CC BY-NC-ND
0Contents lists available at ScienceDirect
Electoral Studies
journal homepage: www.elsevier.com/locate/electstud
Manifesto functions: How party candidates view and use their party's central policy document
CrossMark
Nikolaus Eder a, Marcelo Jenny b' , Wolfgang C. Müller
a IFES Institut für empirische Sozialforschung, Vienna, Austria b Department of Government, University of Vienna, Austria
ARTICLE INFO
Article history:
Received 26 January 2016
Received in revised form
22 August 2016
Accepted 14 November 2016
Available online 17 November 2016
ABSTRACT
Electoral manifestos play a crucial role in visions of party democracy and political science analyses of party competition. While research has focused on the contents of manifestos, we know much less about how parties produce manifestos and the roles they take in campaigns. This paper identifies three campaign-related functions of manifestos: they provide a compendium of valid party positions, streamline the campaign, and are used as campaign material. Based on the characteristics of the candidates, the parties and the campaign, the paper then derives expectations of how party candidates may differ in attributing importance to their party's manifesto. Based on a candidate survey after the 2013 Austrian general election, the paper shows that the key user-group of parliamentary candidates considers manifestos generally important and useful documents. Candidates' policy-centred campaigning and left—right distance from their own party are important in explaining individual differences. While the manifesto's service functions of providing a summary of valid party positions for the candidates and as a campaign means to be handed out to voters are widely appreciated, campaign streamlining is more divisive when it results in constraining candidates.
© 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
1. Introduction
The stylized version of party democracy puts great emphasis on parties' policy programs as structuring the campaign, voters' choices, party coalition building, and government policy. Parties then fight elections rallying behind a manifesto, laying down policy priorities and positions, and a team of leaders committed to them. The victorious party takes government office and implements its policy program. In cases of coalition cabinets, government policy is expected to be some kind of compromise based on the policy programs of the participating parties. In the subsequent election, voters will not only judge parties according to their policy programs for the next term in office, but also retrospectively, focussing on the government's performance and scrutinizing if the parties have kept their promises (see, e.g., Dalton et al., 2011).
The importance of parties' policy programs is reflected in the attention political science has given to them. Indeed, electoral manifestos have become ubiquitous in political science analyses.
* Corresponding author. Department of Government, University of Vienna, Roo-seveltplatz 3/1, Vienna, 1090, Austria
E-mail address: marcelo.jenny@univie.ac.at (M. Jenny).
This is largely due to the regularity with which parties dutifully produce these documents, and the constant effort of the Comparative Manifesto Project (now MAPOR) in turning these texts into data (Budge et al., 2012; Budge et al. 1987; Klingemann et al., 2006; Volkens et al., 2013). This ongoing research program has not only provided the discipline with widely used data, but has also led to a burgeoning methodological debate (Laver, 2001; Volkens, 2007; Daubler, 2012; Laver and Garry, 2000; Benoit et al., 2009; Gemenis, 2013; Budge and McDonald, 2014). Substantively, it is all about the issues that parties emphasise (salience) and the positions they take. While there is more to be found in party manifestos than salience and positions (Dolezal et al., 2016), we see the greatest research gaps not in the analyses of manifesto texts. To begin with, it is not at all clear what manifestos actually are, that is, how manifesto positions — regardless of their measurement — relate to post-election politics. Laver (2001: 67), for instance, distinguishes between ideal policy positions (representing the party's true convictions), stated policy positions (party ideals adapted to what the audience is considered willing to buy), and policy forecasts (what the party claims it will achieve if endowed with government power). Similarly, Ray (2007) sees manifestos as either rather abstract statements of parties' identities and philosophies,
http://dx.doi.org/10.1016/j.electstud.2016.11.011
0261-3794/© 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
sheer party advertising, or contracts between parties and voters. Finally, Harmel et al. (2016) suggest that (voter-directed) party image projection and (party members-directed) identity-building are the core functions of manifestos. Based on manifesto content analyses, they find that there are trade-offs between the two.
Answering this research question empirically, requires information on manifestos' impact on post-election politics. Thanks to pledge scholars, who study how specific party claims translate into government programs and actual policy (Royed, 1996; Thomson, 2001; Naurin, 2011; Schermann and Ennser-Jedenastik, 2014), and other students of the party-government-policy nexus (Budge et al., 2012; Dalton et al., 2011), considerable progress has been made in this research front.
Yet, all other aspects of party manifestos have hardly been touched upon by research. One such lacuna relates to the process of writing these manifestos. Only few studies have addressed this topic (Dolezal et al., 2012; Daubler, 2012; Switek, 2015).
Perhaps the greatest gap exists with regard to the precise role manifestos play in the campaign for which they are written. The present paper addresses this gap, focussing on how party candidates view and use their party's manifesto. To the best of our knowledge, it is the first study on how target 'manifesto users' conceive and use this document that is ubiquitous in political science (see, however, Kercher and Brettschneider (2013: 277-279), who report a survey of German party members).
We focus on the case of Austria, specifically the 2013 general election. Austria has a typical European multi-party system with a broad range of parties in terms of ideology, organization, and age. We develop and empirically test hypotheses relating to the characteristics of candidates and manifestos. We also discuss the relevance of party characteristics such as government status, age or performance in the polls, but due to multicollinearity concerns we exclude them in multivariate analyses. Clearly, there are limits to the generalizability of results relating to one country and one election. Yet, we believe that the expectations and findings of our study are also applicable to other countries and elections that share similarities with the Austrian one.
The paper proceeds as follows: First, we discuss the functions manifestos fulfil in parties' campaigns. Next, we develop expectations about when designated manifesto users, the party candidates, are more or less likely to attribute importance to the manifesto. We then present data and methods before we turn to empirical analysis. We find the manifesto's functions of providing a summary of valid party positions and as a means to be handed out to voters, are widely appreciated. These are the manifesto's service functions. The manifesto's function of streamlining a campaign is more divisive among candidates, in particular when it results in constraining candidates about what they should say. The degree of policy-centred campaigning by candidates, and their left-right distance from their own party, are particularly important in explaining differences between candidates.
2. Manifesto functions
Recognizing that only few voters read party manifestos, the literature has provided a number of arguments why manifestos 'nevertheless do constitute the major indirect influence on what parties are seen standing for' (Budge, 1987: 18). In the words of Ian Budge, '[t]his is because they form the basis of comment in the mass media and provide the cues for questions raised with party candidates at all levels, as well as staple issues for their campaigns' (Budge, 1987: 18). Manifestos, thus aim for indirect effects, at least, as much as for direct ones. In the words of Laver and Garry (2000: 620), manifestos are 'strategic documents written by politically sophisticated party elites with many different objectives in mind'.
We can distinguish, at least, three functions that a manifesto can serve: provide a compendium of valid party positions, establish supremacy over all other policy positions that may be attributed to the party and thereby streamline the party's campaign, and use as a campaign tool to directly inform voters.
