What Explains the Electoral Crisis of Social Democracy? A Systematic Review of the Literature

Abstract While there are some signs of revitalization, social democracy has witnessed a deep electoral crisis over the last decades. The causes for the decline of social democratic parties are highly contested among researchers. This article provides a systematic review of the literature which spans several fields such as party politics, political sociology and political economy. Four kinds of explanations (sociological, materialist, ideational and institutional) are distinguished and scrutinized on the basis of empirical studies published since 2010. The findings indicate that there is not one explanation that stands out but that the electoral crisis of social democracy is a complex phenomenon with multiple causes, such as socio-structural changes, fiscal austerity and neoliberal depolarization. In addition, the findings suggest that a liberal turn on sociocultural issues does not necessarily lead to vote losses. Further research should explore more deeply how short-term and long-term factors have worked together in the process of social democratic decline.

its worst defeat since 1935 in terms of parliamentary seats in the 'Brexit elections' of 2019. One exception to the trend is the Portuguese Socialists, who, against all odds, were able to form a left-wing government in 2015, leading to remarkable electoral successes in 2019 and 2022.
While there is in light of those numbers unanimous agreement about the crisis of social democratic parties in Western Europe (see e.g. Benedetto et al. 2020;  Rennwald and Pontusson 2021;Sassoon 2014), explanations for this development differ not only among social democratic politicians and political journalists but also among researchers. Depending on the author, the blame for the electoral crisis of social democracy is put on social democrats' '"tragic" competitive situation' (Kitschelt 1994: 34) in post-industrial societies, 'the collapse of the postwar economic boom' (Lavelle 2008: 1), 'the global freedom of capital' (Gray 1996: 17), 'the economically neoliberal and institutionally conservative European framework' (Escalona and Vieira 2014: 23), 'the Left's defensive strategy' (Sassoon 2014: xxi-xxii), the 'cartelization' of social democratic parties (Hopkin and Blyth 2019), their third ways' fatal failure to give a 'meaningful voice to insecure, financially anxious, and downwardly mobile people' (Mudge 2018: xv), or their lack of 'attractive messages about how to solve contemporary problems' as well as an 'attractive vision of the future' (Berman 2018: 4). In other words, social democrats' amazement at their own decline is mirrored by researchers' disagreement on the causes of this decline.
Given the numerous, in parts conflicting, explanations, the first goal of this review article is to organize the literature in a sensible way. For this purpose, four kinds of explanation are distinguished: sociological, materialist, ideational and institutional. The second part of the article tests if the derived hypotheses are supported by empirical evidence. This test is based on 51 empirical studies published since 2010. The findings indicate that there is not one explanation that stands out but that the electoral crisis of social democracy is a complex phenomenon which has several causes, such as socio-structural changes, fiscal austerity and neoliberal depolarization.
The crisis of social democracy: four explanations The authors cited above offer very different explanations for the crisis of social democracy, which range from structural changes in economy and society to organizational deficits and strategic errors on the part of social democrats. To systematize the literature, I mirror the work by Fernando Casal Bértoa and José Rama (2020), who distinguish sociological, economic and institutional factors to explain the rise of anti-establishment parties. Adding ideational factors as a fourth category, I propose the following types of explanation for the electoral decline of social democracy: sociological explanations focus on socio-structural changes detrimental to social democrats, confronting them with painful electoral dilemmas. Materialist explanations emphasize economic restrictions that make it difficult for social democratic parties to pursue the policies favourable to their constituencies. Ideational explanations blame the crisis on the adoption of new ideas by party elites and the resulting programmatic shifts. Finally, institutional explanations focus on organizational changes that have contributed to the alienation of voters from social democratic parties.
Sociological: the crisis as a result of a changing social structure The sociological approach is based on the assumption that parties' prospects of success are mainly determined by the given social structure and related voter preferences. Accordingly, the crisis of social democracy results first and foremost from socio-structural changes which have altered party competition to the disadvantage of social democrats. The literature highlights three forms of electoral dilemmas: (1) the well-known but sharpening 'electoral dilemma of socialism'; (2) the dilemma of facing competing demands of fragmented occupational groups in postindustrial societies; and (3) the so-called 'insider-outsider dilemma' originating from opposing interests of different labour market groups.
In its classical form, this explanatory approach points to the shrinking of the working class, the traditional core constituency of social democratic parties, as the main driver of electoral decline (e.g. Dahrendorf 1987;Hobsbawm 1978). The seminal version of this argument is provided by Adam Przeworski and John Sprague (1986). Accordingly, there is a general electoral trade-off for social democratic parties between the working class and the middle class, resulting in the so-called 'dilemma of electoral socialism'. In their efforts to broaden their base by attracting middle-class voters, social democrats 'erode exactly that ideology which is the source of their strength among voters' (Przeworski and Sprague 1986: 55). In times of a shrinking working class, social democrats are more or less forced to turn towards middle-class voters at the cost of losing workers' support (Gingrich and Häusermann 2015; for a nuanced account see Rennwald 2020).
