Life beyond the ballot box: the political participation and non-participation of electoral abstainers

Consistent with the ‘separate lives’ that the fields of electoral participation and social movement research have led (Kriesi 2008: 148; McAdam and Tarrow 2010), experts on electoral politics have not delved deeply into non-electoral forms of political participation of abstainers and social movement scholars have not generally engaged with factors and criteria of institutional behaviour.¹ On the one hand, variables of non-electoral forms of political participation are often disregarded as potential determinants of electoral turnout (e.g. Anduiza 1999; Blais 2006; Pattie and Johnston 1998, 2001). On the other hand, studies on non-electoral behaviour tend to neglect the existence of different channels of mobilisation in general (Hutter and Kriesi 2013: 292), and the role of electoral turnout as a potential predictor of extra-institutional mobilisation in particular (e.g. Dalton et al. 2010; Grasso and Giugni 2016; Schussman and Soule 2005; Vráblíková 2014).² However, in reality different forms of political participation are intimately intertwined and influence each other. In fact, politically motivated actions are a continuum that ranges from the most formal and routinised to the most disruptive and confrontational forms of action. In this article we depart from considering electoral turnout merely as the explanandum. Instead, we see abstainers as not a homogeneous whole and explore their patterns and forms of non-electoral political participation.

Almost five decades ago, Albert O. Hirschman (1970) argued that members of an organisation— understood as any form of human grouping (whether a nation, a business, an interest group, a political party, etc.)— have two basic alternatives at hand when they perceive a decrease in quality or benefit to members of that organisation. Members can either voice their discontent (by communicating their complaint, as well as sometimes acting upon it, in order to redress their grievances and improve their relationship with the organisation) or exit (withdraw from the relationship).³

In a context of generalised discontent, for example with the socioeconomic conditions and political status quo, many citizens choose to voice discontent through formal-institutional mechanisms. Disaffected voters might want to punish the incumbent by voting for established opposition parties, populist challengers, anti-establishment or independent forces (Passarelli and Tuorto 2014: 149; Kriesi 2012). Negative economic conditions not only affect who people vote for but whether they vote at all (Passarelli and Tuorto 2014: 148); abstaining is one further option at hand for disaffected citizens. When institutional channels for voicing grievances are ineffective, parties competing in the elections offer unsuitable alternatives or the next elections are simply too far off to provide an opportunity to voice grievances, then two alternative routes open up. While many might choose to withdraw from the game and disengage from politics (embracing the exit option), some citizens express discontent through alternative, less conventional and non-electoral forms of political participation (opting for the voice route). Whilst the theoretical interpretation of electoral abstention, as a dichotomy between an exit from politics and a voice of grievances through alternative forms of action, may be sound in theoretical terms it has been seldom explored empirically.

Especially from the 1960s onwards, a ‘new set of political activities has been added to the citizens’ political repertoire’ (Barnes et al. 1979: 149) and ‘increasing numbers of citizens have come to affirm the legitimacy of other forms of pressure on governments […] including signing petitions, lawful demonstration, boycotts, withholding of rent or tax, occupations, sit-ins, blocking traffic, and wildcat strikes’ (della Porta and Diani 2006: 166). As arenas for political participation are manifold, so are the alternative channels to voice discontent besides voting. Even though drivers and dynamics of mobilisation inside and outside the voting booths might be interlinked, some authors have argued that there is an inverse association between the opportunities available in the electoral arena and the choice of more unconventional forms of political participation (Piven and Cloward 1977).

In this article, we develop two central arguments. First, we contend that abstention does not exclusively entail an exit. While we observe a significant and positive correlation between electoral participation and engagement in other forms of action, there is a sizeable group of abstainers who choose to engage in other non-electoral political activities. We call this group ‘alternative voicers’, to distinguish them from those for whom electoral abstention means an exit from politics. Some discontinuities notwithstanding, key determinants of non-electoral participation are similar for voters and abstainers, including configurations of political values, social capital and network exposure.