2.1. Compendium of valid party positions
A comprehensive summary of party positions can serve several purposes: It can guide sophisticated voters, who may work out for themselves, which party comes closest to their ideal policy package; it can provide the mass media with input for their reporting; and it can be a useful tool for party candidates and activists. A document that provides economic access to how their party sees the real world, and the resulting blueprints or guidelines for policy change, is certainly advantageous for those who are responsible for spreading the party message to the people in face-to-face contacts in constituencies. Parliamentary work requires MPs to specialize; similarly, other candidates and party activists are likely to be much better informed about the policy areas relating to their own backgrounds and their party's policy, therein, than other domains. Yet, during a campaign party representatives cannot confine to the areas of their expertise, as questions on many others may come up during voter contacts, encounters with journalists, and in direct confrontations with candidates of competing parties. Hence, a single document arming politicians for all these situations can be of tremendous campaign value.
2.2. Streamlining the campaign
Political parties may provide such documents not only to ease the job of their candidates, allowing them to devote more time to actual campaigning rather than preparing for it, but also to streamline the campaign. To begin with, political parties are not confined to the manifesto in making policy statements. In some countries, they issue different types of programs (Budge, 1987: 18). These documents include the ones defining the party's core values and identity ('basic programs'), and the ones that provide the party's policy in a specific domain (e.g. regarding the economy or defence). While the former documents typically have little overlap with manifestos, the latter ones certainly do. Moreover, party leaders may make programmatic speeches, detailing policy plans when addressing, for instance, the party congress, and parties typically take positions on many issues in parliamentary debates. As actions are often seen to speak louder than words, government policy, and the accompanying explication of party goals, may be considered powerful alternative reference points to the contents of the manifesto. According to Laver's (2001) considerations, referred to above, these alternative sources of party policy often distinguish themselves along the principle-practice dimension, have different time horizons, and may be influenced by various context factors. Candidates, thus, may see the electoral manifesto as just one of many party statements and of no particular significance. The claim of the literature on party manifestos, however, is to the contrary. In the words of Klingemann et al. (2006: xvi), the manifesto is 'unique in being the only authoritative party policy statement approved by an official convention or congress'. Thus, the more relevant a manifesto is, the more unique it is in defining the party's policy, and less so if there are alternative (and probably not fully congruent) statements of party policy, from which candidates can freely choose. Manifesto supremacy, thus, means that the manifesto is the one and only authoritative source of party policy in the context of a campaign.
The manifesto should not only establish supremacy over alternative party documents as a 'contract with the voters' (Ray, 2007),
but must also define the party's position when there are different positions on any given issue within the party. If such intra-party tensions exist, it is particularly important to have the 'official version' of the party's stance, to which all sides in the internal debate conform during the campaign. This is important, as 'any great divergence from the line taken by the majority of candidates will be noted by, amongst others, opposing candidates who will make capital out of this sign of dissension' (Robertson, 1976: 52). A party that 'creates the impression of incoherence and unreliability' (Robertson, 1976: 51) certainly carries a competitive disadvantage. Hence, it is the collective goal of a party that candidates conform to the common party line, and thus preserve the party's brand name (Aldrich, 2011; Snyder and Ting, 2002). The manifesto is then the edict of party positions, and the candidates understand that departing from the party line will damage them collectively. This function of the manifesto is the more important the more incentives candidates have to seek a personal vote (Carey and Shugart, 1995; De Winter and Baudewyns, 2015; Eder et al., 2015; Selb and Lutz, 2015). Evidence from studying manifesto writing suggests that the positions taken in these documents are carefully elaborated, and aim at combining electoral appeal with intra-party acceptance - maybe at the price of some ambiguity (Dolezal et al., 2012; Daubler, 2012). Thus, campaign streamlining means that the manifesto is actually constraining what individual candidates, organizational units, and intra-party groups say during the campaign.
2.3. Campaign material
In campaigns, political parties and candidates are often generous in providing voters with printed material. The giveaways at gatherings of target groups, street contacts, or canvassing, include flyers, candidates' photographs, little presents carrying the party label, and all types of programmatic statements. Depending on the manifestos' format and the audience, the latter may include the party manifesto or, if available, a condensed version of it.
2.4. Compulsory exercise
Publishing a manifesto in each election is a time-honoured activity that parties are widely expected to undertake. The presentation of the manifesto to the public is a symbolic act that often signals the beginning of the campaign's hot phase. As such, it guarantees some media attention for the party and its ideas. Although the absence of a manifesto presentation event might lead to snide remarks in the press and a missed opportunity to motivate the party rank-and-file for the campaign, we can easily think of alternatives as a symbolic campaign start such as the presentation of (selected) candidates or a specially advertised speech of the party leader. A manifesto having a function reduced to a kind of compulsory exercise for the party, would thus not be considered particularly important.
Manifestos are then particularly important documents - and hence would conform to the political science attention to this type of text - the more they meet the first three of the criteria listed above. If, on the contrary, writing a manifesto is only a type of compulsory exercise for parties, it would be of lower status. However, note that there are potential tensions between the three important tasks of manifestos. In particular, providing a compendium of all valid party positions may not result in a document that can be handed out to most voters. We also note that manifesto functions are not equally demanding on the candidates themselves. Some are more a service the party provides (compendium of valid party positions, campaign material), while campaign streamlining constrains candidates, and may indeed handicap them in
addressing individual target groups. A manifesto, thus, may not possibly live up to all these tasks simultaneously and may still be an important document well worth analysing.
3. Hypotheses
When will party candidates consider manifestos relevant with regard to the functions widely attributed to these documents? We expect that individual characteristics of candidates, manifesto properties, and party characteristics impact how party candidates attribute functions to the manifesto.
3.1. Candidate characteristics
3.1.1. Policy-centred campaigning
There are different ways candidates can appeal for voter support. They can focus on their group membership, their past constituency services, the leadership qualities of the party's top officials, or on the party's policies. Clearly, real world candidates cannot evade any of these, but can try to focus, more or less, on these components. We reason that candidates with strong policy-centred campaigns find party manifestos more relevant than those who tend to concentrate on other ways to appeal to voters.
3.1.2. Left-right distance to own party
The greater the distance of candidates from the positions taken by their own party, the less they should value and care about the manifesto.
3.1.3. Chance to get elected
Individual candidates may also make their own choices with regard to the manifesto. Given that some of these documents are of book length and are often not written in the most readable style, working through them is costly. It is therefore worth considering the candidates' individual incentive structure. We reason that candidates who have a greater chance of being elected (because their party will be in parliament, and they are running for a reasonably safe seat) are more likely to carry these costs. Also candidates who see themselves as future policy-makers should attribute more relevance to their party's central policy document than candidates who will not be in such a position.
3.2. Manifesto characteristics
Manifestos differ greatly in their length and readability (Benoit et al., 2009; Dolezal et al., 2012). We expect the suitability of manifestos, as a campaign means directly handed out to voters, to decrease as their length increases and readability decreases (Merz and Regel, 2013; Kercher and Brettschneider, 2013).
3.3. Party characteristics 3.3.1. Ideology
With regard to ideology we relate to the greater programmatic tradition of parties of the left. This is also reflected in the relevance of manifestos. It means taking time to carefully elaborate them, producing more substantial documents, legitimising them by party congress vote, giving them an important role in the campaign, and using them as the blueprint for government policy (Dolezal et al., 2012). We, therefore, expect candidates of left parties to consider the manifesto more relevant than those from parties of the right (that traditionally have been more pragmatic and leader-oriented, see Epstein, 1967).
