Publication Cover
Representation
Journal of Representative Democracy
Volume 60, 2024 - Issue 4
139
Views
1
CrossRef citations to date
0
Altmetric
Articles

Squeezing Blood from a Turnip? The Resilience of Social Democratic Governmental Power in Western Europe (1871–2022)

ORCID Icon & ORCID Icon

ABSTRACT

In recent years, a growing body of literature has revived its interest in social democratic parties, emphasising their allegedly irreversible crisis in Western Europe. However, all such accounts focus solely on electoral results, thus neglecting governmental power, the decisive factor to realise social democratic parties’ policy goals. To address this gap, the article tests whether the decline of social democracy is confirmed in terms of governmental power, for which an index that remedies the limitations of existing measures is employed. Through comparative longitudinal analysis on 20 Western European countries and more than 600 legislatures between 1871 and 2022, the article finds that the governmental power of social democracy has remained fundamentally the same as in the golden age of class politics. In particular, in contexts of high party system fragmentation and strong radical left competition, social democratic parties have still managed to secure relevant government positions despite their declining electoral performance.

Introduction

Political science has recently revived its interest in social democratic parties and class politics. In particular, what caught the scholarly attention is the dramatic electoral decline of social democratic parties in Western Europe, leading many to argue that these are now in a profound and hardly reversible crisis (e.g. Abou-Chadi et al., Citation2021; Benedetto et al., Citation2020; Delwit, Citation2021; Diamond, Citation2016; Polacko, Citation2022; Rennwald & Pontusson, Citation2021), mainly as a result of their lessened capacity of mobilising their traditional working-class constituency (e.g. Emanuele, Citation2021; Franklin, Citation1992).Footnote1

Yet, all these works share a primary focus on the electoral performance of social democratic parties and derive their findings from the underlying idea that the electoral arena is what really matters for the assessment of the state of social democracy. Hence, surprisingly enough, they overlooked the performance in the executive sphere, namely the governmental power of social democracy. This represents a fundamental aspect in the assessment of the overall health of a given party or party family (e.g. social democracy), namely the extent to which it still retains its power or falls into a generalised crisis. Indeed, as the governmental arena can be conceived as ‘the principal, even if not the only, prize of party competition’ (Casal Bértoa & Enyedi, Citation2016, p. 266), consistently accessing and controlling government is key for a party or party family to retain its hold over society, even besides electoral trends. This is because it is especially the governmental arena that actually matters for parties to achieve their policy goals, which in the case of social democracy means reducing inequalities and building a fairer society (e.g. Bobbio, Citation1997).Footnote2

On this basis, we aim to fill this gap by testing whether social democracy is in crisis in terms of governmental power, as its electoral results would suggest, or whether certain contextual factors help this party family to be resilient in government. This investigation is pursued through comparative longitudinal analysis that covers a century and a half of social democratic politics. In particular, we rely on a time-series cross-section dataset that includes 20 Western European countries and more than 600 legislatures between 1871 and 2022, hence covering a very extensive spatial–temporal framework. Furthermore, we introduce an index of governmental power that is more fine-grained compared to the available alternatives and remedies several of their deficiencies.

From a substantive viewpoint, we come to a relevant conclusion: whereas the electoral decline of social democracy is confirmed, its governmental power instead does not decrease over time and has remained fundamentally the same as in the golden age of class politics. In particular, our empirical analysis shows that this apparent and growing disconnection between the electoral and executive spheres is associated with contexts of high party system fragmentation and strong radical left competition. Such factors lead to the most impactful conclusion of our analysis, which sets it apart from existing accounts: namely, that social democracy is not in crisis – at least, not yet – when looking at the governmental arena rather than the electoral one.

The article will proceed as follows: the next section provides an account of the literature on social democratic political development and raises our expectations about the disconnect between social democracy’s electoral support and governmental power; the third introduces our index to study the governmental power of social democracy; the fourth section presents the spatial–temporal framework, case selection, and method of the empirical analysis; the fifth section illustrates the results of the analysis. The concluding section discusses the article’s findings and implications.

Theoretical Framework and Expectations

The first social democratic parties emerged in the late nineteenth century, with the progressive extension of voting rights to the original constituency of these parties, namely the working class (e.g. Lipset & Rokkan, Citation1967; Przeworski, Citation2009). Whilst for the first few decades social democratic parties were not competitive electorally, they started gaining traction in the early 20th and, after the introduction of universal male suffrage, they established themselves as the mass parties of the left and became a major electoral force across the continent (Delwit, Citation2021; Keman, Citation2017), gravitating on average around a vote share of 30%.

Subsequently, until the 1970s, the political history of social democracy tells the story of an unrestrainable path to power, with increasing electoral success going hand in hand with a progressive reinforcement of the executive role across Western Europe. However, the 1970s and 1980s marked a slowdown of social democratic electoral performance, only momentarily halted in the 1990s by the economically centrist ‘Third Way’ social democracy (e.g. Delwit, Citation2021; Giddens, Citation1998; Rennwald & Pontusson, Citation2021). Yet, the electoral decline of the mainstream left materialised in the 2000s, with national vote shares no longer reaching, on average, the 30% mark (Abou-Chadi et al., Citation2021; Delwit, Citation2021; Keman, Citation2017). This trend culminated after the global financial crisis in the 2010s, where the electoral support of social democratic parties in Western Europe plummeted to pre-1930s levels, barely above 20% of national vote shares (Abou-Chadi et al., Citation2021; Delwit, Citation2021; Rennwald & Pontusson, Citation2021).

