Left Governmental Power and the Reduction of Inequalities in Western Europe (1871–2020)

Abstract

Despite considerable attention in the literature, existing studies analyzing the effect of left governmental power on inequalities suffer from three main limitations: a privileged focus on economic forms of inequality at the expense of political and social ones, inaccurate measurements of left governmental power, and the analyses’ narrow time spans. This article addresses such concerns through a comparative longitudinal analysis where the impact of left governmental power on different measures of political, social, and economic inequalities is investigated in 20 Western European countries across the last 150 years. Data show that, consistent with previous literature, the Left in government has significantly reduced most forms of inequalities. However, the equalizing effect of the Left in government has decreased over time and has become not significant since the 1980s. The Left is today incapable of accomplishing its historical mission of reducing inequalities. The article discusses the rationale and implications of these findings.

1. Introduction

In recent years, political scientists have found a renewed interest in social democracy and class politics. Besides the endless debate on the decline, persistence, or transformation of class voting (Elff 2007; Franklin, Mackie, and Valen 1992; Oesch 2006), the literature has recently moved in different directions. First, it has focused on the recent electoral crisis of the Left (Benedetto, Hix, and Mastrorocco 2020; Delwit 2021; Emanuele and Trastulli 2023; Loxbo et al. 2019). Second, it has tested the resilience of the link between left electoral mobilization and its historical cleavage roots (Emanuele 2023; 2024). Third, it has empirically investigated the long-term debate on the alleged programmatic shift to the right of the mainstream Left (Abou-Chadi and Wagner 2020; Bremer 2018; Polacko 2022). Finally, it has looked into the composition of the left electorate to test whether its core social constituency has shifted from the working class to urban sociocultural professionals (Abou-Chadi and Wagner 2019; Karreth, Polk, and Allen 2013; Rennwald and Evans 2014).

Within this growing body of work, a comparatively underexplored aspect is the effect produced by left parties in government and its evolution over time. In particular, in view of the recent electoral crisis of left parties and their abovementioned social, organizational, and political transformations, it is worth asking whether left parties are still able to follow their “polar star,” namely to realize the historical goals that constitute the very foundation of the Left. These can, first and foremost, be summarized through a single, overarching objective: the pursuit of equality (Bobbio 1997; Lukes 2003).

Of course, the relationship between left power—which we understand as the strength of left parties in government—and the reduction of inequalities—meaning the uneven distribution of power and resources among individuals in a given society—has received great attention in the social sciences. However, these works suffer from three main shortcomings. First, as this literature mainly comes from fields outside comparative politics (e.g., international political economy and economic sociology), it has almost exclusively focused on economic forms of inequality, thus neglecting the multidimensional nature of this concept. Second, it has often employed inaccurate proxies of the focal predictor—the power of the Left—which are at risk of being misleading in effectively conveying the underlying concept. Third, it has mostly covered a narrow time span, generally a few decades toward the end of the twentieth century. Consequently, it lacks a comparative historical perspective, which is crucial to both gauge the long-term evolution of the impact of left power on the pursuit of equality and capture the potential consequences of recent political, social, and economic transformations in this relationship.

This article aims to fill these gaps in this strand of literature and contribute to this renewed interest in left politics by asking a very simple, albeit fundamental, research question: does the Left in government reduce inequalities? This question is addressed through comparative longitudinal analysis and a time-series cross-section dataset, taking into account 20 Western European countries and more than six hundred legislative terms from 1871 to 2020. Besides its extended spatial and temporal scope, additional added values of this article are the operationalization of our focal predictor—left governmental power—through a measure that is more fine-grained compared to the existing alternatives and the disentanglement of our explanandum through political, social, and economic indicators of inequality.

We find that, in line with theoretical expectations and the bulk of existing evidence, the Left in government has historically reduced inequality, although we show that this is truer in some areas (i.e., access to political power according to socioeconomic position and social group, educational and health disparities, and welfare state universalism) than in others (access to political power according to gender and income inequality). Yet, the most striking finding of our analysis is that the equalizing effect of the Left in government has decreased over time and has become not significant since the 1980s. Moreover, the analysis also shows that in the last 40 years, the Left in government has become indistinguishable from the Center-Right in its (in)ability to reduce inequalities. Therefore, in recent decades, it seems that the Left has lost its raison d’être, as it has no longer been able to uphold its defining principle of reducing inequalities through government action.

