Introduction

In democratic societies, political parties are expected to represent their voters’ preferences (Mair 2008) and as such establish a connection between public preferences and policies (Dahl 1956). However, while the overlap between party behaviour and party voter preferences is substantial, it is far from perfect. Indeed, research shows that parties are not always responsive to their votersFootnote 1 (e.g. Romeijn 2020; Ezrow et al. 2011; Klüver and Spoon 2016). Think, for instance, of parties pursuing unpopular austerity policies, such as welfare cutbacks, against the will of their voters (see Pierson 1996; Vis 2016; Weaver 1986).

One possible explanation for parties being unresponsive to their voters’ preferences lies in the perceptions party decision-makers have of voter preferences being flawed: misperceiving what their voters want, parties would unintentionally go against their preferences (see, for instance, Broockman and Skovron 2013 who studied the accuracy of politicians’ public opinion perceptions). Another explanation may be that parties intentionally cross voter preferences (e.g. Mair 2009; Strøm 1990; Strøm and Müller 1999). This study zooms in on the second, intentional, explanation for the observed mismatch between voter preferences and party positions and decisions. Specifically, we seek to answer the following two research questions: why do parties intentionally go against their voters' preferences? And how do they do it?

The literature on party responsiveness has tackled the issue of parties going against voter preferences. Next to ample theoretical reflections on party behaviour (see for instance Strøm and Müller 1999), scholars have tried to deduce causes for (un)responsiveness by observing when and to what extent there is a link between voter opinions and party behaviour. They show that some parties (mainstream, government parties) are less responsive to their voters’ preferences than others (opposition, niche parties), on some issues (rather on non-salient issues) and especially in the beginning of the electoral cycle (see for instance; Ezrow et al. 2011; Klüver and Spoon 2016; Romeijn 2020). While these observations are highly insightful, the actual considerations that lead parties to refrain from acting on the preferences of their voters are simply not observable. In the USA, some scholars (notably Kingdon 1989) have overcome this problem by interviewing legislators about their motivations for being (un)responsive in their roll-call behaviour. Similar data on party decision-making is scant.

Insights on how parties behave when they decide to go against their voters’ preferences, in turn, can be derived from the literature on blame avoidance. By theorizing about or (in rare cases) observing governmental decision-making, scholars have argued that governments invoke so-called “blame avoidance” strategies when they pursue unpopular austerity policies (see for instance Weaver 1986; Vis 2016; Hübscher and Sattler 2017). Over the years, numerous scholars have come up with typologies of strategies that governments can invoke when pursuing seemingly unpopular policies (see König and Wenzelburger 2014 for an overview of this work). With the exception of Wenzelburger's (2011) anecdotal evidence, though, empirical proof of which strategies parties invoke and with what purpose, is virtually non-existent. Moreover, blame avoidance has so far been discussed almost exclusively in the context of government retrenchment, preventing the phenomenon from being considered in its entirety (see Hinterleitner 2017 for an elaboration of this criticism).

In sum, theoretical and observational studies contributed importantly to our understanding of why and how parties intentionally go against their voters’ preferences. Yet, the considerations and strategic decisions of parties are not always observable. To fully understand the conditionality of voter responsiveness, then, one needs to learn about the motivations of central party decision-makers; how they balance goals, why they want and dare to go against their voters’ will, whether they consider such incongruent behaviour to be electorally risky and if so, how they deal with it (see Wenzelburger and Zohlnhöfer 2020). As Jacobs and Shapiro (2000, p.6) posited; ‘explaining how and why responsiveness to public opinion varies requires understanding the motivations of policymakers’ (see also Weaver 1986; Wenzelburger 2014).

This study aims to complement previous work on party (un)responsiveness by interviewing nineteen key party elites—party leaders and ministers—in Belgium (see Wenzelburger 2014 who argues that interview evidence allows to uncover the reasoning that lies behind political decisions). In particular, we ask elites to reflect on party initiatives they proposed against the will of their voters, to explain why they decided to do so and how they dealt with the pursuit of this unpopular decision. This qualitative, inductive approach allows us to identify what considerations drive party’s deliberate unresponsiveness to voter preferences, to establish that parties think in terms of blame avoidance, and to uncover the different strategies party elites invoke when they anticipate electoral repercussions for unpopular action. As such, we add novel direct insights to the existing work on party responsiveness and the scholarship on blame avoidance.

Party objectives and the consequent lack of responsiveness

The starting point for understanding why parties are not always responsive to their voters’ preferences lies in the objectives parties seek to pursue. Downs' (1957) classical economic theory of party behaviour portrays parties as actors that are solely concerned about maximizing votes. That parties are driven by a desire to win or at least to consolidate voter support shows in their positions shifting towards those of their voters (see Adams 2012; Romeijn 2020), or, for instance, in parties emphasizing issue positions that are popular among their supporters (Rovny 2012). That political actors are induced to act in line with their voters’ preferences because their future hinges on this voter support, has been referred to as anticipatory responsiveness (Mansbridge 2003) or rational anticipation (Stimson et al. 1995).

In response to Downs the rational choice tradition has emphasized that a party’s pursuit of votes is foremost instrumental; they need votes to pursue their idea(l)s (Budge et al. 2010; Wittman 1973). Indeed, next to the predominant desire to win votes, or not to lose votes, (some) parties pursue office-seeking goals—the desire to wield executive power—and policy-seeking goals—the desire to enact certain policies or to influence the policy agenda of others (Strøm 1990). In sum, while electoral success can be a goal in itself, it is foremost considered as a path to policy agenda influence (Manin et al. 1999).