Therefore, all mentioned works come to similar conclusions about the fate of Western European mainstream left parties: that is, these seem to be in an unequivocal and hardly reversible crisis across the continent (e.g. Abou-Chadi et al., Citation2021; Benedetto et al., Citation2020; Delwit, Citation2021; Diamond, Citation2016; Emanuele, Citation2021; Polacko, Citation2022; Rennwald & Pontusson, Citation2021). However, all such accounts focus exclusively on the electoral arena, thus neglecting the performance of social democracy in government: that is, the decisive arena where power is translated into policy outputs. Did the governmental power of social democrats decline in recent decades, hence mirroring its electoral performance? We argue that this is not necessarily the case and that perspectives focusing exclusively on the electoral sphere might overinflate the demise of social democracy.

This is because, besides variations in the national experiences, the trajectory of social democracy in the two arenas might be characterised by the presence of a temporal lag, which delays in the governmental arena the effects shown in the electoral sphere. In particular, two frameworks are useful for our analytical reading. Firstly, Rokkan’s (Citation1970) work traces an explicit distinction between the electoral and governmental arena. In this regard, parties that have already passed the threshold of ‘legitimation’ and ‘incorporation’ by participating in elections can reach two additional thresholds: ‘representation’, when they manage to gain parliamentary seats; and ‘executive power’, when they enter government. The last threshold is clearly more difficult to reach than the first ones: it follows from this that there will usually be a temporal lag between contesting elections and accessing the executive. As shown in Table A1 in the Online Appendix, in Western European countries there is an average lag of 32 years between the first time social democrats run in a general election and the moment in which they first access government.

Complementary to this viewpoint is Pedersen’s (Citation1982) concept of ‘party lifespan’, according to which parties, ever since their birth, usually follow a typical curvilinear, bell-shaped trajectory in the course of their history, with their electoral support increasing over time until reaching a culmination point, and a monotonic decrease afterward.Footnote3 Following Rokkan and Pedersen’s perspectives, one might argue that, consistent with the multiple scholarly contributions already discussed, while social democracy has long passed its culmination point in electoral performance and started its descent, this may not be yet the case as far as governmental power is concerned. Indeed, the illustrated temporal lag between the two arenas suggests that the electoral collapse of social democracy may not be observed yet in government. This may also be reinforced by the fact that, over time, large mainstream parties have become a core part of the national party systems in Western Europe (Smith, Citation1989), and this, of course, also applies to social democrats in each country (Keman, Citation2017). Hence, as core parties of their respective systems, social democrats may have become a stable and recurring governing partner, thus slowing down the decline in terms of governmental power compared to electoral performance. On this basis, this article suggests a different perspective to analyze the evolution of European social democracy. That is, by looking at the governmental arena, the emphasis on the crisis of social democracy in recent decades might be reconsidered.

Indeed, we expect that the social democratic parties’ trajectories between the electoral and the governmental arena have become increasingly differentiated in the last decades. More generally, considering social democratic electoral support and governmental power together, it is possible to build a fourfold typology of social democratic parties’ performance in the two arenas. outlines how the combinations of the two dimensions translate into the four possible types. If social democratic electoral support and governmental power are consistent, social democracy will be either weak or strong depending on whether the two dimensions are, respectively, comparatively low or high. Conversely, two different situations might emerge when the trajectories of the two dimensions diverge. On the one hand, social democrats may show a limited ability to translate their considerable electoral support into relevant executive positions. In such instances, social democratic parties are thus dilapidators of their potential power. On the other, the opposite reasoning applies when social democracy manages to efficiently exploit its comparatively low electoral support and turn it into a powerful governmental status. In this case, social democratic parties are optimisers in that they are much more relevant within the governmental arena than their electoral support would otherwise suggest.

Figure 1. Typology of left electoral support and governmental power.

Figure 1. Typology of left electoral support and governmental power.

Based on the above discussion, we expect a differentiation in recent decades in the trajectories of social democrats within these two realms. As a result, social democratic parties will increasingly deviate from the traditional bottom left-top right diagonal linking weak and strong social democracy in . More specifically, we expect that more and more social democratic parties will efficiently translate their poor electoral results into governmental relevance, thus showing an optimiser profile. Therefore, our first expectation is the following:

Expectation 1: Unlike electoral support, as time goes by, social democratic governmental power has not decreased in the last decades, thus highlighting an increasingly optimizer profile of social democratic parties.

If the above expectation were to be empirically confirmed, the following step of the analysis would be that of asking why social democracy shows an increasingly inconsistent performance between the electoral and the governmental arenas. In other words, what are the contextual factors that may explain why the social democratic left has become increasingly optimiser? We believe it would be worthwhile to address such a research question. In this regard, we raise two additional expectations.

