In May 2011, the “Indignados” of Spain occupied Puerta del Sol in Madrid, setting up a protest that quickly spread to hundreds of other cities across the country (Postill Reference Postill2014). That same month, a massive anti-austerity demonstration took place in Syntagma Square of Athens, Greece, drawing tens of thousands into the streets (Sergi and Vogiatzoglou Reference Sergi, Vogiatzoglou, Fominaya and Cox2013). In October, 100,000 protesters marched in Rome. Rather than being isolated incidents, these demonstrations were emblematic of a protest wave sweeping over the European Union (EU) (della Portua 2015). Given the ability of large demonstrations to captivate the public, it is no surprise that protest has long been a subject of immense scholarly interest. The power of grassroots protest to upend political systems, change societies, and achieve the goals of social movements is well documented (Dalton Reference Dalton2008; Meyer and Tarrow Reference Meyer, Tarrow, Meyer and Tarrow1998). The existing literature generally illustrates that participation in demonstrations (and most social-movement activity) depends on a combination of the level of grievances within a population, the level of resources possessed by that population, and the political opportunities (or lack thereof) encountered along the way (Kriesi Reference Kriesi2012). These factors explain why some groups protest at certain points and in certain places whereas others do not, as well as why some campaigns are more successful than others. Although previous work has greatly advanced the current understanding of protest activity, the study of social movements generally has excluded examination of a special type of protest movement: urban protest (Pickvance Reference Pickvance2003).
It is not surprising that protests often take place in cities, but it remains unclear whether protest is motivated by urban attributes or is merely a byproduct of the same factors that spur demonstrations elsewhere. To this end, I focus on contemporary protest in Europe, which entered a major protest wave in 2009 that has continued to influence European politics and social life to the present (Vassallo and Ding Reference Vassallo and Ding2016). From anti-austerity demonstrations in Athens and Madrid to anticorruption protests in Warsaw, Rome, and Prague and to anti-immigration marches in Leipzig, urban protest has become a pervasive part of the European political sector (albeit far from the only type of protest or social-movement campaign). Using these urban protests as inspiration, I used the sixth wave of the 2012 European Social Survey (ESS) to construct a limited mixed-effect, multilevel logistic-regression model to document how urbanity influences protest participation in the presence of more traditional variables (e.g., grievances and resources). The results demonstrate that whereas grievances and resources indeed predict EU protest, urbanity exerts a powerful positive effect on the likelihood of protest participation.
THEORETICAL BACKGROUND
Diani (1992, 1) defined a social movement as “networks of information interactions between a plurality of individuals, groups and/or organizations, engaged in political or cultural conflicts, on the basis of shared collective identities.” Of course, protests are a common tactic of social movements (Dalton Reference Dalton2008), but contemporary protest is becoming a routine part of political negotiations and bargaining (Jenkins, Wallace, and Fullerton 2014). Here, I make an explicit distinction between urban protest and other forms of protest. The concept of urban social movements is a contribution of Castells (Reference Castells1977, Reference Castells1983), who described them as an outgrowth of capital consumption within densely populated environments that produced inequalities for the city’s diverse collection of residents. Urban social movements became known as coalitions of interest groups operating within cities in pursuit of an expressly urban goal (Pickvance Reference Pickvance2003), such as housing rights (Mayer Reference Mayer, Brenner, Marcuse and Mayer2012), artistic spaces (Novy and Colomb Reference Novy and Colomb2012), social exclusion (Mayer Reference Mayer2013), and other concerns. This analysis is concerned less about the goals of individual protests and more about the context in which they play out. Therefore, urban protest is defined simply as protest events that take place within cities aimed at achieving any goal or target.
The concept of urban social movements is a contribution of Castells (Reference Castells1977, Reference Castells1983), who described them as an outgrowth of capital consumption within densely populated environments that produced inequalities for the city’s diverse collection of residents.
Not everyone or every area is equally likely to protest (Schussman and Soule Reference Schussman and Soule2005; Verhulst and Walgrave Reference Verhulst and Walgrave2009). It long has been known that both grievances (Jenkins, Jacobs, and Agnone Reference Jenkins, Jacobs and Agnone2003) and deprivation (Blau and Blau Reference Blau and Blau1982) can motivate protest, with relative rather than absolute deprivation being more powerful. In addition, protest is influenced by the amount of resources—both material and human—that organizations can marshal (Brady, Verba, and Schlozman Reference Brady, Verba and Schlozman1995). Members and supporters (Cress and Snow Reference Cress and Snow1996) and wealth (Berinsky Reference Berinsky2002; Stern et al. Reference Stern, Dietz, Abel, Guagnano and Kalof1999) also are known to encourage protest.