Given that parties do not necessarily hold coherent views with their voters (see for instance Dalton 2017 or Valen and Narud 2007), a party’s aspiration to pursue the policies it deems necessary could be a probable cause of voter unresponsiveness. After all, policy preferences of parties and voters being out of step, it is unlikely that parties’ policy goals will be satisfied entirely by being responsive to voter preferences. Also, parties in government, because they have to compromise with coalition partners and potentially face other constrains such as budgetary discipline, may pursue initiatives that are unpopular with their voters (Green-Pedersen and Mortensen 2010; Mair 2009; Klüver and Spoon 2016). As such, a party’s policy and office goals may instigate unresponsive behaviour, if voter and party preferences do not coincide (Strøm and Müller 1999).

In sum, that parties may want to advance their own policy agendas, in combination with the empirical reality that they (or their coalition partners) sometimes vision society in a way their voters do not, suggests that responsiveness crucially hinges on how parties balance their policy goals with their re-election goal, if the two collide (Pierson 1996; Strøm 1990; Wenzelburger and Zohlnhöfer 2020). The theory of rational anticipation prescribes that the potential electoral consequences of a decision serve as the crucial deal-breaker (Mansbridge 2003; Stimson et al. 1995; or see Hübscher and Sattler’s 2017 work on fiscal consolidation). Asking U.S. legislators about their voting behaviour, Kingdon (1989) indeed confirms that politicians are less inclined to follow their own opinion and go against their constituents’ will if they think voters care a lot about the issue at stake because they assume chances are high citizens will notice this unpopular behaviour and hold them accountable for it. Jacobs and Shapiro (2000) as well, by means of two case studies, find that political actors’ perceptions of the (electoral) costs and benefits of different policy alternatives play an essential role in their behavioural calculus. Yet, similar direct evidence on how parties arrive at the decision to go against their voters' preferences is scant. Therefore, the first research question this study aims to answer is the following;

RQ1:

According to party leaders and ministers themselves, why do parties go against voter preferences?

How parties try to avoid electoral blame for being unresponsive

Parties are not always responsive to voter preferences (Romeijn 2020). Taking into account early findings on prospect theory (see Tversky and Kahneman 1973), Weaver (1986) argues that political actors are loss averse and, therefore, are motivated to avoid electoral blame for such unpopular actions (König and Wenzelburger 2014; Hübscher and Sattler 2017). Hence, it is commonly assumed that political actors invoke strategies when they anticipate to lose voters over certain unpopular decisions or positions (Hinterleitner 2017; Mayhew 1974). These so-called “blame avoidance strategies” are central to studies on welfare state and retrenchment politics.Footnote 2 After all, scholars working in this field were struck by the observation that governments that invoke widely unpopular retrenchment policies are rarely, or at least not always, punished for it in the elections. One possible explanation, they argue, lies in governments successfully invoking strategies to avoid electoral blame—of course, it might as well be that citizens are simply unaware of these policies being put in place, do not care or do not know whom to blame for it (Giger and Nelson 2011; Vis 2016; Vis and Van Kersbergen 2007). While the absence of accountability for unpopular retrenchment policies does not offer definitive proof for the existence of blame avoidance strategies, it did spark scholarly interest in the topic. As Pierson (1996) argues: ‘if retrenchment policies are enacted in spite of anticipated punishment at the ballot, it is highly likely that blame avoidance strategies will be implemented to circumvent these consequences’ (see also Hinterleitner 2017).

One of the leading works on blame avoidance is authored by Weaver (1986), who identified eight types of strategies that governments can invoke to avoid being held accountable by voters for budget reforms, ranging from compensating voters financially to passing the blame for the budget cuts onto others. McGraw (1990), then, classifies these blame avoidance strategies into justifications (or presentational strategies—see Hood 2010) and excuses (or agency strategies, see Hood 2010). By using excuses, political actors try to weaken the causal link between themselves and the unpopular budget reform (e.g. by claiming that financial institutions asked for it, see Wenzelburger 2011; Cox 2001). Justifications, on the other hand, are invoked to ensure that the consequences of a policy decision are perceived more positively by citizens (e.g. by claiming the policy will keep the country running, see Wenzelburger 2011). Another categorisation of blame avoidance strategies that is made in the retrenchment literature distinguishes between strategies that manipulate procedures (e.g. delegating responsibility to non-state actors), that manipulate perceptions and, finally, that manipulate payoffs (e.g. implementing tough reforms right after the start of the new government) (Vis 2016). In sum, and this is also how König and Wenzelburger (2014) synthesized the work, to avoid electoral repercussions for (seemingly) unpopular policies, political actors can either manipulate the link between an unpopular decision and themselves (i.e. agency/organizational strategies or excuses), and try to increase the endorsement of the policy (i.e. presentational/communication strategies or justifications) (see also Hering 2008; Hood 2010; Vis and Van Kersbergen 2007; Green-Pedersen 2002).