The most prominent variant of the sociological approach can be traced back to Herbert Kitschelt's (1994) seminal work on the transformation of social democracy. Kitschelt (1994: 40-66) rejects the existence of a simple trade-off between workingclass and middle-class voters. Social democrats are rather competing for different occupational groups whose preferences can be mapped in a two-dimensional political space, as those groups not only differ on socioeconomic but also on sociocultural issues (Kitschelt 1994: 8-39;Oesch 2006). Due to the transformation from industrial to post-industrial society, the sociocultural axis is gaining importance; that is, value issues are complementing or even replacing economic issues and thus dividing the social democratic electorate (Bornschier 2010;Kitschelt 2004). The electoral dilemma of socialism is replaced by different forms of electoral dilemmas: Social democrats here face a double trade-off. First, by moving along the vertical libertarian-authoritarian dimension, social democrats must decide whether they will rely on traditional less educated blue collar, working class voters or more on highly educated white collar employees. … Social democrats also face a second trade-off within the working class itself, contingent upon their distributive economic stance. (Kitschelt 1994: 32-33) The resulting constellation of party competition is highly unfavourable to social democratic parties. While they compete with left liberal parties for the votes of the new middle classes, surging parties of the radical right are strong challengers when it comes to the votes of the traditional working class (e.g. Häusermann and Kriesi 2015;Kitschelt and Rehm 2015;Kriesi et al. 2008;Oesch and Rennwald 2018). In this regard, the adoption of a pro-welfare stance by radical right partiesthat is, their embrace of welfare chauvinismposes a particular threat to social democrats, as it helps the former to present themselves as the 'new workers' parties' and to irrevocably rob the latter of their traditional electorate (Harteveld 2016).
A final strand of the sociological approach focuses on splits in the labour market and the resulting insider-outsider dilemma (Bürgisser and Kurer 2021;Rueda 2005Rueda , 2007Schwander 2019aSchwander , 2019b. Here, too, workers are not seen as a homogeneous bloc but rather as divided into two groups: labour market insiders with secure employment and labour market outsiders, such as temporary workers, who are confronted with permanent job insecurity. 1 This divide, which is intensified by deindustrialization and globalization, leads to an insider-outsider dilemma for social democrats as labour market policies which are in the interest of insiders, such as strong employment protection for this segment of the workforce, are often detrimental to outsiders, and vice versa (Rueda 2005(Rueda , 2007. Thus, if social democrats focus on insiders, outsiders react by vote abstention or voting for populist parties. If social democrats, instead, propose policies that benefit outsiders, they risk losing insiders to other competitors such as centre-right parties. In short, whoever profits from splits in the labour market, social democrats are presumably the ones to suffer electorally from the insider-outsider dilemma (for an overview of potential electoral implications see Schwander 2019b).

Materialist: the crisis as a result of changing economic conditions
The materialist approach to the crisis of social democracy claims that changes in capitalist economies limit the policy options available to social democratic parties and complicate or even prevent the pursuit of social democratic goals, such as securing social rights and combating inequality. This explanation is based on the assumption that the electoral success of social democrats primarily depends on certain economic and social policiesthat is, policies that improve the material situation of their traditional core electorate. In this context, we can distinguish three processes which have been highlighted by political economists to explain the crisis of social democracy: (1) globalization; (2) European integration; and (3) the end of the Fordist/Keynesian growth model.
