Protest participation has become a routine part of political engagement in contemporary democracies, complementing more conventional activities such as voting (Aytaç and Stokes Reference Aytaç and Stokes2019; Borbáth and Hutter Reference Borbáth, Hutter, Giugni and Grasso2022; Galais Reference Galais2014; Oser Reference Oser2017; Quaranta Reference Quaranta2018; Schussman and Soule Reference Schussman and Soule2005; Tosun Reference Tosun2025; van Aelst and Walgrave Reference van Aelst and Walgrave2001; Theocharis and Van Deth Reference Theocharis and Van Deth2018). Rather than signalling crisis, protest often reflects demands for greater responsiveness from political elites when formal political channels are perceived as insufficient. A large body of research has linked protest participation to political dissatisfaction and representational deficits (Anderson and Mendes Reference Anderson and Mendes2005; Dalton et al. Reference Dalton, Van Sickle and Weldon2009; Griffin et al. Reference Griffin, de Jonge and Velasco-Guachalla2021). Scholars have also emphasized the importance of institutional contexts, showing that electoral systems, government structures and political opportunities shape citizens’ likelihood of engaging in protest (İlgü Özler Reference İlgü Özler2013; Kitschelt Reference Kitschelt1986; Su Reference Su2015; VanDusky-Allen Reference VanDusky-Allen2017).
Research shows that voters who support winning parties – those forming the government – report higher levels of satisfaction with democracy compared to voters of losing or opposition parties (Anderson and Guillory Reference Anderson and Guillory1997; Blais and Gelineau Reference Blais and Gelineau2007; Curini et al. Reference Curini, Jou and Memoli2012; Janssen Reference Janssen2024; Singh et al. Reference Singh, Karakoç and Blais2012; VanDusky-Allen Reference VanDusky-Allen2017). This satisfaction gap has significant implications, particularly in the realm of political activism. Losers, feeling underrepresented and dissatisfied with political outcomes, are more likely to engage in protest behaviour, demonstrating their willingness to challenge the status quo (Anderson and Mendes Reference Anderson and Mendes2005; Curini and Jou Reference Curini and Jou2016; van der Meer et al. Reference van der Meer, van Deth and Scheepers2009; VanDusky-Allen Reference VanDusky-Allen2017).
Beyond simply winning or losing, various characteristics of political parties contribute to their voters’ satisfaction. Even when in government, parties differ in their ability to represent their voters’ preferences (Aksoy and Tosun Reference Aksoy and Tosun2024; Curini et al. Reference Curini, Jou and Memoli2012; Singh et al. Reference Singh, Karakoç and Blais2012). This is especially relevant in European democracies, where multiparty governments are the norm. In such systems, being a senior or junior coalition partner significantly impacts a party’s influence in policy-making. While junior parties in government can showcase their governance potential and differentiate themselves from non-governing parties, they typically lack the same influence in shaping policy. Consequently, not all winners in the election experience equal levels of representation.
The difference between junior and senior coalition parties in terms of policy influence aligns with Gamson’s Law, which states that government portfolios are proportionally distributed according to each party’s seat share in the coalition. Consequently, the ranking of policy influence can be formulated such that the senior or prime minister’s party becomes dominant in decision-making, followed by junior coalition partners, with opposition parties having the least influence (Fortunato et al. Reference Fortunato, Lin, Stevenson and Tromborg2021). This hierarchical order of the parties creates an opportunity to study how the size of a party affects protest behaviour over and above its winning or losing status.
The role of party size in relation to policy influence within coalition governments is well-documented, showing that as a party’s seat share increases, so does its policy influence (Fortunato et al. Reference Fortunato, Lin, Stevenson and Tromborg2021; Martin and Vanberg Reference Martin and Vanberg2014). However, the implications for opposition parties are less clear. Recent studies show that larger opposition parties are more likely to have a policy influence if they have a sizeable number of seats in parliament, which allows them to control committee chairs (Fortunato et al. Reference Fortunato, Martin and Vanberg2019). What is also evident is that voters attribute more policy influence to larger opposition parties than smaller ones (Fortunato et al. Reference Fortunato, Lin, Stevenson and Tromborg2021). This suggests that voters of smaller opposition parties may experience greater deprivation of representation. Therefore, considering the established rank-ordering of policy influence by party size, it follows that party size could play a significant role in shaping the protest behaviour of small opposition party voters.
