Parties’ Changing Political Projects and Workers’ Political Attitudes

This chapter reviews possible explanations of workers’ new voting patterns. It focuses on the contrast between changes in the way parties mobilise (or do not mobilise) workers’ social class and changes in workers’ political attitudes. In a first step, it discusses significant transformations in social democracy’s political project and attempts by new political parties to mobilise the votes of workers. In a second step, it examines the political attitudes of workers on issues of redistribution and immigration. Workers display a particular combination of political attitudes: they support redistribution and at the same time they are opposed to immigration. Importantly, the chapter shows a strong continuity over time of this particular combination of attitudes. This suggests, therefore, that a transformation in the political supply of parties is more decisive to understand new voting patterns than changes in workers’ demands.


Bringing Parties Back In
The extent to which people vote on the basis of their social class (and hence the extent to which workers support parties of the left) depends on mobilisation by political parties themselves, as was outlined in Chapter 2. In this chapter, I will in a first step review major trends in parties' political projects and their attempts to mobilise workers. In a second step, I will turn to an examination of worker's political attitudes. The logic of the chapter is to contrast changes in the way parties mobilise (or do not mobilise) the social class of workers with changes in workers' political attitudes. I use a two-dimensional conceptualisation of the political space (e.g. Kitschelt 1994;Kriesi et al. 2006Kriesi et al. , 2008, in which a socio-economic dimension (relating to state intervention in the economy and the distribution of resources) is complemented by a socio-cultural dimension (relating to the definition of the community and moral issues).
In the decades following the Second World War, the ambition to reduce the disadvantaged position of workers stood at the core of the political project of social democratic parties. Of course, social democratic parties were active in different political and institutional contexts that shaped their policy positions and strategic decisions. However, it is possible to identify the contours of a relatively coherent political project across European countries, the concrete translation of which into policies could vary from one context to another. One central pillar consisted in developing a universal welfare state-best achieved in Scandinavian countries-that would reduce the dependence of workers on market forces or, put differently, reduce the commodification of workers (Esping-Andersen 1990). Another central pillar concerned the development of collective bargaining, which would allow the (wage) bargaining position of workers relative to that of employers in the corporate arena to be enhanced. Trade unions were the decisive actors in this arena, not parties, but close ties between unions and social democratic parties (Ebbinghaus 1995;Allern and Bale 2017) ensured the coherence of this political project.
Since the 1980s and 1990s, the development of 'third way' ideas and policies marked a crucial change in this political project. Instead of seeking a reformist way to enhance the power of labour relative to the power of capital, third-way social democracy was more accommodative of global market forces. It believed that 'old-style social democracy' could not face various challenges such as those of globalisation and the influence of free market ideas (Giddens 1998). Hence, the social democratic ambition no longer consisted in developing counter-forces to the power of capital. The ambition was instead to find a new path between social democracy and neoliberalism, thus definitively abandoning the long-term goal of a transition to socialism-Giddens evokes the 'death of socialism ' (1998: 3). The 'social investment state' was at the core of this new programme. Priority was given to investment in human capital and less to the compensation of income loss, thus implying a shift from redistribution to the 'redistribution of possibilities' (Giddens 1998: 99-100).
The turn towards third-way social democracy was personified the most by the figures of Tony Blair in Great Britain ('New Labour') and Gerhard Schröder in Germany ('Neue Mitte'). As prime ministers of their respective countries at the turn of the 2000s, they implemented policies that matched this new ideational framework (e.g. Arndt 2013;Merkel et al. 2008;Nachtwey 2009). They also attempted to theorise their new approach in a document (see Blair and Schröder 1999) that, next to Giddens' book, is considered a major text of the third way (Nachtwey 2009: 10). However, transformations in the core ideas of social democracy largely touched the social democratic party family over the entire continent (e.g. Escalona 2018;Green-Pedersen et al. 2001;Huo 2009;Keman 2011). It came as no surprise that many social democratic parties endorsed and implemented austerity in the context of the 2008 economic crisis (Bremer 2018;Bremer and McDaniel 2019;Escalona and Viera 2014).