Since the early 1990s, numerous authors have identified economic globalization as a major constraint for social democratic parties (e.g. Boix 1998;Giddens 1998;Glyn 2001;Gray 1996;Pierson 2001: 64-89). Accordingly, the socioeconomic policies at the heart of social democratic success in the golden age, Keynesian demand management plus welfare state expansion, are doomed to fail under conditions of globalization: The social-democratic economic programmemost centrally, promoting full employment by stimulating investment through a policy of deficit financinghas ceased to be sustainable. … [T]he power of the international currency and bond markets is now sufficient to interdict any such expansionist policies. … It is no exaggeration to say that the global freedom of capital effectively demolishes the economic foundations of social democracy. (Gray 1996: 17) While the globalization of product markets is intensifying international competition among national economies, the deregulation of financial markets provides capital owners with a permanent exit option. As a result, traditional social democratic policy instruments such as progressive taxation and public demand management are no longer feasible due to adverse reactions by capital owners and international investors (Glyn 2001;Hall 2002;Scharpf 2000). Social democrats have been restricted to supply-side policies and, even worse, permanent austerity, thereby alienating voters dependent on the welfare state (Horn 2021). 2 For social democratic parties in Europe, the institutional arrangements resulting from European integration are perceived as an additional constraint (e.g. Bailey et al. 2014;Holmes and Roder 2012;Moschonas 2009;Ross 2011;Scharpf 2009). While globalization mainly affects the autonomy of nation states, European integration goes further by also curtailing their sovereignty through legal restrictions. First, the EU's and the European Monetary Union's (EMU) ordoliberal framework acts like a straitjacket for social democrats (Biebricher 2018: 193-224;Escalona and Vieira 2014: 22-30;Streeck 2017: 97-164). The scope for expansive fiscal policies is restricted by means of the Maastricht convergence criteria, further tightened through the European Fiscal Compact. Second, the European Central Bank with its strong commitment to price stability ensures that EMU members are deprived of the option to pursue an expansionary monetary policy and to stimulate the economy through external devaluation, constraints that proved fatal for deficit countries in the euro crisis (De Grauwe 2013; Scharpf 2011). Finally, social democrats suffer from the asymmetry between positive and negative integration (Scharpf 1999(Scharpf , 2009). While 'integration through law' driven by the European Commission and the European Court of Justice has a mainly liberalizing effect, the corrective positive integration, which is necessary to create the 'social Europe' regularly hailed by social democrats, usually fails due to conflicting national interests in the European Council. The EU has thus become a 'liberalization machine' (Streeck 2017: 103-110) which makes it more difficult, if not impossible, for social democratic parties to defend the interests of workers negatively affected by economic liberalization.
The third strand of the materialist approach, originating from regulation theory (cf. Boyer 1990), attributes the decline of social democratic parties neither to globalization nor to Europeanization but rather to a more general crisis of capitalism: the collapse of the Fordist/Keynesian growth model of the post-war decades. This argument is presented in its harshest form by Ashley Lavelle (2008), who sees the 'death of social democracy' as inevitable due to its dependence on the exceptionally high growth rates of the post-war years: 'A return to low growth in the 1970s removed the economic base of social democracy … [and] the end of the boom rendered impossible the simultaneous pursuit of policies that reduced inequality and raised living standards and which did not undermine capital accumulation' (Lavelle 2008: 1, italics in original). Wolfgang Streeck (2017) makes a similar argument by claiming that the compatibility of (social) democracy and capitalism was restricted to the exceptional situation of the post-war decades. Since then, social democratic parties in power have tried to reconcile the conflicting interests of citizens (social security) and international investors (profits). The socialization of debt in the wake of the financial crisis has shifted the already tilted balance of power even further in favour of capitalists. As a result, social democrats find themselves in a more or less hopeless position when condemned to govern.

Ideational: the crisis as the result of ideological failure
The third explanatory approach rejects purely materialistic explanations and emphasizes the impact of ideas. From this perspective, it is especially the adoption of neoliberal ideas by social democratic elites which led millions of workers to turn away from social democrats. Other authors highlight the additional embrace of cosmopolitan positions on sociocultural issues, such as European integration and migration, which resulted in a 'progressive neoliberalism' (Fraser 2017) out of touch with large parts of the working class.
According to ideational explanations, the crisis of social democracy does not originate from unfavourable circumstances but from the intellectual shortcomings of social democratic elites: 'Upon close examination … the most significant obstacles to a social democratic revival turn out to come not from structural or environmental factors … but from intellectual fallacies and a loss of will on the part of the left itself' (Berman 2006: 210-211). Regarding intellectual fallacies, several authors perceive the initially hesitant, later on sometimes euphoric, adoption of neoliberal ideas as social democrats' original sin (e.g. Berman 2006: 208-218;Blyth 2002;Hall 2003;Hay 1999;Moschonas 2002;Mouffe 2005;Mudge 2018). At the core of neoliberalism is the conviction that individualized, market-based competition is superior to other forms of organization (Mudge 2008: 705-707). In programmatic terms, this means strengthening private property rights and extending the market to all realms of societythat is, policies generally in favour of capital owners in distributive terms (Harvey 2005).