In this study, I follow the previous literature by differentiating parties based on their government status, as this status influences their ability to represent their voters’ preferences. However, beyond government participation, factors such as whether a party is a junior or senior coalition partner, and its size, also contribute to the impact of political inequality on protest participation. I hypothesize that voters of junior coalition parties are more likely to protest than those of senior partners. Additionally, I expect the probability of protesting to increase as party size decreases, due to the limited influence smaller parties have in the decision-making process. However, I anticipate that party size will have a more pronounced effect on the protest behaviour of opposition party voters, given that junior coalition partners, despite their limitations, still wield more influence than opposition parties.
To test these hypotheses, I use the data from the eight rounds of the European Social Survey (ESS) from 2004 to 2018, which provides me with 54 country-year observations of Western democracies. The study uncovers several implications for political representation, party politics and protest behaviour. First, this study provides a novel perspective for understanding the role of political parties in representing their voters by focusing on their capacity to do so. While I take size as one of the most prominent qualities of power, my main focus is on factors that contribute to representational inequality other than the government status of political parties. I therefore differentiate between parties in government and in opposition according to their magnitude in numbers (seats), which may better capture their level of policy influence. By focusing on the magnitude of parties rather than government status alone, the study highlights an important but often overlooked dimension of representational capacity. This has broader implications for understanding dissatisfaction and political engagement in multiparty democracies, where coalition governance is the norm and party competition is increasingly fragmented.
Finally, by linking internal coalition dynamics to protest behaviour, the study sheds light on how complex patterns of party competition shape unconventional forms of participation. As dissatisfaction stemming from limited policy influence translates into protest, these findings have important implications for assessing the resilience of representative democracy in an era of growing citizen activism. Rens Vliegenthart et al. (Reference Vliegenthart, Walgrave, Wouters, Hutter, Jennings, Gava, Tresch, Varone, Grossman, Breunig, Brouard and Chaques-Bonafont2016) show that protests have greater agenda-setting power in majoritarian systems – where policy-making is concentrated – because institutional responsiveness is lower and political actors are less inclined to integrate citizen demands. In contrast, in consensus-based parliamentary systems, political actors are more likely to respond to citizens through institutional means, making protest less necessary and less effective.
The emphasis on institutional responsiveness provides an important bridge between macro-level opportunity structures and the argument I present here: it is voters’ perceptions of their party’s power within these institutions that determines whether they feel heard or compelled to protest. In line with this, studies find that opposition parties are more likely to align with mobilization strategies outside formal channels because they lack access to policy-making (Hutter and Vliegenthart Reference Hutter and Vliegenthart2018). Ruud Wouters et al. (Reference Wouters, Sevenans and Vliegenthart2021) demonstrate that parties’ responses to protests depend on their institutional power, highlighting that larger and governing parties engage with formal channels more than smaller opposition parties. Furthermore, later studies show that elite responsiveness to protests on social media is similarly structured by the institutional power capabilities of parties (Barrie et al. Reference Barrie, Fleming and Rowan2024; Schürmann Reference Schürmann2024; Wouters et al. Reference Wouters, Staes and Van Aelst2023).
These studies focus on parties in government and in opposition to demonstrate the difference between their ability to use formal structures to be responsive to their voters. However, not all opposition parties are equal in this regard. Larger opposition parties benefit from institutional advantages – such as parliamentary resources and committee access – that allow them to challenge government policies through formal means (Fortunato et al. Reference Fortunato, Lin, Stevenson and Tromborg2021, Reference Fortunato, Martin and Vanberg2019). In contrast, the voters of smaller opposition parties, lacking these institutional channels, must rely more on protest to exert political pressure.
Building on this literature, the article suggests that party size plays a central role in structuring not just the responsiveness of parties to protest, but also the conditions under which protest becomes a necessary outlet for political expression. Responsiveness, in this sense, is not only shaped by the government–opposition divide, but also by how large or influential a party is within its institutional context. This representational gap is particularly pronounced for smaller opposition parties, whose voters may feel most excluded from the decision-making processes.
In the current study, I examine how government status and party size shape citizens’ perceptions of representation and, consequently, their likelihood of protesting. The following section builds on this theoretical framework to explore how these dynamics differentiate political winners from losers in terms of protest behaviour.
Political winners, responsiveness and protest
Representational inequality can arise when the parties that individuals vote for are unable or inefficient in carrying out the policies that align with their voters’ preferences. Studies examining protest participation often include variables that account for factors contributing to this inability or inefficiency of parties to represent their voters, such as the government status of parties, their ideological positions, and the power dynamics within the parliament (Anderson and Mendes Reference Anderson and Mendes2005; Quaranta Reference Quaranta2015; Su Reference Su2015; VanDusky-Allen Reference VanDusky-Allen2017).
