Abstract
The functioning of representative democracy is crucially dependent on the representative behaviour of political parties. Large parts of the party representation literature assume that voters expect parties to fulfil the promises of their election programs. What voters actually want from parties, however, remains largely unclear. Within the Australian context, this article investigates the preferences of voters regarding three ideal party representative styles: ‘promise keeping’; ‘focus on public opinion’; and ‘seeking the common good’. Using a novel survey tool, this study finds that voters value promise keeping highly when it is evaluated individually. However, they rate seeking the common good as most important when the three styles are directly compared. A multinomial logistic regression analysis shows that, in particular, voters who have been involved in party grassroots activities prefer promise keeping. These findings have wider implications for our understanding of how representative democracy can and should work.
Introduction
Once parties are elected into parliament or government, which principle of representation should they follow? According to the mandate theory of democracy and most prominent models of party government, the main task of parties following an election is to keep the promises they made before the election (e.g., Naurin, 2011; Thomassen and Schmitt, 1997; Thomson et al., 2017). The argument is that voters give their support to a party on the basis of its election program and expect their selected party to ensure their vote has the intended consequence. While empirical research has investigated whether parties provide meaningful election programs, whether voters make their electoral decisions on the basis of these programs and whether parties indeed fulfil their promises (e.g., Martin et al., 2014; Thomson et al., 2017), there is little research on whether voters share the normative and theoretical assumption that parties should keep their election promises. According to rival theories of democratic (party) representation, 1 voters could prefer two alternative representation styles, either that parties react strongly to public opinion in a constant process of input-updating or that parties engage in a process of identifying and enacting the common good. While other theoretical representational links might exist between parties and voters, promise-keeping, responsiveness to public opinion, and enacting the common good encapsulate the most basic principles of party representation. Thus, focusing on these three styles and by surveying Australian voters and measuring their preferences regarding the representation styles of political parties, this article asks: Which style of party representation do voters prefer?
This article investigates this question utilizing a unique survey among Australian voters that asks respondents, first, to assess the importance of the three theoretical representation styles (promise keeping, public opinion, and common good) independent of each other and rate them against each other in pairs. 2 Unlike previous research, the survey uses a vignette to ask respondents to choose between all three representation styles (see Figure 1). In a subsequent analysis, I show that Australian respondents value promise keeping very highly only when they evaluate the styles individually. Once the respondents need to state which style they prefer over others, a clear hierarchy appears. In contrast to theoretical expectations, Australian respondents value promise keeping least and almost half of them choose the common good over the other options. In a second step, I analyse determinants of respondents’ preferences for different representation styles. The findings suggest that a crucial factor in the overall low valuation of promise keeping is that mainly respondents with first-hand experience in the grassroots organizations of parties prefer this style. As only few respondents have this experience and the low level of engagement in parties is a widespread phenomenon in Western democracies (van Biezen and Poguntke, 2014), it seems likely that similar results can be expected in other established democracies.
Survey vignette, rating three styles against each other.
Representative democracy and party representation styles
There is a long-standing discussion about how representative democracy should work. Its basis is usually the fundamental idea that the people decide on policies by electing individuals or parties ‘who are to assemble in order to carry out its will’ (Schumpeter, 1942: 55). While there are different conceptualizations of representative democracy (e.g., Dahl, 1971; Pateman, 1970), in most modern representative democracies the relationship between parties and voters is central. The theoretical literature on representational styles, that is, the types of activities democratic representatives can undertake to fulfil their role (Eulau et al., 1959), provides possible answers to the question of how parties can, or should, make representative democracy work. For parties, it suggests multiple possible ideals that fall into two broad categories similar to Pitkin’s distinction between delegate and trustee styles of representation (Pitkin, 1967). 3 The delegate conception focuses on representational styles that provide a direct link between the policy preferences of voters and the policy decision-making of political parties. The main difference between the two most common conceptualizations of this party representation is whether its main mechanism is through elections (promise keeping) or a constant communication process from voters to parties (through public opinion). On the other hand, broader discussions of democratic representation call for parties to act in the interest of the common good, which entails the interest of the whole society (including, for example, future generations) and is closer to the trustee style of representation. In modern policy-making, the practice of following expert advice comes closest to implementing this idea. These three styles are discussed in more detail below.
