National parties’ election manifestos and the saliency of EU issues

Parties draft election manifestos for at least three reasons (see e.g. Däubler, Reference Däubler2012; Eder et al., Reference Eder, Jenny and Müller2017; Harmel, Reference Harmel2018; Harmel et al., Reference Harmel, Tan, Janda and Smith2018). Firstly, by summarizing their policy positions and making election promises on important issues, parties provide citizens with a document which might serve as a basis for their voting decisions. Secondly, election manifestos serve intra-party purposes by signalling parties’ main campaign positions and issues not only to voters but also to party members and supporters. Thirdly, these documents are guidelines for parties’ post-electoral bargaining on government formation and can be considered as ‘authoritative’ and ‘representative statements for the whole party’ (Klingemann et al., Reference Klingemann, Volkens, Bara, Budge and McDonald2006, p. 164).Footnote ¹

Manifestos are widely used to analyse parties’ policy positions, their emphasis of specific issues, and their responsiveness to voter preferences (for an overview see e.g. Adams, Reference Adams2012; Budge, Reference Budge2015; Fagerholm, Reference Fagerholm2016). The ways in which parties respond to voter demands have been investigated from two angles: position and salience. Parties try to adopt policy positions that are closely related to voter preferences (see e.g. Stimson et al., Reference Stimson, MacKuen and Erikson1995; Ezrow et al., Reference Ezrow, De Vries, Steenbergen and Edwards2011). Additionally, parties emphasize policy issues in their manifestos that are salient for voters (see e.g. Spoon and Klüver, Reference Spoon and Klüver2014; Klüver and Sagarzazu, Reference Klüver and Sagarzazu2016). In this contribution, we will focus on the second strategy – the varying degree of issue emphasis, that is saliency.

One core assumption of salience theory is that parties strategically choose to manipulate the saliency of issues (Budge and Farlie, Reference Budge and Farlie1983; Klingemann et al., Reference Klingemann, Hofferbert and Budge1994; Budge et al., Reference Budge, Klingemann, Volkens, Bara and Tanenbaum2001). Parties can ‘own’ issues (Petrocik, Reference Petrocik1996) – due to their prior performance and history (Seeberg, Reference Seeberg2017) – or they might promote issues that have been previously ignored (De Vries and Hobolt Reference De Vries and Hobolt2020). We follow this line of research and argue that the emphasis put on EU issues in national parties’ election manifestos will co-vary with several contextual and political variables.

EU issues have undoubtedly gained importance in national election campaigns in recent decades, even though the extent to which these issues are politicized is still contested (see De Vries, Reference De Vries2007; Kriesi, Reference Kriesi2007; Green-Pedersen, Reference Green-Pedersen2012, Reference Green-Pedersen2019; Hutter and Grande, Reference Hutter and Grande2014; Hoeglinger, Reference Hoeglinger2016; Gross and Schäfer, Reference Gross, Schäfer, Bukow and Jun2020). On the one hand, this comes as no surprise since national elections are still mainly about domestic issues. On the other hand, this raises the question of why we would even expect parties to address EU issues in their national election manifestos in the first place. National politics in European countries is more and more intertwined with EU issues. EU directives, EU national and regional transfer of money as well as EU crises have an impact on domestic politics and citizens’ everyday lives. Consequently, we would argue that national parties cannot ignore the European political sphere and must address EU issues in their manifestos.

Previous research shows that parties do emphasize EU issues to a much lesser extent than other issues in their national election manifestos. This is particularly the case for large, mainstream parties (Green-Pedersen, Reference Green-Pedersen2012, Reference Green-Pedersen2019) or dominant parties (De Vries and Hobolt, Reference De Vries and Hobolt2020). The main explanatory factor for why parties put more emphasis on EU issues in their national election manifestos is the salience of EU issues at the party system level (see e.g. Steenbergen and Scott, Reference Steenbergen, Scott, Marks and Steenbergen2004; Netjes and Binnema, Reference Netjes and Binnema2007). Yet these findings based on manifesto data only explain differences in national parties’ EU issue emphasis between countries, not within a country. We therefore add an additional argument that sheds light on the differences in national parties’ EU issue emphasis both between countries and within a country: the temporal proximity of national parliamentary elections and EP elections within the national electoral cycle.