One of the most prominent mechanisms shaping power distribution in parliament is the winning or losing status of political parties. Anderson and Mendes (Reference Anderson and Mendes2005) examine how individuals’ electoral winning or losing status influences their voters’ protest behaviour. They argue that elections not only allocate the majority and minority parties in parliament, but also shape the perceptions of voters towards political institutions. Christopher Anderson and Silvia Mendes (Reference Anderson and Mendes2005) categorize individuals as either winners (majority) or losers (minority) based on whether they voted for a party in government or in opposition. Losers, who are less represented compared to winners, tend to be less satisfied with the functioning of democracy.Footnote 1 Consequently, losers are more likely to engage in protest as a form of unconventional political action to challenge the status quo.
Anderson and Mendes (Reference Anderson and Mendes2005) likewise posit that since the political majority sets public policy, voters of governing parties tend to be more satisfied with the functioning of democracy and are thus less likely to protest compared to opposition voters. However, the authors do not take into account other aspects of the power balance in parliament that might affect the degree of representation for both winners and losers. Notably, the specific composition of the government warrants closer examination, as it can make some parties more or less influential in shaping policy outcomes.
Julie VanDusky-Allen (Reference VanDusky-Allen2017) extends the research on the impact of the winner/loser distinction on protest participation by considering the influence of majoritarian versus proportional representation systems. In majoritarian systems, winning parties can make policy decisions more easily without the need for bargaining with other parties. Because winning parties in these systems hold greater power, governments tend to be less responsive to the demands of losers compared to proportional systems. Consequently, in countries with single-party majority governments, losers are more likely to protest than in countries with coalition governments, where power is more distributed. The research highlights the importance of government composition, which can influence the protest potential of the winning parties’ voters.Footnote 2
While winners in majoritarian systems generally protest less than those in proportional systems, some winners within the same country may still be more inclined to protest if they are dissatisfied with how decisions are made. In most parliamentary democracies, governments are formed through a lengthy process of coalition negotiations and policy compromises between political parties. This process can bring smaller parties with moderate vote shares into government, while keeping out larger parties. Being part of the government can indeed be a factor in being accepted as part of the majority, as decision-making requires the cooperation of smaller parties (Norris Reference Norris2008).
However, coalition bargaining does not guarantee equal influence for all governing parties. Coalition agreements disproportionately favour senior coalition partners, as the prime minister’s party and larger coalition members secure greater control over the policy agenda (Klüver and Spoon Reference Klüver and Spoon2020; Martin & Vanberg Reference Martin and Vanberg2014). Junior coalition parties, despite gaining access to government, often struggle to secure key ministerial positions or policy concessions (Bäck et al. Reference Bäck, Debus and Dumont2011). Empirical studies show that coalition negotiations significantly impact voter perceptions of representation, as junior coalition parties frequently fail to deliver on the expectations of their electorate (Fortunato & Stevenson Reference D and TS2013).
Although junior coalition partners may benefit from being in government, enabling them to gain credibility and prestige in the political arena (Klüver and Spoon Reference Klüver and Spoon2020), their power to shape the policy agenda remains limited compared to that of senior coalition partners. Smaller parties often have less influence in decision-making, as policy control tends to be concentrated in the largest governing party and its key ministers (Fortunato et al. Reference Fortunato, Lin, Stevenson and Tromborg2021; Strøm et al. Reference Strøm, Muller and Bergman2008). Consequently, voters of junior coalition parties often feel politically marginalized, despite their party being in government (Blais et al. Reference Blais, Morin-Chassé and Singh2017).
As the diversity of issues supported by different party members increases, the likelihood of political parties compromising their positions also rises (Laver and Shepsle Reference Laver and Shepsle1990). Junior coalition partners, in particular, face challenges in maintaining their ideological distinctiveness while collaborating with larger parties. This can blur their policy positions, making it difficult for voters to discern what these smaller parties specifically contribute to the coalition (Falco-Gimeno and Fernandez-Vazquez Reference Falco-Gimeno and Fernandez-Vazquez2020; Fortunato and Adams Reference Fortunato and Adams2015; Spoon and Klüver Reference Spoon and Klüver2017). Additionally, junior parties are often held responsible for the coalition government’s actions, despite having less influence over major decisions. This dual challenge – losing their distinct identity while sharing responsibility for decisions they do not fully control – can result in a loss of voter support and increased voter dissatisfaction (Klüver and Spoon Reference Klüver and Spoon2020).