33.2. Newness
Recent research has shed light on the differences between old and new parties that seem particularly relevant in the present context (Bolleyer, 2013; Beyens et al., 2016). New parties lack institutionalization. Part and parcel of the institutionalization process are routinization and value infusion (Bolleyer, 2013; Levitsky, 1998). In the most general terms, routinization means that an organization's members can draw on precedents, and increasingly on party rules, to guide their behaviour. Value infusion means a 'shift in followers' goals from the pursuit of a particular objective', via the vehicle of a party, to the goal of perpetuating the party as an organization (Bolleyer, 2013: 16; Levitsky, 1998: 79). New parties are short of routinization and value infusion; their representatives lack stored information on party positions on which they can draw. Party candidates may have been motivated by a single issue or a few issues, rather than information on and agreement with the party's positions across the board. The manifesto, then, should be an important means to provide information to the candidates, and to ensure that they do not contradict each other during the campaign. In contrast, representatives of old parties can draw on party history and, very often, their own personal experience will provide candidates with many cues about the positions their party takes, so that the manifesto may be less relevant for them as a compendium of valid party positions. These candidates may also have a good intuition of how to handle issues that are 'sensitive' internally. Thus, they may also see less need in a streamlining device.
3.3.3. Inclusiveness of manifesto writing
Here our argument borrows from the literature of candidate selection, which distinguishes between inclusive (i.e. participatory) and exclusive (leadership-directed) processes for such selection (Hazan and Rahat, 2010). Inclusive processes of manifesto writing, allow broad participation of the rank-and-file during the drafting process, and formal decision-making by intra-party referendum or party congresses. We reason that more inclusive manifesto writing leads to manifestos of greater internal acceptance, and hence relevance to the candidates (though not necessarily to voters, see Adams et al., 2011; Spies and Kaiser, 2012).
3.3.4. Party government status
Much more policy-relevant information is available about government parties than opposition parties, and hence government parties may find it more difficult to ensure the centrality of the manifesto. Candidates will be confronted with such evidence in the campaign, and thus cannot afford to focus exclusively on the manifesto. Moreover, government parties often shy back from making firm policy commitments in their manifestos, while opposition parties may feel less constrained (Dolezal et al., 2012). The reasons behind these choices are that government parties, on the one hand, may want to avoid being asked why they had not already acted on these issues during the term that just ended, and, on the other hand, may not want to be too firmly constrained when returning to government. All this suggests that the manifesto is likely to be less important for the candidates of government parties.
3.3.5. Party performance
In a post-election survey, a party's electoral performance may influence the answering behaviour of candidates. Those of under-performing parties may identify their party's (unattractive) manifesto as a factor that contributed to the disappointing result, while candidates of over-performing parties are likely to adopt a friendlier attitude.
As our empirical application covers seven parties in a single election, the set of manifesto and party characteristics discussed above over-determines the patterning of the data. We would
require data from several elections to conduct a proper statistical test of the relative impact of macro-level variables. Therefore the multiple regression models in the next section focus on the explanatory power of individual-level characteristics. The models include no party characteristics apart from manifesto length, which is enough to identify a candidate's party affiliation.
However, descriptively we will show the values of these manifesto and party variables (in Table 1 below) and their bivariate relation with an index of manifesto importance (Fig. 2). We do not consider our results a test of the theorized manifesto and party-related factors, but rather a first check on their plausibility.
4. Data and methods
To increase our knowledge about the importance of manifestos in the campaign, and to provide a first check on the plausibility of our hypotheses, we draw on a candidate survey conducted in Austria following the 2013 parliamentary elections (Müller et al., 2016). All 3946 candidates were mailed a questionnaire shortly after the election. We limited our statistical analyses to the candidates of the seven parties with parliamentary representation, either before or after the election (N = 3204). These were the Social Democratic Party (SPO), the People's Party (OVP), the Freedom Party (FPO), the Alliance for the Future of Austria (BZO), the Greens, Team Stronach (TS) and New Austria (NEOS). The BZO had been in parliament before the election, but failed to gain seats. Team Stronach and NEOS ran for the first time and entered parliament. In contrast to NEOS, Team Stronach was represented in parliament even before the election, as the new party that had been founded outside parliament attracted MPs who had split off from the BZO. The response rate of candidates from the seven parties was 31 percent (n = 992). The representativeness of the sample with regard to gender is excellent. The sample exhibits a minor over-representation of the Green Party and minor underrepresentation of the FPO, a minor overrepresentation of candidates aged 50 years and older and underrepresentation of candidates younger than 40 years and a slight overrepresentation of candidates with an academic degree (see Table A-1 in the appendix).
A number of survey questions focused on the electoral manifesto, probing into the candidates' reading of the document, and its importance in the campaign. The following statements - some in the form of negations - aimed to elicit nuanced responses on manifesto functions:
• The electoral manifesto is an important source of information for candidates on the positions of their own party.
• The electoral manifesto is just one of many sets of political statements of my party during the campaign, and has no particular significance.
• The electoral manifesto is a summary of valid party positions, from which candidates must not deviate.
• The electoral manifesto is well suited to be distributed directly to voters.
• Publishing the electoral manifesto is a compulsory exercise, but it is not particularly useful for the campaign.
In addition to these more general statements, the survey asked them to evaluate the 2013 electoral manifesto's content, more specifically ('The electoral manifesto of my party did not contain much information on topics that became important during the campaign.'). Candidates could rate each statement on a five-point scale ranging from 'agree strongly', 'agree somewhat', 'partly agree and partly disagree' to 'disagree somewhat' and 'disagree strongly'. In the analysis, we first show the percentages for agreement, ambiguity or disagreement with a statement, and then build
Table 1
Party and manifesto characteristics.
Greens
LR position: mean (0 left, 10 right) Government status Newness Inclusiveness (1 highest, 7 lowest) Party performance Previous election - election 2013 Pre-election poll - election 2013 Electoral manifesto Length: sentences words Readability: Flesch scorea
Opp Old
2.0 -0.6
2989 41,086 10
Gov Old 4
-2.5 0.8
15,186
Opp New 1
5.0 5.0
18,009
Gov Old 3
-2.0 1.0
18,686
Opp New 7
5.7 -9.3
Opp Old
-7.2 1.5
Opp Old 5
3.0 2.5
159 1512 22
Note:a To calculate the Flesch score we used the website http://www.leichtlesbar.ch (Bachmann, 2014). A high Flesch score indicates good text readability.
a Likert scale index, from the original scores, for a subset of statements with the most consistent answers.
Our dependent variables include ordinal (the agreement scale) and interval data (the Likert index). We use OLS regression for the index and generalized ordered logit (Fu, 1998; Williams, 2006) for the ordinal variable. The latter does not depend on the 'parallel regression' assumption of the 'common' ordered logit regression model (Long, 1997; Greene and Hensher, 2009). The assumption stipulates identical slopes for the ordered levels, but often fails in empirical tests. The table of regression coefficients in the appendix lists more than one beta coefficient for a covariate. Model estimation was performed in Stata 14 using the gologit2 package (Williams, 2006). Regression modelling was performed without using imputation for cases with missing scores. Thus, the number of cases is lower in the multiple regression models.