Temporal proximity of national and EP elections and parties’ EU issue emphasis in national election manifestos

The rationale for our temporal proximity argument has two aspects. First, even though manifestos tend to get longer over time (see e.g. Dolezal et al., Reference Dolezal, Ennser-Jedenastik, Müller and Winkler2012), there is a limit to how many finely detailed policy issues parties can emphasize. Parties must decide which issues they want to highlight, and which pledges they want to include (also with regard to post-electoral bargaining on government formation). Second, drafting manifestos can be time-consuming for the party organization, especially in parties that have a long process from drafting the first version of the manifesto, to consulting party activists and party factions, to adopting the final manifesto in a congress (see Däubler, Reference Däubler2012; Braun et al., Reference Braun, Popa and Schmitt2019).

Since manifestos are to a large extent not re-used from one election to the other but heavily updated, parties must allocate resources for drafting manifestos for each and every new election, be it at the European, national, or sub-national level (see Dolezal et al., Reference Dolezal, Ennser-Jedenastik, Müller and Winkler2012). We argue that if two elections at different layers of a multilevel political system are close in time, then parties need to decide which issues they will focus on in their respective manifestos because the process of manifesto preparation is not happening simultaneously (Däubler, Reference Däubler2012, pp. 55–56). Regarding national parties’ considerations when drafting EP election manifestos, they ‘are expected to respond to the problem agenda in their immediate national environment, not least because it is on the national level that they are rewarded or punished for their campaign communication’ (Braun et al., Reference Braun, Popa and Schmitt2019, p. 802). Hence, national parties emphasize many national issues in their EP election manifestos. However, we argue that this argument could also be flipped around: if EP and national elections happen close to each other, national parties must also consider European issues in their national election manifestos.

Nevertheless, parties still focus much more on national than on European issues in both their national and EP election manifestos (see e.g. Braun and Schmitt, Reference Braun and Schmitt2020) because they mainly mobilize their voters in first-order election campaigns (Weber, Reference Weber2007) and, therefore, also focus more on first-order than on second-order election issues to maximize their electoral support. Yet, although EP elections are still considered ‘second-order’ elections (Reif and Schmitt, Reference Reif and Schmitt1980; Corbett, Reference Corbett2014; Däubler et al., Reference Däubler, Chiru and Hermansen2022) – and voters continue to use them as an opportunity to punish domestic governments, engage in protest voting, vote sincerely for small parties, or experiment with new ones (Reif, Reference Reif1984; Weber, Reference Weber2007) – the European content of these elections has increased in recent decades compared to the 1980s and 1990s (Steenbergen and Scott, Reference Steenbergen, Scott, Marks and Steenbergen2004; Senninger and Wagner, Reference Senninger and Wagner2015). With regard to national parties’ position-taking on and emphasis of EU issues, parties actually try to ‘demobilize’ voters in domestic election campaigns (Weber, Reference Weber2007, p. 523 emphasis in original). Due to their internal conflicts about European integration issues (see e.g. Hix and Lord, Reference Hix and Lord1997; Lynch and Whitaker, Reference Lynch and Whitaker2013), or because they benefit from not politicizing EU integration, parties intentionally apply a strategy of de-emphasizing such issues in domestic election campaigns when national and EP elections are close in time (Weber, Reference Weber2007, pp. 523–524). Weber’s findings on voters’ perceptual agreement regarding parties’ positions on EU integration (party-level aggregates of individual-level survey data) seem indeed to indicate that the ‘logic of demobilization [on the EU integration dimension] is most effective in the run-up to a first-order election’ (Weber, Reference Weber2007, p. 524).