Party representation styles: three alternatives
The arguably dominant model of party representation in the Western party literature is the idea that elections create a mandate relationship between voters and parties. In a nutshell, mandate theories such as the Responsible Party Model assume that before elections, voters have policy preferences and parties have policy offerings. During the election, voters choose the best-fitting program relative to their own interests. Post-election, parties that are elected then enact their pre-electoral policy proposals (Thomassen and Schmitt, 1997). Similarly, Bach argues that the electoral decision of voters equals voters’ endorsement of the party’s program which then creates the party’s ‘responsibility to enact its program into law’ [emphasis in original]. Thus, the core responsibility of parties after the election is to fulfil their promises by attempting to enact their pre-electoral programs as faithfully as possible. This ideal is also rooted in more general styles of democratic representation, in particular Jane Mansbridge’s ‘promissory representation’ (Mansbridge, 2003), which she identified as the only representation mechanism that establishes an accountability mechanism between parties and voters (Mansbridge, 2003: 515).
Given the importance of the promise-keeping ideal in the literature, a broad range of empirical research investigates whether and under which circumstances parties behave in line with this assumption. These studies, questioning whether parties implement proposed policies, typically find a high degree of program-to-policy congruence (Martin et al., 2014; Moury, 2011; Thomson et al., 2017). In their interpretation, these studies highlight the relevance of this finding for the legitimacy of democracy (e.g., Moury, 2011: 35). While there seems to be wide theoretical agreement that promise keeping is central to modern representative democracies, there is very little indication about whether voters actually agree with this normative assumption. Some research indicates that voters doubt parties’ intention and ability to enact their pre-electoral promises (Naurin, 2011: 71–79). Of course, this does not necessarily mean that voters place the same importance on promise keeping.
A second theoretical party representation style reflects the ideal that voters can give constant input to their representatives and that parties are constantly receptive to this input. In this tradition, Bardi et al. (2014: 237) write about ‘the tendency, and indeed the normative claim, that political parties […] sympathetically respond to the short-term demands of voters, public opinion, interest groups, and the media.’ The core difference from the promise keeping ideal is that parties are constantly updated about which policy decisions are required of them, instead of only at election time. While changes in public opinion can occur and be communicated in a short time frame, such changes are only possible around election time in the promise keeping style. In general, this second representation mechanism pertains to parties receiving constantly updated instructions through public opinion on how they should make decisions. This is not to say that, empirically, this is a one-way communication process. Slothuus (2010) has shown that parties can also influence public opinion by changing the frames in which they discuss policies. However, it seems unlikely that changing public preferences can be a principle mechanism that parties should adhere to in order to fulfil their representative task. As the question at hand is about what parties should base their decision-making on, the relationship between public opinion and parties is conceptualized as unidirectional.
In the democratic representation literature, we find a third prominent option to conceptualize democratic representatives, which is based on the theoretical rejection of the idea that democracy includes voters binding their representatives’ decision-making (e.g., Dahl, 1971; Schumpeter, 1942). Instead representatives are perceived as trustees. In the traditional Burkean sense, trustees are individual members of parliament (MPs) gaining decision-making capabilities through elections and being tasked to find the policy solution that they agree is best for the nation or, more generally, for the common good (Brito Vieira and Runciman, 2008; Pitkin, 1967). This idea, while distinct, is compatible with the normative demand for representatives to not only focus on the interest of their particular voters, but to find policy solutions that consider the needs of non-voters (e.g., children and future generations).