Empirical studies further demonstrate that coalition complexity itself lowers voter satisfaction, particularly when smaller governing parties fail to deliver on their electoral promises due to coalition constraints (Aarts and Thomassen Reference Aarts and Thomassen2008; Banducci and Karp Reference Banducci and Karp2003). Research also finds that coalition governments can reduce voters’ sense of efficacy, as compromises dilute ideological clarity and weaken policy responsiveness (Blais et al. Reference Blais, Morin-Chassé and Singh2017). Consequently, those who vote for junior coalition parties are not politically equal to those who vote for the major parties in government, because junior parties have limited power to shape decisions: voters of these junior parties face political inequality because the majority party holds more influence. Therefore, despite being labelled as winners, these voters are likely to be less satisfied with the democratic process. This dissatisfaction increases the likelihood of protest action among the voters of junior coalition parties. Accordingly, the first hypothesis can be formulated thus:
Hypothesis 1: Voters of junior coalition parties participate in protests more than the voters of senior coalition parties.
The role of party size
Gamson’s Law suggests that the more seats a party holds, the more influence it wields in policy-making – a pattern consistently supported by research on coalition governments (Fortunato et al. Reference Fortunato, Lin, Stevenson and Tromborg2021; Laver and Shepsle Reference Laver and Shepsle1990; Martin and Vanberg Reference Martin and Vanberg2014). A larger seat share strengthens a party’s bargaining power, which in turn increases its ministerial allocations, policy influence, and ability to deliver on voter expectations (Klüver and Spoon Reference Klüver and Spoon2020; Warwick and Druckman Reference Warwick and Druckman2001). Consequently, voters of smaller governing parties may perceive their party’s limited power as a form of political underrepresentation, making them more likely to engage in protest.
While coalition politics makes size a relevant factor for government parties, size is even more critical for determining political influence among opposition parties. Unlike governing parties, whose influence is directly linked to cabinet participation, opposition parties’ power depends on their ability to influence policy-making through institutional tools such as legislative committees and agenda-setting power (Fortunato et al. Reference Fortunato, Martin and Vanberg2019; Martin and Vanberg Reference Martin and Vanberg2014). Larger opposition parties are more likely to secure committee chairs, shape legislative debates and exert indirect influence on policy outcomes, making them more responsive to their voters’ preferences. In contrast, smaller opposition parties lack these institutional levers, limiting their ability to translate voter demands into political action (Fortunato et al. Reference Fortunato, Martin and Vanberg2019, Reference Fortunato, Lin, Stevenson and Tromborg2021).
In addition to institutional factors, party linkages further condition the effect of party size on protest behaviour. Michael T. Heaney and Fabio Rojas (Reference Heaney and Rojas2015) argue that stronger partisan ties constrain protest participation when a party is well-established and politically integrated. This is particularly relevant for larger opposition parties, which tend to have more organized partisan networks and a greater institutional presence. As a result, their voters may be less likely to mobilize against the system through protest. In contrast, smaller opposition parties, which have weaker ties to formal institutions and fewer pathways for institutional influence, may actively encourage protest as a political strategy.
Thus, party size should have a more pronounced effect on the actions of parties in opposition compared to those in government, as larger opposition parties can channel dissatisfaction through formal mechanisms, whereas smaller opposition parties rely on protest as a primary means of influence. This reinforces my expectation that size conditions protest behaviour most significantly among the voters of opposition parties, whose level of institutional representation is already lower than that of parties in government.
Considering party size alongside government–opposition status, I conceptualize political parties along two intersecting dimensions: institutional power and governing responsibility. Larger coalition parties are typically senior governing partners that hold substantial policy-making authority and have greater control over the legislative agenda. Smaller coalition parties, by contrast, are junior partners with limited influence, although they still benefit from some access to government decision-making. Within the opposition, larger opposition parties possess significant parliamentary representation and institutional tools – such as committee leadership – which grant them influence even outside government. However, smaller opposition parties lack both legislative power and institutional responsiveness, leaving their supporters with few formal channels for political influence.
While government status affects access to policy-making, party size determines a party’s ability to represent its voters effectively within its institutional constraints. Smaller coalition parties, despite their limited power, still benefit from access to government decision-making, making their voters less deprived than those of smaller opposition parties, who experience the most significant representational disadvantage.