Second, we develop a more complex, nuanced typology of non-electoral participation among abstainers based on their patterns of non-electoral behaviour. While a majority of abstainers tend to be exiters (they neither vote nor engage in other political activities), we distinguish three further subsets of alternative voicers: those who participate exclusively in media-related activities (e-activists), those who boycott and buycott (consumerists), and those who get involved more deeply in all kinds of non-electoral forms of political participation (super-activists). Further, we analyse the determinants of belonging to each group of abstainers. Although to a differing extent, we find that employment status, political values and social capital/ network exposure account for membership of one subset or another.

In order to conduct our empirical analysis, we used data from an original cross-national online survey conducted in 2015 across nine European countries in the context of the ‘Living with Hard Times’ (LIVEWHAT) research project funded by the European Commission under the auspices of their 7th Framework Programme— including France, Germany, Greece, Italy, Poland, Spain, Sweden, Switzerland and the United Kingdom (N = 18,367).⁴ The specialised polling agency YouGov conducted the online surveys between June 2015 and August 2015, using balanced country quotas in terms of sex, age, region, social class and education level in order to match national population statistics.⁵

The article is structured as follows: in the next section, we discuss some relevant literature and introduce our research questions; after that, we present our data, its operationalisation and empirical design; and then analyse the main results of the paper. In the conclusion, the main arguments of the paper are summarised and we signal further avenues for inquiry.

Scholars have studied the drivers of political participation broadly defined— i.e. looking at both electoral and extra-institutional arenas— in Europe in the shadow of the Great Recession and in other contexts (e.g. Kern et al. 2015; Dalton 2004). Specifically, levels of electoral participation— and, conversely, electoral abstention— are key ‘indicators of the democratic health of the systems themselves’ (Cazorla et al. 2017: 32). Although abstention ‘is perhaps one of the least studied phenomena in the field of electoral behaviour’ (Cazorla et al. 2017: 32), there are a wealth of studies that focus on the determinants of electoral turnout (Anduiza 1999, 2002; Blais 2000, 2006; Campbell et al. 1960; Franklin 2004; Pattie and Johnston 1998). However, in the words of Ragsdale and Rusk (1993: 721), while ‘researchers have analysed factors that distinguish voters from non-voters, […] they have not studied factors that reveal distinctions among non-voters’. In fact, we know very little about the patterns of political participation among abstainers beyond the ballot box. Yet, abstainers are not a monolith: as some contributions have already emphasised they respond, and fail to respond, to politics in different ways.

Based on a small sample of people from Vancouver in the 1963 federal election, Jean A. Laponce (1967) identified five types of voluntary non-voters:⁶boycotters (those who do not vote in order to question the legitimacy of the political system or of those who control it), retired electors (those who might have voted previously, but do not do so because of their age or health), barbarians (those less knowledgeable about politics), and spectators, who might be either poorly or well informed about politics (although they have an average interest in politics, they abstain due to unspecified reasons). Focusing on individuals’ attention to campaigns, Ragsdale and Rusk (1993) performed a cluster analysis that allowed them to distinguish among five distinct types of non-voters in the 1990 midterm Senate elections in the US: politically ignorant non-voters (they are uninformed about the campaign), indifferent non-voters (they do not observe a clear difference between the candidates), selectively aware non-voters (they are aware of only one candidate), dissatisfied non-voters (dislike one or both candidates), and the conditionally inactive (those who do not participate because they are socially and economically disadvantaged).⁷ More recently, de Waele and Vieira (2011) identified three different types of non-voters (the underprivileged, alienated and isolated) according to whether they did not vote because they could not, because they did not want to, or because nobody encouraged them to do so. Finally, Bagozzi and Marchetti (2017) distinguished among occasional and routine non-voters, depending on whether they failed to cast a vote due to temporary factors (e.g. distaste for the candidates running in a given election or poor weather conditions) or a consistent disengagement from the political process. As far as they go, all these proposed typologies are enlightening: they look at different motivations and reasons for non-voting. However, they share a major shortcoming: they overlook the differential effect of non-electoral political participation on patterns of abstention. In this paper we aim to fill this void, categorising abstainers not according to their motives or habits but to their different choices in terms of non-electoral political participation.