This original understanding of MPs as trustees, however, is not directly applicable to modern party representation. First, parties are not individuals who can form beliefs about the common good but collective actors with complex internal decision-making processes. Second, decision-making takes place in the increasing complexity of a globalized world. Thus, one prominent, if not uncontroversial, implementation of this idea is to outsource the decision-preparation to groups of policy experts that are tasked to deliberate and propose policy reflecting the common good (e.g., Fischer, 2009). The idea of evidence-based policy-making falls into the same category (Sanderson, 2002) and empirical findings indicate some voter affinity with this representation style, sometimes called ‘stealth democracy’ (Hibbing and Theiss-Morse, 2002; Ruostetsaari, 2017). The deliberation among experts, here, takes the place of the Burkean ideal of deliberation among parliamentarians, who were understood as experts in their own right. It is important to note that this mechanism does not rest on the idea that experts decide; elected representatives still make the democratically legitimized decision. 4 However, instead of following their pre-election promises or the preferences of the public opinion, parties pursue the common good by making decisions on the basis of policy experts’ advice.
Of course, these three representation styles constitute ideals. In reality, parties might follow one style in one policy decision and another if faced with a different situation. This leads authors like Goot to argue that ‘promise-keeping is a political act; it is neither a question of morality nor a question of logic’ (Goot, 1999: 342). Similarly, individual voters might want parties to behave one way or another, depending on the decision at hand, or even purposefully use institutional settings to restrict a government’s mandate in case there are unpopular individual promises (see Sawer, 2001: 50 for an Australian example). However, the purpose of this article is not to explain real governmental behaviour or test all possible preferences voters might have towards party behaviour in all possible circumstance. Rather, this article investigates the ideal preference of voters in the light of common assumptions in the literature.
These ideal representative styles are important because the mode of decision-making would become irrelevant if we accept that voters only care about the individual policy decision in light of their preferences and their salience judgement. However, we know that voters do care about on which basis parties make their decisions (Evans et al., 2014: 6; Naurin, 2011: 87) and media reports about parties’ promise keeping (Costello and Thomson, 2008). Further, these three alternative preferences play a prominent role in political life and are often pitted against one another. For instance, parties face the accusation of lying if they did not stick to their promises but are criticized for being inflexible and ignoring the public if they are seen to ‘cling’ to their policy programs. However, parties that listen to the public might be accused of ‘pandering to opinion polls’. Also, parties that follow experts’ opinions might be considered to be outsourcing decisions to unelected experts instead of doing their work, that is, sticking to their promises or listening to the public, but those that do not follow the experts might be charged with ignoring political guidance.
Voter preferences towards representative styles: previous evidence
Empirically, the relatively young literature on the representative preferences of voters shows that they value all three representation styles. 5 Most of this literature, however, focuses on how voters want MPs to behave, independent of whether the primary representative actors in the countries are MPs (see Campbell and Lovenduski 2015; Carman, 2006, 2007; Vivyan and Wagner, 2016) or whether they are parties (see Bengtsson and Wass, 2010, 2011; Méndez-Lago and Martínez, 2002; von Schoultz and Wass, 2016). Just as Thomassen (1994) ascertained that the representation literature in general focuses too little on parties, parties have so far been ignored by scholars of voters’ attitudes towards representative styles. Still, some of these results are applicable. For instance, Bowler (2017) analyses Europeans’ preferences for governments being delegates or trustees in 2012 and finds that only about 18% of respondents preferred governments to stick to their policies. Finally, most established voter surveys include questions about voters’ preferences regarding specific policies. While such questions allow analysing the congruence between party positions and voter preferences, they do not enable direct inferences about voters’ preferences towards parties’ representative styles.