The first one is related to party system fragmentation. As we know from multiple scholarly accounts, over the last decades, and especially since the Great Recession, a large number of new parties has successfully emerged in the electoral arena (Bolleyer & Bytzek, Citation2017; Emanuele & Sikk, Citation2021; Hernández & Kriesi, Citation2016). Consequently, party systems in Western Europe have become increasingly fragmented (Best, Citation2013; Casal Bértoa & Enyedi, Citation2021; Mair, Citation1997). As the number of political competitors increases, all else equal, the vote share of established parties becomes smaller. Yet, the multiplication of options emerging at the electoral level may not directly translate into more options for the government formation process, as established mainstream parties continue to hold the lion’s share in the process (Bäck & Bergman, Citation2015; Bäck & Dumont, Citation2008). The role of political formateur may continue being in the hands of parties with a consolidated government experience. As social democratic parties can certainly be considered as belonging to an established mainstream party family with decades-long government experience, they may benefit from their position in the ‘core’ of their respective party systems as far as governmental power is concerned. As a result, our second expectation is the following:

Expectation 2: The higher the party system fragmentation, the more ‘optimizer’ the profile of social democratic parties

Besides fragmentation, another contextual feature that may explain the increasingly optimiser profile of the social democratic left has to do with the characteristics of the party environment in each country. Following classics like Duverger (Citation1954) and Sartori (Citation1976), we expect that ideologically radical competition brought about by centrifugal political forces located at the fringe of the ideological spectrum will reduce the number of options available for government formation, thus boosting the chances of social democrats to maximise governmental power in spite of declining electoral performances. This is because most of the time radical left or right parties are a priori excluded from government due to the vetoes of the established mainstream parties (e.g. the so-called convention ad excludendum for the Italian Communist Party or the cordon sanitaire for the French Front National). The rise of challenger and radical parties in Western Europe in recent decades is a well-known and largely documented phenomenon (Hino, Citation2012; Hobolt & Tilley, Citation2016; Zulianello, Citation2019) and may have contributed to maintaining social democrats to power, sometimes even forcing them to lead or become part of ‘grand coalition’ governments with mainstream conservative parties. Consequently:

Expectation 3: The higher the vote share of ideologically radical parties, the more optimizer the profile of social democratic parties

Measuring Social Democratic Power: The Governmental Power Index (GPI)

The power of left parties has been analyzed extensively by the ‘power resource theory’ (PRT) strand (for an overview, see Brady et al., Citation2016). This literature is situated within the fields of international political economy and economic sociology; hence, it is found mainly outside of comparative politics, which instead largely neglected this topic besides electoral accounts. Moreover, PRT mainly treats the power of the left as a predictor of economic outputs rather than as a phenomenon to be studied per se. Most importantly, its key limitation is the inaccurate operationalisation of such power. Indeed, some works within this strand do not even go beyond looking at the electoral or the parliamentary arena. For instance, they take into consideration the share of parliamentary seats of left parties, which does not convey any information whatsoever on their government status (Braga et al., Citation2013; Hewitt, Citation1977). PRT contributions that do look at the governmental arena consider the overall ideological position of cabinets (e.g. Cusack, Citation1997; Iversen & Soskice, Citation2015; Mahler, Citation2004) or explicitly whether the left is in government and look at the share of their parliamentary seats vis-à-vis those of the other cabinet partners (e.g. Bradley et al., Citation2003; Huber & Stephens, Citation2000); or, finally, the share of cabinet seats of the left (e.g. Allan & Scruggs, Citation2004; Garritzmann & Seng, Citation2016; Jensen, Citation2010). Yet, these measures, even those that somehow look at the governmental arena, are inaccurate and incomplete insofar as they do not translate important theoretical aspects of social democratic power, such as the parliamentary status of the cabinet (majority or minority) and the party status within the cabinet (i.e. whether it is a leading or junior partner) in operational terms.

As illustrated, all such operationalizations are unsatisfactory. In order to properly measure governmental power, it is necessary to develop an index that moves conceptually from all these elements and, especially, is able to properly operationalise them (Adcock & Collier, Citation2001). To do so, this article proposes a governmental power index (GPI) through a revision of Bartolini’s original proposal on this matter (Citation1998). The index measures the power of a party at a given point in time by considering three fundamental characteristics. First, it accounts for the government status of the party, distinguishing between cabinet and opposition status, as well as considering if a party provides an abstention or external support that are necessary for the survival of a minority cabinet. Second, it operationalises the cabinet status in parliament, that is, whether or not it has a majority status and if the party governs alone or as part of a coalition. Third, it considers the party status in cabinet by looking at its cabinet seats to distinguish its leading or junior status in the cabinet vis-à-vis the other formations that are part of it.Footnote4 The index provides a rank-ordering of the possible levels of governmental power by considering the various combinations between cabinet and governing parties characteristics (see ). As a result, it ranges between 0 (when the party is in opposition) and 10 (when the party holds a single-party majority cabinet).Footnote5 Because of its conceptual and operational characteristics, the GPI is an accurate and encompassing measure of governmental power in general and of social democratic power in this specific instance, constituting a significant improvement over available alternatives.