With regard to the nexus between electoral abstention and non-electoral political participation, two alternative approaches can be identified in the literature. On the one hand, some authors argue that norms of civic duty are weakening which leads to a decrease in community engagement and as a result general levels of political participation are on the decline (Pharr and Putnam 2000; Putnam 2000). From this perspective, electoral and non-electoral participation are qualitatively different forms of political behaviour (Vráblíková 2014: 212), and complement one another. As the resource mobilisation and civic voluntarism approaches show, not only are enhanced resources associated with increased prospects for conventional participation (Verba and Nie 1972), but also favour engagement through extra-institutional means (McCarthy and Zald 1977; Verba et al. 1995).⁸ Given the weakening power of the vote, citizens who feel alienated and disillusioned might not only refrain from voting but also might demobilise— or simply, not mobilise— in the non-electoral arena, opting to exit political participation. As an implication of this stream of literature, one would expect a negative association between abstentionism and non-electoral forms of participation.

On the other hand, citizens who believe that their vote is no longer influential (i.e. the power of the voting resource diminishes) might resort to non-electoral forms of action (Piven and Cloward 1977). People who give up on institutional forms of participation could be more likely to embrace non-electoral forms of political participation in order to channel and redress their grievances, eventually engaging in grassroots and non-institutional forms of action. According to this approach, non-electoral participation increases as electoral turnout decreases. Some scholars have emphasised that the political actions of some groups of citizens (especially those who are highly educated) are motivated by self-expressive values and are more critical of those in power and elite-defined forms of action (Norris, 1999). Similarly, Dalton (2008) argues that while traditional duty-based norms might be on the decline, those who value other aspects of democratic citizenship might be engaged as good citizens, resorting to non-electoral forms of political participation. Hence, ‘“engaged” or “critical” citizens are thus expected to shy away from traditional electoral acts such as voting, while they should have a much more positive attitude toward engagement beyond the electoral arena, such as protecting the environment or working within their own community’ (Hooghe et al. 2016: 117). Moreover, in the wake of the Great Recession disengagement might have ‘become increasingly difficult to sustain as an exit route, as austerity measures and labour market restructuring have restricted the possibility that people might avoid the more disciplinary elements of neoliberal capitalism’ (Bailey et al. 2018: 7). It is not coincidental that we have seen a proliferation of mobilisations by disengaged and disaffected actors, which have often resorted to extra-conventional and more rebellious forms of action.

In short, there are two alternative interpretations of the nexus between electoral abstentionism and non-electoral forms of political participation. Citizens who feel alienated from the political system as a whole tend to exit the political game. However, those who perceive the electoral procedure as not strategically effective or legitimate in comparison to other alternative forms of political participation would be more likely to become alternative voicers in order to advance their claims.

Our first research question consists of testing whether such two-way alternatives (exit or alternative voice routes) exist, and compare the drivers of non-electoral political behaviour for abstainers and voters. Using a latent class analysis of the patterns of non-electoral political participation among abstainers, we then go beyond the exit/ alternative voice pairing and come up with a four-fold typology of abstainers. Finally, we ascertain the determinants of these routes, testing which factors can account for differential belonging to the subsets in our typology.

By means of latent class analysis and regressions, we group non-voters’ and ascertain the determinants of each group. Latent class analysis (LCA) is a powerful but underused (at least in political science) statistical method to identify unobserved groups of cases in our data— though there are some recent exceptions in the field that make an extensive use of LCA techniques; e.g. Beissinger (2013); Hooghe et al. (2016); Portos and Masullo (2017). This finite mixture model gives a probability value for each individual to be assigned to a given group. The groups— more accurately, ‘classes’— are the unobserved (latent) variables. These could be different attitudes or political behaviour-sets of people which lead to certain response patterns in a survey (Beissinger 2013). We use LCA to identify unobserved groups among electoral abstainers on the basis of their records of non-electoral political participation.⁹