Thus, while seniority versus juniority in coalitions captures within-government variation, party size is a broader and continuous measure that extends to both government and opposition parties, shaping how effectively they can translate voter preferences into policy. I therefore expect that size will have a general negative effect on protest participation but will be especially pronounced among opposition party voters, who lack formal policy-making power.
Accordingly, the next hypotheses are formulated as follows:
Hypothesis 2: As the size of a coalition party decreases, its voters’ probability of protesting increases.
Hypothesis 3: The negative effect of party size on protest participation is stronger for voters of opposition parties.
Research design
In this study, my aim is to understand the relationship between individuals’ party preferences, the influence these parties have on policy formation, and individuals’ protest behaviour. To do so, I rely on an individual-level dataset to test my hypotheses and uncover the causal mechanisms between party politics and protest behaviour.
At the individual level, I use data from eight waves of the ESS to gather information on individuals’ protest participation and voting preferences.Footnote 3 The propositions are heavily reliant on regular party competition supported by democratic institutions. In democratically advanced countries, citizens are provided with stable and legitimate party options, ensuring democratic accountability through political institutionalization (Dalton and Weldon Reference Dalton and Weldon2007). In contrast, in many Eastern European countries, democratic transformation was not accompanied by strong democratic institutionalization or vigorous party competition. As a result, political representation by parties in these countries is influenced by factors beyond the scope of this study. Consequently, although the ESS provides data for most European countries, the analysis is limited to Western European countries.Footnote 4
In all waves of the ESS, respondents were asked about their previous protest experience. Within a broader framework, they were inquired about their participation in seven different forms of political action other than voting. Although the dataset provides information on respondents’ participation in various protests, such as signing petitions and joining boycotts, I focus specifically on the question regarding respondents’ previous participation in public demonstrations. Accordingly, the dependent variable is a binary variable coded as 1 if an individual participated in public demonstrations in the last 12 months, and 0 otherwise.
In the sample of 66,578 respondents, only 8% of individuals answered yes to participating in a demonstration. Among the countries in our sample, France scored highest with 20.2%, and Finland scored lowest with 1.1%.Footnote 5
The first independent variable in the study is a categorical variable that takes the value 0 for opposition parties, 1 for junior coalition partners and 2 for senior coalition partners, based on the parties respondents voted for in the previous elections. To determine the status of the parties, I used the Parliament and Government (ParlGov) dataset, which compiles data on European parties and cabinets from the 1920s onwards. After processing the data to generate this variable, I found that 33.2% of the sample had voted for senior coalition partners, 19.7% for junior coalition partners and 47.1% for opposition parties. The second independent variable is a binary version of the first, coded as 0 if the respondent voted for an opposition party and 1 if they voted for a government party.
The third independent variable in the study is a continuous variable that measures party size. I consider party size to be a key factor in understanding a party’s policy influence in parliament. Grounded in Gamson’s Law, which posits that the distribution of cabinet positions among coalition partners is proportional to the number of seats each party holds in the legislature, I measure party size by seat share rather than vote share.Footnote 6 The information is obtained from the Comparative Manifesto Project (CMP) dataset.
In constructing these three primary independent variables, I carefully examined the timing of elections, cabinet periods and fieldwork dates. To accurately match individuals’ party preferences from the previous elections with the CMP and ParlGov datasets, I created a dataset that provides information on the ESS fieldwork periods for each country and the date of the last election before this fieldwork. Since the protest participation question asks whether respondents participated in demonstrations within the previous 12 months, I ensured that the elections were held more than a year before the start of the fieldwork. I also verified that there was no interim cabinet change during this period.Footnote 7
In addition to the primary variables, I included several independent variables relevant to the literature and necessary to account for compounding factors. In the literature, trust in people (Bernhagen and Marsh Reference Bernhagen and Marsh2007) and political parties (Braun and Hutter Reference Braun and Hutter2014) are essential variables that positively affect protest participation. To account for trust, I include a question asking respondents’ degree of trust in political parties, ranging from 0 (no trust at all) to 10 (complete confidence). For the role of individual resources on political participation (Brady et al. Reference Brady, Verba and Schlozman1995), I include the political interest variable, which takes a value between 1 (not at all interested) and 4 (very interested). Additionally, I account for the role of ideology by including a question asking respondents to place themselves on an 11-point left-right scale, ranging from 0 to 10 (Barnes and Kaase Reference Barnes and Kaase1979; Bernhagen and Marsh Reference Bernhagen and Marsh2007; Kostelka and Rovny Reference Kostelka and Rovny2019; van der Meer et al. Reference van der Meer, van Deth and Scheepers2009).