It would be wrong to understand social democracy's political project as a purely socio-economic project. Socialist ideas have also involved an important component of cultural emancipation, where freedom constitutes a key value. The struggle for full democratisation of political life was at the core of the agenda of socialist parties when they were created at the end of the nineteenth century. They were also the first parties to work for women's voting rights. In countries with a significant church-state cleavage, they generally adopted a clear secular position on issues relating to the role of the church in society. There is no doubt that social democracy has always adopted a clearer position in favour of cultural liberalism as compared to Conservatives, Christian Democrats or even mainstream liberal parties. However, these issues played a somewhat secondary role in the social democratic political project, since the ambition was first of all to reduce socio-economic inequalities.
The transformation of social democracy's project towards more accommodation to market forces gave a new importance to 'cultural' issues. If social democracy was less willing or able to deliver results on the socioeconomic dimension, it became increasingly critical to achieve results on the socio-cultural dimension-a dimension that is generally of increasing salience in political competition (Kitschelt 1994;Kriesi et al. 2006Kriesi et al. , 2008. On the basis of detailed case studies on four West European countries (Germany, Great Britain, France and Sweden), Escalona (2018: 409) demonstrates that new cultural issues (such as the promotion of women's rights and the elimination of discrimination against minorities) have played a crucial role in the transformation of social democracy's political project since the mid-1970s. He observes a common trend towards an increasing salience of these issues in the manifestos and policies of social democratic parties in the post-1968 context-and not so much a change in their positions. If one follows Escalona's analysis, one cannot really capture social democracy's transformation by focusing solely on the thirdway turn. With the same logic, Rennwald and Evans (2014) underline the importance of new cultural issues in the transformation of the political offer of some social democratic parties. Martin (2018: 169) suggests even taking a step further in the relation between the socio-economic and socio-cultural components of social democracy's new project. He argues that the promotion of cultural liberalism had the function of 'compensating' for a lack of social democratic achievements while at the same time allowing social democracy to respond to the aspirations of the new middle classes.
Overall, the third-way turn of social democracy and the rise of new issues on its political agenda weakened the possibility of a class-based mobilisation of workers. One must not forget that social class is only one possible source of identity, alongside gender, religion and ethnicity (see Heerma van Voss and van der Linden 2002). Class was always in competition with other cleavages for votes (Lipset and Rokkan 1967). A worker was therefore particularly likely to be mobilised on the basis of 'cultural' identities if he/she was not mobilised on the basis of social class. This was the argument developed by Przeworski and Sprague (1986), who emphasised the importance of the mobilisation of class relative to other identities. In their view, the weakening of a class-based appeal would leave room open for competing appeals based on religious, ethnic or regional identity (Przeworski and Sprague 1986: 45-46). As a result, workers may not cast a class-based vote (and support social democratic parties) but instead a vote based on religious or ethnic loyalties.
During the post-war decades in Europe, religion was in strong competition with social class for votes (Knutsen 2004;Rose and Urwin 1969). Christian Democratic trade unions and parties aimed to mobilise workers on the basis of religious loyalty (Kalyvas and van Kersbergen 2010). In contrast, no major party explicitly mobilised workers on an ethnic or nationalist basis in this period. Extreme right-wing parties, which were often the direct heirs of the fascist regimes, remained at the margins of politics for a long period after the Second World War. No party really dared to exploit anti-migrant resentment. It is as if there was an agreement among the political elite. This changed in the 1970s and 1980s, when extreme right-wing parties turned to an anti-immigration strategy and tried to demonstrate a clear distance from the fascist legacy (Carter 2005;Ignazi 1992;Mudde 2007).
One can therefore conceptualise today's competition for the workers' vote in the following way. Workers can be mobilised on the basis of their social class by left-wing parties to improve redistribution (on this issue social democratic parties are in competition with radical left parties) or they can be mobilised by far right parties on the basis of their nationality to restrict immigration. Clearly, parties' political projects have changed in recent decades and this has massively affected the political choices available to workers (see Evans and De Graaf 2013). However, on the 'demand-side' it is not clear whether the political attitudes of workers have also undergone important changes. In the next sections, I therefore analyse the distribution of preferences by social class.

Between Pro-redistributive and Anti-immigration Worker Preferences in the 1970s
One can expect that social classes diverge on the extent to which they prefer redistribution and state intervention in the economy. Workers are more in favour of redistribution, because this is a way of reducing their more disadvantaged position in the labour market (see Evans and De Graaf 2013). There is also a clear connexion between social class and attitudes towards immigration. Workers, and particularly lower-skilled workers, are likely to be the most affected by an increase in labour supply. Hence, they should be more likely to be in favour of restricting an increase in the supply of labour. Workers are therefore expected to combine pro-redistributive preferences with anti-immigration preferences.