Importantly, with the exception of Wenzelburger's (2011) anecdotal evidence, proof of which strategies, when and with what purpose are invoked by parties when they go against their voters’ preferences, is virtually non-existent. Recently, Hinterleitner (2017) has argued that the presence of blame avoidance is a black box, because of its limited focus on retrenchment politics, and especially because it is assumed rather than measured (see Vis 2009). After all, most of the observational literature suffers from omitted variable bias; it is unclear whether the lack of punishment for certain policy decisions is due to the decision not being electorally risky, or because this risk has been eliminated by blame avoidance strategies. Similarly, certain (invisible, for instance) strategies simply cannot be identified by observing the behaviour of political actors. Therefore, the second research question this study seeks to answer is;

RQ2:

According to party leaders and ministers themselves, how do parties try to avoid electoral retribution for going against their voters’ preferences?

Methods

This study relies on interview data collected among party elites in Flanders, the Dutch-speaking part of Belgium.Footnote 3 Studying party decision-making in Belgium has two particular advantages. For one, Belgium is the archetype of a partitocracy with powerful and rather homogeneous parties (Deschouwer 2012). The fragmentation of the party system, and the necessity of strong multilevel coordination, have even enhanced the position of political parties in Belgium in recent years (De Winter 2019). In this party system, we study the politicians who ultimately determine the party line, namely party leaders and ministers. Belgian party leaders are the undisputed party decision-makers and while Belgian ministers are expected to serve the cabinet’s interests and decide by governmental consensus, they are also loyal party servants that defend the party’s interests at the government table. Ministers keep close contact with the party leader and other party officials to discuss their position within the cabinet and have large cabinets working on safeguarding the party’s interests. Similarly, ministers have a strong bearing on party decisions because they attend the weekly meetings of their party executive, a select group of party elites that meets regularly to decide on the party line (De Winter and Dumont 2006). While the position of party leader and minister is different—Belgian ministers ultimately have a collective responsibility in serving the cabinet by executing the government agreement—and while there may be differences in how loyal individual elites are to their party (see for instance Alexiadou 2015 on different ‘types’ of cabinet members), it is beyond doubt that both are key actors in party decision-making (De Winter 2019). In short, we chose to interview party leaders and ministers because they, as key witnesses and actors, can provide essential insights into how party decisions are made (Wenzelburger and Zohlnhöfer 2020).

The second advantage of this study being conducted in Belgium is the simple fact that the rather open political culture, in combination with our quite established relationship with Belgian political elites,Footnote 4 allowed us to get high-ranking politicians to participate in our research. By interviewing party leaders and ministers, we circumvent one major issue researchers tend to struggle with: the need to rely on circumstantial observational evidence to make claims on the conditionality of party responsiveness. Having access to top politicians in Belgium offers a rare opportunity to study the cognitive processes underlying party unresponsiveness.

To examine why and how parties deliberately go against their voters’ preferences, we conducted semi-structured face-to-face interviews with the leaders of all six parties currently holding seats in parliament and thirteen ministers—seven ministers were not able to participate.Footnote 5 The interviews with nineteen party elites were conducted between March and June 2018 by one experiencedinterviewer in their offices in Brussels. During approximately one hour,Footnote 6 elites were questioned about the media, about representation and about candidate selection. The questions on intentional unresponsiveness that we rely on in this study were asked at the beginning of the interview, and all elites answered them.

The interviewer asked the exact same questions to all politicians—questions that were developed out of sheer interest in understanding the mismatch between party action and voter preferences. Keeping the inductive approach of the study in mind, the interviewer allowed politicians to reflect and elaborate freely on the questions asked: he did not present elites with existing theories of unresponsiveness or typologies of blame avoidance strategies because it would harm the main goal of our study, which is to see what considerations, and what strategies prominent elites spontaneously come up with. Only when the interviewer felt that respondents, advertently or inadvertently, did not answer, he repeated the question, sometimes formulated slightly differently, but not leading the interviewees towards certain responses via these interjections. Also, we recorded the interviews (which none of interviewees objected) to facilitate the analysis afterwards (see Harvey 2011).

After a brief introduction in which the interviewer emphasized that we were interested in party decision-making he asked; ‘Does it happen that your party proposes an idea or initiative of which you know it is not supported by a majority of the voters of your party? Can you give an example?’. Making interviewees think of a particular situation makes their reflections more tangible and as such increases the likelihood of receiving honest answers (see Kingdon, 1989 for empirical evidence substantiating this claim). Most respondents referenced multiple examples of situations where they deliberately went against their voters’ preferences, and gave a coherent explanation for why they took the initiative against the will of their voters. As a follow-up question, the interviewer came back to a specific example of an incongruent party decision that the politician had just mentioned and asked; ‘How did you handle this situation?’. Interestingly, most elites spontaneously mentioned how they dealt with other unpopular initiatives they had pursued or were pursuing. Our anonymity agreement prevents us from zooming in on the exact proposals elites mentioned, but the policy issues were very diverse, ranging from pension reforms, air travel taxes, abortion laws, the reintroduction of military service, to, for instance, the introduction of noise standards on music festivals.