The LIVEWHAT (2014) survey has some useful information to empirically pin down abstainers. In line with standard empirical inquiries on abstentionism, we alternatively define as abstainers those who have not voted in the last national elections in the nine countries: in order to qualify as an abstainer, the interviewee must have responded that ‘was eligible to vote but did not do so’ in the last national elections (n = 2,205, which accounts for 13.26% of the total sample). While the main models reported in the manuscript are based on abstentionism (exclusively) in the last national elections, we also report in the Appendix the results with a more restrictive definition of non-voters.¹⁰

Table 1 represents the distribution of the subset of people who did not vote in the last national by country (for the subset of abstainers both in the last national and European elections, see Appendix, Table A1).¹¹ While the European countries that have suffered the impact of the recession only to a moderate or low extent give polarised figures, with either low (Sweden, UK) or high levels of abstainers (France, Germany, Poland, Switzerland), the most severely affected countries have intermediate values in terms of non-voter size (Italy, Spain, Greece)— see Zamponi and Bosi (2016).

In order to build the dependent variable, we rely on information from 16 dummies, which capture whether the respondent has engaged in some of the following non-electoral political activities in the last 12 months (Table 2). It includes the items that capture non-electoral forms of political participation, which are similar to those in the questionnaires of the most important longitudinal and cross-sectional datasets, such as the European Social Survey, the World Values Survey and the International Social Survey Programme. Among these variables, we can find: contacted a politician, donating money to political organisations, boycotting, petitioning, attended a political meeting, using social media for political purposes, demonstrating and striking. The LIVEWHAT survey extends the traditional repertoire of actions, considering different platforms (both online and offline) and degrees of disruption. Thus, it also asks respondents whether they have displayed or worn a logo or badge of a political campaign or organisation, resorted to ethical consumerism (i.e. buycotting), performed occupations, followed a political organisation on social media platforms, visited the website of a politician or party, used the Internet to search for political information, damaged public goods (e.g. broke windows, removed traffic signs) and used violence (e.g. clashed with the police). In line with extant literature (Hooghe and Marien 2013; Vráblíková 2014; Kern et al. 2015; Braun and Hutter 2016), we perform a principal component analysis in order to construct a summated rating scale for non-electoral political participation.¹² The summated rating scale is based on a tetrachoric correlation matrix (Table 3), and is robust.¹³

Following recent literature on the determinants of non-electoral political participation, we distinguish between four main groups of determinants: biographical availability, grievances, political values and network exposure (Schussman and Soule 2005; Dalton et al. 2010; Vráblíková 2014). We include them in order to strengthen our arguments against alternative explanations (Table 4).

First, certain biographical features and personal constraints increase the costs and risks of non-electoral participation, such as age, gender, having children and educational level (McAdam 1986: 70). Gender is measured through a dummy variable (1 = female; 0 = otherwise). Similarly, we use a dummy to capture whether the respondent (or his/her partner) currently have children under 18 years old living in the household (1 = yes; 0 = otherwise). A continuous indicator captures the age of the respondent— we also include age’s squared term to account for potential quadratic effects. A 9-point interval-level variable measures the highest educational level attained by the respondent, ranging from ‘primary school or less’ to ‘PhD or equivalent’.

Second, the respondent’s personal experiences of hardship and adverse economic circumstances could determine not only whether she decides to punish the incumbent by voting for the challenger (Lewis-Beck and Stegmaier 2000; Passarelli and Tuorto 2014) but also whether she engages in extra-electoral forms of action or not. A 1–10 increasing scale is used to capture the net monthly income level of the respondent’s household. Also, we consider whether the respondent reports being currently unemployed or not (1 = unemployed; 0 = otherwise). Although material conditions may be relevant, grievances are socially constructed thus considering how they are perceived is also important (Kern et al. 2015). A dummy variable measures financial self-sufficiency (whether the respondent declares to have periods of the month when s/he is confronted with ‘real financial difficulties’ or not). Finally, taking into account both objective and subjective aspects, a 0–11 summated indicator accounts for the worsening job conditions of those who remained employed.¹⁴