The hypotheses presented in the previous section relate to characteristics of candidates, manifestos and parties, which constitutes a hierarchical data structure (Steenbergen and Jones, 2002). We include binary party variables as fixed effects, but except for manifesto length no party-level characteristics are included in multivariate analyses. We had no particular expectations for additional candidate characteristics such as gender, education and age. They were included simply as control variables. Continuous variables are standardized by dividing the values by two times their standard deviation to put the beta coefficients of these variables on the same scale as the coefficients of binary variables (Gelman, 2008).
Table 1 provides some context information on the Austrian party system and the 2013 campaign, Table 2 the descriptives of candidate characteristics subsequently used in the multiple regression models. The party's left-right position is the mean of candidates' locating their own party on an eleven-point scale. There are two left parties (Greens and SPO), and five parties located from centre-right to the right (NEOS, OVP, TS, BZO and FPO). Government (1) versus opposition (0) indicates each party's status before the election; newness is whether a party is running for the first time (1) or not (0) in a national election. Two indicators measure the parties' performance in the 2013 campaign. One is more long-term and compares a party's previous election result of 2008 to its result in 2013 (difference in vote shares). For the new parties, Team Stronach and NEOS, it is the 2013 election result. The other measure relates to the short term, and shows the improvement or decline from the level of support a party held half a year before the election (measured as the difference between the potential share in a poll, ATV Osterreich Trend/Hajek, 22 March 2013, n = 1000) and the actual vote share in the election.
In terms of the inclusiveness of manifesto writing, Austrian
parties had a rank-order from Team Stronach, the foundation of the octogenarian entrepreneur Frank Stronach (the least inclusive), to the other newly founded party NEOS. The NEOS manifesto had been put together in a very broad participatory process involving its members and registered party sympathizers. The inclusiveness variable negatively correlates with manifesto length: the more exclusive the process of writing a manifesto the shorter was the text presented in the election of 2013 (Pearson r = -0.85 with both sentences and words). Although we note the theoretical relevance of this characteristic, we only use manifesto length. Manifesto length is given as the number of sentences and words, which, as it turns out, are perfectly proportional (Pearson r = 0.999). We selected the number of sentences for the regression models. The Flesch score is a measure of text readability (Flesch, 1948; Bachmann, 2014), with higher values indicating better readability. The lowest values in Table 1 are at the level of the readability scores of academic papers.
From the candidate survey, we selected the following list of individual characteristics: Policy-centred campaigning indicates how strongly a candidate emphasized the party's 'program' in his or her campaign activities. The chance to get elected is based on his or her own subjective expectation prior to the election. Distance to own party is the gap between a candidate's own position on the left-right scale, and his or her location of their party on this scale. The remaining three variables - Education, Age and Female - are socio-demographic control variables.
5. Results
We begin by showing how many candidates have actually read their party's electoral manifesto. Next, we present to what extent the manifestos of the seven parties lived up to the functions generally ascribed to them. We then proceed from bivariate analyses to multiple regression modelling.
5.1. Reading the electoral manifesto
Table 3 summarizes whether and how carefully candidates read their party's electoral manifesto ahead of the campaign. The candidates' reading of these documents strongly varies by party. Overall, most of them recorded that they had read the manifesto, at least parts of it. This, of course, is the normative expectation candidates may be reluctant to deviate from by admitting that they did not read their own party's electoral manifesto. Three percent of the OVP candidates, and about the same percentage of SPO and Green candidates, nevertheless, stated that they had not read the electoral manifesto. Careful reading of the complete manifesto was most
% agree ■ % disagree EJ % undecided Fig. 1. Agreement and disagreement with manifesto-related statements.
prevalent among candidates of the two recently founded parties, Team Stronach and NEOS, as well as among BZO candidates. OVP and Green candidates were most inclined to partial careful reading. The pattern is suggestive and slightly surprising at the same time.
The high share of new party candidates reading the electoral manifesto carefully may be due to the lack of information stored by the candidates or information shortcuts about their party's stance.
In contrast, candidates from old parties can draw on a party history of well-known, entrenched policy positions or party records in government offices, to compensate for not reading the party's latest electoral manifesto. Extrapolating policy positions from the past provides an estimate of what the party stands for in the current election.
The BZOa excluded some prominent politicians from its list of
Fig. 2. Bivariate boxplots of manifesto importance.
Note: The first five panels show bivariate boxplots of predicted values of the manifesto importance index with covariates included in the regression model, the remaining four show covariates that were excluded due to multicollinearity constraints.
candidates ahead of the 2013 election, and tried to re-position itself as an economic liberal party oriented towards small businesses. The party's re-positioning may have motivated enlisted candidates to carefully read the manifesto. A potential explanation for the low percentage of candidates of the OVP, who read the complete manifesto, is the party's factional character. The OVP's ambition is to represent a broad spectrum of social and economic groups ('the people'), and frequently participates in government. Its electoral manifesto is expected to cover all policy areas, but in the eyes of various intra-party groups and their representatives these policy areas vary strongly in importance. At first glance, and somewhat surprising, is the low share of careful readers among Green candidates, given that they stand for a leftist party (i.e. a party with greater programmatic tradition) and many of them are highly educated. Yet, we should not forget that the Greens also published, by far, the longest manifesto of all parties, which may have discouraged Green candidates from reading it carefully.
5.2. Manifesto functions
What are the functions ascribed to electoral manifestos, and what is their overall significance for party candidates? To answer these questions, we refer to manifesto-related statements that candidates could agree or disagree with. Simplifying the five-point scale, we present the percentages for agreement and disagreement in the upper part, and the percentage of 'partly yes, partly no' answers in the lower part of each panel in Fig. 1 (see also Table A-2 in the appendix).
The first item asks whether the party's electoral manifesto constitutes an important source of information for candidates on the party's policy positions. An overwhelming majority across parties agreed. For these candidates, the electoral manifesto fulfils the function as a useful compendium of valid party positions. The next two items deal with the manifesto's role as a streamlining device. Ninety-one percent of NEOS candidates, but only 55 percent of OVP candidates disagreed with the statement 'The election manifesto is just one of many sources of political statements of my party during the campaign and has no particular significance'. For an additional 29 percent of the OVP candidates, it was partly true. The next statement establishes a strict benchmark: 'The election manifesto is a summary of valid party positions, from which candidates may not deviate'. Candidates' opinions split across parties into sizable camps of pro, contra and ambiguous. A relative majority of the candidates of Team Stronach (47 percent), the FPO (44), BZO (40) and SPO (35) agreed, whereas a plurality of candidates of the OVP (44), Greens (42) and NEOS (40) disagreed. Overall, OVP candidates exhibited the most sceptical attitude towards a view that the election manifesto is a substantive guideline, or even a constraint, on candidate campaigning. This may reflect the character of a party, in which leaders of factions frequently voice divergent policy positions (Müller and Steininger, 1994). At the other end of the spectrum are the Team Stronach candidates. They showed a highlevel of respect to the electoral manifesto as a disciplining device, but seemed to acknowledge that headline grabbing interviews of party founder and financier Frank Stronach, during the campaign, were an additional source of authoritative
Table 2
Candidate characteristics.
