Scholar; Blackbourn, David and Eley, Geoff, The Peculiarities of German History (Oxford: Oxford University Press, 1984)CrossRefGoogle Scholar; Anderson, Margaret, Practicing Democracy (Princeton: Princeton University Press, 2000)Google Scholar; and Berman, Sheri E.“Modernization in Historical Perspective: The Case of Imperial Germany” World Politics 53 (April 2001)CrossRefGoogle Scholar. For the most recent intervention, see Smith, Helmut Walser, “When the Sonderweg Debate Left Us” German Studies Review 31 (May 2008)Google Scholar.
28 For a sample of the current and rich historical work on barriers, problems, and prospects of suffrage reform in German states after 1871, see Kühne, Thomas, Dreiklassenwahlrecht und Wahlkultur in Preussen, 1867–1914 (Düsseldorf: Droste, 1994)Google Scholar; Laessig, Simone, Wahlrechtskampf und Wahlreform in Sachsen, 1895–1909 (Weimar: Boehlau, 1998a)Google Scholar; Lässig, Simone “Wahlrechtsreformen in den deutschen Einzelstaaten,” in Lässig, Simone, Pohl, Karl Heinrich, and Retallack, James, eds., Modernisierung und Region (Bielefeld: Verlag für Regionalgeschichte, 1998b)Google Scholar; Retallack, James, The German Right, 1860–1920: Political Limits of Authoritarian Imagination (Toronto: University of Toronto Press, 2006)CrossRefGoogle Scholar; idem, “Wahlrechtskämpfe in Sachsen nach 1896,” Dresdner Hefte 22, no. 4 (2004), 13–24Google Scholar.
29 Two notable recent exceptions of social scientists explicitly placing agrarian structure against other hypotheses are Rössel, Jörg, Soziale Mobilisierung und Demokratie: Die preussischen Wahlrechtskonflikte 1900 bis 1918 (Wiesbaden: Deutscher Universitäts-Verlag, 2000)CrossRefGoogle Scholar; and Bernhard, Michael, “Democratization in Germany. A Reappraisal,” Comparative Politics 33 (July 2001)CrossRefGoogle Scholar.
30 This is one indirect implication of Berman's (fn. 27) important note of the declining economic significance of agriculture in Germany (p. 442). Additionally, recent important works have questioned the conventional conception of the Junkers' local social power. See Wagner, Patrick, Bauern, Junker, und Beamte (Göttingen: Wallstein Verlag, 2005)Google Scholar; and Hagen, William, Ordinary Prussians (Cambridge: Cambridge University Press, 2002)Google Scholar.
31 Ertman, Thomas, “Liberalization and Democratization in Nineteenth and Twentieth Century Germany in Comparative Perspective,” in Lankowski, Carl, ed., Breakdown, Breakup, Breakthrough: Germany's Difficult Passage to Modernity (New York: Berghahn Books, 1999), 47Google Scholar.
33 For a discussion of single-outcome studies, see Gerring (fn. 32). On within-case comparison, see Brady, Henry and Collier, David, eds., Rethinking Social Inquiry (Lanham: Rowman and Littlefield, 2004)Google Scholar; and George, Alexander and Bennett, Andrew, Case Studies and Theory Development in the Social Sciences (Cambridge: MIT Press, 2005)Google Scholar. On the strategy of using “new measures and new units” to increase the number of theoretically relevant observations, see King, Gary, Keohane, Robert, and Verba, Sidney, Designing Social Inquiry (Princeton: Princeton University Press, 1994), 224-25Google Scholar.
34 For front-page press coverage of the vote, see Berliner Tageblatt, May 21,1912.
35 The definitive work on the three-class voting system is Kuhne (fn. 28). See also Rossel (fh 29). The older literature on the subject includes Pateman, Reinhard, “Der Deutsche Episkopat und das Preussische Wahlrechsproblem, 1917/1918,” Vierteljahrsheftefür Zeitgeschichte 13 (October 1965)Google Scholar; Bergsträsser, Ludwig, Die Preussische Wahlrechtsfrage im Kriege und die Entstehung der Osterbotschaft 1917 (Tübingen: Verlag von Mohr, 1929)Google Scholar; Dietzel, Hans, “Die Preussischen Wahlrechtsreformbestrebungen von der Oktroyierung des Dreiklassenwahlrechts bis zum Beginn des Weltkrieges” (Ph.D. diss., University of Cologne, 1934)Google Scholar; Grünthal, Günther, “Das Preussiche Dreiklassenwahlrecht: Ein Beitrag zur Genesis und Funktion des Wahlrechtsoktrois von May 1849,” Historische Zeitschrift 226 (1978)Google Scholar.