The varying impact of the temporal proximity of first-order (national) and second-order (EP) elections is a recurrent topic in the literature on voter behaviour and electoral cycles in multilevel systems, because EP elections ‘take place at different moments of their respective national (first-order) electoral cycles’ (Reif, Reference Reif1984, p. 248; see also Weber and Franklin, Reference Weber and Franklin2018, p. 833). This literature and its empirical insights, however, primarily deals with national parties’ strategies during election campaigns, whereas our argument is based on national parties’ strategies for writing manifestos. We acknowledge that national parties may have incentives to decrease their EU issue emphasis in their national election campaigns because voters mainly care about domestic issues in these first-order elections; however, due to the increase in the intertwining of politics at the national and supranational level, and because national government parties will be represented in the various EU institutions, we argue that national parties must address EU issues in their national election manifestos more than they do during their campaigns. The German parties’ election campaigns for the federal elections in 2021 provide an example of this divergence in manifesto issue emphasis and campaign strategy. Even though, on average, parties’ manifestos contained around three percent of content referring positively to European topics – and this value is even higher when factoring in the Alternative for Germany’s additional six percent of manifesto content that refers to European topics negatively (see Burst et al., Reference Burst, Lehmann, Regel and Zehnter2021, pp. 5–6) – European issues did not play a role during the electoral campaigns (see Faas and Klingelhöfer, Reference Faas and Klingelhöfer2022).

The closer national elections are to EP elections, the more parties are inclined to concentrate more on European issues in their national election manifestos. When an EP election is on the horizon, the media will increase its coverage of parties’ positioning on and emphasis of EU issues in their national election manifestos, and parties do want to shape the public agenda with their electoral programme (on this manifesto-media link see Merz, Reference Merz2017a,b).Footnote ²

The close succession of national and EP elections can make EU issues more salient to voters because of priming by media, opinion leaders, or political elites (Rauh, Reference Rauh2015; Schulte-Cloos, Reference Schulte-Cloos2018). This priming often takes the form of emphasizing the national executive’s EU-related responsibilities and powers. Since member state governments decide on EU constitutive issues, both the media and more informed voters could reasonably expect to learn of parties’ plans for the overall pace of integration, enlargement, and other EU polity-related issues from national election manifestos. This increased saliency of the EU for voters, facilitated by increased media coverage of EU issues (Hutter and Kriesi, Reference Hutter and Kriesi2019), could also alter parties’ calculations of the costs of ignoring EU issues in national contests that happen close to EP elections. The plausibility of such adaptation is strengthened by the finding that temporal proximity of national and European elections leads to an increase in parties’ responsiveness to citizens’ issue priorities as indicated by parties’ issue emphasis in manifestos (Spoon and Klüver, Reference Spoon and Klüver2014).

Temporal proximity could also affect EU saliency through a different mechanism. Unlike other issues, such as health care or the environment, the party ownership of EU-related issues is contested in several member states (Guinaudeau and Persico, Reference Guinaudeau and Persico2013). When an EP election takes place in close succession to the national election, all parties with a stake in EU issue ownership would anticipate a higher level of EU attention from competitors and would be incentivized to reassert their ownership claims via increased EU salience in their manifestos. This could spill over even for parties which do not compete over EU issue ownership. As Seeberg (Reference Seeberg2022, p. 291) has shown: ‘high and consistent attention from all rival parties and proximity to an election’ makes parties engage with their rivals’ agendas. While his findings are restricted to one type of party communication (press releases) the same logic should apply to manifestos.