Policy-centred campaigning (not at all, a little, fairly, a lot): median Fairly
Chance to get elected (no, low/some, high/safe seat): median no chance
Distance to own party (11 point scales for own/party position): mean (sd) 0.9 (1.1)
Education (secondary, vocational, college, university, academic degree): median College
Age: mean (sd) 47.7 (131)
Female (%) 35.0
party statements. Candidates of the other new party, NEOS, held the electoral manifesto in high esteem, but showed no willingness to bow to it (see Fig. 1).
The fourth item probes whether a party's electoral manifesto is suitable for direct distribution to the voters. Candidates may factor in characteristics of the manifesto, such as length, when considering its suitability for informing voters about their party's positions. Relying on the transmission of the party's message via mass media may appear more attractive, but parties clearly differ in their a priori chance of getting media attention. According to conventional news value criteria, government parties and large parties, should do better than opposition parties and small parties. In five of the seven parties, a plurality or the majority of candidates saw the electoral manifesto as a suitable giveaway for voters. The pattern is intuitive if we look at the differences in manifesto length and readability. The BZO (65 percent agree), FPO (62) and TS (60) issued very short manifestos in plain German, whereas the manifestos of the OVP (51 percent disagree) and SPO (46 percent disagree) were long and more difficult to read. Two parties stand out as odd cases: almost two thirds of NEOS and 39 percent of Green candidates considered their manifestos as suitable campaign material for handing out directly to voters. The manifestos of these two parties were challenging and long reads, especially the tome the Greens presented as their manifesto.
Overall, most candidates rejected calling the manifesto a useless exercise. NEOS candidates overwhelmingly (80 percent) disagreed with the notion that publishing an electoral manifesto is only a compulsory exercise. Within the OVP, agreement and disagreement was evenly split (36 percent for each). In other parties, disagreement was between 40 to slightly above 50 percent.
The last item invited candidates to reflect whether their manifesto covered the important topics of the last election. We expect that the degree of satisfaction with the outcome of the election might influence the judgements about a party's campaign personnel, issues, strategy and means. The result of the election provided ample reasons for candidates of Team Stronach and BZO to voice discontent ex post. The chaotic campaign of Team Stronach saw a continuous, steep downward slide in the polls. The BZO failed the 4 percent threshold and dropped out of parliament, while NEOS was the surprising new party in parliament. Half a year before the election few observers considered the newly founded party a serious contender in the national election, and some polls did not include the party in questions on voting intentions. NEOS candidates uniformly disagreed with the statement that their electoral
manifesto did not contain much information on important campaign topics, but 29 percent of Team Stronach, 23 percent of BZO and 13 percent of OVP candidates strongly agreed. An additional 41 percent of the OVP candidates who partly agreed with the statement, suggest widespread discontent in the party. Partly due to the success of NEOS, the Greens fared below their own aspiration level, and despite the party's very long manifesto a large majority of the candidates (79 percent) partly agreed with the statement, implying a mismatch between the manifesto and the important issues in the 2013 election.
Our main dependent variable is an additive index from the candidates' answers. To arrive at that index, we first constructed the Likert index to check on the consistency of the answers. A test of the inter-item correlations with Cronbach's alpha led us to drop the statement 'the election manifesto is a summary of valid party positions, from which candidates may not deviate'. Alternative multivariate methods, such as principal component analysis or factor analysis, also attest that the answers to that specific item stand out from the rest, and produce a two-dimensional data pattern. We also excluded the evaluative statement on the 2013 electoral manifesto, as it is of a less general nature and does not specifically relate to one of the theorized manifesto functions. To create the additive index from the remaining four items, we reversed the two negations. The index has a Cronbach alpha of 0.67. Its scale ranges from 0 to 16 points. A candidate with four 'partly yes, partly no' is at the midpoint of 8 points. Higher values indicate higher appreciation of the electoral manifesto's importance in terms of the theoretically assigned functions. The mean index values by party are 9.1 for OVP, 9.9 for SPO, 10.7 for Greens and Team Stronach, 11.2 for FPO, 11.3 for BZO and 13.2 for NEOS candidates. The aggregate values reinforce the findings from the previous item by item discussion. NEOS candidates, on average, value the relevance of their manifesto considerably more than the candidates of the OVP.
5.3. A multiple regression model of manifesto importance
We use the index as a dependent variable and introduce manifesto length, the policy-centeredness of a candidate's campaigning, the distance to own party and the chance to get elected, as cova-riates, for which we have some theoretical expectations. Gender, level of education and age are control variables. Party affiliation is included as fixed effect and standard errors are clustered by party.
Table A-3 in the appendix provides detailed results of the OLS
Table 3
Party candidates reading their party's electoral manifesto (%).
Party NEOS TS BZO FP° SPO OOVP Greens
Carefully read the complete manifesto 77.1 65.5 64.9 51.9 40.8 24.9 23.5
Carefully read parts and glanced over the remainder 22.9 27.3 23.4 34.6 41.9 47.6 46.3
Skimmed through the manifesto 0.0 5.5 10.4 12.8 14.7 24.3 27.5
Did not read the manifesto 0.0 1.8 1.3 0.8 2.7 3.2 2.7
(n) (35) (55) (77) (133) (184) (189) (298)
regression model, Table A-4 contains party-by-party results. In Fig. 2, we use the predicted values from the statistical model, and create bivariate boxplot graphs. Variables that have a large impact are easy to understand when presented graphically. The graphs also allow displaying results for party-level covariates that due to multicollinearity were excluded from the statistical model.
The strongest explanatory factors are the policy-centeredness of the candidate's campaign and the perceived distance to the party a candidate represents in the campaign. The more 'programmatic' a candidate's campaigning, the more positive is the evaluation of the electoral manifesto. The larger the perceived gap between the position of the candidate and the candidate's party, the less important is the manifesto. Neither the chance to get elected nor any of the socio-demographic control variables, contribute much to explain the empirical variation in candidates' appreciation of manifestos.
The length of a manifesto does not seem to relate strongly to the importance candidates attributed to the manifesto functions, nor does its readability, a trait that is highly negatively correlated with manifesto length. Candidates of opposition parties, on average, seem to hold a somewhat higher opinion of the manifesto than candidates of the government parties, SPO and OVP (1.3 index points difference). The difference between the mean index scores of candidates from new parties and old parties is about the same, but the variance of the evaluations within these groups is larger. Candidates from parties that performed well, relative to the preelection poll results, had a friendlier opinion than candidates from parties that remained below expectations.
5.4. A generalized ordered regression model of manifesto as a constraint
We include the same list of covariates in a generalized ordered regression model with the item that we excluded from the index ('the election manifesto is a summary of valid party positions from which candidates may not deviate'). The variable is recoded in three categories indicating disagreement, ambiguity or agreement with the statement. Detailed results of the regression model are provided in Tables A-5 and A-6 in the appendix. Here we summarize its results as follows: individual-level variables do not account for much of the variation. The sign of some coefficients makes intuitive sense, for instance candidates subscribing to policy-centred campaigning more often express a deference to the manifesto. However, so do candidates who are more distant to their party's position, which is more difficult to rationalize. The weak explanatory power of the covariates suggests that the explicit prohibition on expressing a deviating opinion was a slight to personal autonomy and peeved a number of candidates from all quarters.