36 For a more detailed account of voting practice, see Kühne (fn. 28), 129–32.
37 Geheimes Staatsarchiv, Preussische Kulrurbesitz, I HA Rep90 Staatsministerium St. M 9317, Berlin, May 7,1849.
39 It should be noted that any effort to reform the three-class voting system would also have had to win approval of the Prussian upper chamber and the king. Though high hurdles to pass, the king's cabinet had itself proposed reform legislation earlier in the decade, suggesting the willingness to endorse some modernization of the electoral system. See Kühne (fn. 28).
40 For discussion of the vote, see Rossel (fn. 29), 158,284–85; and Schuster, Dieter, “Das Preussische Dreiklassenwahlrecht, der politische Streik und die deutsche Sozialdemokratie bis zum Jahr 1914” (Ph.D. diss., Universität Bonn, 1958), 270Google Scholar. It is worth noting that post-World War I roll-call votes also occurred (May 1918). See the analysis in Rössel, Jörg“Eliteninteressen und soziale Konfliktlinien in Demokratiserungsprozessen: Die soziale Konstruktion von Interessen und das Abstimmungsver-halten von Abgeordneten bei der Reform des Dreiklassenwahlrechts in Preussen 1918,” Historische Sozialforschung 25 (2000)Google Scholar.
41 For an analysis of this election, see Bertram, Jürgen, Die Wahlen zu Deutschen Reichstag vomjahre 1912 (Düsseldorf: Droste, 1964)Google Scholar.
42 See, e.g., Eley, Geoff, Reshaping the German Right (New Haven: Yale University Press, 1980), 316-30Google Scholar.
43 For a sense of the evolution of this debate, see, briefly, Kehr, Eckart, Battle Ship Building and Party Politics in Germany, 1894–1901 (1930; Chicago: University of Chicago, 1975)Google Scholar; Eley (fn. 42); Ferguson, Niall, “Public Finance and National Security: The Domestic Origins of the First World War Revisited.” Past and Present 142 (February 1994)CrossRefGoogle Scholar.
44 For historical evidence that state politicians compared the results in their own elections with the results in their districts in federal elections when considering suffrage reform, see Lassig (fn. 28,1998a).
45 Verhandlungen, Haus der Abgeordneten, 77 Sitzung, May 20,1912, 6428–32.
46 The voting record lists these forty-four representatives as “excused,” “sick,” or “on vacation”; Verhandlungen (fn. 45), 6432.
47 See Schuster (fn. 40). Schuster reports that fifteen National Liberals left the chambers, but according to the roll-call results there were thirteen National Liberals and thirty-three Center Party representatives who had “unexcused” absences; Verhandlungen (fn. 45), 6432.
48 In an additional analysis below, I model the “partisan affiliation” of individual legislators as a way of evaluating the causal mechanisms at work in the other models. The data for this outcome draw on a biographical handbook of Prussian legislators. See Kühne, ThomasHandbuch der Wahlen zum Preusischen Abgeordnetenhaus, 1867–1918 (Düsseldorf: Droste, 1994)Google Scholar.
50 Kaiserliches Statistisches Amt. 1898. Statistik des Deutschen Reids, vol. 112 (Berlin: Verlag des Königlich Preussichen Statistischen Bureaus, 1898), 351–413Google Scholar, table 9. The five million farms identified for the survey included all officially designated “landwirtschaftliche Betriebe,” 40 percent of which were operated by their owners and the remainder of which were rented or communal land or took some other form. For a discussion of what qualified as an agricultural unit for the survey, see discussion, pp. 8–9. For a discussion of the type of ownership structure of units involved in the study, see pp. 15–19.