The Australian case
Previous research on voters’ preferences focuses especially on the representative behaviour of MPs, mainly in the United States (US) or single Western European countries. The current study was conducted in Australia, which is comparable to the earlier cases as an established democracy with a consolidated party system. At this general level, it represents a typical case among Western democracies. At the same time, Australia’s political system includes some distinguishing institutional features. First, Australia has a very short election cycle of three years at the federal level and compulsory voting, which forces all citizens with voting rights to engage with electoral and party politics on a very frequent basis. Second, the electoral system is based on single-member districts, which could lead to weak parties if the system were dominated by individual MPs, as in the US. Instead, Australia is a country with strong parties, exhibiting a very high degree of party unity in parliament (Carey, 2007), and party identification is a strong predictor of vote decisions (McAllister et al., 2015: 341). Thus, Australia is in this regard comparable to most Western European democracies. Third, Australia has a bicameral parliamentary system, which makes government mandate claiming difficult if the two houses are dominated by different parties (Bach, 2003; Goot, 1999). Thus, voters might discount the possibility of promise keeping and adjust their expectations. Whether this is true can only be tested in a comparative setting.
Measuring voter expectations
When constructing survey questions regarding voters’ preferences for representation styles most studies have respondents rate representational styles against each other (Bengtsson and Wass, 2011; Bowler, 2017; Carman, 2006). Bengtsson and Wass (2010) rely on survey respondents’ evaluation of the importance of each possible option by itself. This article combines these approaches, asking respondents, first, to individually assess the importance of the three representation styles. In the second step, the respondents are asked to rate the three styles against each other in pairs. Finally, in the third step, the respondents are confronted with a situational vignette that leads them to choose between all three styles. These questions were fielded in a purpose-made survey experiment run among Australian voters. 6
The first survey question asks respondents to rate the importance of each representation style individually on a four-point scale from ‘very important’ to ‘not important at all’. 7 From previous studies, we would expect most respondents to rate all three styles as important (Bengtsson and Wass, 2010). However, in any given policy decision – and especially when the three styles result in different policy decisions – a party can logically only behave in accordance with one of them. Therefore, the second question asks the respondents to rate the importance of the styles in pairs. For each combination of the three styles, the respondents are asked whether they agree that one style should be adhered to even if this means that another representation style needs to be disregarded. The response scale has four steps, with a stronger and a weaker version of rating one style over the other.
The final question asks respondents to choose between all the three ideal styles (Figure 1). This question is framed in the form of a decision-making vignette (Hughes and Huby, 2004). Vignettes ‘refer to text, images or other forms of stimuli to which research participants are asked to respond’ (Hughes and Huby, 2004: 37) and are particularly useful to investigate individuals’ beliefs, values, and perceptions. Instead of being asked directly for their evaluation of a specific concept, which might be too abstract or open to interpretation, respondents indicate their reaction to a picture or a situation and researchers infer respondents’ evaluation from their answer. As this study is the first to approach the question of voter expectations from a clear trade-off perspective and is targeted at parties instead of MPs, the purpose of the vignette needs to be limited. The main aim of this vignette is to test whether respondents distinguish between the three representation styles and which choice they make on the level of principles. The trade-off between the three styles in a real-life policy decision would likely be influenced by the complexity of the policy issue, individuals’ preferences regarding the policy and individuals’ attitudes towards the acting party.
As the purpose of this study is to investigate the fundamental preference of respondents regarding party representation styles, these confounding factors were eliminated from the survey. To this end, an abstract vignette was created that focuses on the essence of the three styles and contains a neutral situation without any regional or temporal information. The policy decision at hand is about the placement of a new school and, thus, of very low party-political contention as there is no reference to the level of education spending or similar resource distribution issues. The vignette states that a new school will be built and the decision is only regarding its location in one of three streets. Furthermore, the placement of the school was not put into a personal context of the respondent, to avoid ‘not in my backyard’ considerations.