To formulate an answer to our research questions, then, we performed a conventional qualitative content analysis of the interview transcripts. Qualitative content analysis is a research method for the interpretation of content through the systematic classification process of coding and identifying patterns, or put differently; of organizing text with similar content (Cho and Lee 2014; Kuckartz 2014). The aim of this study is to inductively build a typology of party unresponsiveness, which implies that categories were drawn from the data instead of starting off with preconceived categories (following the approach of Elo and Kyngäs 2008). The analysis of the interview data followed several steps. First, the interviews were transcribed by student coders and we started with analysing the responses to the first and second question separately. Systematically reading through the relevant parts of the transcripts, we developed categories from the material; similar arguments (i.e. sentences or even paragraphs with a similar meaning) were grouped together—that is: given the same code in NVivo. Reading through subsequent transcripts, similar arguments got the same code, and new codes were added if other content was encountered. A second reading was imposed on all the transcripts to verify the first round of coding. In a next step, similar arguments identified by codes were placed into broader categories. For instance, arguments such as ‘I will not showcase this decision’ and ‘we won’t make it public’, were classified together under the category “obfuscating”. Finally, the categories withheld from the data—that are discussed in the result section by means of exemplary quotesFootnote 7—are analysed in connection to existing literature (notably on blame avoidance and responsiveness, theories that came in only after the data were collected and analysed).

Some methodological decisions warrant additional explanation. A first matter to address is the possibility of elites giving ‘desirable’ or ‘strategic’ answers. We have good reason to believe politicians gave us honest answers in the sense that they match their actual thinking. For one, the interview context discouraged strategic responses; politicians were assured anonymity and knew that their answers would never be made public in an identifiable format—in previous waves of elite interviews we did in 2013 and 2015 the same rule applied and no such information reached the public realm. Importantly, these repeated contacts allowed us to establish a trust relationship with some of the elites (Harvey 2011). Another reason why we feel confident about the trustworthiness of elites’ answers is that they were at times surprisingly honest, stating, for instance: ‘What I will now tell you should never be published.’ or ‘Don’t tell this to my colleagues’. Of course, we cannot rule out that politicians’ responses were subject to post-hoc rationalizations. Still, what the elites told us is largely consistent with the findings of observational studies on party responsiveness and blame avoidance.

A second point is that this study focuses primarily on why and how parties in general go against their voters’ preferences. With just six parties in our sample, it makes sense to focus on considerations and strategies that exist across parties rather than to explore differences between types of parties. There is one exception, though. Given that there are substantial differences between opposition and government parties—for instance because the former have more leeway to focus on popular issues while the latter are constrained by the coalition agreement (a compromise between all governing parties) and other (international, budgetary) constraints, which may weaken the link between their decisions and their supporters’ preferences (Green‐Pedersen and Mortensen 2010; Romeijn 2020; Klüver and Spoon 2016). Therefore, we do highlight differences between government and opposition parties as they emerge.

Why parties go against voter preferences

What first of all stands out in party elites’ reflections is that pleasing, or rather not offending, their voters if top of mind because, they unanimously argue, public support is a precondition to ‘survive’ in politics: ‘The “business capital” of each party is its voters. Your members are your shareholders and your voters are your customers. No shop can survive without customers. Therefore, no party is going to take many initiatives that may scare away customers’ (PL5-M). The opinions of their voters clearly guide parties in choosing which initiatives to propose, which to emphasize and, importantly, which not to propose. As one minister puts it; ‘It does not happen too often that we propose things that could potentially offend our voters. We depend on electoral success for survival, and we obviously have a desire to survive.’ (M4). Parties are reluctant to launch ideas that might meet a great deal of resistance from their voters; ‘You should always weigh the anticipated benefits against the potential repercussions. Sometimes it is just not worth going against your voters.’ (M6). Clearly, all parties feel constrained by their voters’ preferences, which explains, at least in part, why observational work has identified a sizable overlap between party positions and party voter preferences (see for instance Adams 2012; Romeijn 2020).

Even though parties are careful not to act against their voters’ preferences all too often, all elites emphasize that this does not mean they blindly let their behaviour be dictated by their voters’ will. Some even posit that ‘Leaders should not follow. Leaders must lead.’ (M11). Asking elites to tell us about situations in which they intentionally did not follow their voters, we learn that the electoral incentive to keep voters on board does not always outweigh other considerations of the party. In particular, three considerations are commonly referred to as to why parties decide to disregard what they believe their voters want—even though they are, in general, careful to not put off these voters.

Ideology is key

First, parties go against their voters’ will if it drastically opposes their core ideology. Parties want to unfold their own ideological story, we learn from the interviews, and their vision on society does not always coincide with their party electorate’s preferences. As one party leader puts it; ‘We would never betray our core ideals.’ (PL3-M). Or as another minister points out; ‘Across the ideological spectrum, certain things are a no-go zone for our party’ (M3). In sum, and confirming Budge et al.’s (2010) hypothesis, a party’s ideological vision sometimes takes priority over keeping voters satisfied; ‘This [..] ethical issue goes to the heart of our ideology, so we simply did not care whether our voters were ready for this… We felt we had to do ground-breaking work.’ (PL4-M).

The greater good should prevail

Second, elites posit that voter opinions can, or rather need to, be disregarded when they contradict the general public interest. Elites emphasize that ‘the greater good’ (M4) or ‘national interests’ (PL3-M) should sometimes prevail over the desires of their voters. Some emphasize that parties are obliged to take decisions in the best interest of all citizens rather than just their own voters (see for instance Ezrow et al. 2011 on mean voter representation versus partisan constituency representation), while others stress that citizens are inherently egoistic and that their preferences sometimes contradict what is best for society. As one minister argues; ‘I knew some of our voters would resent this… But yeah, there are some serious economic interests at stake here… We just had to do it.’ (M3). Parties feel they have an obligation to the broader community (Mair 2009). Or as another minister bluntly puts it; ‘It is our duty to move opinions in a direction that benefits the general public interest.’ (M11). This second argument may be related to the previous one; parties likely consider their ideology to be consistent with the general public interest. Yet, this need not be the case; party positions may as well serve the interests of particular groups (of voters).