In contrast to their critics, social democrats perceived their 'third ways' as an alternative to neoliberalism (see e.g. Giddens 1998). However, empirical studies show that the programmes of social democratic parties have become increasingly permeated by the neoliberal ethos of competitiveness since the 1980s (Amable 2011;Manwaring and Holloway 2022;Mudge 2011). The result is the emergence of a 'social democratic variant of neo-liberalism' (Hall 2003), 'neoliberalized social democracy' (Mudge 2018) or 'market social democracy' (Nachtwey 2013). The argument is put forward in the most drastic way by Chantal Mouffe (2005Mouffe ( , 2019, according to whom social democracy, with its subordination to neoliberal hegemony, has lost the ability to present an attractive political alternative to the electorate. As voters realize that they have nothing to gain from this languid social democracy, they turn to more energetic populist parties on the left and the right. This trend accelerated substantially when social democratic governments clung to the neoliberal rulebook even after the global financial crisis (Blyth 2013: 132-177;Hopkin and Blyth 2019).
A second ideational explanation complements this economic turn to the right with a turn to cosmopolitanism and identity politics (Cuperus 2018; Fraser 2017; Fukuyama 2019). More precisely, this argument comes in two forms. According to the first version, social democratic elites' turn to identity politics with the focus on the emancipation of women, LGTBQ groups, migrants and other minorities resulted in the neglect of the struggle for the preservation of social rights essential for workers, regardless of their ethnic background or sexual orientation (Fukuyama 2019: 105-123;Lochocki 2018). The second version of the argument goes further in pointing out that the 'progressive neoliberalism' hailed by 'third way' social democratscombining the political fight for equal opportunities with strong individualism and a rather critical attitude towards the welfare stateis at odds with the more communitarian and welfare-friendly positions taken by most workers (Cuperus 2018;Fraser 2017). In each case, it is not so much social cleavages or economic constraints but social democrats' ideological shifts that have led to voter estrangement (Moschonas 2002: 120-122).

Institutional: the crisis as a result of organizational deficits
The final explanation focuses on party institutions to explain the crisis of social democracy. According to this approach, social democratic parties have undergone an organizational transformation from mass integration party to cartel party. This transformation, which is connected to the inadequate social representation of important voter groups within social democratic parties, has resulted in an electorally fatal convergence with mainstream parties of the right. The parallel decline of trade unions is supposed to have an additional negative impact on social democratic parties.
According to several party scholars, the mass integration parties of the past have turned into cartel parties (Blyth and Katz 2005;Mair 1995, 2009;Moschonas 2002: 123-144). Echoing Robert Michels's (2001[1915) 'iron law of oligarchy', social democratic parties have thus over time turned from organizations representing their members and voters into governing agencies: With the development of the cartel party, the goals of politics become selfreferential, professional and technocratic, and what substantive inter-party competition remains becomes focused on the efficient and effective management of the polity. Competition between cartel parties focuses less on differences in policy and more … on the provision of spectacle, image and theatre. (Mair 2013: 83) While this transformation is not without problems for other mainstream parties, it is seen as particularly harmful for social democrats for three reasons. First, social democrats' success has always relied heavily on social democrats' ability to mobilize workers and related social groups under the banner of class struggle, social justice and other egalitarian principles (Moschonas 2002;Sassoon 2014). The technocratic nature of the cartel party quenches this vital energy, regardless of whether it comes from the working class or the new middle classes. Second, the participation in the cartel of established parties means collaboration with centre-right parties as well as shrinking policy polarization. The convergence among mainstream parties, in turn, is seen as one of the main reasons for voters abandoning these parties in favour of more radical challenger parties (Berman and Kundnani 2021;Grant 2021;Meguid 2007). Finally, political economists have argued that those problems have been aggravated by the fact that the party cartel mainly operates on the basis of the neoliberal ideas outlined above (Blyth 2003;Blyth and Katz 2005;Hopkin and Blyth 2019).
A final process that has to be taken into account is the weakening of trade unions, best illustrated by declining membership numbers. While this development can partly be attributed to socio-structural changes, the liberalization of labour market institutions in which social democrats participated has contributed to union decline (Baccaro and Howell 2017;Streeck 2009). Driven by social democrats' attempts to broaden their voter base as well as unions' disappointment about social democratic policies, the traditionally close ties between social democrats and unions, once the 'Siamese twins' of the labour movement (Ebbinghaus 1995), have eroded, though there remain substantial differences across advanced democracies (Allern et al. 2007;Allern and Bale 2017). Union decline can, however, come at a substantial electoral cost for social democrats, as strong unions are supposed to boost voter turnout among the working class as well as support for social democrats (Arndt and Rennwald 2016;Flavin and Radcliff 2011;Rennwald and Pontusson 2021).
Are the four explanations supported by empirical evidence?
The four explanations are summarized in Table 1. They will now be assessed based on empirical research. For each explanatory approach, two hypotheses have been formulated which are presented in the bottom row of the table. For instance, Hypothesis 1 states that the electoral crisis of social democracy is the result of a shrinking working class, while increasing electoral trade-offs are the cause of the electoral decline according to the two variants of Hypothesis 2.