Third, certain political attitudes related to political engagement could likewise be important predictors of non-electoral political participation (Beissinger 2013: 575; Norris et al. 2005; Dalton et al. 2010). Political interest is captured through an ordinal variable that consists of a 1–4 increasing scale. A 10 point left-right interval-level variable measures ideological self-placement. We also take into account satisfaction with democracy, which is measured on an increasing 0–10 scale. In order to capture internal political efficacy, we use an index that results from combining three different indicators that happen to be moderately inter-correlated (0.51 < Pearson’s r < 0.65).¹⁵ Based on a 5-point scale that ranges from strong disagreement to strong agreement, external efficacy measures the extent of agreement with the statement ‘people like me don’t have any say about what the government does.¹⁶ Besides ideological self-placement, political interest and perceived (both internal and external) efficacy, trust in political institutions can be an important predictor of extra-institutional political participation (Braun and Hutter 2016). The LIVEWHAT survey has information on 10 indicators related to institutional trust, each of them being measured on a 0–10 increasing scale.¹⁷ As these 10 items are moderately-to-highly correlated (0.25 < Pearson’s r < 0.83), we perform a principal component analysis and include an institutional trust index as predictor of non-electoral political participation.¹⁸

Fourth, social capital and network exposure are important predictors of participation (Verba et al. 1995; Schussman and Soule 2005; Beissinger 2013; Laurison 2016). To capture these aspects, we incorporate three dummy variables on whether the respondent thinks that most people cannot be trusted, whether s/he is member of a political party or trade union, respectively. Also, a 1–4 ordinal variable that rages from ‘less than once’ to ‘almost every day’ captures how often the respondent met with friends who do not live in the same household during the last month.

While having resources, bearing certain attitudinal configurations and increased political awareness may affect individuals’ decision to engage in action, this decision is also likely to be shaped by the contexts within which individuals operate. Besides individual-level predictors, factors concerning system-level institutional frameworks (voting process, electoral and party systems) have been proved to be important predictors non-electoral participation (Vráblíková 2014; Braun and Hutter 2016). We use three types of country-level indicators: 1) 2014 GDP growth, borrowed from Grasso and Giugni (2016); 2) an index of cultural differences calculated by Braun and Hutter (2016) which combines information on whether societies are statist or not, aggregated party responsiveness and atmosphere of reliability in a society; 3) building on the work by Braun and Hutter (2016), three further variables capture institutional variation, including the Lijphart's executives-parties dimension, fiscal decentralisation and a direct democracy index.¹⁹

In order to answer the first research question, we ascertain whether the twofold exit/alternative voice route exists and study the determinants of non-electoral participation. Different OLS regression models are run, with the rating index of non-electoral political participation as the main dependent variable. Although our target group are abstainers, in order not to sample on the dependent variable, we proceed as follows.

First, we take the whole population and include electoral turnout in the last national election as the main predictor variable (models 1 and 4, Table 5). These specifications allow us to examine whether electoral abstainers are more or less keen to engage in non-electoral participation, thus embracing the exit or alternative voice routes. Second, we split the general population into two sub-samples depending on their turnout in the last national elections: abstainers (models 2 and 5) and voters (models 3 and 6). All models follow similar specifications. While we take into account a number of determinants that concern biographical features, grievances and political values in models 1-2-3 (Table 5), in models 4-5-6 (Table 5) we also incorporate variables related to social capital and network availability.

As expected, models 1 and 4 show a strong and positive association between voting turnout and non-electoral political participation. Taking model 4 as a reference, we can observe that casting a vote (in both the last European and national elections) leads to a 0.27 unit increase on average in the summated scale of non-electoral participation, keeping the other predictors constant. In line with the exit hypothesis, abstainers do not only refrain from participating at the voting booths but also are less prone to get involved in political activities beyond the ballot box. Additionally, we find that a number of factors make people keener to engage in non-electoral political action, such as being a woman, having more formal education completed, not having children, feeling socially and economically disadvantaged, and suffering worse job conditions during the crisis (models 1 and 4, Table 5).