A control for economic grievance is also included, as people’s perceptions of the discrepancy between their value expectations and value capabilities lead them to participate in mass protests (Gurr Reference Gurr1970). This takes the form of a four-scale feeling about income variable at the individual level. I also use the Gini inequality index for the countries in the sample to show the effect of country-level inequality on protest participation.Footnote 8
In the political opportunity framework, more open structures increase mobilization activity and the recruitment capabilities of mobilizing organizations (Kriesi et al. Reference Kriesi, Koopmans, Dyvendak and Giugni1995; Tarrow Reference Tarrow2011). However, from a comparative perspective, there are limited indicators to measure the interconnectedness between political opportunities, political organizations and individual participation in non-electoral forms of political participation. Katerina Vrablikova (Reference Vrablikova2014) focuses on organizational membership to account for one facet of this relationship. In this study, I also account for the role of organizational resources with a binary variable that takes a value of 0 when respondents are not members and 1 if they are trade union members.
The disproportionality index is used as one indicator of the quality of representation (Gallagher Reference Gallagher1991; Lijphart Reference Lijphart2012; Powell Reference Powell2000), while at the country level, the disproportionality variable calculated in the CMP dataset is used, which considers the difference between parties’ vote shares and seat shares. Lastly, I include additional individual-level control variables of gender, age and education.
Empirical findings
All the models reported in Table 1 employ logistic regression with robust standard errors in parentheses.Footnote 9 Models 1 and 2 exclude all variables that have a theoretically important effect on protest behaviour. The results show that the main findings remain robust across model specifications. In both Models 1 and 2, party size and government status remain significant predictors of protest participation, in the expected direction and magnitude. Even after the inclusion of post-treatment variables in Model 3 and beyond, the coefficients for party size and government status remain statistically significant and substantively meaningful, though – as expected – slightly attenuated.
Logistic Regression Estimates
Notes: Country fixed effects in Model 3, Model 4 and Model 5 are omitted from the table. Robust standard errors in parentheses. Two-tailed tests.
* p<0.1, ** p<0.05, *** p<0.01
In the subsequent models I incrementally include control variables and fixed effects to account for potential confounders. Model 3 includes only individual-level predictors alongside party size and government status but does not incorporate any fixed effects. Models 4, 5 and 6 introduce country fixed effects to account for unobserved heterogeneity across national contexts. Model 4 adds country-level variables, specifically, disproportionality and income inequality. Model 5 builds on this by including a categorical variable distinguishing between junior and senior coalition partners. Finally, Model 6 extends the analysis by incorporating an interaction between party size and government status to examine whether the effect of size varies across governing and opposition parties.
To show the effect of the junior/senior coalition partner variable in Model 5, I plot the predicted probabilities in Figure 1 by setting all other variables to their means and modes for the values of the main independent variable. The coefficient estimates and the plotted predictions show distinguishing results for the protest behaviour of opposition, junior and senior coalition party voters. According to Figure 1, the probability of participating in a protest for opposition party voters is around 9%, while the probability decreases to 7% for junior coalition party voters. For voters of the senior coalition partners, the probability decreases to less than 6%. The difference between opposition and senior coalition party voters is 3%. Considering the baseline probability of participating in a protest is 8%, this finding shows that the probability of participating in a protest decreases by nearly 40%. When the predicted probabilities between junior and senior coalition party voters are compared, the 1.5% decrease indicates approximately an 18.5% decrease in protest potential. This finding supports the hypothesis that junior coalition party voters protest more than senior coalition party voters. Additionally, the predictions show that losers – opposition party voters – protest more than government party voters, which supports the findings of previous studies on protest behaviour.
Predicted Probabilities of Protest Participation for Opposition, Junior Coalition and Senior Coalition Parties
To comment on the substantive significance of the logistic regression estimates in Model 6, I plot predicted probabilities by allowing only the primary independent variables to vary, as shown in Figure 2. Lastly, I include an overlaid histogram to display the distribution of the party size variable.