6. Conclusion
Electoral manifestos are essential documents in political science analyses. This is attested by the many theories, normative and empirical, and the large number of studies in which manifestos figure either as cause or effect. Yet, we know more about the contents of these documents than how they are written. An even greater lacuna exists with regard to how key groups of targeted manifesto users actually view and use this document. This article is the first study to address these questions.
We have outlined four different functions that electoral manifestos may fulfil. First, they collect a party's policy positions in one place for easy reference. Second, they streamline the campaign. This involves, on the one hand, defining and framing the party's key campaign themes and, on the other hand, containing the populist impulses of factions or individual candidates who opportunistically
might want to take a different stance on issues. To the extent that a manifesto can do both, it helps a party preserve its distinct brand. Third, manifestos can be used as campaign means to be handed out to voters. Alternatively, the manifesto may have no particular importance and can be seen as a compulsory exercise, a recurring ritualistic element of electoral campaigns.
We find that most party candidates consider their party's manifesto a relevant guidepost in their own campaigning. While candidates are largely appreciative of the manifesto as a kind of service their party provides (summarizing party policy positions, a document to be handed out to voters and providing guidelines for campaigning), opinions are much more divided about the constraining element in campaign streamlining (i.e. ruling out individual deviations from manifesto positions). Even when excluding this divisive item, candidates differ in their appreciation of manifesto importance. Candidates who conduct a policy-centred campaign have a higher opinion about the manifesto than others. We also empirically show that the degree of substantive congruence between candidates and the party impacts the value attributed to the party's central campaign document: ideologically more extreme candidates are less appreciative of the manifesto.
While we cannot conduct a proper statistical test based on data from a single election, the results do not suggest a strong divide between leftist parties as being more programmatic, and rightist parties as being more leadership-oriented and pragmatic. How a party fared in the election seems to have more impact on candidate views.
This article is the first attempt at studying how target manifesto users actually view and use these documents in the campaign. It relates to a single election in one country, and by necessity our findings require comparison between systems and elections. Given the noted importance of manifestos in political science research comparative efforts of such kind, for instance in the context of the Comparative Candidate Survey, could be of great use to the discipline. In the present article, we have provided a set of expectations that could be fully tested and further enriched in such contexts. As always in a new research field, not all of the initial expectations are confirmed when confronted with data. This, however, does not rule out that such questions would show relevant differences between countries if applied comparatively. Nevertheless, the Austrian experience suggests that careful consideration would be the order of the day. We have not reported here on items probing into whether and how candidates collected information about the stances of other parties (one in five read rival parties' manifestos) and whether they mentioned rival parties, thereby engaging in negative campaigning in the runup to the election (9 out of 10 candidates did so). We have also asked a few questions which merit probing deeper in further surveys. In terms of manifesto writing, for instance, it might be worthwhile to ask about the participation and early information of (presumptive) candidates. In terms of manifesto evaluation, in turn, it might be interesting to ask separate questions about the attribution of coverage and salience to issues. At a different level, another extension would be to study how other groups of manifesto users, such as journalists and societal opinion leaders, view and use these documents, and how this feeds back to the parties.
Acknowledgement
This research was carried out under the auspices of the Austrian National Election Study (AUTNES), a National Research Network (NFN) sponsored by the Austrian Science Fund (FWF) (S10903-G11).
Appendix
Table A-1
Representativeness of the candidate survey.
Absolute Frequency
Population
Sample
Relative Frequency
Population
Sample
a) Party affiliation
Social Democratic Party of Austria (SPO) Austrian People's Party (OVP) Freedom Party of Austria (FPO) Alliance for the Future of Austria (BZO) Greens (GRÜNE) New Austria (NEOS) Team Stronach (TS) Total
Duncan-Index of Dissimilarity1
b) Gender Male Female Total
Duncan-Index of Dissimilarity1
c) Age group <30 years 30-39 years 40-49 years 50-59 years
60 years or older Total
Duncan-Index of Dissimilarity1
d) Education
without academic degree with academic degree Total
Duncan-Index of Dissimilarity1
636 331 904 149 223 3535 7.13
2260 1275 3535 1.09
2429 1106 3535 4.40
645 347 992
105 145 274 287 118 992
638 354 992
0.186 0.180 0.180 0.094 0.256 0.042 0.063 1.000
0.639 0.361 1.000
0.124 0.175 0.278 0.280 0.l43 1.000
0.687 0.313 1.000
0.190 0.196 0.136 0.078 0.307 0.038 0.055 1.000
0.650 0.350 1.000
0.106 0.146 0.276 0.289 0.182 1.000
0.643 0.357 1.000
Notes: 1Duncan-Index of Dissimilarity: 0 [perfect sample representativeness] to 100 [maximum dissimilarity] (Duncan and Duncan, 1955).
Table A-2
Candidate opinions on electoral manifesto by party (%).
1. The electoral manifesto is an important source of information for candidates on the positions of my party.
agree 83.7 81.4 87.2 85.5 86.8 100 77.8
disagree 4.9 6.4 3.0 11.8 4.4 0 3.7
ambiguous 11.4 12.2 9.8 2.6 8.8 0 18.5
(n) (184) (188) (133) (76) (295) (35) (54)
2. The electoral manifesto is just one of many sets of political statements of my party during the campaign and has no particular significance.
agree 14.1 16.1 17.3 17.1 11.6 2.9 23.6
disagree 62.5 54.8 64.7 63.2 70.0 91.4 61.8
ambiguous 23.4 29.0 18.1 19.7 18.4 5.7 14.5
(n) (184) (186) (133) (76) (293) (35) (55)
3. The electoral manifesto is a summary of valid party positions, from which candidates must not deviate.
agree 35.3 23.5 43.6 40.0 32.3 28.6 47.3
disagree 31.5 44.4 21.1 24.0 42.3 40.0 25.5
ambiguous 33.1 32.1 35.3 36.0 25.4 31.4 27.3
(n) (184) (187) (133) (75) (291) (35) (55)
4. The electoral manifesto is well suited to be distributed directly to voters.
agree 27.2 20.3 62.1 64.5 38.9 65.7 60
disagree 45.7 50.8 12.1 9.2 32.8 8.6 20
ambiguous 27.2 28.9 25.8 26.3 28.3 25.7 20
(n) (184) (187) (132) (76) (293) (35) (55)
5. Publishing the electoral manifesto is a compulsory exercise, but it is not particularly useful for the campaign.
agree 27.7 35.8 22.7 21.3 24.3 8.6 32.7
disagree 41.9 35.8 50.8 45.3 52.4 80.0 45.5
ambiguous 30.4 28.3 26.5 33.3 23.3 11.4 21.8
(n) (184) (187) (132) (75) (292) (35) (55)
6. The electoral manifesto of my party did not contain much information on topics that became important during the campaign.