Table 1. Governmental Power Index (GPI): dimensions and values.

Spatial–temporal Framework, Case Selection, Data and Method

Our analysis of the evolution of social democratic electoral support and governmental power over time requires a prior definition of what social democracy is and which parties we consider in the context of this study. Firstly, in terms of spatial and temporal scope, we focus on 20 Western European countries and adopt a comparative longitudinal perspective by covering a period stretching from the very first electoral appearance of social democracy to date (1871–2022).Footnote6 Spatially, this framework allows the largest possible generalizability of our findings across the geographical region of Western Europe. Temporally, the vast timeframe analyzed here allows us to fully cover the whole century and a half in which social democratic parties participated in national elections in Western Europe. Moreover, this longitudinal extension avoids the risk of suffering from the ‘short-termism’ typical of many studies on this topic, which ‘judge contemporary politics by the exceptional standards of mid-20th Western European politics’ (Enyedi & Deegan-Krause, Citation2007, p. 13). After specifying the spatial and temporal scope of our analysis, the definition of social democracy needs to suit a comparative longitudinal analysis. We select the main historical social democratic formation that traditionally mobilised the working class in each country (Bartolini, Citation2000; Bartolini & Mair, Citation1990). The resulting list of parties is reported in Table A3 in the Online Appendix.

In terms of empirical analysis, we first show the evolution of our dependent variables – the electoral support and GPI of social democratic parties – over time. The electoral support of social democrats is measured by the country’s respective social democratic party’s vote share in the general election (Lower House). The GPI, introduced above, is calculated at the legislature level for the sake of comparability with electoral support.Footnote7 If a legislature is made up of a single cabinet, the GPI value for that cabinet is attributed. Otherwise, when there are multiple cabinets within a legislature, the GPI of the legislature is calculated by summing the GPIs of every single cabinet within the legislature, which are all weighted for the respective cabinet duration in number of days.Footnote8 Secondly, we empirically analyze which factors are associated with the disconnect between electoral support and GPI of social democratic parties. The latter is operationalised as the difference between the standardised versions of, respectively, GPI and vote share. It is a continuous variable assuming increasingly positive values as the social democratic left becomes more optimiser (i.e. comparatively stronger GPI than vote share) and increasingly negative values as it becomes more dilapidator (i.e. comparatively stronger vote share than GPI). The variable has been rescaled to a 0–100 range in order to ease the interpretation of the results.

Moving to the independent and control variables, we measure party system fragmentation through the Effective Number of Electoral Parties (ENEP) by Laakso and Taagepera (Citation1979). The electoral support for ideologically radical parties has been measured as the aggregate vote share in general elections (Lower House) of parties belonging to the radical left and right party families, according to ParlGov (Döring & Manow, Citation2022). We will also disentangle the ideologically radical vote by testing the separate effects of radical left and radical right support.Footnote9 Moreover, we will also control for another factor that may affect the process of translation of votes into governmental power, namely the disproportionality of electoral systems, measured through the Gallagher index (Citation1991).Footnote10 Finally, following Emanuele (Citation2023), we have modelled time through a categorical variable, which divides our temporal framework into six periods of substantive interest: (1) the ‘Pre-WWI’ period (1871–1917), going from the very first time a social democratic party contests an election in Western Europe (i.e. the Social Democratic Workers’ Party, the predecessor of the Social Democratic Party (SPD), in Germany) to the end of World War 1; (2) the ‘Inter-war’ period (1918–1944); (3) the ‘Golden Age’ phase of mass parties and party system stability in Western Europe (Janda & Colman, Citation1998), starting in 1945 ending in 1967, the year of Lipset and Rokkan’s (Citation1967) ‘freezing hypothesis’; (4) the ‘Post-L&R’ (Lipset and Rokkan) period, which runs from the socially and politically fundamental year of 1968 to the watershed year of 1989, with the fall of the Berlin Wall; (5) the ‘Post-Wall’ period, lasting from 1990 to the outbreak of the Great Recession in Europe (2009); (6) and the ‘Great Recession’ period, lasting from 2010 until the end of the considered temporal framework, that is the years in which the economic and political consequences of the global financial crisis manifested.Footnote11

Our models are specified by taking into account the time-series-cross-section nature of our data, with repeated observations over time (623 legislatures) over the same fixed units (countries). As diagnostic tests pointed to expected issues such as heteroskedasticity and autocorrelation,Footnote12 we select a proper model specification by performing Prais-Winsten regressions with panel-corrected standard errors (PCSE) and country-fixed effects (Beck & Katz, Citation1995).Footnote13

Analysis and Results

Before moving to the regression models, we look into the bifurcation of the paths of social democratic electoral support and governmental power over time. Based on the typology outlined in the theoretical framework, reports the mean values of social democratic vote share (x-axis) and GPI (y-axis) by country over each of the six outlined temporal phases, with the horizontal and vertical reference lines corresponding, respectively, to the mean GPI and vote share values (respectively, 2.87% and 26.64%) across our spatial–temporal framework.