Predicted Probabilities of Protest Participation by Party Size for Government and Opposition Parties
While estimating the predictions, the party size variable is allowed to vary between 0% and 50%. However, while the maximum size of a party in government is off the scale at 63%, the maximum size of an opposition party is only 41%. Calculated from the estimates of Model 4 in Table 1, the plotted predictions support Hypotheses 2 and 3. For individuals who voted for government parties in previous elections, the probability of having participated in a protest is around 8% for the lowest value of the party size variable. The probability decreases with increasing party size. For the highest value of party size, the protest participation is less than 6%. The 2% decrease from 8% to 6% indicates a reduction from one-third to one-fourth of the baseline probability of protesting among individuals who voted for a government party. In that regard, we can state that the findings are substantively significant in rejecting the null of the second hypothesis.
While the effect of party size has a negative influence on protesting for both groups of voters, the plotted prediction shows that the effect of party size is more substantial for losers. This finding indicates that voters of smaller opposition parties are more likely to protest than voters of smaller government parties. In that regard, the results support the idea that losers protest more than winners.Footnote 10
The plotted predictions for opposition parties provide strong support for the third hypothesis: while the probability of participating in a protest is at its highest when the party that the individual voted for has no seat in parliament, the probability decreases with the increasing size of the party. The effect of party size is substantively significant, considering the 5% decrease in the probability of protesting between the voters of small and bigger parties.
It might be important to consider the decreasing effect of government status on protest participation when distinguishing between majority and minority parties. These findings indicate that, although not in government, parties with considerable size and support still provide channels for their voters to be part of the political decision-making process. However, for voters of smaller parties, their parties’ government status significantly affects their representation. Therefore, based on the findings, we can state that the true ‘losers’ in politics are the voters of smaller parties, which have limited opportunities to represent their supporters in conventional politics.
Conclusion
In Greece, the 2019 Macedonia name agreement provoked mass demonstrations in Athens, driven by nationalist opposition to the government’s diplomatic deal. Just days before the protests, Panos Kammenos – leader of the junior coalition partner, the Independent Greeks (ANEL) – resigned from his post as defence minister and withdrew his party’s support from the ruling coalition. Although the senior government party Syriza survived the resulting confidence vote, the episode illustrates how a junior party can wield outsized influence in shaping political crises and potentially galvanize protest activity aligned with its base. The resignation was not merely symbolic; it reshaped the power dynamics of the coalition and underscored the precarious position junior partners occupy – caught between government responsibility and ideological commitment.
In contrast, Germany’s 2021 climate protests illustrate the reverse dynamic: how the anticipation of coalition compromises may lead to disillusionment among core supporters of junior parties. During negotiations to form the ‘traffic light’ coalition, the Green Party was already facing criticism from grassroots climate movements – particularly the emergent Klimaliste – for signalling insufficient ambition on climate policy. Protestors framed their mobilization as a necessary push to ensure more ambitious policies, anticipating that the party’s likely compromises in coalition government would fall short of the urgent action needed. This dynamic underscores the dual pressures junior coalition parties face: even when ideologically aligned with protestors, their limited influence within government can leave core supporters unconvinced – intensifying representational gaps and sustaining protest activity.
Together, these examples support the article’s core theoretical claim that junior coalition parties are neither entirely powerless nor fully influential. Their position grants them access to formal decision-making structures, but their constrained policy-making capacity can simultaneously disappoint supporters and trigger protest. This dual role – as both insiders and outsiders – renders their voters particularly susceptible to representational inequality, making protest a uniquely relevant outlet for expressing political demands.
In studies of protest behaviour, political parties play a crucial role because their dual functions influence protest participation. On the one hand, parties provide a channel between individuals and conventional politics which, according to common perceptions, decreases the potential for protest. On the other hand, parties may increase protest participation by becoming potential allies of protestors within conventional politics, thereby supporting their demands. These two contrasting roles are often performed simultaneously by parties, which can obscure the underlying mechanisms of political representation and protest participation.
By demonstrating the effect of party size on protest participation, I address these two roles of parties. While smaller parties in government may better represent their constituents compared to smaller parties in opposition, due to their strategic position in government formation, they still fall short of full representation because of their limited influence within the government. However, these smaller parties may serve as allies for their supporters in conventional politics, which increases the likelihood that their voters will protest to ensure their preferences are heard during policy-making.
To consider the contrasting roles of parties, this study approaches the relationship between parties and protest by examining party size. By differentiating between senior, junior and opposition parties, it has shown that voters of junior coalition partners are more likely to protest than those of senior coalition partners. The analysis finds that as the size of a party decreases, the probability that its voters will participate in protests increases. This effect is particularly pronounced for voters of smaller opposition parties, as the study shows that the influence of government status diminishes the propesnity to protest for voters of larger parties.