Importantly, political values are key predictors of political participation outside the ballot box. In line with prior research, being politically interested, reporting low levels of trust in political institutions²⁰ and high standards of internal political efficacy are core factors motivating non-electoral political involvement. In addition, people who are left-wing, democratically dissatisfied and with high perceptions of external efficacy are (albeit to a small extent) more willing to engage in action (Beissinger 2013; Dalton et al. 2010; Hooghe and Marien 2013; Verba et al. 1995). Similarly, network availability and social capital favour non-electoral political participation: meeting with friends, being socially trustful and being a member of a party and trade union is positively associated with individual-level engagement in non-electoral political activities (model 4, Table 5). Although most of these results hold for the general population and the sub-sets of abstainers and voters, we can observe some discontinuities. For instance, models 2 and 5 show that education, having children and perceived deprivation do not explain non-electoral participation among abstainers.²¹ Conversely, interpersonal trust and democratic satisfaction do not help to predict non-voters’ NEP.

Based on the OLS regressions that have been specified in models 4-5-6 (Table 5), we have performed additional multi-level regression analyses as robustness checks (models 7-8-9, Table 5). They will allow us to analyse the impact of contextual factors of the individual’s likelihood of engaging in non-electoral participation. At the aggregate level, political-institutional factors do not seem to affect non-electoral behaviour, but country-level cultural differences do: internal cultural heterogeneity is positively correlated with non-electoral behaviour. However, this variable does not account for non-electoral participation among the abstainers’ subset. In a similar way, an increase in the country’s GDP seems to be associated with higher NEP, although this effect is constant neither across groups nor across the definitions of abstainers (see Table 5 and Table A2, Appendix). As expected, the effects of the individual-level predictors are robust.

Although alternative voicers deliver all sorts of non-electoral political activities to a larger extent than exiters, their patterns of political behaviour are far from constant (Figure 1). By looking at online political activities and consumerist behaviour (e.g. searching for political information on the Internet, discussing politics in social media, purchasing products for ethical reasons, etc.), alternative voicers are much more mobilised than exiters. However, consistent with the descriptive statistics in Table 4, neither exiters nor voicers are very keen to resort to the more disruptive tactics in absolute terms, such as occupying, damaging goods or engaging in violent clashes with the police. Additionally, semi-institutional activities report a mixed trend: relative to exiters, voicers tend to engage in actions such as signing petitions, but not in other forms of action (e.g. displaying logos and badges of a political campaign or organisation).

In short, we have observed a positive relationship between electoral turnout and non-electoral political participation. However, electoral abstention does not only entail exit: alternative voicers, who are about one third of non-voters, engage in non-formal political activities in order to redress their grievances. Going beyond the exit/alternative voice coupling, we next will try to propose a more nuanced typology of abstainers, and study the determinants of belonging to each class and its implications for the dynamics of abstentionism.

On the basis of their patterns of non-electoral political participation, we can distinguish four classes of abstainers. Class 1, 2, 3 and 4 encompass 63.3%, 4.5%, 16.7% and 15.5% of the respondents, respectively. Similar to the LCA with concomitant variables (Figure 1), we can observe that the class that includes most abstainers —about two thirds of the whole sub-sample— are political outsiders (class 1; exiters). Besides refraining from voting, about two thirds of the abstainers tend not to engage in other forms of political participation (neither in confrontational nor in semi-institutionalised, consumerist, or online activities). In sharp contrast, the small numbers of abstainers who fall within the super-activist class tend to resort to all kinds of actions (class 2). However, although the probability of engaging in the most radical activities such resorting to violence or damaging goods is higher than for the other classes, it remains small in absolute terms. Besides embracing more radical repertories than other abstainers, this subset is also more willing to engage in media-based and semi-institutionalised repertories of action as well as activities related to consumerist behaviour.