The abstract nature of the vignette has the advantage of focusing the respondents’ attention on the three representative styles. While there was a risk of respondent detachment from the question, the results show a relatively low rate of ‘don’t know’ answers and that 12% of respondents chose to invest the extra time to note down their own answer. The alternative approach of confronting respondents with a more salient, conflictual or moral decision would have risked measuring policy preferences instead of attitudes towards party representation styles. This could have only been mitigated by including a whole series of policy decisions and measuring respondents’ salience and policy preferences for all of them. Such an approach would likely have triggered survey fatigue and was generally outside the possibilities of this study. Furthermore, if representative styles do not matter when nothing is at stake for the respondent, it is unlikely that they matter when the decision is highly salient or conflictual.
Part of the vignette is that a party needs to make the decision regarding the school’s placement. This party was given an artificial name that has no resemblance to any existing party to avoid the confounding effect of the respondents’ party preferences. Within the vignette, all three representation styles are established, and the respondents are asked according to which of them the party should decide. Furthermore, the respondents were given the opportunity to express their preference for alternative ways of decision-making.
Determinants of preferences
Existing studies of voters’ preferences for representation styles and foci have provided evidence that we can expect variation in these preferences (Bowler, 2017; Esaiasson et al., 2017). With regards to the style of representation, meaning the decision between promise keeping, public opinion, and common good, a number of individual level factors have been shown in earlier studies to influence voters’ expectations toward the style of MP and government representation.
First, an individual’s level of interest in politics should influence their preference for representative styles. Those with a high level of interest are likely to have opinions about policies and prefer them to be heard (Bowler, 2017: 773). Individuals with little interest in politics are not bothered with the details of political decision-making and, thus, prefer their representatives to do what they promised or what experts advise (Bowler, 2017: 774). Second, education levels influence individuals’ understanding of the political process and, thus, might influence individuals’ representative preferences (Esaiasson et al., 2017). Higher education gives an individual the cognitive tools to understand the complexity of political decision-making (Carman, 2006: 110) and, thus, should lead to a preference for an expert-driven representation style. Individuals with low levels of education are likely less knowledgeable about political processes and thus opt for promise or public opinion-based representation styles that give them more control. Finally, previous studies have found standard demographic factors to affect representation preferences. These factors are: (1) age and gender; (2) class (Carman, 2006); (3) vote choice (Bengtsson and Wass, 2010; Bowler, 2017); and (4) rural or urban living (Bengtsson and Wass, 2011; Carman, 2007). To avoid omitted variable bias, these are added as control variables.
Furthermore, I introduce a new factor that has so far been neglected: whether voters have first-hand experience in parties’ grassroots activities. It stands to reason that individuals within the organizations that provide representation might have different experiences, views, and understandings than the general population. Within parties, individuals become familiar with political processes such as majoritarian decision-making and will be exposed to (and possibly take part in) internal struggles about the programmatic development of the party. As such first-hand experiences may have socialization effects (Newton, 1997) they might lead to different attitudes towards representation styles.
Analysis: what do Australian voters want?
The questions about voters’ preferences for specific representation styles were fielded in a tailor-made online survey of randomly selected Australian voters aged 18 and older. The survey was run on 4–10 May 2016 by the professional Australian survey company OmniPoll. OmniPoll has the contact and demographic information of multiple ten thousand Australians that have agreed to participate in surveys. The company contacted potential respondents, randomly selected from their panel within basic demographic strata, and invited them to participate in the survey. Further potential respondents were contacted during the fielding time to reach a representative sample. OmniPoll provides weights for gender, educational level, region of residence, and vote intention where the sample does not match the last available census data. The final sample contains 1,222 responses and the results reported below are population estimates.
Voters’ preferred representation style
The upper panel of Table 1 shows the results of the respondents’ evaluation of each individual representation style. At first, Australian respondents agree with the mandate argument that it is vital for parties to keep the promises they make in their electoral programs. 92% of the respondents deem this representation style important, and three of every four respondents as very important. Neither of the other two styles have higher rates of being deemed very important. At the same time, Australian respondents also value the other two representation styles as important with 83% rating receptiveness to public opinion and 91% of the respondents rating the common good as important. From this first impression, it seems that the answer to ‘What do Australian voters want?’ is ‘They want it all.’