In order to analyse the preferences of workers and social classes, I select one question on redistribution and one question on immigration. I first present findings for the mid-1970s 1 and then for the early 2010s. I assess whether the configuration of attitudes by social class is relatively stable over time. In the Political Action Survey conducted in the mid-1970s, respondents were asked how much responsibility the government has to reduce wealth differences between people. 2 They had to choose an answer on a four-point Likert scale ranging from 'an essential responsibility' to 'no responsibility at all'. I standardise the answers to the range from 0 to 1, where 1 means the strongest government responsibility. Figure 5.1 shows the preferences of workers relative to a few other classes on what I name the 'pro-redistribution scale'. I select production workers, service workers, socio-cultural professionals and managers. I also present the overall score for the total electorate (including non-voters). The choice of these groups-with diverse segments of the working and (salaried) middle classes-makes it possible to put the preferences of production workers into perspective and to get an overall view of class configurations.
Overall, the respondents were in favour of redistribution. As expected, production workers were always (slightly) more in favour of redistribution than the average respondent. As we can observe for social democratic Names of classes: Prod = production workers, Serv = service workers, Socio = socio-cultural professionals, Manag = managers voting in the 1970s, service workers were relatively close to the position of production workers. Managers were the least in favour of redistribution in comparison with the other classes. Interestingly, socio-cultural professionals were in all the countries located closer to the position of production workers than to that of managers, with the exception of Austria. However, one should notice that class differences are relatively small. The largest class differences can be found between managers and production workers in Great Britain and Switzerland: differences of 0.14 points and 0.15 points respectively on the pro-redistribution scale. Turning now to immigration, I start with a cross-national analysis and then limit the analysis to Switzerland, where the issue of migration came onto the political agenda very early. Starting with the cross-national perspective, in the Political Action Survey, respondents were asked their position on the provision of equal rights for guest workers (Gastarbeiter) 3 -the question referred to coloured immigrants in Great Britain. One should notice that the question does not tackle attitudes towards immigration per se (towards the admission of immigrants or the benefits of immigration, for example) but instead captures attitudes towards the integration of immigrants. These are indeed two different dimensions of migration preferences (Afonso 2013;Tichenor 2002). It is well possible to combine a restrictive attitude towards the admission of migrants with an allocation of equal rights to migrants (in terms of access to the labour market, to the welfare state, to citizenship, etc.), while the contrary is also possible. However, in the absence of any other question, I consider this a viable proxy. Figure 5.2 indicates that the respondents were on average relatively open to immigration (with the exception of Austria). In most of the countries, production workers were slightly more restrictive towards migration (or slightly more against the idea of according equal rights to immigrants) than the average respondent. However, the differences are not always significant. As for redistribution, service workers were located close to production workers. Managers were also located close to the position of production workers and to that of the average respondent. The largest difference between production workers and managers was 0.08 points in Germany. Socio-cultural professionals were clearly different to all the other classes. They were clearly more pro-immigration than the other classes. Switzerland offers a unique opportunity to further study attitudes towards immigration in the 1970s. The movement against 'overforeignization' (Überfremdung ) that emerged in this country in the 1960s can be considered a forerunner of anti-migrant movements in Europe (Skenderovic 2009). Since the 1960s, newly formed radical right parties mobilised on the issue of immigration, playing on resentment against guest workers, who had been massively recruited in the Swiss economy in the post-war years. These parties launched several popular initiatives against 'over-foreignization' demanding a drastic change in immigration policy and several initiatives on foreign policy (Skenderovic 2009: 68). Between 1970 and1977, no less than four anti-immigrant initiatives were submitted to the vote and provoked heated public debates. The first one launched the most intensive discussions in the public sphere. It was entitled 'Initiative against "over-foreignization"' (gegen 'Überfremdung' ), but it is sometimes known as the Schwarzenbach initiative after the name of the leader of the National Action Party, James Schwarzenbach. It demanded restriction of the number of foreign residents to 10% of the population in each canton. It was rejected in 1970 with a small majority of the votes (54%), while the later initiatives were more firmly rejected-support did not exceed 35% (Skenderovic 2009: 65-68).
