BOWLING FOR FASCISM: SOCIAL CAPITAL AND THE RISE OF THE NAZI PARTY* Shanker Satyanath Nico Voigtländer Hans-Joachim Voth NYU UCLA, NBER, and CEPR University of Zurich and CEPR Forthcoming in the Journal of Political Economy Abstract: Using newly collected data on association density in 229 towns and cities in interwar Germany, we show that denser social networks were associated with faster entry into the Nazi Party. The effect is large – one standard deviation higher association density is associated with at least 15% faster Nazi Party entry. Party membership, in turn, predicts electoral success. Social networks thus aided the rise of the Nazis that destroyed Germany’s first democracy. The effects of social capital depended on the political context – in federal states with more stable governments, higher association density was not correlated with faster Nazi Party entry. Keywords: social capital, democracy, institutions, associations, networks JEL Classification: N44, P16, Z10 * We thank the editor, Jesse Shapiro, four anonymous referees, as well as Daron Acemoglu, Laia Balcells, Eli Berman, Sheri Berman, Johannes Buggle, Davide Cantoni, Nick Crafts, Joan Esteban, Ray Fisman, Nicola Gennaioli, Akos Lada, Stelios Michalopoulos, Massimo Morelli, Torsten Persson, Giacomo Ponzetto, James Robinson, Yannay Spitzer, Enrico Spolaore, Ann Swidler, Debraj Ray, Dominic Rohner, James Robinson, David Strömberg, Peter Temin, Mathias Thoenig, Romain Wacziarg, David Yanagizawa-Drott, and Luigi Zingales for helpful comments. Seminar audiences at the Barcelona GSE Summer Forum, CUNY, CREI, FGV Sao Paulo, Harvard Economics, the IGIER workshop on conflict, IIES-Stockholm, PUC-Rio, SOFI- Stockholm, the University of Copenhagen, and the Warsaw-Penn workshop offered useful criticisms. We are grateful to Hans-Christian Boy, Casey Petroff, Inken Töwe, and Michaël Aklin for outstanding research assistance. We would also like to thank Jürgen Falter for kindly sharing his data on Nazi Party entry. Voigtländer acknowledges financial support from the Hellman Foundation. Voth thanks the European Research Council (AdG FP-7-230515). 1 Hitler's seemingly mysterious mass appeal could hardly have been so extensive without the unplanned propaganda of daily social life ... -Rudy Koshar (1986, p. 202) 1 Introduction Social capital is typically associated with the emergence and persistence of good institutions and favorable economic outcomes.1 Tocqueville argued that American democracy thrived because of a vibrant civic society; conversely, Putnam (2000) concluded that a decline in social capital threatened it. On the other hand, social capital can also be associated with negative outcomes such as organized crime (Durlauf and Fafchamps 2005; Field 2003), and it can serve as a means of control, thereby entrenching the power of autocratic rulers (Acemoglu, Reed, and Robinson 2014). In this paper, we study the role of social capital during one of the key discontinuities of the 20th century – the Nazi Party’s (NSDAP’s) rise to power. In 1933, Germany went from a pluralistic, tolerant democracy to one of the most repressive dictatorships in history. A vast literature has sought to explain the Nazi “seizure of power.” Answers currently range from a history of deep-rooted anti-Semitism (Goldhagen 1996) to the social changes engendered by German industrialization, hyperinflation, and the structural flaws of the Weimar constitution interacting with weak political leadership before 1933 (Bracher 1978). We emphasize a different channel – that Germany’s vibrant “civic society,” its dense network of social clubs and associations, facilitated the rise of Hitler by bringing more people into contact with his party’s message (Berman 1997). Mass membership was crucial for the Nazi rise to power. Long before it became a force at the polls, the Nazi Party developed a mass following of often fanatically devoted members. The electoral success of the NSDAP after 1930 would have been impossible without massive organizational support by thousands of local chapters and hundreds of thousands of dedicated members who campaigned for the party all over Germany, paid dues, and influenced friends and family (Brustein 1998). The party’s vast size was also essential in bargaining for power after 1930 – thanks to its mass appeal, the party controlled a huge paramilitary force of storm troopers (SA). By 1932, it had grown so strong that, 1 Knack and Keefer (1997), Grootaert and Bastelaer (2002), Guiso, Sapienza, and Zingales (2008). 2 according to a war game conducted by the German Army, the SA had a good chance of defeating the regular armed forces in the case of civil war (Winkler 1987).2 Our empirical analysis focuses on one aspect of social capital – dense networks of clubs and associations. We combine individual-level records of Nazi Party membership from Falter and Brustein (2015) with newly collected information on civic associations from a cross-section of 229 towns and cities from all over Germany in the 1920s. We demonstrate that the Nazi Party grew more quickly where association density – measured by number of civic associations per capita – was higher. Figure 1 summarizes the basic pattern in the data: in towns and cities with above-median association density, Germans were substantially more likely to enter the Nazi Party than in towns with below-median club density. The