51 Gerschenkron (fn. 2).
52 A key source in aggregating county units was Kühne (fn. 48), which lists which towns and counties were in which electoral district. Using 1895 census data on the size and number of farms at the county level, I aggregated these data to correspond with the 276 Prussian electoral districts. This involved identifying the administrative units of each electoral district and adding the corresponding values for both number and area of farms to obtain data for the larger electoral district. For each constituency, I calculated a Gini coefficient to measure landholding inequality. To ensure accuracy, it was necessary to record changes in district areas after 1895. Major changes such as those resulting from the redistricting reform of 1906 could mostly be incorporated in the aggregation of data. It was only in Berlin, Teltow-Beeskow-Storkow-Charlottenburg, and Tarnowitz-Beuthen-Zabrze-Kattowitz that administrative units were not precise enough to completely allow for that. In these cases, I approximated Gini coefficients by using the same county data for all newly created electoral districts. Other territorial changes were minor, concerning twenty-five exchanges of two or three municipalities or estates between two districts, and were ignored in the analysis.
53 United Nations, United Nations Human Development Report 2006 (New York: United Nations Development Program, 2006)Google Scholar.
54 This variable measures the “base level of support” for a legislator's party under the two systems and is measured by contrasting the percentage of vote received in a district in the state election for a legislator's party with the percentage of vote received in the same district in the first round of the federal elections for that same party. Data source: for federal elections, see ICPSR, German Reichstag Election Data [Computer File] (Ann Arbor: ICPSR, 1991)Google Scholar; Kühne (fn. 48) for state elections.
55 Since we are interested here in the extent to which the rule change would change the outcome of an election (crossing 50 percent of the vote), this is measured by taking the absolute value of (50-local vote share won) and multiplying it by -1. If a party is at 50 percent, this number will be 0. If a party is at 10 percent or 90 percent in the Prussian election, this number will be -40. The closer a party gets to 0, the closer the election, and thus the more important the rule change.
56 The product of “base level of support” and “how important” an election is gives us the interaction of these two terms.
60 According to my theory, party affiliation and constituency characteristics are not competing explanations for the 1912 vote. Instead party affiliation was a vehicle for constituency characteristics (and their associated political preferences), and therefore it was not included as a separate explanatory variable in the first set of models. I later provide empirical support for the assertion that characteristics like inequality and economic modernization are, in large measure, determinants of party affiliation. Other scholars have demonstrated that constituency characteristics operate both directly on legislators and indirectly through party. See Schonhardt-Bailey, Cheryl, “Parties and Interests in the ‘Marriage of Iron and Rye,’” British Journal of Political Science 28 (1998).CrossRefGoogle Scholar
61 For example, a Center Party abstention is coded 0, a vote against democratization, since the rest of the party voted for the bill.
62 Because income inequality data are available for only approximately two-thirds of the cases, I exclude it from two of the models. The chief variables of interest, landholding inequality and electoral incentive, are unaffected by its exclusion.
63 Acemoglu and Robinson (fn. 3,2006). In analysis not reported here, I also included squared and cubed terms of “income inequality,” which did not affect the statistical significance of the landholding inequality variables.
64 Landholding inequality has a negative but insignificant effect when the model includes either (1) only a first-order inequality term or (2) first- and second-order inequality terms.
65 P(0 or 1) when inequality is at its minimum is 0.04. P(0 or 1) when inequality is at its maximum is 0.90). P(2) drops from 0.96 to 0.10 over the same range of inequality. Probabilities are calculated by adjusting inequality—along with its squared and cubed terms—and holding all other variables at their mean values. Probabilities and probability changes presented in this paper can be calculated in Stata 9 using the prchange and prvalue commands available as part of the SPost package created for Stata 9 by Jeremy Freese and J. Scott Long. The simulations all use model 3.
66 P(0) decreases from 0.57 to 0.53. P(l) increases from 0.17 to 0.18. P(2) increases from 0.25 to 0.29.
67 Boix (fn. 3); Acemoglu and Robinson (fn. 3,2000,2006)
68 Since the variable measuring “how important” and its interaction with “electoral incentive” were not significant in any specification, I dropped them from the analysis and do not report the findings here.
69 Again, probabilities and probability changes were calculated in Stata 9 using the prchange and prvalue commands available as part of the SPost package created for Stata 9 by Jeremy Freese and J. Scott Long.
70 That forty-seven representatives departed from the party line to abstain from the vote, determining the outcome, indicate that party line cannot explain the full variation in the dependent variable. However, the relatively small number of defectors make it difficult to assess through a regression analysis whether the decision to abstain was a result of underlying constituent characteristics, electoral incentives, or other factors.