The ignorant should not (always) be followed

Finally, and related to the previous consideration, parties sometimes consider voter preferences to be uninformed and inconsistent. Elites stress that citizens are swayed by issues of the day; ‘One day they are in favour of a certain policy and the next day they are against.’ (M5) or ‘Public opinion can change in a jiffy’ (PL3-M). Not only do they consider these preferences to be rather fluid, elites also emphasize that voters ‘often have preferences that contradict each other’ (M7). Party elites seem to believe that they can ignore their voters’ opinions because they do not really know what is best for themselves anyway. Not that citizens are necessarily incapable of holding informed opinions, but they simply lack the information necessary to form such opinions; ‘Their opinions are often built on wrong assumptions, on misunderstandings. For some issues, they cannot help it, though. They lack the information to form an opinion, information that we as politicians have access to’ (M2).

There is a fourth reason, cross-cutting the other three, for government parties to deliberately pursue initiatives that are unpopular among their voters. The (obliged) loyalty towards other coalition partners sometimes forces government parties to implement policies that are not supported by their own voters—which explains why government parties tend to be less responsive to their voters’ preferences (see for instance Klüver and Spoon 2016). As one minister puts it; ‘Some issues are raised by the coalition partner and we know they are not liked by our voters. You calculate your losses then. But, in turn, of course, we also raise issues that we know we can please our own voters with’ (M9).

Looking at the four considerations for parties to go against their voters’ preferences, a clear difference between government and opposition parties becomes apparent. While government elites commonly argue that they opposed their voters’ will because of coalition dynamics (coalition partners trying to please their own voters in combination with the fact that governments take decisions by consensus), opposition party leaders tend to stress that they disregarded their voters’ preferences because they were seen as uninformed or driven by self-interest, and because the policy issue at hand mattered a lot to the party (i.e. an issue they ‘own’ see Lefevere et al. 2015).

We can conclude that parties care deeply about their voters’ preferences; they pander towards their positions to keep them aligned, and they refrain from proposing things that may scare them away to other parties. In line with theoretical accounts on party behaviour (see for instance Strøm and Müller 1999), we find that the necessity of voter support is top of mind for all parties and instigates responsive behaviour. At the same time, all elites give examples of policy proposals, initiatives and ideas of their party that, in their perception, went against their voters’ desires. Hence, party unresponsiveness not only originates from parties misunderstanding their voters’ preferences (see Broockman and Skovron 2013), parties also intentionally go against these preferences from time to time. In particular, they do so when voters’ desires are considered unfounded or inconsistent, or if they feel their preferences contradict the ideological vision of the party or the general public interest. As one minister eloquently puts it;

Our voters’ will is one of the elements we take into account. We also need to look at what is financially possible, what is feasible and whether it is in line with our ideology. These opinions are just one of the objective elements that help us to choose a direction and take decisions, but it is certainly not the only thing that matters (M2).

How parties try to avoid electoral retribution for going against their voters’ preferences

When the position of the party coincides with the preferences of its voters, as one party leader notes, they are golden; ‘This is a goldmine, because our views are already in line with the sensitivities of our voters. We just have to formulate the right supply.’ (PL5-M). At the same time, we showed, parties occasionally want, or have to in the case of government parties, pursue a course of action that does not correspond with their voters’ opinions. What is crucial in their decision (not) to go against their voters, elites argue, is the amount of retribution they anticipate; ‘Our main concern is not losing votes… Winning support is always nice, and every politician would like the party to grow, but no one wants to lose, right?’ (M8). Or as another minister meticulously describes it; ‘When we are considering ideas, we rarely say ‘yes, let’s do this because 90% of the electorate would like it’, the discussion always is; ‘we really need to watch out because some of our voters will not approve of this.’ (M3). These two exemplary quotes clearly indicate that party elites are risk averse.

Still, elites do not always consider it electorally hazardous to go against the will of their voters. Some stress that voicing unpopular positions or pursuing unpopular policies might sometimes even positively affect their election results. Opposing voter preferences and being honest about it, they can show their courage, demonstrate ‘statesmanship’, which they presume to be rewarded on election day. Their rationale holds that a party’s reputation matters a great deal in elections. A party leader addresses this apparent contradiction; ‘It often is a paradox. On the one hand people appreciate it when we know what’s on their mind. But often they also appreciate it when we have a clear vision, an idea of where we want to go’ (PL3-M). Or as another minister puts it; ‘sometimes people will vote for you even if you’ve done something they did not agree with. Because they respect you. Sometimes you meet people who say; “I disagree with what you’ve done, I was even mad and I still don’t agree, but I appreciate that you’ve done it”’. (PL4-M).

Usually, though, going against voter opinion is considered risky behaviour, electorally speaking. Especially so when it concerns issues that belong to the core of the party, that are ‘owned’ by the party (e.g. environmental issues to the Green party) (see Lefevere et al. 2015 for more information on this concept). The rationale holds that these issues are more important to party voters, and therefore, parties anticipate more electoral sanctioning when offending their voters on these issues that (often) made them vote for the party in the first place. As a party leader describes it; ‘The closer an issue comes to your core themes, the more you need to handle it with care; the more you need to think about it, deal with it more strategically. Why? Because the more important the issue, the more going against your voters potentially costs you.’ (PL5). Interestingly, that (perceived) issue salience increases the fear of electoral retribution explains why observational studies have found that often, issue salience is positively related to responsiveness to public opinion (e.g. Lax and Phillips 2009). At the same time, other observational work suggests that parties dare to oppose voter preferences more on these ‘owned’ issues because they have a ‘trust advantage’ on these issues. Whether or not parties are more or less responsive on issues they own, depends on the estimation of electoral consequences made by elites.