Figure 2. Mean social democratic vote share and governmental power in each country per phase.

Figure 2. Mean social democratic vote share and governmental power in each country per phase.

The ‘Pre-WWI’ and, to a lesser extent, the ‘Inter-war’ periods see very low levels of GPI and no optimisers at all, as access to government for the electorally surging social democracy is complicated by both the aforementioned temporal lag experienced between the first electoral contestation and the executive threshold, and the emergence of right-wing dictatorships and war cabinets across the continent. The only significant exceptions are Sweden and Denmark, showing a strong profile already in the Interwar period. During the ‘Golden Age’, the average GPI values increase considerably, and this situation is even reinforced in the ‘Post-L&R’ period. Graphically, this seems to be the period where the association between the two analyzed variables is closer, as most countries fall close to the bottom left-to-top right diagonal. The picture starts changing during the ‘Post-Wall’ years, as social democratic electoral support starts declining and rarely reaching the 40%s. Here, however, the institutionalisation of social democrats as key governmental actors begins to pay off, as for the first time, optimisers outnumber dilapidators (seven to two). Finally, during the ‘Great Recession’ these processes accelerate even further as social democratic vote share decreases dramatically, with national averages above 30% being now the absolute exception. On the one hand, this generalised electoral collapse entails the almost complete emptying of the right side of the graph, with the British and Norwegian social democrats being the only dilapidators. On the other, this period attains the maximum number of optimisers, nine: a remarkable difference compared to the early phases of our timeframe. This phenomenon involves a group of social democratic parties in a very heterogeneous pool of national contexts: from Scandinavia (Denmark, Finland) to Southern Europe (Italy); from larger (France, Germany) to smaller Continental European polities (Austria, Belgium, Luxembourg, and Switzerland). Such a picture confirms the institutionalisation of social democrats within the respective party system across the region and their ability to remain center stage as a credible coalition partner despite the reduced electoral support. This finding conveys the idea that looking only at the electoral arena is not sufficient to discriminate properly between the different development paths of social democrats.

Consistent with the preliminary pieces of evidence emerged in , the analysis reported in confirms recent studies showing that the electoral support for social democrats has notably decreased in the Great Recession period (Benedetto et al., Citation2020; Delwit, Citation2021; Polacko, Citation2022).Footnote14 Model 1 illustrates precisely the classical ‘party lifespan’ development posited by Pedersen (Citation1982), with social democratic vote share following a curvilinear trajectory over time. Indeed, taking the ‘Golden Age’ as the baseline category, we see that this period corresponds to the electoral peak of social democracy, with a negative coefficient for the pre-1945 period. Conversely, after the peak, the electoral decline of social democracy is shown by the negative coefficients of all three subsequent phases, whose effects increase in size the further away we move in time from the ‘Golden Age’.Footnote15

Table 2. Social democratic vote share and governmental power across temporal phases.

While social democracy is clearly in crisis from an electoral viewpoint, the story changes once we look at the GPI. Indeed, in Model 2 while the ‘Pre-WWI’ and ‘Inter-war’ phases show negative coefficients compared to the ‘Golden Age’ consistent with the vote share trajectory, the subsequent periods display a pattern of fundamental stability in social democratic governmental power. In other words, while vote share has been declining remarkably since the beginning of the Great Recession, if we look at the governmental arena the power of social democrats in the 2010s is not fundamentally different from what it was in the 1960s and in the following decades.Footnote16 Moreover, the analysis clearly shows that social democratic electoral support and governmental power appear as diverging in their evolution over time, thus confirming our first expectation. Consequently, if we adopt the so far overlooked governmental perspective and hence focus on the arena that matters the most to realise parties’ political goals, we would conclude that social democracy is not in crisis.Footnote17

This finding entails an important consequence. Indeed, the relevance of social democracy within the executive has become more and more independent from its performance at the polls. Resorting to Pedersen’s (Citation1982) ‘party lifespan’ framework, whilst social democracy has passed its culmination point after the ‘Golden Age’ from an electoral viewpoint, the same has not yet occurred in the governmental sphere, where it seems to be in a substantial plateau since the 1960s. This is clearly shown by , which plots the predicted probabilities of, respectively, social democratic vote share and GPI at different phases. The divergence between the vote share and GPI trajectories might be due to the previously discussed temporal lag between the electoral and governmental arenas (e.g. Rokkan, Citation1970), which would then lead us to assume that it will only be a matter of time before social democracy enters a crisis also from a governmental viewpoint. At the same time, however, the divergence might also be determined by the progressive institutionalisation over time of social democrats within the ‘core’ (Smith, Citation1989) of Western European party systems and mainstream politics (Keman, Citation2017), which may contribute to their resilience in power for a long time despite their ascertained electoral decline.

Figure 3. Predicted levels of social democratic vote share and GPI at different points in time.

Figure 3. Predicted levels of social democratic vote share and GPI at different points in time.