effect is quantitatively important, corresponding to a 27% difference in Nazi Party entry rates over the period January 1925-January 1933. All types of associations – civic and military clubs, “bridging” and “bonding” associations – positively predict Nazi Party entry. The historical record suggests that associations facilitated Nazi recruitment by helping to spread the party’s message, and by increasing trust in its intentions and officials. Party membership, in turn, predicts electoral success. Our results are robust to a wide range of alternative specifications and group definitions. In a panel analysis, we exploit membership growth over time and include city fixed effects to capture local unobservables that may be related to both association density and Nazi Party entry. We show that the marginal effect of existing party members on subsequent membership growth was significantly larger in cities with higher association density. This is in line with historical evidence that existing Nazi Party members successfully exploited local associations to proselytize. An obvious concern with our analysis is that the drivers of association density could directly affect NSDAP penetration. We therefore use the historical record to guide our empirical analysis. Nazi party success correlated with religion, the share of industrial workers (Childers 1983, Falter 1991), and city size. These also predict association density – bigger cities offered more clubs and associations, and sociability was more likely to become formalized; Catholics organized many social activities through the Church; and 2 The regular armed forces were limited to 100,000 as a result of the Versailles Treaty. The head of the army, General Streicher, realized that the NSDAP could not be repressed by violent means. This led him to seek out the party’s moderates in a bid to forge a compromise in December 1932. This was one of the first steps in a process of mutual accommodation between Germany’s traditional right-wing parties and the Nazis Party that culminated in the latter’s entry into government in 1933. 3 workers often had their own associations and clubs, making them less likely to participate in “bourgeois” activities. In all of our analysis we control directly for these factors. However, and crucially, we also argue that some drivers of association density are unlikely to be related to factors affecting the rise of the NSDAP. Associational life was partly defined in the run-up and aftermath of 1848, and its vigor in the interwar period still reflected this earlier period’s opportunities and restrictions. All German states heavily restricted the formation of clubs and societies prior to 1848. These restrictions were dismantled as a result of citizens demanding the right of free assembly during the 1848 revolution. The way restrictions were abolished varied by location, depending on local conditions and individual officials. Despite the revolution’s ultimate failure, the freedom of assembly was never again curtailed on the same scale; as a result, early clubs and associations persisted. Crucially, the revolutionaries of 1848 – Germany’s only attempt at a democratic, “bourgeois” revolution – shared none of the Nazis’ racist, militaristic, expansionist ideology. We show that these democratic associations predict the density of local associations, both in the 1860s and in the 1920s. Based on this fact, we use 1860s associations in an IV exercise that shows a strong link between historical club density and Nazi Party entry. We also examine where and when local club density was particularly strongly associated with Nazi Party entry, and thereby shed light on the conditions under which social capital can become corrosive. To this end, we exploit variation in the stability of government across Germany’s federal states. Unstable governance and higher association density combined were particularly conducive to Nazi Party entry. Our results also indicate that only the Nazi Party benefitted from social capital in Weimar Germany, and it did so to a greater extent during early stages of its development – local contact through social networks counted the most when the Nazi Party itself still had few members. Our paper is the first to show – on the basis of detailed cross-sectional data – that social capital can undermine and help to destroy a democratic system. Our findings complement and extend the results by Acemoglu, Reed, and Robinson (2014), who conclude that powerful chiefs in Sierra Leone “build social capital as a way to control and monitor society” [p.363]. In addition to entrenching autocratic rulers, social capital may also contribute to the rise of autocratic regimes in the first place, by providing a pathway 4 for radical parties to spread and garner support.3 These findings modify our understanding of the relationship between the rise of dictatorships and social capital. Theories of “mass society” and the origins of totalitarianism in the spirit of Ortega y Gasset (1993), Arendt (1973), and Bendix (1952) argued that economic modernization uprooted individuals and dissolved traditional social ties. Confronted with a major economic crisis, the faceless masses could then be easily swayed by demagogic agitators like Hitler, or by dreams of a Communist utopia. In line with the predictions of mass society theory, Shirer (1960) saw marginal loners as the core