So far, we’ve learned that parties occasionally pursue ideas they assume are not supported by their voters, and that they sometimes consider this unresponsiveness electorally risky. When parties anticipate that such behaviour may put off voters, they invoke strategies to avoid or at least reduce the odds of electoral punishment; ‘When you feel you cannot just promote your position, you have to change reality in your advantage. It asks for a subtle quest for support, which is a matter of carefully building a strategy around certain issues.’ (PL5-M).

Make sure they don’t see it

One way parties try to deflect electoral sanctioning for unpopular decisions is by simply hiding them. If it is not visible, it cannot be held against the party on election day, the rationale holds; ‘We will not make it public. We might write a few sentences on it in our program, but we will for sure not showcase it.’ (PL1-O). Or as another party leader puts it, ‘We will shut up about it. We won’t send out press statements to promote it. We will not bring it up ourselves in debates, for instance, but won’t lie about it either’ (PL6-O). Another minister argues; ‘we will not communicate about it, otherwise people might take to the streets,’ (M12). Or, as a party leader claims; If your view is not shared by your supporters then you best remain silent. Selective mutism happens, not only in parties. Trade unions, for instance, have been silent about the migration/identity issue for the past four years […]. They are trying to position themselves on non-salient leftist themes, but you do not hear the union leaders speak about asylum or identity because they know their supporters disagree’ (PL3-M). Besides passively hiding their decisions, parties also actively try to divert attention; ‘We proposed a package of policies that were quite unpopular but in our communications we emphasized one specific part we knew people would appreciate, and care about.’ (PL4-M).

Hiding unpopular decisions for the public eye is not always feasible. Elites stress that not attracting attention is fairly easy for technical issues that do not generate much interest among voters. It is more challenging, though, with regard to initiatives that attract media and consequentially public attention. Also, and importantly, parties do not always want to disguise initiatives they believe are legit; ‘We were well aware that our idea was unpopular at first but we had credible, clear objectives. We knew why we wanted to propose it and therefore had no problem publicly defending it.’ (M6).

Explain it

When party elites anticipate that a proposal may jeopardize the party’s election result, and they can or will not cover it up, they try to explain it to their supporters. Government and opposition elites alike stress the importance of communicational strategies to reduce voter resistance; ‘The crucial question always is; can we explain it?’ (PL4-M). Interestingly, explaining unpopular ideas is not necessarily the same as manipulating citizens to endorse it. Sometimes, interviewees argue, public resistance is founded upon false assumptions, assumptions that parties can remedy in their explanations; ‘They simply did not grasp the technicities so we needed to explain them.’ (M2). Yet, strategic communication often goes beyond factual explanations—confirming Jacobs and Shapiro’s (2000) finding that US presidents try to lower electoral costs of going against public opinion by crafting their arguments to change public opinion in favour of their policy. If parties feel that a particular proposal or position may come back to haunt them at the ballot box, they try to reduce initial voter resistance by offering three types of explanations.

First, parties highlight the societal fairness of the proposal; ‘You need to convince voters that the initiative is in their best interest, that you are proposing it for the greater good.’ (PL5-M). Moreover, some stress that it is easier to convince citizens that their proposition is fair if they understand that the status-quo is unfair and untenable, and therefore needs to change; ‘A taxi home from the airport is often more expensive than the plane ticket. How on earth is that possible? We need to stress this is wrong. People are receptive of our argument once they realize this is wrong (PL1-O).

Second, parties emphasize the future positive consequences of their proposal to reduce voter resistance. Often, the interviewees posit, initiatives that are only beneficial in the long run are perceived by voters as unpleasant. It is up to the party to make sure voters understand the (future) benefits, which they sometimes do by referring to ‘successful examples in other countries.’ (M1). As a party leader meticulously describes it;

Even our hard-core party militants show behaviour that is not in line with our vision. We can blame them and say ‘you cannot do this or that [..]’ and lose their support, or we can think on it and try to get them along. We have an ideal, and it is pretty easy to communicate it straightforwardly, but we always have the reflex to think of ways to get our voters along. We emphasize the damaging consequences of [..], explain how it [the initiative] will help us in the long run. We make it ‘digestible’, so that they eventually realize it is not a punishment, but that it is in their best interest. (PL1-O)

Third, parties try to increase voter endorsement by claiming their proposal ‘fits the ideological line of the party.’ (M5). In doing so, we learn from elites' reflections, they seek to appear trustworthy and consistent, which in turn can have a positive impact on their election results.

Interestingly, politicians stress that all of these justifications are more effective when they are backed by influential societal actors; ‘Finding support for the initiative was crucial, so we made stakeholders do part of this work.’ (M10). Opinion leaders, interest groups or authorities in a certain policy domain may help the party to promote an idea that is not (yet) supported by their voters. Parties seem to be well aware of the potential mobilizing power of societal actors; ‘The CEO of […] for example backed our last initiative on […]. That this CEO says something like that is golden. We suddenly had an ally to help us sell our idea.’ (PL1-O).