Besides what lies ahead in the future of social democracy, it is now worthwhile addressing our ‘why’ research question about the increasingly inconsistent social democratic performance between the electoral and governmental arenas. reports the results of the analysis of the effects of the above-discussed contextual factors on the degree of disconnect of social democratic parties between the two arenas. Model 1 presents a baseline specification with only the categorical time variable and country-fixed effects. It confirms that, over time, social democratic parties have become more optimiser, and especially in the Great Recession period the positive disconnect between GPI and vote share (i.e. comparatively stronger GPI than vote share) is about 7.2% larger than in the Golden Age. Model 2 adds the substantive independent and control variables and shows the positive and statistically significant effect of both party system fragmentation and ideologically radical parties’ vote share.Footnote18 On the one hand, a unitary increase in the ENEP (e.g. from 3 to 4 effective parties) increases by 1.6% the positive disconnect between GPI and vote share (p < 0.01). On the other hand, a 10-point increase in the electoral support of ideologically radical parties leads to a similar substantive result (p < 0.05). As a result, both factors contribute to boosting the optimiser profile of the social democratic left, thus confirming both Expectations 2 and 3.Footnote19

Table 3. Determinants of social democratic parties’ disconnect between GPI and Vote share.

Interestingly, after the inclusion of independent and control variables in Model 2, the statistical significance of the Great Recession phase displayed in Model 1 disappears. This result suggests that the detected increase of the optimiser profile of social democracy observed in recent years is well explained by the rise of contextual factors explaining this outcome, such as the increase in fragmentation (the average ENEP grew from 4.1 to 5.6) and radical competition (the average electoral support for ideologically radical parties grew from 8.3% to 17.4%) experienced by Western European party systems in the Great recession period. Indeed, in a crowded party environment with multiple new alternatives and where ideologically radical parties exert a strong blackmail power, social democratic parties successfully succeed in maintaining their role as an established mainstream governing force in spite of their declining electoral performance. This also entails a change in the nature of social democratic government experience, as the new fragmented and radicalised context often imposes to seek compromise for power through the formation of coalitions rather than single-party minority or majority cabinets. Our data on GPI shows that the number of legislatures characterised by social democratic single-party minority or majority cabinets has declined from 39% to 25% of all instances of social democratic government experiences between the 1945 and 2009 period and the Great Recession period (2010-2022). At the same time, the number of legislatures characterised by social democrats leading a coalition government (multi-party minority, surplus, or minimum winning coalition) has risen from 28% to 46%.Footnote20

In order to delve deeper into the findings emerging from Model 2, Model 3 disentangles the variable measuring the ideologically radical parties’ vote share between radical left and radical parties so as to provide a more specific understanding of the context fostering a more optimiser social democratic left. What strikingly emerges from the analysis is that it is the presence of a strongly supported radical left formation to push the social democratic left towards overrepresentation in government compared to its vote share. While the electoral support for the radical right does not seem to play any role in the story, a 10-point increase for the radical left is associated with an increase of 3.1% of our dependent variable. In empirical terms, this result means that when a radical left party is absent or weak (e.g. < 5%), the social democratic party receives, on average, 26.9% of the vote share and a GPI of 2.4. Conversely, when a radical left party is electorally strong (e.g. > 10%), the social democratic party receives, on average, four percentage points less but 0.9 more in the GPI (22.9% and 3.2, respectively).

Interestingly, this finding shows that the presence of a radical competitor at the same end of the ideological spectrum not only – as one would have expected – reduces the electoral support for social democratic parties but also boosts its chances of maximising governmental power. Hence, the interpretation we draw is that the blackmail power exerted by the radical left competitor helps the social democratic left to play a win-win strategy in the government formation process. This is because, notwithstanding the lower electoral support compared to the past, the social democratic party may either govern together with the mainstream right in a grand coalition or lead a left-wing alternative by including its radical competitor in the coalition government.

This opportunity is a peculiarity of the mainstream left. Indeed, by replicating the analysis with the disconnect between GPI and vote share of the main conservative or Christian democratic party as the dependent variable, the presence of strong ideologically radical competitors does not exert a significant role for the mainstream right, not even in the case of radical right competitors (see Table A13 in the Online Appendix).Footnote21 Therefore, this result seems to suggest that the mainstream left has an edge over the mainstream right as far as bargaining for government formation is concerned: while radical right parties are generally excluded from negotiations due to the reciprocal vetoes of the other parties, radical left ones, at least in certain contexts (e.g. Scandinavia, Portugal, Spain) are allowed to access power and their presence indirectly helps the resilience of social democracy in government.Footnote22

Conclusion

Recently, a large number of studies have labelled social democratic parties’ performances in the last few decades by using a series of expressions that unequivocally point to their demise, such as ‘decline’ (Polacko, Citation2022; Rennwald & Pontusson, Citation2021), ‘crisis’ (Abou-Chadi et al., Citation2021), ‘fall’ (Benedetto et al., Citation2020), and ‘endgame’ (Diamond, Citation2016). All these studies share a common perspective, namely the focus on the electoral performance of social democratic parties. However, if we want to understand the evolution of political parties’ strength intended as their power to change the status quo by influencing policy outcomes, elections are only a part of the story. In this article, we have shifted the focus to a so far neglected, albeit fundamental, political aspect: social democrats’ governmental power. Our empirical analysis leads us to say that current scholarly claims emphasising the demise of social democracy are overall exaggerated. Whilst our analysis confirms that social democratic parties have been experiencing a deep electoral crisis that started in the 1970s and massively accelerated in the last years, the picture is different if we look at the governmental arena. Indeed, the article shows that social democratic parties in Western Europe are not in crisis as far as their governmental power is concerned.