The diminishing effect of government status aligns with the theoretical assumption about the significance of party size in political representation. As has been emphasized, larger opposition parties, despite being in opposition, are better able to represent their voters in parliament. Additionally, these parties have the potential to become future government parties or may have previously held government positions. Consequently, their voters are not as deprived of political representation as the voters of smaller opposition parties.
Despite the robustness of the findings, there is a potential concern of endogeneity – particularly that individuals with a higher propensity to protest may be more likely to support smaller or opposition parties. I therefore implemented several strategies to address this potential endogeneity. Most notably, in Table A5 in the online Supplementary Material, the sample was restricted to respondents interviewed at least one year after national elections, ensuring that their party’s role in government or opposition was not merely nominal, but reflected real governing experience. This produced similar findings. Additional analyses in Table A7 in the online Appendix further test the robustness of the results using long-term ‘winners’ and ‘losers’, as well as interaction models showing that the effect of government status is most pronounced among smaller parties. Together, these strategies reinforce the article’s core argument: that protest participation is shaped not just by individual-level predispositions, but by the institutional position and representational capacity of political parties.
These findings carry broader implications for understanding how citizens engage with democratic institutions. Importantly, the results should not be interpreted as evidence of instability in coalition governments. It is not my view that protest participation is inherently destabilizing; rather, it can be seen as a routine and legitimate form of democratic engagement – particularly when institutional responsiveness is perceived to be insufficient. In this light, protest is not a rejection of the system but a complementary channel through which citizens communicate unmet demands. The finding that voters of junior coalition parties are more likely to protest underscores the dual role these parties play. While they succeed in gaining access to government, they are often constrained in shaping the policy agenda, which may create a gap between voter expectations and actual influence. This duality helps explain why protest persists even among voters of governing parties. Overall, the present study highlights that voting alone may not fully satisfy democratic participation when representational inequalities persist. By drawing attention to how internal hierarchies within coalition governments shape citizens’ perceptions of influence, the findings contribute to a more nuanced understanding of political representation and the dynamics of protest in contemporary democracies.
It is important here to underline certain limitations of this study. First and foremost, the dependent variable only asks respondents whether they participated in any public demonstration in the last 12 months. Although the wording of the question in the ESS indicates that this political action was taken to improve conditions in the respondent’s country, it is impossible to know whether the protest in which the individual participated had policy-related demands. Considering that most demonstrations target the government, the assumption is that these protests were aimed at influencing political decision-making. However, a more detailed investigation of protest behaviour in survey studies is clearly needed.
This study examined the role of one of the most prominent factors for understanding a party’s ability to influence policy-making: party size. However, the strategic position of parties in coalition governments is shaped by many considerations that warrant thorough investigation in future studies. One of the first steps in this direction could be to explore various coalition possibilities between parties, which might strengthen or weaken a smaller party’s bargaining power with larger parties. Voters may adjust their expectations based on pre-election promises, coalition compromises and the resulting policy agenda. When junior parties are forced to concede core policy positions in order to enter government, their voters may feel disillusioned. Additionally, the ideological positions and party families of these parties could provide another avenue for exploring the role of coalition dynamics on protest behaviour.Footnote 11
In addition to the demand-side dynamics of protest behaviour emphasized here, the supply side also plays a critical role. Political parties – especially those in opposition – are not merely passive recipients of public dissatisfaction; they can actively shape, coordinate, or even initiate protest events. Opposition parties may strategically use protest as a tool to expand their influence, mobilize supporters or increase media visibility. This strategic dimension of protest politics suggests that voters of smaller or less influential parties may be embedded in broader mobilizational structures, where the parties themselves encourage protest as a form of political pressure. While the data here capture individual participation, future research should explore these party-driven mechanisms more directly.
Lastly, it is important to acknowledge the limited scope of this study, which focuses exclusively on Western European democracies. Given the differences and more complex dynamics in less democratically institutionalized countries, I chose not to base my theoretical framework on these contexts. However, future research should expand the sample to include more diverse political settings, enhancing the generalizability of the findings.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/gov.2025.10029.
Data availability
The datasets generated and analysed during the current study are available from the author upon request.
Acknowledgements
I am grateful to Mert Moral, Özge Kemahlıoğlu, Çiğdem Bağcı, Faruk Aksoy and the three anonymous reviewers for their valuable feedback and constructive comments. Any remaining errors are my own.
Financial support
No funding was received for conducting this study.
Disclosure statement
There are no conflicts of interest to disclose pertaining to this work.
Open access
