Are social democratic parties insider parties? Electoral strategies of social democratic parties in Western Europe in the age of dualization

Abstract

The article analyses the electoral strategies of social democratic parties in the context of increased and new inequalities among its electorate. The literature on the politics of dualization argues that social democratic parties are essentially insider parties. By contrast, I argue that they also target outsiders by promoting policies that facilitate the integration of outsiders into the labour market and enhance their social protection. There are two reasons that make outsiders electorally attractive for social democratic parties: first, the share of outsiders has been growing strongly, while the share of insiders has been decreasing. As a consequence, social democratic parties cannot hope to win elections solely on the basis of their insider constituency. Second, outsiders are not as politically inactive as the original insider–outsider literature suggests which makes them electorally more attractive to social democratic parties. To analyse the electoral strategies of social democratic parties, I rely on original data on party statements as reported in newspapers during election campaigns between 2007 and 2010 in four continental European countries (Spain, France, Germany and the Netherlands). The evidence clearly suggests that social democratic parties are no insider parties but propose policies in the interests of outsiders.

Appendices

Appendix 1

See Fig. 3.

Fig. 3

Labour market situation in France, Germany, the Netherlands and Spain in the election year

Appendix 2: Coding scheme

I code the policy statements according to their reference to four electoral groups: lower-skilled outsiders, lower-skilled insiders, higher-skilled outsiders and higher-skilled insiders. Each statement can either improve the situation of a specific electoral group or reduce an existing social right (or make no statement about it). Hence, the statements have a direction, which is measured by a three-point scale: 0 means a neutral, + 1 a positive and − 1 a negative relationship between the political actor and the electoral group.

Obviously, policies have effects on more than one group, and for some policies one could argue that they are beneficial for all or most of all individuals. However, most statements concern some groups more immediately than others and, from a rational-choice perspective, should matter more for them than for others. Hence, I code the distributive effects only for those group(s) whose situation is directly affected. Nevertheless, a statement can change the situation of more than one group directly. For example, redistribution employment is in the interest of both higher- and lower-skilled outsiders, while lowering the retirement age is beneficial for lower-skilled insiders and lower-skilled outsiders. It follows that a statement can generate more than one observation. Also, everyone wants the fight against unemployment to be won (it is a so-called valence issue). I do not include these valence issues.

In addition, we should differentiate between positive and negative activation. Incentives to work can be increased in a punitive way, for example, by reducing social benefits or pushing unemployed to accept any jobs at the risk of remaining working poor or losing their job again. Despite having an activating effect, these punitive workfare policies are not in the interests of outsiders, so I assign − 1 for statements that aim to push outsiders to work by reducing social benefits. I do not code statements referring to the minimum wage for its ambivalent effect on low-skilled outsiders: minimum wage protects low-skilled outsiders from wages below the poverty level. At the same time, minimum wages can be seen as a way to protection low-skilled insiders from competition from low-skilled outsiders (Table 4).

Table 4 Coding scheme for the classification of statements into welfare directions and interests of distributive groups

As database for the content analysis serve the quality newspaper and the tabloid with the highest print-runs. Because no genuine tabloid exists in the Netherlands, a widespread tabloid-style newspaper was selected (see Kriesi et al. 2008). A further advantage of the data is that it relies on the media coverage of national election campaigns. For a long time, party positions have been coded mainly from party manifestos (Klingemann et al. 1994; Benoit and Laver 2006). In the time of media democracy, however, most voters obtain their information through the media. Moreover, voters’ perceptions about the relevance of an issue are shaped by the media (Petrocik et al. 2003). Thus, a recent strand of research on party positions has argued that parties’ manifestos have become less relevant for transporting information of party positions and we should focus on media coverage to assess party positions (Kleinnijenhuis and Pennings 2001; Kriesi et al. 2008; Bornschier 2010).

Table 5 shows the selected newspapers, the election year and the number of statements derived from the coding.

Table 5 Description of selected years, newspapers and number of statements