The analysis is based on a systematic review of empirical studies published in peer-reviewed journal articles from January 2010 to October 2021. The selection process was oriented at the PRISMA guidelines (Moher et al. 2009) and proceeded in four steps (for a more detailed account see Figure A1 in the Online Appendix): 1. Extensive literature search via Web of Science: 621 studies 2. Screening the identified studies for eligibility: 30 studies 3. Snowball research based on eligible studies: additional 21 studies 4. Categorization of the 51 included studies A study was assessed as eligible if it offered empirical evidence on at least one of the outlined hypotheses. The resulting 51 studies provide evidence on three different levels. Out of those studies, 14 deal with the determinants of social democrats' vote share (aggregate level) and thus offer the most direct test of the hypotheses. Twenty of the studies provide cross-country evidence on the voter level (i.e. individual level). Those studies generally rely on survey data, which are particularly useful in studying the reasons for turning away from social democrats as well as electoral trade-offs. In addition, 27 studies include analyses on individual social democratic parties, with most studies covering Germany, Britain and the Nordic countries. 3 The review builds primarily on the cross-country studies due to the high generalizability of the findings, but it also includes strong pieces of evidence provided by country studies (for an overview of the included studies see Tables A1 and A2 in the Online Appendix).

Sociological explanation
According to Hypothesis 1, social democrats' electoral losses result from the shrinking of the working class. Basic evidence comes from Giacomo Benedetto et al. • Turn to the right on economic dimension (H5) • Liberal turn on cultural issues (H6) • Governing with centre-right/declining policy polarization (H7) • Decline in union density (H8) (2020), who combine an analysis of the determinants of declining vote shares with a look at individual-level support for social democrats. The authors offer qualified support for H1 by concluding that 'most of the fall in support for social democratic parties in recent years is correlated with a decline in the number of industrial workers as well as a reduction in the propensity of these parties' core supporters … to vote for them' (Benedetto et al. 2020: 928). 4 This is supported by Robin Best (2011), who shows that group-size effects have strongly contributed to the decline of social democrats' vote share among workers, but that reduced loyalty of workers also plays a role. On the country level, several studies on class voting also stress the latter point by highlighting social democrats' role in the demobilization of the working class (for Germany see Roßteutscher 2011, 2017; for the UK see Tilley 2012a, 2012b;Heath 2015Heath , 2018. Even with this qualification, the empirical evidence clearly indicates that the shrinking of the working class over time accounts for a part of social democrats' vote losses. The second hypothesis highlights the role of electoral trade-offs concerning occupational groups (H2a) or labour market insiders and outsiders (H2b) in social democracy's decline. Here, evidence comes exclusively from individual-level studies on voter attitudes and behaviour. Support for the trade-offs between working-class and middle-class voters (H2a) is provided by Line Rennwald and Jonas Pontusson (2021), though the authors find that this trade-off is mitigated by high unionization. In contrast, Tarik Abou-Chadi and Markus Wagner (2019) find that strong unions exacerbate the trade-off between manual workers and professionals. There are more conflicting results. Focusing on sociocultural issues, Abou-Chadi and Wagner (2020) find no trade-off between workers and professionals. Studies focusing on social democrats' positions on immigration indicate the opposite. While Jane Gingrich (2017) shows that the potential left coalition of working-class and new middle-class voters is divided on immigration, country studies on Germany (Chou et al. 2021) and Denmark (Hjorth and Larsen 2020) point to related electoral trade-off between those groups. On the one hand, the conflicting findings indicate the substantial impact of the methodical approach on the results. On the other hand, the electoral trade-off seems to be conditional on country-specific factors, such as union strength (Abou-Chadi and Wagner 2019; Rennwald and Pontusson 2021) and the electoral system (Arndt 2014b).
Concerning the insider-outsider dilemma (H2b), the literature yields similar results. The extent of differences in distributional and political preferences between labour market insiders and outsiders depends on the operationalization of the cleavage, with labour market status being a stronger indicator than labour market risks (Marx 2014;Rovny and Rovny 2017). The only study that directly tests H2b comes from Johannes Lindvall and David Rueda (2014). Based on five elections between 1994 and 2010, the authors demonstrate that the Swedish Social Democrats were indeed confronted with the insider-outsider dilemma. Further research is needed to test if H2b holds for other countries.