Evaluation of representation styles, individually and in pairs, relative shares in percentages.
Note: values are population estimates based on a random sample of 1.222 Australian voters, aged 18 or older. Post-stratification for region, gender, education, and vote intention. Deviation from 100% due to rounding.
Source: Author’s own data.
This result might, of course, stem from voters not regarding these three representation styles as principally mutually exclusive. In other words, they might agree with one style for a particular policy decision and with another style for a different decision. Thus, the lower panel of Table 1 is more instructive to determine which representation style voters principally prefer and parties should principally adhere to. It shows the results for the question that makes respondents rate the styles against each other. The lower panel in Table 1 should be read as follows: to the left in the first row, we can see that 8% of respondents rate promise keeping as a lot more important than the common good and 17% say that keeping promises is more important than finding and enacting the common good. In the same line to the right we see that 41% and 26% deem the common good to be more or much more important than keeping promises. The results indicate a clear contrast to how highly Australian respondents regarded promise keeping seen on its own. While 75% of Australian respondents say that promise keeping is ‘very important’ when this style is evaluated by itself, it is valued as less important than the common good by two-thirds and as less important than public opinion by about 55% of the voters. This shows the importance of having respondents evaluate representation styles against each other. Furthermore, this result raises doubts about the priority assigned to promise keeping in much of the party representation literature.
The comparison between the importance of the common good and public opinion does not yield a clear result. 40% of Australian respondents value the common good higher, while 50% of them value public opinion more. To draw conclusions between these two styles, the results of the most complex question must be taken into account. Figure 2 shows the results for the survey question that asked the respondents to choose between three policy options that stand for the three representation styles.
Share of Australian respondents preferring representation styles.
The results in Figure 2 are surprisingly clear. First, the findings from Table 1 are confirmed that promise keeping is not very high on Australian respondents’ wish list. Only 9% of them chose this option. 22% of the respondents opted for the street supported by public opinion, while twice as many respondents (44%) decided that the party should follow the expert panel’s advice on what is in the general public’s interest. 8 Thus, when pressed to make a choice, Australian respondents overwhelmingly reject the party literature’s focus on promise keeping, choosing an alternative instead. Further, policy decisions that are marked out as optimal by a set of experts can be evaluated as preferable to public opinion and past proposals. Finally, the question remains whether there are any patterns underlying these decisions that might help explain why respondents opt for one of the three representation styles. The next section, thus, analyses the determinants of the decision shown in Figure 2.
Determinants of expectations
The factors proposed above to help explain respondents’ choice in representative styles were their interest in politics, education and first-hand experience with party grassroots organizations. Interest in politics was measured through a standard four category question (‘not at all’ to ‘very interested’) and education as a five-level variable ranging from primary school to university. Regarding individuals’ first-hand experience with the grassroots organization of parties, the most straightforward question at first glance is whether a respondent is a party member. However, membership has largely lost its relevance as numbers waned over the last decades (van Biezen and Poguntke, 2014) and official membership is no longer the only way of participating in a party (van Haute and Gauja, 2015). Thus, the variable used here measures whether respondents have ever participated in a local party branch meeting with answers ranging from ‘yes, many times’ to ‘no, never’. These are recoded into a binary variable, coding ‘0’ for respondents with no or only one attendance and coding ‘1’ for respondents with more experiences. The operationalization of the control variables is mostly straightforward (e.g., gender, age, and residence in an urban or rural area). Respondents’ yearly income (Carman, 2006) and whether they work for the government are used as proxies for class, and their vote choice is measured through a vote intention question. A final aspect that the analysis controls for is the temporal setting of the survey during the election cycle. The survey presented here was conducted two months before the 2016 Australian election. While a general election was anticipated for weeks before the survey was conducted, 9 the election was officially called in the middle of the surveying period. Therefore, the analysis controls for whether the respondents answered the survey before or after this event.