Therefore, we expect parties to increase the space assigned to EU issues in national election manifestos when an EP election is imminent. The latter point is crucial for our argument. We expect parties to draft their manifestos mostly in a present- and future-oriented manner, meaning that they will include the time span to the next EP election in their considerations, but not previously held EP elections. The general idea that manifestos are forward looking has recently been corroborated empirically by an analysis of the temporal focus of national parties’ election manifestos in various countries, showing that, on average, less than ten percent of a manifesto focuses on the past (see Müller, Reference Müller2022). Thus, our first hypothesis reads as follows:

Euroscepticism and the temporal proximity of national and EP elections

We also argue that the temporal proximity of European and national elections contributes to the reduction of differences between Eurosceptic and non-Eurosceptic parties’ EU issue emphasis. Eurosceptic parties should emphasize EU issues more than non-Eurosceptic parties at all types of elections because they want to signal that they provide a political alternative for voters unsatisfied with the consensus of mainstream parties in favour of further European integration (Hoeglinger, Reference Hoeglinger2016). It is in their direct interest to ‘keep the [EU integration] issue alive and on the political agenda’ (Steenbergen and Scott, Reference Steenbergen, Scott, Marks and Steenbergen2004, p. 189). Since most Eurosceptic parties are not ‘traditionally successful parties’, they also ‘benefit from emphasizing their relatively distinctive, and potentially, extreme issue positions’ (Basu, Reference Basu2020, p. 456) on the EU. This strategic choice also allows them to blur their positions on other policy dimensions, on which their stances might be less popular, or on which they might be perceived as less competent than other parties. Even though second-order elections, such as sub-national or supranational elections, are a good opportunity for Eurosceptic parties to attract voters (Gross and Jankowski, Reference Gross and Jankowski2020; Schulte-Cloos, Reference Schulte-Cloos2018), the mere fact of ‘being Eurosceptic’ is a brand claim for such parties which they should play out at all political levels on which they compete. Therefore, at the national level as well, Eurosceptic parties should emphasize EU issues in their manifestos more than other parties (see also Senninger and Wagner, Reference Senninger and Wagner2015).Footnote ³ This assumption is complemented by the notion that non-Eurosceptic, mainstream parties do not want to put EU issues on the political agenda and de-emphasize these issues as much as they can (see e.g. De Vries and Hobolt, Reference De Vries and Hobolt2020; Green-Pedersen, Reference Green-Pedersen2012, Reference Green-Pedersen2019).

Yet when national elections take place right before EP elections, we would expect these differences between Eurosceptic and non-Eurosceptic parties to be diminished. Eurosceptic parties know that their electoral gains in national elections are larger, the closer in time national elections occur to EP elections (Schulte-Cloos, Reference Schulte-Cloos2018; also see Somer-Topcu and Zar, Reference Somer-Topcu and Zar2014). This is explained by scholars with reference to the increased saliency of EU issues and the large visibility enjoyed by such challengers in the context of the proximity of elections. However, we argue that Eurosceptic parties will always be ‘Eurosceptic’ in their manifesto content, just as green parties will always be ‘green’, because it is their essential feature which distinguishes them from their competitors (cf. De Vries and Hobolt, Reference De Vries and Hobolt2020; Green-Pedersen, Reference Green-Pedersen2019). Therefore, the temporal proximity of national and EP elections should not matter for Eurosceptic parties’ manifestos drafting strategies as much as it does for non-Eurosceptic parties.

Consequently, we argue that the temporal proximity of national and EP elections matters more for non-Eurosceptic than for Eurosceptic parties’ issue emphasis strategies. Non-Eurosceptic parties will put more emphasis on EU issues if national elections are close in time to the next EP elections, but they will display lower levels of EU issue emphasis if the next EP elections are still far away. We expect the differences between Eurosceptic and non-Eurosceptic parties regarding their varying degrees of EU issue emphasis in national election manifestos to be more pronounced the further away national elections are from the next EP elections. Thus, our second hypothesis is as follows:

Research design and data

To evaluate the hypothesized effects of the temporal proximity of national and EP elections on parties’ emphasis of EU issues in national election manifestos, we rely on the Manifesto Project Database (version 2019b) (Volkens et al., Reference Volkens, Krause, Lehmann, Matthieß, Merz, Regel and Weßels2019). In a first step, we restricted this dataset to only include manifestos from countries that were EU member states at the time of the national elections. In a second step, we deleted all national manifestos that are coded with missing values in the EU-related categories per108 and per110. In a third step, we excluded all party lists running for national elections which comprised more than one party, that is, alliances and coalitions. In a fourth step, we excluded parties which did not gain parliamentary representation in the national elections. The Manifesto Project comprises many parties that did not gain parliamentary representation following national elections and we concentrate on relevant parties that had a realistic chance of crossing the threshold of parliamentary representation.Footnote ⁴ Lastly, we dropped all observations when no EP election took place between two national elections (see Table S1 in the supplemental material). Overall, our dataset comprises 956 observations from 340 parties running for national elections in 27 countries between 1979 and 2019.

Operationalization of the dependent, independent, and control variables

The dependent variable is the salience of EU issues (EU issue emphasis) in party manifestos drafted for national elections. Using the aforementioned data set from the Manifesto Project (Budge et al., Reference Budge, Klingemann, Volkens, Bara and Tanenbaum2001; Klingemann et al., Reference Klingemann, Volkens, Bara, Budge and McDonald2006), which developed a classification scheme with 56 categories via human coding of national election manifestos, we compute a party’s emphasis of EU issues as the percentage of quasi-sentences in the manifestos by summing the two shares of ‘European Community/Union: Positive’ (per108) and ‘European Community/Union: Negative’ (per110).

To test the effect of the temporal proximity of national and the next EP elections on national parties’ EU issue emphasis, we calculate the time span between national elections and the next EP elections, relative to the length of the country-specific legislative period at the national level (cf. Reif, Reference Reif1984; Weber, Reference Weber2007). The Cycle position is calculated as the number of days between a national election in t and the next EP election (x), divided by the total number of days between the national elections in t and t + 1 (y), thus ranging from 0 to 1 (see Figure 1). Low values indicate that a national election starting the legislative period in a country is close in time to the next EP election. Following the electoral cycle literature, this operationalization has three advantages: (1) using the national electoral Cycle position takes the varying lengths of legislative periods in the countries under study into account; (2) when drafting their national election manifestos parties already know when the next EP elections will happen, and they cannot manipulate the election date because this is decided at the supranational level; (3) whereas ‘national elections occur on different cycles’, EP elections are held every five years and ‘fall at different points in these cycles’ and this allows us to analyse variations both between and within countries (Weber and Franklin, Reference Weber and Franklin2018, p. 838).

Figure 1. Illustrative example for calculating the independent variable Cycle position.

Notes: The figure displays the calculation of the independent variable Cycle position. Cycle position is calculated as the number of days between a national election in t (NPEt) and the next EP election (EPE) (x), divided by the total number of days between the national elections in t (NPEt) and t+1 (NPEt+1) (y), thus ranging from 0 to 1.

Figure 1 demonstrates why we exclude observations where no EP election took place within a national electoral cycle: otherwise, we would end up with values greater than ‘1’ if we took the national elections in t + 2 instead of the national elections in t + 1. For example, we include the Danish parliamentary elections in 1984 and 1990 in the analysis but exclude the elections in 1987 because there were no EP elections within the national electoral cycle between 1984 and 1987. Furthermore, the Belgian example highlights why Cycle position can take on the values ‘0’ and ‘1’. We deliberately set the value to ‘0’ when there are concurrent national and EP elections at the beginning of a national electoral cycle because we assume that concurrent elections force parties to emphasize EU issues in their national election manifestos, and also because the media will increase its coverage of parties’ emphasis of EU issues leading parties to try to shape the public agenda with their electoral programme (see Merz, Reference Merz2017a, b). Therefore, Cycle position takes on the value ‘0’ for the observations of the Belgian parliamentary elections in 1999, which were held concurrently with EP elections on 13 June 1999. If, however, an EP election is held concurrently at the end of a national electoral cycle, such as has been the case for the Belgian electoral cycle starting with the parliamentary elections in 1995, Cycle position takes on the value ‘1’ because the national electoral cycle is exactly as long as the time span between the Belgian parliamentary elections on 21 May 1995 and both the parliamentary and the EP elections on 13 June 1999.