agree 8.2 13.0 7.6 22.7 6.2 0 29.1
disagree 75.5 46.2 78.6 56.0 79.0 100 45.5
ambiguous 16.3 40.8 13.7 21.3 14.80 0 25.5
(n) (184) (184) (131) (75) (291) (35) (55)
Table A-3
OLS regressions of manifesto functions and index.
is an important source of information on positions of own party is just one of many party statements during the campaign is well suited to be directly distributed to voters is a compulsory exercise but not particularly useful for the campaign Index of manifesto importance
Manifesto
Length (sentences) 0.06** -0.27*** -0.30*** -0.11** 0.23***
(0.01) (0.02) (0.02) (0.02) (0.03)
Candidate
Policy-centred 0.53*** -0.49*** 0.43*** -0.45*** 1.93***
campaigning (0.04) (0.08) (0.05) (0.05) (0.14)
Left-right distance -0.15 0.24* -0.11 0.17 -0.64**
to own party (0.07) (0.06) (0.07) (0.12) (0.13)
Chance to get 0.02 -0.04 0.18* -0.05 0.27*
elected (0.04) (0.08) (0.04) (0.05) (0.11)
Female -0.03 -0.10 0.23* -0.17 0.48
(0.07) (0.10) (0.06) (0.11) (0.24)
Education -0.04 -0.11 -0.21** -0.16 -0.00
(0.06) (0.08) (0.04) (0.07) (0.21)
Age -0.00 0.20 0.27** 0.16* -0.10
(0.06) (0.09) (0.06) (0.06) (0.18)
Party controls:
included
Constant 2.64*** 1.94*** 1.77*** 2.32*** 8.06***
(0.20) (0.24) (0.21) (0.16) (0.69)
Log Likelihood n
0.13 -1006 836
0.11 -1180 834
0.23 -1211 835
Notes: *p < 0.05; **p < 0.01; ***p < 0.001.
Table A-4
OLS regressions of manifesto importance by party.
SPO OVP FPO Greens BZO TS NEOS
Policy-centred 2.07*** 1.81*** 2.55*** 1.85*** 2.26** 1.32 -1.11
campaigning (0.50) (0.42) (0.60) (0.34) (0.73) (1.13) (1.25)
Left-right -0.67 -1.05* -0.46 -0.29 -0.26 -0.70 -2.52
distance (0.36) (0.49) (0.54) (0.36) (0.81) (0.97) (1.73)
to own party
Chance to get 0.56 0.21 0.03 0.31 0.12 0.48 -0.53
elected (0.37) (0.39) (0.47) (0.48) (0.71) (1.20) (0.70)
Female 0.25 -0.35 0.80 0.79* 0.43 2.67 0.47
(0.43) (0.43) (0.67) (0.38) (0.80) (1.44) (0,97)
Education 0.91* -0.06 0.53 -0.45 -0.30 -1.13 -0.13
(0.45) (0.45) (0.56) (0.43) (0.70) (1.01) (1.02)
Age 0.97 -0.36 -0.19 0.15 -0.44 -0.59 -0.47
(0.54) (0.41) (0.51) (0.38) (0.70) (1.11) (0.76)
Constant 4.21*** 8.21*** 7.10*** 8.46*** 8.94*** 10.06** 16.50***
(0.76) (1.16) (1.57) (1.14) (2.07) (3.19) (2.87)
R2 0.18 0.16 0.18 0.16 0.19 0.16 0.18
Log Likelihood -376 -370 -287 -563 -167 -126 -66
n 159 159 119 239 69 49 33
Notes: *p < 0.05; **p < 0.01; ***p < 0.001.
Table A-5
Generalized ordered logistic regression of manifesto as a constraint on candidate behaviour.
Manifesto is a summary of valid party positions, from which candidates may not deviate
Ordinal level
positive to ambivalent
ambivalent to negative
Manifesto
ManifestoLength Candidate Policy-centred campaigning Left-right distance to own party Chance to get elected Female Education Age
Party controls: included Constant
0.02 (0.07)
2.54*** (0.47)
0.47* (0.23)
-0.40 (0.23)
-0.26 (0.13)
-0.07 (0.13) 0.15 (0.14) -0.04 (0.14) -0.64*** (0.13)
0.83 (0.45)
Pseudo R2 Pseudo LL n
Notes: *p < 0.05; **p < 0.01; ***p < 0.001; Beta coefficients are from a generalized ordered logistic regression.
ch a n
h ic ih w m
le ti al at
> it bi
a it bi
e n 9) .4 4) .5 8) .4
to (0. (0. (0.
nt le 3 .9 0. - 14 0. - 3 .2 0. -
.2 > it bi
a sit bi
le ti al at
1 a am to
- > it bi
to n te
e al it bi
e g o tr in o
n ni t t
e g ht e -ce aig igh nce
8582 .9 .0 0. .7
S » 3 ^
.9 .5 .6 0. (0. (0. (0 0
3 6 .4 6
1 5 .3 )
.0 .1 0. 9
0 0 0- .4
) ) 7) (0. 6) 9) 27 (
ss§§ 72(
.1 .0 6 0. 0. 0.1 0) --08
0 .8 0.
2) .4 (0 0) .4 (0 2) .3 (0. 8 .3 1 0) .4 - (0
7 .4 0. .61 0. 5 .3 0. 5) .41 0.
) (0. ) 2) (
.3 0.3 (0 .70 (0 (0 2
5 18 6 6) .4 0. .2 .9 0 - 1. 0.
6) .2 .31 2) .3 9) .3 8 .4 0.
(0. (0. (0. (0. -
4 5 2 8 .4 0. -
.0 0. 0. .01 .0 2) .0
I j-j ^ U ^ y
ft st r a ec m u e äg
22 .1 3 3
References
Adams, J., Ezrow, L., Somer-Topcu, Z., 2011. Is anybody listening? Evidence that voters do not respond to European parties' policy statements during elections. Am. J. Polit. Sci. 55 (2), 370-382.
Aldrich, J.H., 2011. Why Parties? A Second Look. University of Chicago Press, Chicago.
Bachmann, C., 2014. www.leichtlesbar.ch. (Last accessed 21.12.15.).
Benoit, K., Laver, M., Mikhaylov, S., 2009. Treating words as data with error: uncertainty in text statements of policy positions. Am. J. Polit. Sci. 53 (2), 495-513.
Beyens, S., Lucardie, P., Deschouwer, K., 2016. The life and death of new political parties in the Low Countries. West Eur. Polit. 39 (2) http://dx.doi.org/10.1080/ 01402382.2015.1066589.
Bolleyer, N., 2013. New Parties in Old Party Systems. Oxford University Press, Oxford.
Budge, I., 1987. The internal analysis of election programs. In: Budge, I., Robertson, D., Hearl, D. (Eds.), Ideology, Strategy, and Party Change: Spatial Analysis of Post-war Election Programmes in 19 Democracies. Cambridge University Press, Cambridge, pp. 15-38.
Budge, I., Robertson, D., Hearl, D. (Eds.), 1987. Ideology, Strategy, and Party Change: Spatial Analysis of Post-war Election Programmes in 19 Democracies. Cambridge Unversity Press, Cambridge.