Table 2 shows the results of a multinomial logistic regression for the choice of the respondents between promise keeping, public opinion, the common good, and another solution. The baseline category is the answer that underpins the party representation literature: promise keeping. Thus, Table 2 reports the relative risk ratios (RRR), meaning the odds of a respondent choosing the other representation style over preferring that parties enact their program. A relative risk ratio of ‘1’ means this explanatory factor does not change the odds between a representation style and the baseline category. If the RRR is smaller than 1, an increase of the explanatory variable decreases the odds of choosing the respective representation style. If the RRR is larger than 1, this indicates an increase in the odds of Australian respondents choosing the style.
Determinants of preferences for representation styles.
n = 766; F(45, 721) = 1.91, Prob > F = 0.00.
Note: multinomial logistic regression using svy and mlogit in Stata. Relative risk ratios reported. Baseline category: preference for promise keeping. Results in bold: p < 0.05 and confidence interval (not shown) does not include 1.
Source: Author’s own data.
First, Table 2 shows that whether a respondent has first-hand experience with parties has, next to age, the strongest and most consistent effect on the representation styles preferred. Those respondents that have attended a few or many party branch meetings, thus learning about the inner workings of their parties, have 50% to 80% higher odds to prefer that parties keep their promises than follow any other representation style. 10 This finding is illustrated in Figure 3, which shows the marginal effect of having attended party branch meetings on the probability to prefer one of the representation styles. For the preference of common good and other considerations, first-hand experience with parties does not have an effect. However, those with party experience have a lower preference for parties following the public opinion and a higher preference for promise keeping. This is likely because they have a better understanding of the intraparty decision-making processes and, thus, might value the election program as a blueprint for party behaviour to a greater extent.
Predicted probabilities for the representation styles by first-hand experience with parties.
Table 2 also shows that most other factors do not have an effect on which representation style Australians prefer. Neither education nor interest in politics has an influence, thus indicating that neither political knowledge nor understanding political complexity plays a substantive role. Furthermore, most demographic variables do not have an effect. Neither income nor working for a government institution is influential for Australian voters (in contrast to Carman, 2006) nor is whether a respondent lives in a rural or urban area (in contrast to Bengtsson and Wass, 2011). Also in contrast to the findings of Bowler (2017), Bengtsson and Wass (2010), and Carman (2006), but confirming Carman (2007) as well as Vivyan and Wagner (2016), party preferences do not play a role for their preferred representation style either. 11 As neither of these previous studies placed their survey question in a specific policy context, it seems unlikely that the partisan neutrality of the vignette can explain these effects. It is conceivable that this is a country effect or dependent on certain political developments or events in the countries, which future studies could investigate.
Furthermore, gender only has an influence on the comparison between the baseline category and the option of adding additional steps to the decision-making process. Female respondents are 63% more likely to add further consultation or voting processes before coming to a decision than male respondents. For the choice between promise keeping and the two other representation styles, on the other hand, gender is also of no consequence. Age, on the other hand, plays a clear role on whether respondents prefer promise keeping or any of the representation styles. The older the respondent, the more likely they are to prefer one of the alternatives to promise keeping. Whether this is an age or a generational effect cannot be determined within this study. Finally, the official call for elections does not have an effect on respondents’ expectations. As it was general knowledge in the Australian public that the official call was only a matter of time, this result cannot be interpreted as a dismissal of the idea that expectations might depend on the temporal location within the electoral cycle. At this point, we can only infer that the official election call did not compromise the data collection.