The Swiss component of the Political Action Study included some specific questions, among them was one on the vote choice on the second anti-immigrant initiative in 1974 (which was accepted by only 34.2% of the electorate). A post-electoral survey related to the Swiss national election of 1971 also included the vote choice on the 1970 Schwarzenbach initiative. Only men were allowed to vote at the federal level until 1971 in Switzerland. Figure 5.3 presents the preferences of classes in these two popular votes. Higher values indicate stronger support for the initiatives.
Preferences on immigration measured using support for these popular initiatives indicate a stronger polarisation of classes than preferences on immigration measured using answers on the provision of equal rights to migrants. Production workers were now much more in favour of restrictive migration policies. In contrast, socio-cultural professionals and managers were more opposed to restrictive immigration policies (they were twice or more than twice as opposed to restrictive migration policies as production workers were). Practically, this meant that on average production workers supported the popular initiatives, while the salaried middle classes voted against the initiative on average. In 1974, the level of support for the initiative dropped, but the relative support of production workers remained relatively similar (a ratio of 1.36 in 1970 and of 1.38 in 1974).
If one isolates the social democratic electorate, it can be observed that, on average, social democratic voters gave the initiative in 1970 slightly stronger support (0.45) than the total electorate (0.41) but still rejected  it. This was in line with the party's recommendation to reject the initiative. It is possible to further differentiate within the social democratic electorate by social class. However, the number of cases becomes relatively small and one should remain cautious about the results. Among the social democratic electorate, production workers accepted the initiative (0.53), although the degree of support was smaller than among all workers irrespective of their party choice (0.56). In 1974, support for the anti-immigration initiative was this time slightly lower among the social democratic electorate (0.31). However, support for the initiative was still relatively strong among social democratic production workers (0.52).
Working-class support for the initiatives was in line with analyses based on aggregate results that showed that support was strong in workingclass neighbourhoods (Gilg 1972). Moreover, the initiatives provoked heated debates in the ranks of Swiss trade unions (Steinauer and von Allmen 2000). Trade unions fought against the initiative submitted to the popular vote in 1970 (and all other anti-immigrant initiatives) but large segments of the rank and file disagreed with this position. This indicates that the anti-immigration initiatives aroused sympathy, or at least contributed to spreading confusion among organised workers. However, it is interesting to note that at this time an important share of production workers voted for the Swiss Social Democratic Party. There was therefore a unique combination of anti-immigration attitudes (and support for anti-migration initiatives) and social democratic voting in national elections. In the arena of direct democracy, the small radical right parties that launched the anti-migrant initiatives were major players at the time but they remained fringe forces in the electoral arena (Skenderovic 2009). They reached an electoral peak in the 1971 national election (7.5%) but then experienced losses in later elections (Skenderovic 2009: 60). It was only from the 1990s onwards that the Swiss People's Party managed to obtain strong electoral successes in national elections and definitively transformed the Swiss party system.

Continuity in Class Preferences in the 2010s
Reproducing the same exercise for the first half of the 2010s leads to similar conclusions about combinations of preferences by social class. Workers combined a pro-redistributive position with an anti-immigration position. For redistribution, I use a similar question on the role of the government in income inequality. In the ESS, respondents were asked if they agreed or disagreed with the following statement: 'the government should reduce differences in income levels' on a five-point Likert scale (from 'strongly agree' to 'strongly disagree'). I standardise the answers to the range from 0 to 1. On immigration, respondents were asked to take a position on various dimensions of immigration. I use three questions where respondents were asked to take a position on an eleven-point scale: 'the country's cultural life is generally undermined or enriched by immigrants', 'it is bad or good for the country's economy that people come to live here from other countries', and 'the country is made a worse or better place to live by people coming to live here from other countries'. I standardise the responses between 0 and 1 and calculate the average value for these three items. I recode both items so that 1 means a proredistributive position on redistribution and an anti-immigration position on immigration. I use the same selection of elections and survey rounds as in the previous chapter. was not strong polarisation among social classes, production workers and service workers were systematically more in favour of redistribution than the average respondent. Managers were consistently less in favour of redistribution than the average respondent and socio-cultural professionals were located between. Their position was generally close to that of the average respondent. In Germany and Switzerland, they even displayed a more pro-redistributive attitude than the average respondent and came close to the position of production and service workers. Hence, the same structuration of classes on redistribution attitudes as in the 1970s can be observed. On average, respondents wanted more redistribution in the 2010s than in the 1970s, but the class differences remained relatively stable. The case of Switzerland provides an interesting comparison with the 1970s. Support for redistribution among the total electorate reached exactly the same level. Support among workers and socio-cultural professionals was almost identical. This is not the case in all the countries. In Austria, production workers had become less distinct from the average citizen in their redistribution attitudes over time. In contrast, in the Netherlands, production workers had become more distinct from the average citizen.