Although these factors still have relevance, this article contends that urbanity also encourages protest activity beyond the influence of these other factors. Nicholls (Reference Nicholls2009), for example, theorized that cities act as “movement spaces” in which their density and spatial layout facilitate social and political activism. Because cities are bigger and more densely populated than rural areas (Walton Reference Walton1998), they become a political “theatre” for potentially larger audiences (Anthony and Crenshaw Reference Anthony and Crenshaw2014). This population base also gives protest movements a greater platform from which to draw participants; creates advanced social networks for recruitment at the local level (Gould 1991); and facilitates the diffusion of political ideas (Hedstrom Reference Hedstrom1994). Additionally, cities tend to be wealthier than the hinterlands, which gives protest movements a greater resource base (Bruhn Reference Bruhn2008). Contemporary issues such as immigration, social polarization, and the decline of welfare states have created newly perceived competition for jobs, resources, and services. In Europe, this risk is felt most strongly in cities (Taylor-Gooby Reference Taylor-Gooby2004).
The recent literature on urban movements—specifically, urban protest—implies two styles of thinking. First, higher levels of urban (as compared to nonurban) protest results from the greater abundance of factors already known to spur protest: aggrieved populations, a wealthier resource base, and increased political opportunities. In contrast, others argue that cities present a unique stage for political protest and conflict to play out independent of these other factors. Therefore, we must empirically examine traditional predictors of protest in the presence of urban indicators to separate effects from one another (Schoene Reference Schoene2017).
The EU provides an ideal case study. First, the EU entered a major protest wave in 2009 that continues to shake the continent (Vassallo and Ding Reference Vassallo and Ding2016). Although specific goals of each country’s protest movement were linked to the national context, commonalities exist. Most of these protests took place in major cities. For example, more than half of all Greek anti-austerity protests occurred in Athens (Rüdig and Karyotis Reference Rüdig and Karyotis2014). Furthermore, ESS data demonstrate the dominance of urban protest. Figure 1 displays the percentage of urban residents who reported participation in a lawful demonstration during the past 12 months compared to nonurban residents in the same country.
Figure 1 Rates of Protest Participation of Urban and Nonurban Residents by Country, European Social Survey 2012
Figure 1 shows that a higher percentage of urban residents reported participating in a demonstration than their nonurban counterparts in all 20 countries examined. The trend is clear, but the determinants of this trend are less clear. Delanty (Reference Delanty and Isin2000) argued that whereas European integration was centered on states, it is European cities that are emerging to destabilize this project. European politics may be primarily state-based, but cities are increasingly the centers of political, social, and economic life. The integration process and related forms of globalization create myriad challenges for European cities that have major implications for all types of issues, from the place of the nation in the EU to various urban-quality concerns (Giffinger et al. Reference Giffinger, Fertner, Kramar and Meijers2007). Footnote 1
DATA AND METHODOLOGY
For this article, I used the 2012 sixth wave of the ESS. Of the available countries, I excluded Albania, Switzerland, Israel, Russia, Norway, Kosovo, Ukraine, and Iceland for the simple reason that they are not EU members. I also was regrettably forced to omit the United Kingdom (UK) because it does not provide all necessary regional variables. Therefore, this study examines the 20 remaining EU countries. Footnote 2 I used them to construct a multilevel, mixed-effects logistic-regression model. Hierarchical analysis is preferable when respondents are not randomly distributed but instead form natural groupings that make individuals within a group more alike than those of other groups (Raudenbush and Byrk Reference Raudenbush and Byrk2002). In cross-national analysis, people are nested within countries. However, I wanted a smaller, more-specific grouping variable to better understand the role of urbanity (Schoene Reference Schoene2017). I therefore classified respondents according to their ESS region, a geographical designation derived from Eurostat’s Nomenclature of Territorial Units for Statistics. This was crucial because it allowed certain variables to be measured at the individual level and others at the regional level. Footnote 3 The ESS region, or the group, was the second-level geographical variable for this analysis.