group of NSDAP supporters; Stern (1972) claimed that interwar German civic society was weaker than in other European countries, and that the country lacked “the kind of voluntary, civic activity that attracted their English and American counterparts” [p.xxix]. In other words, an important strand of the literature on the rise of totalitarianism has argued that the weakness of German civic society facilitated the rise of the Nazis. Our results demonstrate that the opposite is closer to the truth. In this way, we reinforce evidence by Riley (2005) for Italy (discussed in more detail in Section 2.2).4 We also corroborate the conjecture by Berman (1997), who had argued that Weimar Germany as a whole actually had comparatively dense networks of clubs and associations, and that the NSDAP successfully exploited these structures. In contrast to our study, neither Riley nor Berman used detailed quantitative data to test for a systematic link between association density and the rise of fascism. We connect with work on social dynamics and network effects in politics. Recent work has emphasized the importance of influential individuals shaping beliefs in networks (Acemoglu and Jackson 2015).5 Madestam et al. (2013) analyze the rise of the Tea Party in the US. They find evidence for a “social multiplier,” with many more people favoring a radical movement if they see support in large numbers. The paper proceeds as follows. Section 2 discusses the historical context. Section 3 presents our data, and Section 4, our main empirical results. Section 5 asks under what conditions social capital was more beneficial for the Nazi Party’s rise, and Section 6 offers robustness checks and demonstrates the plausibility of our findings. Section 7 concludes. 3 In this sense, adverse political consequences need to be added to the list of negative aspects of high social capital in social settings, such as social exclusion and its enabling role for organized crime (Durlauf and Fafchamps 2005; Portes and Landolt 1996; Field 2003). 4 Note that there is also an ongoing re-evaluation of the Italian evidence in Putnam’s work (Tarrow 1996, Goldberg 1996). 5More generally, Zuckerman (2005) and Lohmann (1993) emphasize the role of group interactions in spreading new political ideas. 5 2 Historical Context In this section, we first argue why, for historical reasons, variation in association density is arguably exogenous to the rise of the Nazi Party, conditional on some key controls. We then describe our key dependent variable – Nazi Party entry – and discuss the party’s social origins. We also summarize related research on the link between association membership and Nazi Party entry. 2.1 Associations in Germany after 1815 and the Source of Identifying Variation What are the origins of associations in Weimar Germany – and thus of the spatial differences that we exploit in our empirical analysis? A close reading of the historical literature suggests that political and social conditions at a “critical juncture” played a key role in determining the strength of associational life in any one location – and once clubs and associations were established, they tended to last a long time. A confluence of largely accidental factors allowed clubs and associations to form after the restoration of 1815 until the early 1850s, with the 1848 revolution as a particular turning point. After the end of the Napoleonic Wars, associations grew in number and scope all over Germany, but were often repressed by the authorities; political associations were banned altogether. Both associations and larger gatherings needed government approval, which was routinely denied. For example, gymnast associations – inspired by Friedrich Ludwig Jahn, and often a focal point for liberal nationalists – were outlawed from 1820 until 1848. Singers’ associations never suffered a blanket ban, but were closely watched by the police. Student fraternities (Burschenschaften) also grew after 1815. They agitated in favor of German unification. Following a political murder, most of the student fraternities were suppressed.6 Over time, restrictions on the formation of associations were repealed or ignored in most German states, and many clubs and associations played a role in 1848 (Botzenhart 1977). The way restrictions were abolished varied by location, and partly reflected differences in attitudes of local and state officials. We argue that the state- and city-level factors driving variation in the repeal of restrictions are plausibly exogenous to NSDAP entry in the 1920s and early 1930s. Germany’s early associations were often both liberal and nationalist in character. They mostly favored the formation of a unified fatherland and 6 The fraternity movement split into a political and a non-political branch, and never recovered its wider political significance (Wentzcke 1965). 