To test the conditional effect of Cycle position on the varying strategies of Eurosceptic and non-Eurosceptic parties, we use a dummy variable: being a Eurosceptic party is measured as a binary variable (value 1) if a party covered in the Manifesto Project Database is considered as Eurosceptic by The PopuList (Rooduijn et al., Reference Rooduijn, van Kessel, Froio, Pirro, de Lange, Halikiopoulou, Lewis, Mudde and Taggart2019) and by De Vries (Reference De Vries2018) and Szczerbiak and Taggart (Reference Szczerbiak and Taggart2008).

We control for several factors that are regularly considered to influence parties’ issue emphasis strategies when drafting their manifestos: first, we control for a party’s status of being in government or in opposition when drafting a national election manifesto (Government party). One might expect a positive effect of Government party on EU issue emphasis because (a) government parties are directly involved in EU politics via the various EU institutions (Gross, Reference Gross2021), and (b) government parties tend to write longer manifestos and include more issues than opposition parties, not least because ‘[g]overnment parties simply have more resources available’ (Dolezal et al., Reference Dolezal, Ennser-Jedenastik, Müller and Winkler2012, p. 872). The data on a party’s status of being in government or in opposition are obtained from the ParlGov database (Döring and Manow, Reference Döring and Manow2020).

Second, we control for whether the manifesto was issued by a party campaigning in Central and Eastern Europe (CEE party) or in Southern Europe (South European party). We expect CEE parties to emphasize EU issues in their national election manifestos less than parties from other European regions (see Whitefield and Rohrschneider, Reference Whitefield and Rohrschneider2009; Haughton, Reference Haughton2014). Election campaigns and manifestos in CEE are usually dominated by bread-and-butter issues and anti-corruption appeals, while EU themes remain marginal. Bulgaria, Croatia, Czechia, Estonia, Hungary, Latvia, Lithuania, Poland, Romania, Slovenia and Slovakia are coded as CEE countries.Footnote ⁵ Furthermore, the literature suggests different patterns and levels of EU politicization in Southern Europe compared to Northern Europe and CEE (Kriesi, Reference Kriesi2016; Otjes and Katsanidou, Reference Otjes and Katsanidou2017; Hutter and Kriesi, Reference Hutter and Kriesi2019). Cyprus, Greece, Italy, Portugal and Spain are coded as Southern European countries.

Third, we control for concurrent parliamentary and EP elections (Concurrent elections), which is a hard test for our data because such elections impact the calculation of both the lower and upper limit of the independent variable Cycle position. Fourth, we control for two ‘critical junctures’ with regard to party competition over European integration: the Maastricht treaty, marking the largest authority transfer to the supranational level, and the ‘Euro crisis’ from 2010 on, which further intensified party competition over European integration and EU matters (see Schäfer et al., Reference Schäfer, Popa, Braun and Schmitt2021). Pre-Maastricht receives the value ‘1’ for national elections before 1994 (we expect a negative effect on EU issue emphasis), whereas Euro crisis displays the value ‘1’ for all national elections from 2010 on (we expect a positive effect on EU issue emphasis).

Sixth, we include a party’s general left-right position (MARPOR’s Rile) to control for the empirical finding that it is mainly parties from the far-left and the far-right that are opposed to further European integration (see e.g. Hooghe et al., Reference Hooghe, Marks and Wilson2002), which might also be reflected in greater EU issue emphasis. Lastly, we use a variable to account for a party’s EU issue emphasis in its previous manifesto (EU issue emphasis _t−1 ). Some parties may have a stronger tendency to emphasize EU issues in their manifestos, which may not be captured by other variables in our models (e.g. preferences by party activists or political elites, see Spoon Reference Spoon2012).