Materialist explanation
Hypothesis 3 highlights external material constraints, such as globalization (H3a) and fiscal and monetary restrictions in the EU (H3b). Concerning the impact of globalization, the empirical evidence is ambivalent. Based on the KOF globalization index, Benedetto et al. (2020) find no general negative impact of globalization on the social democratic vote share. In contrast, a recent study by Helen Milner (2021: 2288 which focuses on, among other things, import shocks on the regional level concludes that 'globalization, especially trade, is associated with declines for mainstream left parties' and that 'these effects accelerated with the financial crisis'. The comparison of both studies highlights the impact of the selected globalization indicator. In other words, import shocks on the meso level seem to be a better indicator than globalization indices to capture the negative electoral effects of globalization on social democrats. Turning to the EU and the eurozone (H3b), there is mounting evidence that existing restrictions are indeed detrimental to social democrats. Focusing on election results before and during the euro crisis, Sonia Alonso and Rubén Ruiz-Rufino (2020) show not only that social democrats paid an electoral price for implementing austerity measures but also that this price was considerably higher in countries facing external intervention by the so-called Troika. This finding is corroborated by Sara Hobolt and James Tilley (2016), who also point to the impact of the euro crisis and fiscal constraints on voters' defection to challenger parties. Finally, Stuart Turnbull-Dugarte (2020) shows that left-leaning voters are more likely to abstain from voting than other voters when countries are exposed to EU intervention. In sum, H3b is corroborated for Southern European debtor countries.
Hypothesis 4 highlights the negative impact of austerity measures and other 'structural reforms' on social democrats. In addition to the studies by Hobolt and Tilley (2016) and Alonso and Ruiz-Rufino (2020), Gijs Schumacher et al. (2013) and Alexander Horn (2021) confirm the hypothesis that social democrats are punished for welfare retrenchment. According to Horn (2021), the negative effects are permanent and materialize in the long run, irrespective of positive economic effects. In contrast, Nathalie Giger and Moira Nelson (2011) as well as Abel Bojar (2018) find no electoral punishment. This is, however, clearly at odds with findings from country studies. In a comparative study on Germany, UK, Denmark and Sweden, Christoph Arndt (2013) demonstrates that social democratic parties in all countries suffered electoral setbacks as a consequence of welfare state reforms, with the scale of those setbacks contingent on the electoral system. For Germany, Hanna Schwander and Philip Manow (2017) show that the negative long-term effects of the 'Agenda 2010' for the SPD resulted mainly from the establishment of a strong competitor on the left. More generally, there is substantial evidence that social democrats tend to benefit from high levels of social spending and, more specifically, spending on labour market policies (Benedetto et al. 2020;Fossati and Trein 2021;Kweon 2018). Karl Loxbo et al. (2021) find, however, a curvilinear relationship, with diminishing social democratic vote shares at higher levels of welfare generosity. On balance, the empirical evidence supports H4, especially in terms of the negative effects of spending cuts.

Ideational explanation
Hypothesis 5 highlights social democrats' neoliberal turn, that is, their rightward shift on the socioeconomic dimension. Here, studies which find no such effect can be contrasted with research that offers general or at least qualified support for this hypothesis. Focusing on the adoption of the 'third way' and a rightward shift on the state-market dimension, neither Hans Keman (2011) nor Abou-Chadi and Wagner (2019, 2020) find support for H5. This stands in stark contrast to the findings by Johannes Karreth et al. (2013) on the long-term effects of moving to the right. Based on survey data from Germany, Britain and Sweden, the latter study shows that 'the gains [social democratic] parties derived from the policy shift toward the middle in the 1990s were short-lived and came at the expense of electoral success in subsequent decade' (Karreth et al. 2013: 792). This negative longterm effect is in line with Schwander and Manow (2017), who demonstrate that the negative electoral effects of the German SPD's neoliberal turn took some time to unfold. Other studies indicate that the effect is contingent on additional factors, such as a proportional electoral system (Arndt 2014b), as this facilitates the voter defection witnessed in the German case, high income inequality (Polacko 2022) and low to moderate levels of welfare generosity (Loxbo et al. 2021). In sum, the evidence indicates that the negative electoral effect of moving to the right unfolds over time and that its strength is contingent on the political and economic context.