6 an end to the rule by princes over often tiny territories, as well as parliamentary representation, a bill of rights, and freedom of assembly, speech, and religion. Importantly, the liberal nationalist part of this agenda was neither militaristic nor xenophobic; it differed substantially from the later nationalism under Bismarck in 1871, and especially from the ideology of the Nazi Party (Eley 1980). Instead, Vereinsnationalismus (nationalism of the associations) mainly emphasized the need to unify all Germans in a nation state similar to France and England, where all could interact as equals (Dunn 1979). After the failure of the 1848 revolution, many associations became increasingly apolitical, focusing on folklore and local traditions (Düding 1984). In addition to the original associations, new ones brought together pigeon breeders, rabbit owners, stamp collectors, and supporters of a plethora of other causes. Student associations, on the other hand, became increasingly nationalistic and militarist, and several of them adopted xenophobic and anti-Semitic ideas in the late 19th century (Haupt 1925). Despite the revolution’s ultimate failure, earlier prohibitions never returned with full force. Once formed, clubs and associations were sticky – as reflected in the fact that many integrated their date of founding into their name. A local culture of associational life persisted, and it influenced the extent to which people continued to gather and pursue like- minded activities into the interwar period (Bösch 2005, Hardtwig 1984). One way to illustrate this argument is to show that involvement of democratic associations in the 1848 revolution is a strong predictor of association density in the 1920s. In 1848, associations in part acted as precursors of modern parties in many German states, organizing the collective expression of political beliefs for the first time (Langewiesche 1978). Good examples are the Democratic Congresses in 1848, representing the left-wing of the revolution. Where local associations had formed, their delegates participated in these gatherings, which included the first promulgations of universal human rights in Germany. Sending local delegates to the Congresses required three things – a sufficient number of people interested in a distinctly left-wing agenda, the ability to organize locally, and the ability and right to do so. We find that – for the limited subset of towns and cities with available data on delegates – involvement with Democratic Congresses is strongly positively correlated with both the vigor of associational life in 1860s Germany and in the 1920s (see Appendix F.1). In other words, our main explanatory variable – association density in the 1920s – is strongly predicted by clubs that represented the political left during the 1848 revolution. This makes it unlikely that the historical origins of associational life in Germany reflect 7 local, unobserved Nazi-compatible ideology – few supporters of universal human rights admired Adolf Hitler.7 During the interwar period, membership in associations soared. The main singers’ association’s membership tripled, to 1.2 million; the German gymnasts’ association registered a 50% rise in membership. Most associations saw themselves as apolitical, and did not support particular parties. In the Catholic Rhineland, all ranks of societies often joined Carnival associations, organizing revelry during the annual “silly season.” While some organizations were explicitly Catholic or Protestant, almost every town and city also had a large number of non-denominational associations (Reichardt 2004). Associations reflected the views and biases of German civic society in general; where politics were not deliberately kept out of the club, there was a society for every political grouping. Workers gathered in workmen’s singing associations; Communists reminisced about their frontline experiences together; and members of the nobility and rich industrialists conferred in gentlemen’s and equestrian clubs (Zeiss-Horbach 2008; Koshar 1986). While many clubs and societies catered to a particular social group, others transcended divisions of class and education – such as the many sports and hiking clubs, chess clubs, and associations for the preservation of local customs and culture. In our later analysis, we will explicitly distinguish between “bonding” and “bridging” social capital (Putnam 2000). 2.2 Historical Correlates of Association Density and Nazi Party Appeal In the following, we discuss historical factors that may confound our empirical analysis – variables that arguably influenced both association density and the recruitment success of the Nazi Party. These include the share of Catholics, the share of blue-collar workers, and city size. Catholics in Germany were initially less inclined to support the Nazi Party. This is not because Catholics, or the Catholic Church, were not immune to the appeal of fascism. The Catholic Church built amiable relationships with fascist regimes in Italy and Spain (and eventually in Germany). Before 1934, the situation in Germany was different. 7 One could alternatively argue that what persisted was a local tendency to go against the established system. This is difficult to examine empirically. The only truly anti-regime party of the pre-1914 period, the Social Democratic Party (SPD), was the political home of the worker movement. It was outlawed under Bismarck for 12 years (1878-90), and persecuted by the authorities as 'rabble without a fatherland' (Rovan 1980). Unsurprisingly, areas with an “anti-regime” bias in this regard before 1914 were not more likely to support the Nazi Party. We show in Section E.7 in the online appendix that SPD votes in 1890-1912 do not predict Nazi Party entry or NSDAP votes; they are also not correlated with association density in Weimar Germany. 