The article is structured as follows: the next section critically reviews the literature on the relationship between left power and equality and introduces our hypotheses; the third section introduces our focal predictor, left governmental power, while the fourth section focuses on the operationalization of the dependent variables, namely the measures of political, social, and economic inequalities; the fifth section presents the control variables and the method to be employed in the analyses; the sixth section shows the empirical results and, finally, the concluding section discusses the article’s findings and implications.

2. Theoretical Background and Hypotheses

The mission of the political Left is described by Norberto Bobbio (1997, 96) as “the destruction of a despotic regime founded on inequality between those at the top and those at the bottom of the social scale, perceived as unjust precisely because it is inegalitarian and constituted hierarchically.” Generally speaking, the view that the Left pursues social change in an egalitarian direction to rectify socially determined disparities is widely shared in the literature (Anderson 1998; Bartolini 2000; Lukes 2003; Noël and Thérien 2008). This is because of the Left’s conception of human nature, whereby what makes people similar is greater than what sets them apart (Bobbio 1997; White 2011). Hence, in political contexts determined by electoral competition, the Left seeks power to be in a position to actively pursue this goal. It is, then, legitimate to ask and investigate whether this goal is actually achieved. Of course, the capacity of left parties in government to pursue this objective is conditional on several contextual factors, from international security and stability to the current state of the economy, to external and internal economic and institutional constraints (Bircan, Brück, and Vothknecht 2017; Fernández-i-Marín et al. 2023; Jahn 2018; Mair 2009). In particular, seminal works underline that weak labor-market institutions, the lack of institutionalized collective bargaining, and the maturation of previous welfare commitments negatively impact the capacity of partisan actors to pursue egalitarian policies (Garrett 1998; Pierson 2001).

The link between the Left and the reduction of inequalities has been analyzed by several works, especially within the strand of “power resource theory” (for an overview, see Brady, Blome, and Kleider 2016). However, this literature has evident shortcomings. ¹ For a start, substantively, only a few of these works specifically look at the impact of left power on the reduction of inequalities as their core focus (Garritzmann and Seng 2016; Kauder and Potrafke 2013) or as part of it (Busemeyer 2009; Hewitt 1977; Jensen 2010). Further, such contributions most often rely on rather idiosyncratic research designs, especially in terms of the measurement of left power, the selection of countries, and the (usually short) time span covered, making an overall historical assessment of the Left’s impact on inequality difficult. Finally, albeit from a theoretical viewpoint, equality is a multidimensional concept that also includes political and social aspects (see, e.g., McKnight, Mendes Loureiro, and Vizard 2019), yet this literature has almost exclusively focused on economic forms of inequality. This translates into widespread neglect of the Left’s impact on social and especially political forms of inequality.

Indeed, political inequality is only considered marginally within the theoretical strand of power resource theory. Studies explicitly investigating the link between left power and political aspects of inequality are limited to assessments of the positive effect that the parliamentary strength of left parties has on women’s access to elected office (Kenworthy and Malami 1999), hence substantively overlapping with the existing contributions on gender inequality at large (e.g., Avdeyeva 2009; Esping-Andersen 1993). However, the relevant literature does not provide insights into the association between left power and equality of access to political power according to socioeconomic status and social group, including ethnicity, race, and religion. This is a theoretical gap that this article aims to fill.

Existing research is also limited in its assessment of how the Left impacts social forms of inequality. With regard to education, the Left is generally found to expand educational access once in power (Ansell 2008; Braga, Checchi, and Meschi 2013; Busemeyer 2009; Hewitt 1977; Kauder and Potrafke 2013). Whether an expansion of education spending occurs when the Left is in power is instead debated: while some argue that this is the case (Busemeyer 2009; Schmidt 2007), others contend that the impact of left power on education spending is unclear (Ansell 2008; Jensen 2011) or absent (Garritzmann and Seng 2016). Beyond this form of inequality, little research exists on the effect of left power on healthcare, with available works (e.g., Huber and Stephens 2000) indicating that the stronger the Left, the lower the health inequalities.