This article provides three contributions to the party politics literature. First, it stresses the importance of looking at political parties’ fortunes through a more comprehensive perspective and, following Bardi and Mair’s (Citation2008) lesson, putting the governmental arena, namely the most valuable prize of party competition (Casal Bértoa & Enyedi, Citation2021; Mair, Citation1997), back into the analysis.

Second, it proposes an index of parties’ governmental power that, by taking into account different aspects of cabinet characteristics and the position of a party within the cabinet, represents an important step forward that remedies the several illustrated deficiencies of existing measures. The index has been employed in this article to measure the governmental power of social democrats but can be applied to any kind of political party or party family, thus potentially constituting a yardstick for future party politics research focusing on government.

Third, besides theoretical and methodological innovations, the article contributes to the understanding of the current evolution of social democracy. Through comparative longitudinal analysis covering a century and a half of social democratic politics (1871–2022) and more than 623 legislatures in 20 Western European countries, it shows that whilst social democratic electoral support substantively declines after the 1960s, in the governmental arena the power of social democracy is still the same as it was during the ‘Golden Age’ of class politics. Moreover, and linked to this finding, after decades of being tied together, social democratic electoral support and governmental power have taken their separate ways. In this context, our analysis of national variations shows that, especially in the last few decades, the most remarkable trend going on is a progressive turn towards what we have defined as an optimiser profile of social democratic parties. As vote shares dramatically declined, an increasing number of social democratic parties managed to ‘squeeze blood from a turnip’, being able to secure or maintain relevant government positions despite their mediocre electoral performances. Our explanatory analysis shows that the increasing disconnect between electoral support and governmental power of social democratic parties is mainly driven by party system fragmentation and ideologically radical political competition. More specifically, a crowded party environment and the presence of electorally strong radical left parties boost the chances of social democratic parties to maintain their governmental power in spite of their declining electoral support. Both such factors are on the rise in Western European party systems, which have become increasingly fragmented and characterised by strong ideologically radical parties over the last decades and especially after the Great Recession. Consequently, these party system changes have created a more favourable context for social democracy to exploit its centrality as a core institutionalised component of the party system.

To conclude, does this mean that currently all is well for Western European social democracy and it can rest assured of its governmental status? Will inertial processes such as the governmental institutionalisation of social democrats as a core component of Western European party systems prevail, hence keeping them in government despite their declining electoral performance? Or will the traditional ‘peak-and-decline’ ‘party lifespan’ trajectory perspective, which is already evident in electoral terms, be reproduced soon in the governmental arena, too? Only time will tell which of these scenarios will come to fruition. Of course, if the electoral decline of social democracy continues even further, going beyond the decisive acceleration that was recorded in the 2010s, it seems as if its governmental power is also bound to decrease over time. Indeed, when the main social democratic party gets less than 10% in general elections, as was recently the case in France, Iceland, and the Netherlands, it becomes very difficult to secure a continuous and considerable presence in the respective national governments. In other words, depending on the specificities of the national contexts, there is an electoral threshold below which being resilient in terms of governmental power would simply not be possible. At any rate, further research is needed to delve deeper into this critical puzzle, as well as to understand whether social democracy is capable of exploiting its optimiser status in government to deliver the policy outcomes requested by its constituency, especially in contexts where responsiveness towards voters’ demands is often set aside due to the increasing external economic and institutional pressures for responsibility experienced by ruling parties (Mair, Citation2009). In this regard, the findings of this article indicate the urgency to integrate the governmental perspective in the analysis of the political development of social democracy and party politics at large.

Supplemental material

Supplemental Material

Download MS Word (63.6 KB)Supplemental Material

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Vincenzo Emanuele

Vincenzo Emanuele is Associate Professor in Political Science at Luiss, Rome. His research has appeared – among others – in Comparative Political Studies, European Journal of Political Research, Perspectives on Politics, West European Politics, Party Politics, and Government and Opposition. His monograph, Cleavages, institutions, and competition. Understanding vote nationalisation in Western Europe (19652015) has been published by Rowman & Littlefield/ECPR Press.

Federico Trastulli

Federico Trastulli is a Postdoctoral Researcher at the University of Verona and a Research Affiliate at the Italian Centre for Electoral Studies, Luiss, Rome. His research interests concern political parties, voting behaviour, and the dimensions of political contestation in Italy and Western Europe from a multi-method perspective.

Notes

1 More generally, ample literature has also investigated the decline of social democracy in other dimensions, such as organizational strength (e.g. Bailey, Citation2009), linkages to trade unions (e.g. Howell, Citation2001), and the ability to deliver traditional social democratic public policies in government (e.g. Merkel & Petring, Citation2007).