Budge, I., McDonald, M., 2014. Getting it (approximately) right (and center and left!): reliability and uncertainty estimates for the comparative manifesto data. Elect. Stud. 35 (1), 65-77.
Budge, I., Keman, H., McDonald, M., Pennings, P., 2012. Organizing Democratic Choice. Party Representation over Time. Oxford University Press, Oxford.
Carey, J., Shugart, M.S., 1995. Incentives to cultivate a personal vote: a rank ordering of electoral formulas. Elect. Stud. 14 (4), 417-439.
Dalton, R.J., Farrell, D.M., McAllister, I., 2011. Political Parties and Democratic Linkage: How Parties Organize Democracy. Oxford University Press, Oxford.
DaOubler, T., 2012. The preparation and use of electoral manifestos: learning from the Irish case. Ir. Polit. Stud. 27 (1), 51-70.
De Winter, L., Baudewyns, P., 2015. Candidate centred campaigning in a party centred context: the case of Belgium. Elect. Stud. 39 (3), 295-305.
Dolezal, M., Ennser-Jedenastik, L., Müller, W.C., Winkler, A.K., 2012. The life cycle of party manifestos: the Austrian Case. West Eur. Polit. 33 (4), 869-895.
Dolezal, M., Ennser-Jedenastik, L., Müller, W.C., Schermann, K., Winkler, A.K., 2016. Beyond salience and position-taking: How political parties communicate through their manifestos. Party Polit. (forthcoming).
Duncan, O.D., Duncan, B., 1955. A methodological analysis of segregation indexes. Am. Soc. Rev. 20 (2), 210-217.
Eder, N., Jenny, M., Müller, W.C., 2015. Winning over voters or fighting party comrades? Personalized constituency campaigning in Austria. Elect. Stud. 39 (3), 316-328.
Epstein, L.D., 1967. Political Parties in Western Democracies. Praeger, New York.
Flesch, R., 1948. A new readability yardstick. J. Appl. Psychol. 32 (3), 221.
Fu, V.K., 1998. sg88: estimating generalized ordered logit models. Stata Tech. Bull. 44, 27-30.
Gelman, A., 2008. Scaling regression inputs by dividing by two standard deviations. Statistics Med. 27, 2865-2873.
Gemenis, K., 2013. What to do (and not to do) with the Comparative Manifestos Project data. Polit. Stud. 61 (1S), 3-23.
Greene, W.H., Hensher, D.A., 2009. Modelling Ordered Choices. Cambridge Unver-sity Press, Cambridge.
Harmel, R., Tan, A.C., Janda, K., Smith, J.M., 2016. forthcoming. Manifestos and the "two faces" of parties: addressing both members and voters with one document. Party Polit (OnlineFirst).
Hazan, R.Y., Rahat, G., 2010. Democracy within Parties: Candidate Selection Methods and Their Political Consequences. Oxford University Press, Oxford.
Kercher, J., Brettschneider, F., 2013. Wahlprogramme als Pflichtübung? Typen, Funktionen und VerstOndlichkeit der Bundestagswahlprogramme 1994-2009. In: Weßels, B., Schoen, H., Gabriel, O.W., Wahlen, WOhler (Eds.), Analysen aus Anlass der Bundestagswahl 2009. Springer SV, Wiesbaden, pp. 269-290.
Klingemann, Hans-Dieter, Volkens, Andrea, Bara, Judith, Budge, Ian (Eds.), 2006. Estimates for parties, electors, and governments in Eastern Europe. Oxford University Press, Oxford. European Union and OECD 1990-2003.
Laver, M., 2001. Position and salience in the policies of political actors. In: Laver, M. (Ed.), Estimating the Policy Position of Political Actors. Routledge, London, pp. 66-75.
Laver, M., Garry, J., 2000. Estimating policy positions from political texts. Am. J. Political Sci. 44 (3), 619-634.
Levitsky, S., 1998. Institutionalization and Peronism: the concept, the case and the case for unpacking the concept. Party Polit. 4 (1), 77-92.
Long, S., 1997. Regression Models for Categorical and Limited Dependent Variables. Sage Publications, Thousand Oaks.
Merz, N., Regel, S., 2013. What are manifestos for? Selecting and typing documents in the Database. In: Volkens, A., Bara, J., Budge, I., et al. (Eds.), Mapping Policy Preferences from Texts III: Statistical Solutions for Manifesto Analysts. Oxford University Press, Oxford, pp. 146-168.
Müller, W.C., Steininger, B., 1994. Party prganization and party competitiveness: the case of the Austrian People's Party, 1945-1992. Eur. J. Polit. Res. 26 (1), 1-29.
Müller, W.C., Eder, N., Jenny, M., 2016. AUTNES Candidate Survey 2013: Public Use Version 2.1. University of Vienna, Codebook, Vienna.
Naurin, E., 2011. Election Promises, Party Behaviour and Voter Perceptions. Palgrave Macmillan, Basingstoke.
Ray, L., 2007. Validity of measured party positions on European integration: assumptions, approaches, and a comparison of alternative measures. Elect. Stud. 26 (1), 11—22.
Robertson, D., 1976. A Theory of Party Competition. Wiley, London.
Royed, T.J., 1996. Testing the mandate model in Britain and the United States: evidence from the Reagan and Thatcher eras. Br. J. Political Sci. 26 (1), 48—80.
Schermann, K., Ennser-Jedenastik, L., 2014. Coalition policy-making under constraints: examining the role of preferences and institutions. West Eur. Polit. 37 (3), 564—583.
Selb, P., Lutz, G., 2015. Lone fighters: intraparty competition, interparty competition, and candidates' vote seeking efforts in open-ballot PR elections. Elect. Stud. 39 (3), 329—337.
Snyder Jr., J.M., Ting, M.M., 2002. An informational rationale for political parties. Am. J. Polit. Sci. 46 (1), 90—110.
Spies, D.C., Kaiser, A., 2012. Does the mode of candidate selection affect the representativeness of parties? Party Polit. 20 (4), 576-590.
Steenbergen, M., Jones, B.S., 2002. Modeling Multilevel Data Structures. Am. J. Political Sci. 46 (1), 218-237.
Switek, N., 2015. Viel Arbeit für Nichts? Die Programmprozesse der Parteien auf europäischer Ebene. In: Kaeding, M., Switek, N. (Eds.), Die Europawahl 2014. Springer VS, Duisburg, pp. 115-124.
Thomson, R., 2001. The programme to policy linkage: the fulfilment of election pledges on socio-economic policy in The Netherlands, 1986-1998. Eur. J. Polit. Res. 40 (2), 171-197.
Volkens, A., 2007. Strengths and weaknesses of approaches to measuring policy positions of parties. Elect. Stud. 26 (1), 108-120.
Volkens, A., Bara, J., Budge, I., et al., 2013. Mapping Policy Preferences from Texts III: Statistical Solutions for Manifesto Analysts. Oxford University Press, Oxford.
Williams, R., 2006. Generalized ordered logit/partial proportional odds models for ordinal dependent variables. Stata J. 6 (1), 58-82.