In line with existing typologies in the blame avoidance literature—Wenzelburger (2011) speaks of communicational strategies, McGraw (1990) of justifications and Hood (2010) of presentational strategies, we find that government and opposition parties alike, by explaining the seemingly unpopular proposal, stressing its fairness, future positive impact, and ideological consistency, try to ease party voter resistance.

Organize it so that it does least damage

Besides strategically tailoring their explanations, parties invoke organizational strategies (or in Hood’s terms agency strategies) hoping to decrease the likelihood of being held accountable for unpopular propositions on election day. First and foremost, and this strategy is mentioned by government party elites only, parties strategically time the announcement of a policy (confirming Vis 2016). It is electorally safer, they argue, to announce unpopular initiatives at the beginning of the electoral cycle—which explains why observational work finds that unpopular budget cuts are often implemented in the first year of the electoral cycle (see for instance Fernández (2012) on pension policy retrenchments). Indeed, a party leader argues; ‘This is why ‘tough’ decision are mostly taken in the beginning of the term’ (PL4-M). The rationale holds that citizens cannot hold parties accountable immediately after the unpopular idea is announced, so the further away from elections parties announce it, the less likely it is that voters will recall it at the ballot; ‘There are elections next year. Everyone realizes that now is not the right time to put an initiative on the table that is disliked. The distance from elections plays a big role… always.’ (M2). Or; ‘I know that most people that vote for our party […] and would therefore be offended by this proposal. We decided to go ahead anyway, but only because we were able do it in the beginning of the legislature.’ (PL5-M).

In a similar vein, government parties try to postpone the unpleasant effects of an initiative. By lifting these effects over the election, they reason, citizens will not (yet) punish them for it. Interestingly, as the following quote exemplifies, postponing the immediate unpleasant effects of a policy influences the (rationality of the) public debate on the policy;

This [policy initiative] has hardly any immediate impact on citizens. It is only effectively put in place by 2030, and human beings in general have a hard time imagining the future; they can look one year, perhaps in rare cases three years ahead. Therefore, discussions about these future policies are more rational. By making an abstraction of time, you get a different discussion, it is less emotional. (PL4-M)

Organize it in such a way that you can explain it

Not only do parties propose unpopular initiatives in the beginning of the electoral cycle to reduce the prospects of electoral accountability (see Hood 2010; König and Wenzelburger 2014), it additionally gives parties more time to explain their ideas. As such, the strategic organization of the implementation or announcement of a seemingly unpopular policy is related to the strategic communication; it simply gives parties a better shot at increasing the endorsement before the next elections take place. And, some interviewees stress, it is easier to convince voters of the benefits of what they propose if they have had the opportunity to actually experience the new policy before they are summoned to vote. As one minister elaborately describes it;

We recently decided to […], which received a lot of negative criticism. Journalists were covering the presumed drawbacks extensively, and the […] sector reacted quite vehemently. They claimed that we were deliberately boycotting them. We continued either way, and now everyone is on board. That it turned out positively, is probably due to the fact that the next elections were quite far away. There was time for the industry, but also for people, to get used to the new guidelines and now they actually experienced it, they can see the benefits. If we proposed this right before the elections, people would only see the negative sides (M2).

A final and related strategy is to announce an idea in different phases because it is less intrusive; citizens have time to get used to parts of it rather than having to process it all at once. In addition, a step-wise introduction allows for more fine-grained strategic communication. Parties, we learn from the interviews, first try to convince voters about a rather unobtrusive aspect of the proposal, and as soon as they have their support, they move on to propose and promote other, more tricky, parts, which is easier because citizens are already convinced about the baseline idea; ‘Our voters are not ready yet for the radical ban on […] we want to propose. So we have to do it in different stages; subsidize alternatives […] for example, and get everyone on board by tacitly explaining why such a ban is necessary. If this works out well, we can introduce a total ban.’ (PL1-O).

Party strategic unresponsiveness

Summarizing the above findings, Fig. 1 visualizes a party’s decision-making calculus. First of all, we find that parties want to please their voters, and in particular, that they are careful not to offend them. Still, all party elites come forward with multiple examples of initiatives, ideas or policy proposals they did pursue against the will of their voters. The reasons for ignoring party voter preferences are in the first grey arrow; sometimes, voter preferences go against the party’s core ideology (i.e. policy-seeking goals), go against what is considered best for society or are simply considered flawed, unfounded or inconsistent. Additionally, government parties need to compromise with coalition partners and take decisions by consensus and, therefore, they are sometimes forced to implement policies that go against their voters’ preferences. These four considerations explain (in part, of course) why party decisions and positions do not always match with their voters’ desires.

Fig. 1
figure 1

Strategic unresponsiveness: party’s decision-making process

Full size image

How parties behave in this situation of intentional incongruence, hinges on their estimation of the electoral consequences they might face (second grey arrow). Interestingly, we find that parties do not necessarily anticipate electoral losses when opposing their voters’ preferences. They show statesmanship in doing so and this too may be appreciated by voters, elites reason. Often, though, parties do anticipate electoral repercussions for unresponsive behaviour—and especially on salient issues, and in that case they invoke different strategies to avoid, or at least to contain the expected electoral retribution. If possible, parties try to hide these initiatives from the public eye, or even actively distract attention. Usually, however, this kind of obfuscating behaviour is considered either undesirable or unfeasible (see third arrow). Assuming their voters are or will become aware of the unpopular initiatives, parties invoke two types of strategies. For one, they try to decrease voter resistance by communicating about the fairness, future positive effects and ideological consistency of their initiatives. Organization-wise, then, government parties propose unpopular initiatives at the beginning of the electoral cycle and postpone the most immediate effects to avoid electoral accountability. Interestingly, this strategic timing foremost facilitates (strategic) communication: the earlier an initiative or idea is announced, the more time parties have to ease public resistance by unfolding communication strategies.