According to Hypothesis 6, it is the adoption of a more libertarian position on the sociocultural dimension that leads to vote losses. The strongest rejection of this hypothesis comes from Abou-Chadi and Wagner (2020). Based on their analysis of survey data from 13 West European countries, they conclude that 'more progressive and more pro-EU positions are if anything electorally beneficial to Social Democratic parties' (Abou-Chadi and Wagner 2020: 247-248). Inconsistent with H6, workers are not found to be less likely to vote for culturally progressive social democrats. In combination with an investment-oriented economic stance, culturally liberal positions are even found to be beneficial to social democrats if unions are weak (Abou-Chadi and Wagner 2019). Focusing on the issue of immigration, Jae-Jae Spoon and Heike Klüver (2020) instead come to the conclusion that centreleft parties can profit from 'going tough on immigration'. Results from country studies on this issue are mixed but mainly not in line with H6. Survey data from Britain indicate that a liberal stance on immigration indeed leads to the defection of voters from Labour (Evans and Mellon 2016). In contrast, Carl Dahlström and Anders Sundell (2012) find no support for H6 for elections in Swedish municipalities. For Germany, the empirical evidence shows that a restrictive stance on immigration is either not sufficient to attract radical right voters (Spoon and Klüver 2020) or comes at the cost of alienating its own voters (Chou et al. 2021). 5 This corresponds to the findings of a recent survey experiment among Danish voters which shows that Danish Social Democrats face a trade-off between antiimmigration and pro-immigration voters, though a tougher stance on immigration might be beneficial from an office-seeking perspective (Hjorth and Larsen 2020).

Institutional explanation
Hypothesis 7 is derived from cartel party theory and highlights political collaboration and policy convergence with mainstream parties on the right. The empirical evidence mainly supports this hypothesis. On the aggregate level, Benedetto et al.
(2020) find a significant negative effect for social democrats participating as junior partner in coalition governments. On the individual level, further empirical support comes from studies focusing on policy convergence among mainstream parties. Spoon and Klüver (2019) find that those parties' convergence on the left-right scale leads to vote losses as their voters switch to non-mainstream parties. This is in line with Zack Grant's (2021) finding that de-polarization between establishment parties fuels the support of anti-system parties in times of economic crisis (see also Hobolt and Tilley 2016). In Sweden, policy convergence between social democrats and centre-right parties on welfare issues has benefited the latter (Arndt 2014a), while convergence on the general left-right dimension and migration have been beneficial to the radical right (Oskarson and Demker 2015). In sum, the literature indicates that policy convergence with the centre-right has indeed negative electoral consequences for social democrats. 6 Hypothesis 8 highlights the negative impact of union decline on social democrats. On the voter level, the literature paints a unanimous picture, since all studies controlling for individual union membership find that union members are more likely to vote for social democrats than non-union members (Abou-Chadi and Wagner 2019, 2020; Arndt and Rennwald 2016;Kweon 2018;Marx 2014;Milner 2021;Mosimann et al. 2019;Polacko 2022;Rennwald and Pontusson 2021). Somewhat surprisingly, the effect is quite similar among working-class and middleclass voters (Mosimann et al. 2019: 78-80). But does this effect translate to the aggregate level in times of union decline? Here, the results are at first sight rather inconsistent. While Abou-Chadi and Wagner (2019), Loxbo et al. (2021) and Matthew Polacko (2022) find no significant effect of unionization on the social democratic vote share, Rennwald and Pontusson (2021: 40) estimate that 'a one-percentage-point decline in union density is associated with a vote-share decline of nearly half a percentage point' and demonstrate that strong unions mitigate the trade-off between working-class and middle-class voters. A fine-grained analysis by Arndt and Rennwald (2016) reveals that the union effect is highly contingent on union confederations, with strong blue-collar and public sector confederations but not white-collar confederations in the private sector having a positive impact on social democrats' vote share. The literature thus lends a qualified support to H8.

Summary
As Table 2 demonstrates, there is not one explanation that stands out. Rather, several factors seem to have contributed to the electoral crisis of social democracy, especially the shrinking of the working class (H1a), austerity measures implemented by social democrats (H4), their neoliberal turn (H5) and their policy convergence with centre-right parties (H7). The first factor reflects long-term socio-structural changes, whereas the three latter factors are parts of a different story which can be summarized as the ideational and institutional neoliberalization of social democracy (Bandau 2021). I will return to this point in the conclusion. H6 receives the least support by the literature, since movements in either direction on the sociocultural dimension risk alienating voters.
The support for the other hypotheses is more qualified. One important qualification is that negative electoral effects are conditional on other factors. For instance, the supposed electoral trade-offs between different groups of voters are contingent on the electoral system and unions (Abou-Chadi and Wagner 2019; Arndt 2014b; Rennwald and Pontusson 2021). In addition, the literature shows that the interaction of certain factors, such as austerity and EU intervention or policy convergence with the mainstream right in times of economic crises, are highly detrimental to social democrats. These findings indicate that the causes for social democrats' electoral losses are complex and thus might differ among Western European countries. This point is supported by the different images of social democratic decline presented in Figure 1.
Finally, inconsistent evidence for some of the hypotheses is at least in part the product of divergent operationalizations and statistical techniques. The impact of operationalization is best illustrated by the diverging results on electoral behaviour for different operationalizations of labour market outsiders (Rovny and Rovny 2017). Concerning statistical techniques, models on the levels of social democratic support can yield different results than first-differences models focusing on short- term vote changes. By applying both kinds of models, Benedetto et al. (2020) find that the shrinkage of the working class affects the levels of the social democratic vote share but not the short-term changes between elections (see Table A1 in the Online Appendix). Covering levels and changes thus provides a way to distinguish between long-term and short-term effects.