Conclusion
Much of the party representation literature assumes that voters perceive their relationship to parties as one of mandating, meaning that voters elect a party so that they fulfil their pre-election policy program. The analysis in this article shows that this assumption is likely misleading. This analysis is built on three ideal types of representative party behaviour and tests which voters prefer: (a) promise keeping as the mechanism of voters’ electoral mandate; (b) following opinion polls as a more imminent responsiveness; and (c) targeting the common good with the help of expert advice as a modern mechanism for the trustee relationship. The innovative vignette-based survey measuring voters’ preferences asks respondents to first rate the importance of the three ideals independently of each other and then to evaluate their importance against each other. The analysis shows that about half of Australian respondents want their parties to first and foremost enact the common good based on expert advice. Importantly, less than one in ten respondents agree that faithfully enacting electoral programs is the most important representation style. Thus, only a small minority of respondents agrees with the dominant conception of mandate-based party representation.
One wider implication of this finding becomes clear in the subsequent analysis of explanations for Australian respondents’ preferences for representation styles. Consistently, promise keeping is preferred by those who have first-hand experiences with the inner life of parties. Being involved with a party on the ground seems to influence voters’ understanding of intraparty procedures as well as the centrality of programs to election campaigns and the party’s self-perception. Both groups that value promise keeping highly – young voters and voters with first-hand party experience – are, however, comparatively small, both in Australia and in Western democracies generally. Given current demographic trends, as well as the generally low tendency of voters to be actively engaged in these countries’ parties, the low preference for promise keeping is likely a wider phenomenon as well.
The conclusions drawn from this study are necessarily limited. In particular, Australia’s bicameralism obfuscates mandate attribution. Whether either this or compulsory voting influence voters’ attitudes towards party representation can only be studied comparatively. At the same time, Australia shares several important features with many Western democracies. It has a well-established, relatively stable party system and, even though the electoral system is in part built on single member districts, Australian parties are strong and the dominant political actors. Thus, the present study does allow for some insight into how voters in established democracies generally may want representative democracy to work.
One stream of democracy theory focuses on voters mandating their representatives, mainly parties, to enact certain pre-determined policies and much of the empirical literature on how democratic parties behave builds on this theory (Naurin, 2011; Thomson et al., 2017). However, studies like those by Achen and Bartels (2016) have shown, albeit for the US, that voters might not be able and interested in mandating their representatives to act out certain policy programs. If people (a) do not have the interest and ability to make policy decisions, (b) are aware of this, thus rejecting the notion of their direct involvement, and (c), as this study has shown, also put little value on the mechanism of promise-based accountability, this raises further doubt about the empirical and normative value of mandate-based theories of democracy. If voters are far less interested in ‘promissory accountability’ than many democracy theorists and party scholars, alternative forms of representation might need to be moved into focus (Mansbridge, 2003). Finally, the findings indicate that while public opinion should definitely play a role in political decision-making, people agree with experts playing a prominent role in advising parties in this process. That a sizable portion of people choose expert judgements over those of parties and, importantly, the general public runs counter to recent claims about the spread of ‘post-truth’ beliefs as well as calls to put more political decision-making power directly in the hands of the people.
Supplemental Material
IPS787330_Appendix – Supplemental material for Voters’ preferences for party representation: Promise-keeping, responsiveness to public opinion or enacting the common good
Supplemental material, IPS787330_Appendix for Voters’ preferences for party representation: Promise-keeping, responsiveness to public opinion or enacting the common good by Annika Werner in International Political Science Review
Supplemental Material
IPS787430_French_and_Spanish_abstract – Supplemental material for Voters’ preferences for party representation: Promise-keeping, responsiveness to public opinion or enacting the common good
Supplemental material, IPS787430_French_and_Spanish_abstract for Voters’ preferences for party representation: Promise-keeping, responsiveness to public opinion or enacting the common good by Annika Werner in International Political Science Review
Footnotes
Acknowledgements
A previous version of this article was presented at the 2016 Australian Political Studies Association Conference. I would like to thank all participants for their comments and insights. I would also like to express appreciation for the invaluable suggestions received from Anika Gauja, Duncan McDonnell and the People, Elections and Parties Group at Griffith University as well as the reviewers for the International Political Science Review.
Funding
This research was supported by Griffith University, New Researcher Grant 47404.
Notes
Author biography
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