71 Schotuiardt-Bailex (fn. 60).
72 For example, in the National Liberal Party, if two-thirds of the members agreed on a common party position in prevote meetings, then this was considered a “priority” item for the party (Parteisache),demanding unanimous party loyalty and abstentions from those who insisted on defecting from the party line; Nipperdey, Thomas, Die Organisation der deutschen Parteien vor 1918 (Düsseldorf: Droste, 1961), 160Google Scholar. Similarly, in the Center Party, prevote internal negotiations regularly occurred in which the party leadership established consensus among its parliamentary representatives, usually requiring defectors to stay away from the vote rather than vote against the party line (p. 289). See also Mittmann, Ursula, Fraktion und Partei: ein Vergleich von Zentrum und Sozialdemokratie im Kaissereich (Düsseldorf: Droste, 1976).Google Scholar
73 Though Social Democrats and the Danish Party were included in the analysis, Table 2 does not report the coefficients for these two parties that had less than seven representatives.
74 One possible explanation of this anomaly is the unusually urban character of the constituencies the Left Liberal Parties represented. The mean level of agricultural employment for the other parties' districts was 35 percent, while the Left Liberals were, along with the very small Social Democrats, the most urban party, representing districts with, on average, 16 percent of employment in the agricultural sector.
75 It is also worth noting that although the decisive Catholic Center Party was programmatically in favor of suffrage reform, the party split on the 1912 vote in a way that reflected broader cultural and ideological divisions within the party. Center Party representatives from urban and more religiously mixed districts, who tended to be more progressive, disproportionately voted for the reform and Center Party representatives from rural and more homogeneously Catholic districts abstained, in effect voting against the reform. This reflected divisions on a range of issues between urban and rural factions inside the Center Party itself. See analysis in Rössel (fn. 29). See also discussion in Bachem, Karl, Vorgeschichte, Geschichte und Politik der Deutschen Zentrumspartei (Cologne: J. P. Bachem, 1932)Google Scholar; Loth, Wilfned, Katholiken im Kaiserreich: der politische Kathohzismus in der Knse der Wilhelmischen Deutschlands (Düsseldorf: Droste, 1984)Google Scholar. On the varied political consequences of the urban-rural split more generally, see Kühne (fn. 28), 52–55; 133–41, 575–77.
76 Boix (fn. 3); Ansell and Samuels (fn. 3).
77 This is a long-standing debate in German historiography. See, most prominently, Kehr (fn. 43). My finding supports Grant's (fn. 57) contention that a decline in income inequality by 1912 suggested the surprisingly diminishing relevance of industrial or class conflicts as a determinant of German politics in this period.
78 See, for example, Ross, Michael L., “Does Oil Hinder Democracy?” World Politics 53 (April 2001)CrossRefGoogle Scholar; Smith, Ben, “Oil Wealth and Regime Survival in the Developing World, 1960–1999,” American Journal of Political Science 48, no. 2 (2004)CrossRefGoogle Scholar; Fish, M. Steven, Democracy Derailed in Russia (Cambridge: Cambridge University Press, 2005)CrossRefGoogle Scholar. For an account that focuses on the distribution of rents from resources that is arguably compatible with the Gerschenkronian logic, see Dunning, Thad, “Does Oil Promote Democracy? Regime Change in Rentier States” (Ph.D. diss., University of California, Berkeley, 2006)Google Scholar.
80 First-wave democratization has not been analyzed through the analytical lens of “durable authoritarianism.” Yet this conclusion suggests a convergent line of future inquiry with recent work that examines in comparative perspective, the institutional loots of the persistence of authoritarian regimes. See, for example, Slater, Dan, “Iron Cage in an Iron Fist: Authoritarian Institutions and Personalization of Power in Malaysia,” Comparative Politics 3 (October 2003)Google Scholar; Brownlee, Jason, Authoritarianism in an Age of Democracy (Cambridge: Cambridge University Press, 2007)CrossRefGoogle Scholar; Przeworski, Adam and Gandhi, Jennifer, “Authoritarian Institutions and the Survival of Autocrats,” Comparative Political Studies 40 (November 2007)Google Scholar; Magaloni, Beatriz, “Credible Power Sharing and the Longevity of Authoritarian Rule,” Comparative Political Studies 41 (April 2008)CrossRefGoogle Scholar.
Full text views reflects PDF downloads, PDFs sent to Google Drive, Dropbox and Kindle and HTML full text views.