As the dependent variable, “protest” refers to participation in a lawful demonstration within the past 12 months, a common measure of recent protest activity (Dalton Reference Dalton2008; Jowell et al. Reference Jowell, Roberts, Fitzgerald and Eva2007).
Protest
As the dependent variable, “protest” refers to participation in a lawful demonstration within the past 12 months, a common measure of recent protest activity (Dalton Reference Dalton2008; Jowell et al. Reference Jowell, Roberts, Fitzgerald and Eva2007). Footnote 4 This is a dichotomous outcome where 1 = participated in the activity in the past 12 months and 0 = did not.
Satisfaction (Economy)
I included a measure of respondents’ satisfaction with the economy for their country of residence. This variable is measured on an 11-point scale, with 0 representing total dissatisfaction and 10 representing complete satisfaction.
Satisfaction (Government)
Next, I supplemented my economic-satisfaction measure with a variable gauging satisfaction with the national government. Like the previous measure, this is an 11-point scale ranging from zero to 10.
Income
Measuring income across 20 countries—only some of which use a common currency—was especially challenging. Rather than include a continuous variable, I constructed two dichotomous variables that classify respondents as middle or high income, with low income omitted to serve as a reference category (Schoene Reference Schoene2017). The ESS provides respondents with 10 deciles and asks them to categorize themselves relative to their fellow residents. I classified deciles 1–3 as low income, deciles 4–7 as middle income, and deciles 8–10 as high income. Footnote 5
GDP Per Capita
The ESS provides a measure of GDP per capita at the level of the ESS region. I then logged this measure to standardize it and for ease of interpretation.
Urban
I recoded the question, “Which phrase on this card best describes the area where you live?,” into a dichotomous measure, classifying urban as a “big city” and “suburbs or outskirts of a big city.” I defined not urban as those residing in a “town or small city,” “country village,” or “farm or home in countryside.” Urban = 1; not urban = 0.
Population Density
I also measured the effect of population density. This variable was measured at the regional level and logged for ease of interpretation.
Controls
Finally, I controlled for an individual’s age, sex, and immigrant status because these factors are known to influence protest participation.
As previously stated, I constructed a multilevel logistic-regression model with a limited number of predictors to test the likelihood of participation in a protest. To justify the need for multilevel modeling, I constructed a one-way ANOVA with random-effects model, which included only the dependent variable (i.e., protest) and no predictors (table 1). This was to confirm that there was sufficient variation of the outcome across level-two groups to warrant the increased complexity of a multilevel model (Raudenbush and Byrk Reference Raudenbush and Byrk2002). Footnote 6
Table 1 One-Way ANOVA with Random Effects Regression of Protest
***p < 0.001.
The significant coefficient confirmed that a multilevel strategy indeed captured more variation than a nonhierarchical model. Footnote 7 Results of the full-regression model follow.
RESULTS
Table 2 displays the coefficients, standard errors, and odds ratios for each independent variable included. I interpreted the odds ratios, which refer to the change in the odds of participating in a protest given a one-unit change in the predictor. In the case of dichotomous predictors, it was the change in odds of participating in a protest for each group compared to its reference group, as noted at the bottom of the table. A value above 1 denotes an increase in the odds, whereas an odds ratio below one refers to a decrease in the odds, with all other effects held constant.
Table 2 Mixed Effects Multilevel Regression Results of All Independent Variables on Participation in a Lawful Demonstration, European Social Survey Sixth Wave
†p < 0.10 *p < 0.05 ** p <0.01 ***p < 0.001.
(a): Reference group is low income.
(b): Reference group is nonurban.
(c): Reference group is female.
(d): Reference group is native-born citizen.