Similar to super-activists, members of the e-activist class (i.e. class 3) also report relatively high values in digital-based political activities, such as: searching information about politics online, discussing politics on a social network site and visiting the website of a politician or a political party. However, in contrast to super-activists, the e-activists encompass individuals who are almost exclusively engaged in digital activities. The predicted values of cases in the e-activist class are also relatively high for petitioning. Since we cannot disaggregate online and offline petitioning, this form of action tends to be Internet-based these days and thus considered part of digital repertoires. Together with engaging in some Internet-based activities such as searching political information on the Web (and also petitioning), there is a fourth subset among abstainers who tend to engage in activities such as boycotting and the ethical purchase of products (i.e. buycotting). Thus, the most salient and distinctive feature of individuals that fall upon class 4 is that they tend to engage in consumerist-related behaviour.

All in all, when focusing on the non-electoral behaviour of non-voters, there does not appear to be one single exit/alternative voice route. Although a majority of abstainers tend to exit the political game, there are different classes of non-voters that are active by other means: the e-activists tend to exclusively participate in media-related activities, the super-activists get involved to a deeper extent in all sorts of non-electoral forms of political participation and the consumerists tend to participate in boycotting and buycotting. Drawing on the estimated values of posterior probability membership for each class, we build one 4-category variable that classifies abstainers according to their NEP features. This is used as the dependent variable in the multinomial logit regressions (models 1-2, Table 6) and the two-level multinomial logistic model (model 3, Table 6) we run next.²⁵ Multinomial regressions allow us to shed light on the factors that underpin belonging to each class, taking the exiters (class 1) as the reference category.

Overall, we observe that different predictors account for each class’s membership. In other words, the determinants of e-activism, super-activism and consumerism (always relative to exiting the political game) are not homogeneous. Consistent with the literature on the mobilisation of excluded and marginalised groups (Baglioni et al. 2008), indicators of grievances helps us to distinguish exiters from the other classes of non-voters. Relative to abstainers who do not engage in non-electoral political activities, both super-activists and consumerists have seen their job conditions worsen during the recession. From this standpoint, e-activists and exiters are people who have been less severely hit by economic hardship.

The most important predictors of class membership are related to political values and attitudes. Right-wing ideological self-placement and political trust negatively affect the probabilities of falling within the super-activists and e-activists subsets. Perceptions of external political efficacy are lower among super-activists. Importantly, political active abstainers (either super-activists, e-activists or consumerists) are more politically interested and have higher perceptions of internal political efficacy than exiters. These results are consistent with the defining features of Norris’s (1999) critical citizens. There is a group of critical abstainers (some groups, to be more accurate) who might show disaffection towards political institutions but are politically sophisticated and engaged. Although they choose not to cast a vote, alternative forms of political participation may provide a channel for these wary— yet politically involved— citizens to redress their grievances. Conversely, more passive and less politically involved abstainers (in terms of perceptions of internal efficacy, self-reported political interest and political trust) would tend to exit the political game. Exiters not only refrain from engaging in elite-defined forms of action but also withdraw from alternative, elite-challenging political activities. Given the willingness and capabilities needed to use the Internet for political purposes, these findings also are in line with existing literature on e-activism, which has found political interest and efficacy as strong determinants of online participation (Van Laer 2010).

Belonging to the super-activists class is also explained on the basis of social capital and network availability (Norris et al. 2005; Verba et al. 1995): relative to exiters, members of a political party are much more willing to fall upon the super-activists’ subset. In turn, biographical availability explains belonging to the consumerist class: relative to class 1, being a woman and having more education increase the chances of falling within the class that mostly involves engaging in buycotting and boycotting activities. Although low levels of political trust increase the chances of resorting to consumerism in relation to taking the exit option, interpersonal trust has the opposite effect. As Neilson and Paxton (2010: 18) posit, ‘positive social interactions provide motivation and information that empower consumers to take action for social change through their purchase decisions’. Finally, some exceptions notwithstanding, aggregate-level predictors do not seem to explain belonging to the three activist classes. Among these exceptions, we find that an increase in GDP growth increases the probabilities of falling within the consumerist class relative to being an exiter.