8 Catholics were traditionally in a minority in Germany in its pre-WWII borders, and their loyalty to the German state was historically suspect. There was an important political tradition on the political right of the spectrum (including under Chancellor Bismarck) of pursuing an anti-Catholic agenda (Wehler 1994). Hence, Catholics had a historically rooted distrust of far-right nationalist parties. The Nazi Party was thus less appealing to Catholics – especially since they also had their own party, the Zentrum. The formal organization of social life in independent clubs and societies was also less common among Catholics; the Catholic Church provided a natural focal point and offered many activities under its own aegis. We do not include religious clubs and associations in our dataset.8 Consequently, cities with more Catholics tend to have lower measured association density in our sample. The situation for workers was similar to Catholics in some ways. Long suspected of a lack of patriotism, the Communist Party was the main party of protest (rather than the Nazi Party) for those workers disenchanted by the Weimar regime (Winkler 1987). At the same time, the inclination of workers to form clubs and associations was lower than amongst the rest of the population. While they often had their own sports associations and the like, most workers lived in large cities, where association density tended to be lower; also, forming clubs and societies is a prototypical “bourgeois” activity, performed by those with higher educational attainment and broad networks of contacts. Finally, we also need to control for population size. Larger cities had lower association densities, arguably because of both economies of scale and because bigger cities contained more workers. We also know that rural areas and smaller towns were particularly given to supporting the Nazi movement (Brustein 1998, pp. 102-08). For this reason, we also include city size in our baseline controls. After controlling for city size, the share of Catholics, and the proportion of workers, we believe that differences in the density of associations are reasonably exogenous for the purpose of our study (i.e., driven by deep historical factors that have no direct link with Nazi Party entry). 2.3 The Rise of the Nazi Party and Fascism in Europe There is a large literature analyzing the Nazi Party’s success at the polls and as a mass movement (e.g. Childers 1983, Hamilton 1982, Falter 1991, King et al. 2008). Initial theorizing focused on isolated members of the masses – marginal loners for whom the party 8 This is because we are interested in the “bottom-up” characteristic of grassroot organizations, not in ready- made sociality created by members of the Church hierarchy. 9 represented a group where they finally belonged (Shirer 1960). An alternative literature interpreted the rise of the Nazi Party as a form of class conflict (Winkler 1987). Recent research on voting behavior emphasizes “ordinary economic voting” – with the working poor particularly susceptible to the Nazi message (King et al. 2008). Our paper is closely related to research emphasizing group membership as a pathway to Nazi involvement, which gained wider currency from the 1970s onwards (Linz 1976). This strand of the literature assigns crucial importance to the “conquest of the bourgeois infrastructure” (Mommsen 1978, p. 186), i.e., the infiltration of existing high-level national and regional lobbying groups (Verbände) representing farmers and other special interests. Berman (1997) points out that Weimar Germany as a whole had many civic associations. She argues that “… had German civil society been weaker, the Nazis would never have been able to capture so many citizens for their cause ...” (Berman 1997, p. 402). Koshar (1986), in a detailed study of Marburg, demonstrates that NSDAP members were active in many local groups. Anheier (2003) shows how well-connected individuals acted as political entrepreneurs. Using their social connections and professional standing, they attracted new members for the party, leading to the founding of new local chapters. Our work also follows earlier historical research on interwar politics in Europe. Riley (2010; 2005) analyzes the role of civic associations and the rise of fascism in Italy and Spain. Based on evidence from 20 Italian regions, he argues that associations fostered the rise of fascism. In Spain, associational life was dominated by the Catholic church, and was largely compatible with a more traditionalist form of fascism. Riley contends that in countries without strong hegemonic organizations – i.e., well-established parties – social capital can undermine the development of democracy. In a similar spirit, Wellhofer (2003) examines the rise of fascism in Italy, focusing on election results. In contrast to Riley, he finds that civic society blunted the rise of fascism, but only in certain elections. 2.4 Nazi Party Membership The Nazi Party deliberately competed with leftwing parties for mass support, replacing their class-based ideology with nationalist and racist ideals (Shirer 1960). From the party’s early days, Hitler and his associates viewed organization-building as crucial for the rise to power. The party’s initial growth was slow. Eventually, membership grew to 850,000 members in January 1933 – on par with the Social Democratic Party (SPD), and nearly three times higher than Communist membership (Childers 1983).