On immigration, production and service workers were systematically more against immigration than the average respondent (see Fig. 5.5). In contrast, socio-cultural professionals and managers were consistently less against immigration than the average respondent. Socio-cultural professionals generally displayed the most favourable position on immigration. When focusing the cross-time comparison on popular votes on immigration in Switzerland, production workers had not become more antiimmigrant over time. The class structure that we could observe for the 1970s-with socio-cultural professionals and production workers occupying two distinct positions on this axis-had not changed. More systematically, using ratios to check changes in the average position on immigration, we can even conclude that class differences in immigration had become smaller over time.
For the other countries, the comparison leads to another conclusion. Relative to the average citizen, production workers had become more distinctively opposed to immigration over time. If one compares production workers to managers, there is an increasing divergence between the two classes on immigration issues. However, there is more stability if one compares production workers to socio-cultural professionals, as the latter had also become less strongly pro-immigration over time.
This divergence in cross-time comparisons between Switzerland and the other countries is not entirely surprising. The survey questions are different, as is the degree of politicisation of migration issues. The popular votes in Switzerland were the early outcome of agenda-setting strategies by radical right parties. In the other countries, migration was less an issue on the political agenda at the time, although one should not forget early attempts to deploy an anti-immigration discourse in several countries, such as in Great Britain around the Conservative shadow cabinet member Enoch Powell (see Schofield 2012).
Finally, it is possible to gain a finer view by examining different subgroups (see Table 5.1). Restricting the sample to those who voted for a social democratic/socialist/Labour Party in the last election generally decreases the opposition to immigration. The same is also true if we restrict the sample to social democratic supporters who were production workers-this is important to underline. However, within the social democratic electorate, class differences remained and production workers continued to display a relatively stronger anti-immigration position than the average social democratic voter. Again, if we consider workers who had not participated in the last election (and thus who might at some point again vote for social democracy), we observe that they were relatively more opposed to immigration than the average non-voter in all the countries-but the difference was small for the Netherlands and Switzerland. Finally, if one isolates voters for radical right parties, 4 class differences were almost absent. Production workers who voted for radical right parties were very much aligned with the position of the average radical right voter, with the exception of France.

Summing Up
In a first step, this chapter has reviewed major trends in social democracy's political project. In recent decades, one can observe both an increasing accommodation to market forces by social democratic parties and an increasing salience of new issues on their political agenda. By substantially revising their original political project, social democratic parties therefore contributed to a demobilisation of workers on the socio-economic dimension of the political space. At the same time, social democratic parties faced a more difficult context to develop class-based mobilisation. Populist radical right parties started to mobilise workers for a political project with a restriction of immigration at its core. In a second step, the chapter has analysed attitudes among social classes on redistribution and immigration issues. It has demonstrated that workers display a particular combination of political attitudes: on the one hand, they support redistribution more than average; on the other hand, they are more than average opponents of immigration. Most importantly, the chapter has demonstrated that this combination of attitudes was already present in the 1970s. This was especially the case in Switzerland, where immigration issues were present on the political agenda very early.
The contrast between changes in parties' political supply and voters' demands that has been at the centre of this chapter indicates that transformations in the former are more important than transformations in the latter. Hence, the ways in which parties propose political choices to workers are crucial to understand workers' new voting patterns. The demobilisation of class by social democracy went along with a new mobilisation of nationality by populist radical right parties. Hence, these two changes constitute two sides of the same coin. The result is that social class became less relevant for voting behaviour, but more relevant for abstention.

Notes
1. I restrict the analysis to four countries for this period because of a lack of data for France. The results for this country in the 1980s indicate a similar configuration of preferences. 2. In the survey, respondents are confronted with political issues and then for each political issue asked the importance of the problem, the government's responsibility for dealing with the problem and the performance of the government in handling the problem. One political issue concerns wealth differences. 3. I select the question on the importance of the problem of according the same rights to migrants. Using the question on government responsibility leads to similar conclusions about class differences. 4. One should be cautious in the interpretation of the findings for Great Britain and Germany since the number of radical right voters in the survey is particularly low.