Leading proponents of participatory theories of democracy have long argued that a more extensive, generalised type of political involvement is necessary in large and complex polities (Pateman 1970; Barber 1984). These scholars contend that a more inclusive democratic model is needed, characterised by broadened participation and deliberative practices, which allows citizens to actively participate in the decision-making process— both inside and outside institutions. As della Porta (2013: 189) suggests,

We have qualified Hirschman’s 1970s thesis, that when people are unhappy about the political status quo (i.e. not merely about the incumbent) they may choose to exit the political game (i.e. neither cast a vote nor engage in further political activities) or voice their discontent (through extra-electoral forms of participation). We argue that even though abstainers tend to exit the political game— i.e. there is a positive association between voting and engaging in further non-electoral political activities—, there is a critical flank that gives up institutional activities while embracing alternative forms of action. Although some predictors related to grievances and biographical availability are useful to explain which route abstainers take, compared to exiters, alternative voicers encompass the most politically involved and critical abstainers. Moreover, we have gone beyond the exit/voice pairing by proposing a four-fold typology of abstainers on the basis of their patterns of non-electoral behaviour: exiters, e-activists, super-activists and consumerists. While reporting higher levels of education and being a female are positively associated with falling in the consumerist class, both consumerists and super-activists are more aggrieved than exiters. Political values and attitudes are crucial to understand all kinds of political activism among abstainers, especially political interest and perceived internal efficacy. Although consumerists are defined by their high standards of social trust, right-wing ideology and political trust negatively affects the chances of belonging to the consumerist (and super-activist) classes, as well as some country-level features.

This article may contribute to open up avenues for further research. First, it cannot be discounted that patterns of abstention differ in changing electoral contexts. Additional inquiries could test whether the proposed typology of abstainers holds in backgrounds other than Europe during the (post) Great Recession. Second, further research could try to ascertain the determinants of adopting the alternative voice route (i.e. abstaining, but engaging in non-electoral forms of action) relative to embracing institutional routes, to show discontent with the political status quo, such as voting for anti-establishment, populist and/or challengers parties. Third, we have taken exiters as a homogeneous whole. However, if focusing exclusively on this subset of passive authors, we could have found different motivations behind the decision of exiting, as some earlier contributions have shown. Drawing on unique data on Swedish adolescents, Amnå and Ekman (2014) find that non-participation (in non-electoral political forms of action) is not a uni-dimensional phenomenon: some young people do not participate in politics either because they are unengaged or disillusioned, but there is also a constituency of highly-informed ‘standby citizens’ who are prepared for political action should circumstances warrant (Amnå and Ekman 2014). These limitations notwithstanding, we have tried to make an uncommon move in the field by not only developing a new typology of electoral abstainers on the basis of their forms of non-electoral political participation, but by shedding light on the determinants that drive belonging to each of these sub-groups.

No potential conflict of interest was reported by the authors.

Lorenzo Zamponi is an assistant professor at the Scuola Normale Superiore in Florence (Italy), where he is part of the COSMOS (Centre on Social Movement Studies) research team. His research interests include memory, contentious politics and media analysis. His publications focus mainly on the recent wave of anti-austerity protest in Europe, on the cultural elements of social mobilisation and on the emergence of non-protest based forms of collective action.

Lorenzo Bosi is Assistant Professor in Political Sociology at the Scuola Normale Superiore. He received his Ph.D. in politics from Queen’s University, Belfast, and is the past recipient of the ECRC (University of Kent), Jean Monnet and Marie Curie (EUI) postdoctorate fellowships. He is a political sociologist pursuing comparative analysis into the cross-disciplinary fields of social movements and political violence.

Martín Portos is Postdoctoral Fellow at the Department of Political and Social Sciences, Scuola Normale Superiore (Florence, Italy). He holds a PhD from the European University Institute. Winner of the Juan Linz Best Dissertation Award in Political Science and the ISA's Seventh Worldwide Competition for Junior Sociologists, he has published on political participation, social movements, inequalities and nationalism.