Discussion

The current study identifies the conditions under which parties deliberately pursue ideas that are not endorsed by their voters by means of in-depth interviews with nineteen ministers and party leaders in Belgium.

Our qualitative approach adds insights to the existing observational and theoretical literature on party responsiveness and blame avoidance. With regard to the former, we show that all parties are keen on pleasing their voters and, in particular, are careful not to offend them, especially so on salient issues. Yet, although parties are overall reluctant to propose initiatives that might meet a great deal of resistance among their voters, they sometimes intentionally do. In particular when they deem their voters’ desires uninformed, ideologically flawed or a threat to the general public interest. Parties are not mere agents of voter preferences, nor do they simply bring their ideological preferences to voters on election day. In balancing their electoral ambitions with the pursuit of policy goals, the anticipation of electoral sanctions plays a vital role. Yet, the expectation of future elections constrains rather than paralyzes parties; when parties go against the preferences of their voters and fear that they will be punished for it on election day, they invoke strategies to reduce voter resistance and decrease the likelihood that they will be sanctioned. Thus, these strategies allow to reconcile conflicting party goals; advancing policies parties deem necessary while minimizing the likelihood of electoral repercussions.

Adding to the blame avoidance literature, then, we first of all show that parties think in terms of blame avoidance—something that has often been assumed, but that has rarely been shown empirically. Indeed, we show that all parties in their daily decision-making—not only government parties implementing tough retrenchment policies (e.g. Vis and Van Kersbergen 2007)—invoke strategies to contain electoral losses when going against their voters’ will, and we add more nuance as to what strategies are used. After all, we find that parties mostly accept responsibility for unpopular proposals and try to justify them rather than to run away from their responsibilities—in that sense concept ‘blame avoidance’ is ambiguous because parties do seem to accept responsibility and try to move voter opinion in their advantage. It is true that parties try to avoid electoral retribution by proposing unpopular ideas in the beginning of the electoral cycle, but they also use this time until the next elections to find support for their ideas by carefully explaining why they adhere a certain position or pursue a particular initiative by stressing its fairness, benefits and/or ideological consistency. In that regard, the term “coping strategies” may be better suited to grasp how parties strategically handle being unresponsive to voter preferences.

These findings matter for democratic representation. It shows that Belgian parties tend to adhere to a Burkean view of representation in that they cross voter preferences when they feel that these preferences conflict with society’s best interest. In light of a trustee view of representation, then, where political actors supposedly act in in the interest of citizens rather than following their wishes (see Pitkin 1967), the finding that parties occasionally oppose their voters’ preferences for the sake of the ‘greater good’ could be considered as positive for democracy. That political actors want to avoid electoral sanctioning by invoking coping strategies (which is in line with Mansbridge’s theory of anticipatory representation), in a way helps positions being chosen and decisions being taken that might otherwise fail because the risk of losing support is too high (Mayhew 1974). Hence, a positive read of the findings would be that party politics is not capitalized entirely by short-term responsiveness to voters’ desires and a refrainment from enacting politically costly (responsible) initiatives (Mair 2009). From a delegate view on representation, then, the finding that parties dare to go against voter preferences is more problematic in that political actors are supposed to dutifully translate public opinion into public policy. Additionally, the fact that parties perceive voters as educable, or even manipulable, might be worrisome too; parties hiding certain initiatives, for instance, deny their voters a fair chance to hold them to account on election day.

The novelty as well as the limitation of this study is that we rely on accounts from party elites that are inherently subjective (Harvey 2011). Even though interrogating key party decision-makers is the only way to grasp the considerations that drive party behaviour, we are well aware that this method has its limitations. For one, our explorative approach enabled us to identify considerations and strategies that are top of mind for party elites while it does not offer definitive proof about how regularly parties oppose their voters’ will, nor about how often they fear electoral backlash and invoke certain strategies. In that sense, it is important that our typology of party (un)responsiveness is validated (and elaborated on) in a more systematic, observational setting. By means of a content analysis of party communications and party initiatives, scholars could establish how often (certain) parties use (certain) strategic explanations for being unresponsive, or, for instance, what characterizes policy proposals that are announced early on in the electoral cycle. In general, that we have identified what strategies parties use, may help scholars to look for them in observational data, data that would allow to examine variation between parties, issues and even political systems. Another way forward is to conduct (survey-embedded) experiments with party decision-makers; scholars could manipulate the degree of voter resistance and see whether it affects their (hypothetical) communicative reactions, and the likelihood to pursue responsible policies. Finally, this study asks elites about party voters only—a deliberate choice, in that parties are first and foremost expected to represent their own voters, which makes it interesting to understand why they intentionally go against their preferences. Yet, it would be interesting if future work could zoom in on how parties deal with the preferences of potential party voters, and those of the general public (e.g. Ezrow et al. 2011).

Overall, this study has taken a modest step in unravelling one particular aspect of the chain of representation; it has identified why and under what conditions parties want, or rather dare to, pursue unresponsive policies and take unpopular positions.