Conclusion: avenues for future research and the future of social democracy Based on the findings, there are three starting points for future research: the inconsistency of some of the findings, the relation of individual factors in explaining the electoral crisis of social democracy and the portability of the findings to crisisridden left parties in Eastern Europe. Concerning the first point, conflicting results are, as shown, often the product of the used indicators. Here, researchers have not only to look for the best indicators available but also be careful with the interpretation of their findings. For example, recent evidence shows that the sociocultural divide is much stronger for issues such as immigration than for other sociocultural issues (Lancaster 2022). Such findings have to be taken into account when analysing the electoral consequences of social democrats' programmatic shifts.
Regarding the interplay of various factors, I propose two main avenues for future research. First, researchers should distinguish long-term and short-term causes of social democrats' electoral crisis. As outlined above, estimating models for social democratic vote shares as well as electoral changes between elections offers one promising strategy. A more sophisticated approach is to heed Paul Pierson's (2004: 102) advice: 'The connection between "triggers" and "deeper", more longterm causes suggests that seemingly rival explanations may often be complementary.' In the case of social democracy, this means combining long-term processes, such as socio-structural changes, with 'triggers', such as policy decisions by social democrats in office. Second, the finding that some of the hypotheses only hold under certain conditions points to complex causality at the heart of the crisis of social democracy. Qualitive comparative analysis (QCA) presents an appropriate method to deal with complex causality and related equifinality (Schneider and Wagemann 2012). QCA can thus be applied to determine which constellations of sociological, material, ideational and institutional factors cause social democratic decline. Previous applications of QCA on electoral performance, which have focused on populist and anti-establishment parties (Fernández-García and Luengo 2019; van Kessel 2015), provide blueprints for this kind of research.
Finally, this review has centred on the crisis of social democratic parties in Western Europe. However, left parties in Central and Eastern Europe have also suffered substantial electoral defeats in recent years. Though those parties have a special history, some of them being the successors of communist parties, their electoral calamities may have similar causes (Berman and Snegovaya 2019). Indeed, recent studies demonstrate that neoliberal stances and policies have strongly contributed to the left's electoral crisis in Eastern Europe (Bagashka et al. 2022;Snegovaya 2022). While this is in line with the findings on Western Europe, further research is needed to get a better understanding of similarities and differences between Eastern and Western European parties. This leads to the final question: what does all of this mean for the future of social democracy? In general, the review holds a positive and a negative message for social democrats. The negative message is that the shrinkage of the traditional working class reduces social democracy's core electorate, while post-industrial electoral trade-offs change partisan competition and complicate the creation of new electoral coalitions. A return to previous electoral heights is thus all but impossible. The positive message is that the electoral crisis is to some extent the result of the neoliberalization of social democracy. Social democrats' recent adoption of more leftist economic positions and the slightly growing policy divergence between them and centre-right parties thus offer some signs of hope for social democracy (Manwaring and Holloway 2022). However, to permanently halt the downward trend, social democrats have to present a clear alternative to today's ailing neoliberalism.
Notes 1 A closer look at the literature reveals divergent operationalizations of insiders and outsiders, with some focusing on labour market status and others on occupational groups (Rovny and Rovny 2017). 2 The rise of populist parties is also often linked to globalization processes, including migration. According to Rodrik (2018), left-wing and right-wing populism are both forms of protest against different aspects of globalization, i.e. against the free movement of goods and money in the former case and against the free movement of persons in the latter. Whether social democrats tend to be confronted with populist challengers from the left or the right thus depends largely on whether free trade or labour migration is perceived as the major economic threat.
3 Some studies include analyses on the aggregate and the individual level or combine cross-country analyses and analyses on individual social democratic parties. 4 However, deindustrialization has no explanatory power when it comes to electoral changes between elections. 5 Other studies on Germany indicate that a high salience of the immigration issue is detrimental to the SPD (Pardos-Prado et al. 2014;Wurthmann et al. 2021). 6 Additional support for H7 comes from country studies on social and policy representation. For Britain, the work by Tilley (2012a, 2012b), Heath (2015Heath ( , 2018 and O'Grady (2019) highlights how decreasing social representation and the increasing policy divergence between Labour and its supporters has led to growing alienation. A similar picture emerges in Belgium, where 'Flemish social democracy's left-universalistic elites have largely lost connection with their left-particularistic base' (Gaasendam et al. 2021).