First, the odds ratios confirmed the role of both grievances and resources as meaningful frameworks. With regard to grievances, I found that satisfaction in institutions correlated with a lower likelihood of protest: a one-unit increase in economic and governmental satisfaction was correlated with a 0.97 and a 0.94 times lower likelihood of protesting, respectively. Although this is only a sampling of potential grievances, this perspective clearly has merit. With regard to resources, I found that material wealth at the individual and regional levels predicts protest. Middle- and upper-income individuals are 1.12 and 1.11 times, respectively, more likely to protest compared to lower-income individuals. Because the upper-income coefficient was only marginally significant (p < 0.10), I interpreted this as evidence that protests are driven by the middle class and above, and that there likely is not a major difference in the protest behavior of the middle and upper classes. Wealth also has a contextual effect. Each unit increase in the logged regional GDP correlates with a 1.58 times greater likelihood of protest participation. As with grievances, I found support for the role of resources in explaining these puzzles. My three demographic controls were all significant and expected: people are less likely to protest as they grow older, men protest more than women (Kuumba Reference Kuumba2001), and native-born citizens protest more than immigrants (Koopmans Reference Koopmans2005).
With regard to grievances, I found that satisfaction in institutions correlated with a lower likelihood of protest: a one-unit increase in economic and governmental satisfaction was correlated with a 0.97 and a 0.94 times lower likelihood of protesting, respectively.
The major contribution of this article, however, is whether urban residence predicts protest in the presence of these other well-supported factors. My results indicated that it does. Urban residents are 1.47 times more likely to protest than nonurban residents. The magnitude of this coefficient is quite significant, given other findings. I found somewhat weaker but still significant support for population density: a one-unit increase in this variable correlated with odds of protest 1.11 times greater. Even when controlling for grievances, resources, and demographic factors, there is strong support for the idea that cities stimulate protest behavior beyond traditional predictors of protest.
DISCUSSION AND CONCLUSIONS
In summary, this study indicates that grievances, resources, and urbanity all predict European protest participation. Whereas the first two factors are widely accepted, urbanity is in need of further theorizing. Clearly, the greater prevalence of urban protest in the EU cannot be fully explained by the greater prevalence of grievances and resources. Our understanding of how space operates may be in need of rethinking. How, then, should we conceptualize urban protest?
Cities should be thought of as networks of individuals, where urban location facilitates the formation of coalitions that spur protests and other forms of activism. Nicholls (Reference Nicholls2009) argued that cities should be thought of as providing a “movement space” for the formation of these social networks. However, Nicholls (Reference Nicholls2008) also argued that whether these networks and relationships actually develop heavily depends on local political-power relationships between the social-movement organizations responsible for planning protests and the local authorities. He then showed how France’s immigrant-rights movement developed from activist networks centered in Paris, which sustained large-scale mobilization of individuals. The idea of cities as networks is found in other studies of European urban movements, such as Novy and Colomb’s (2012) study of creative-class mobilization in Hamburg and Berlin. Furthermore, Mayer (Reference Mayer2003) argued that the main shortcoming of the urban-movement literature is how it ignores social capital as a driving force behind protest and related forms of activism. However, these analyses do not attempt to theorize beyond their city in question. Future research should take greater steps to study urban networks and social capital and to generalize these findings outside of only specific cities. Conceptualizing the city as a collection of dense social networks clearly is not mutually exclusive with other predictors of protest. Recent work on the European protest wave focuses on the idea of “protest potential,” or the likelihood that certain people or areas may be pushed into protest participation. It may be that networks interact with factors such as grievances and resources to produce networks with great protest potential.
Of course, this does not imply that nonurban areas are powerless in the face of urban economic, social, and political power. This is clearly demonstrated by the UK’s recent referendum on EU membership, in which a largely nonurban coalition successfully dragged the country out of the EU. The Standing Rock protests in the United States provide another example of nonurban protest potential. However, social-movement studies have always exhibited a bias toward successful movements, and these recent protest movements (powerful as they may have been) do not detract from the overall trend. It may be that whereas urban protest is more likely, it is national social movements that remain more likely to succeed. More theorizing is needed in this area but, as these results indicate, European protest will continue to be driven by the continent’s cities.
SUPPLEMENTARY MATERIAL
To view supplementary material for this article, please visit https://doi.org/10.1017/S1049096517001780
ACKNOWLEDGMENTS
Direct correspondence to Matthew Schoene, Department of Anthropology and Sociology, Robinson Hall, Albion College, Albion, MI 49224. I thank Dr. Francesca Vassallo and the anonymous reviewer for their helpful comments and feedback during the process of writing this article. I also thank my research assistant, Taylor Anhalt, for her keen eye in editing this article. This manuscript emerged from a presentation at the 2016 Third European Social Survey Conference in Lausanne, Switzerland.