2 We are, of course, aware that, on the one hand, social democratic parties have achieved important political objectives while in opposition thanks to blackmail power and extra-institutional actions, especially during the early phases of socialist mobilization (see Bartolini, Citation2000); on the other hand, being in government is not necessarily a guarantee of achieving political results favorable for the left constituency, as external economic and institutional constraints often force ruling parties to prioritize responsibility at the expense of responsiveness to voters’ demands (see Mair, Citation2009).

3 On parties’ electoral trajectories, see also Emanuele and Sikk (Citation2021).

4 When two or more parties share the same number of cabinet seats, the tie is broken by looking at the following criteria: (1) whether the Prime Minister belongs to any of the parties, which will then be the leading party; (2) if not, by considering which of these formations has the largest share of parliamentary seats.

5 The comparison between our revised GPI and Bartolini’s original version is reported in Table A2.

6 The included countries are Austria, Belgium, Cyprus, Denmark, Finland, France, Germany, Greece, Iceland, Ireland, Italy, Luxembourg, Malta, the Netherlands, Norway, Portugal, Spain, Sweden, Switzerland, and the United Kingdom.

7 Data on social democratic vote share is compiled by relying on Nohlen and Stöver (Citation2010), where available, and the pertinent electoral authority of each country. Information on cabinet composition is taken from Casal Bértoa (Citation2021) and, in case of missing data, Sonntag (Citation2015).

8 For current legislatures, the index is calculated by considering only those cabinets that were terminated by December 31, 2021.

9 The two families are labelled, respectively, ‘Communist/Socialist’ and ‘Right-wing’.

10 Data for ENEP and Disproportionality are taken from Gallagher (Citation2021) and, for earlier elections, are the results of our calculations.

11 We have also replicated the analyses through the periodization by Benedetto et al. (Citation2020). The results are reported in Table A5 in the Online Appendix and fully confirm our findings.

12 Specifically, we performed an LR test of panel-heteroskedasticity and a Wooldridge test of serial correlation. Both are significant at p < .001, confirming that panels are heteroskedastic and observations serially correlated.

13 We include country-fixed effects to detect the within-country evolution over time of left electoral support and governmental power properly and control for the entry into the sample of the younger Cypriot, Greek, Portuguese, and Spanish democracies in the 1970s.

14 Descriptive statistics are reported in Table A4 in the Online Appendix.

15 The results are substantially identical without country-fixed effects (Table A6 in the Online Appendix) and by replacing vote share with parliamentary seat share, thus accounting for the effect of the different electoral systems (Table A7 in the Online Appendix).

16 These results are fully confirmed by replacing the GPI with a measure reporting, for each legislature, the weighed share of cabinet seats held by social democratic parties (Table A8). The latter can be considered the most accurate measure of social democratic power employed by the current literature (Allan & Scruggs, Citation2004; Garritzmann & Seng, Citation2016; Jensen, Citation2010). The fact that this additional robustness test corroborates our results further confirms that they are not driven by the specific measure used.

17 Notice that the results are substantially the same by replacing the correction for autocorrelation of the Prais-Winsten specification with a lagged dependent variable (Table A9 in the Online Appendix) and considering the main left party (i.e. the left party that received the largest vote share in the election of interest) or the whole left bloc rather than the social democratic party (see Tables A10 and A11 in the Online Appendix).

18 Conversely, the disproportionality of electoral systems shows no effect. The result is analogous if disproportionality is replaced by a categorial variable assuming a value of 0 for proportional electoral systems, 1 for mixed, and 2 for majoritarian systems.

19 The result is robust by replicating the analysis with an alternative operationalization based on a dichotomous variable that takes the value of 1 when the social democratic left displays an optimizer profile (i.e. it shows a GPI above the overall mean and a vote share below the overall mean of our spatial-temporal framework), and 0 otherwise (see Table A12 in the Online Appendix).

20 The data reported refers to single-cabinet legislatures only.

21 Conversely, party system fragmentation is a strong positive predictor of the disconnect between the GPI and vote share of conservative parties. Therefore, while the presence of strong ideologically radical parties is not an important factor for conservative parties, the extent to which they show an optimizer profile is mainly a consequence of the number of parties competing in the party system.

22 However, the ongoing and thorough process of radical right normalization (e.g. De Jonge, Citation2021) is likely to alter this situation in the near future. In turn, this normalization may provide the mainstream right with a specular situation, whereby they could alternatively build government coalitions either with the mainstream left or with the radical right.

References

Reprints and Corporate Permissions

Please note: Selecting permissions does not provide access to the full text of the article, please see our help page How do I view content?

To request a reprint or corporate permissions for this article, please click on the relevant link below:

Academic Permissions

Please note: Selecting permissions does not provide access to the full text of the article, please see our help page How do I view content?

Obtain permissions instantly via Rightslink by clicking on the button below:

If you are unable to obtain permissions via Rightslink, please complete and submit this Permissions form. For more information, please visit our Permissions help page.