Finally, as already mentioned, most of the available contributions focus primarily on economic inequalities. Within this area, existing works distinguish between inequality of access to welfare provisions and income inequality. The Left appears to have had a positive effect on welfare and social expenditure expansion in OECD countries (Allan and Scruggs 2004; Huber and Stephens 2000; Iversen and Soskice 2015; Jensen 2010), while on income inequality evidence from the literature is more mixed. Several studies (e.g., Beramendi and Cusack 2009; Bradley et al. 2003; Huber and Stephens 2014; Iversen and Soskice 2006; Pontusson, Rueda, and Way 2002; Rueda and Pontusson 2000) highlight a direct and positive effect of left power on the reduction of income inequality. Yet, many contributions (e.g., Hicks and Kenworthy 2003; Lupu and Pontusson 2011; Mahler 2010; Morlino 2020; Rueda 2008; Yi 2013) challenge this side of the debate, arguing that there is no significant effect of left power on income inequality. A prime example is provided by Kristal (2010), who demonstrates that left cabinets have a negative short-term effect on labor’s national income share while having no long-term effect at all.

Overall, despite some mixed findings within the literature, our general theoretical expectation is that the power of the Left, based on its whole political history and consistent with its mission (Bobbio 1997), is positively associated with the reduction of political, social, and economic inequalities. Hence, we can derive the following hypothesis:

H1: The higher the governmental power of the Left, the lower the political, social, and economic inequalities.

Besides this general hypothesis, we are also interested in understanding whether the relationship between left power and inequalities has changed over time. This is because, during the last 150 years, since the class cleavage was first politicized through the mobilization of the working class (Lipset and Rokkan 1967), left parties and the surrounding political, social, and economic context of Western European societies have experienced remarkable changes. Once in government, the Left sought to build more egalitarian societies, mainly by constructing inclusive welfare state structures. However, especially in recent decades, a number of phenomena that could potentially transform this relationship have emerged.

First, left parties, and particularly social democratic formations, have recently experienced a dramatic electoral decline (Benedetto, Hix, and Mastrorocco 2020; Delwit 2021; Loxbo et al. 2019). For instance, just to mention one of the most astonishing cases, the vote share of the French Socialist Party (PS) collapsed from 29.4% to 7.4% in the 2017 legislative elections. Similar record-low results have recently occurred in Italy, the Netherlands, Sweden, and Switzerland. This electoral crisis has resulted in reduced access to government for the Left and, therefore, may have restricted its opportunities to produce egalitarian policies.

Second, the massive social and economic development experienced by Western European societies during the twentieth century may have led to a sense of accomplishment vis-à-vis the historical mission of the Left. Indeed, the long march toward equality in its various forms may have today reached such high standards that there may be little room for further improvements. For instance, as recently shown by Alvaredo and colleagues (2018, 9–11), since the second half of the 1990s, Europe has become the world’s least unequal region in economic terms.

The third potential phenomenon contributing to undermining the relationship between left power and the reduction of inequalities regards the increasing external constraints deriving from membership in supranational institutions and embeddedness in globalized markets. These factors put governing parties under tremendous pressure from the “responsibility vs. responsiveness dilemma” (Mair 2009). The increasingly limited discretion available to national governments on social and economic policy has led left parties in government to privilege responsibility toward external constraints at the expense of responsiveness toward their traditional constituency and, therefore, to assume more centrist economic positions. This positional shift has meant that the Left has moved away from its traditional goals in a convergence with the mainstream Right (Evans and Tilley 2017; Karreth, Polk, and Allen 2013; Mair 2008; Polacko 2022).

Additionally, we must also consider the possibility that not only the ability of left parties to reduce inequalities but also their willingness to do so has gradually waned over time. Indeed, the structural and behavioral dealignment constituted by the shrinking and transformations of the working class and the emergence of postmaterialist issues and values cutting across traditional class loyalties (Best 2011; Goldberg 2020) have meant that, in order to remain electorally competitive, left parties have had to progressively shift their reference constituency away from industrial workers to sociocultural professionals, and their programmatic focus away from traditional left goals (e.g., Abou-Chadi and Wagner 2019; Oesch and Rennwald 2018). ² This transformation may have also been reinforced by the decline in size and political relevance of unions (e.g., Rennwald and Pontusson 2021), which traditionally compel left parties to pursue a leftist policy agenda (Garrett 1998; Piazza 2001). Based on such considerations, we expect that the impact of left power on inequalities decreases over time. As a result, our second hypothesis is as follows:

H2: The impact of left governmental power on inequalities decreases over time.

3. Left Governmental Power

Moving toward the empirical part of the study, the first task is to outline and measure our focal predictor, namely left governmental power. To do so, a preliminary step is defining what the political Left is and which political formations can be considered left parties. As mentioned in the introduction, we focus on 20 Western European countries through a comparative longitudinal perspective. ³ Indeed, we go back in time to include all cases in which left parties contested elections in Western Europe since the class cleavage was first politicized. While many studies on the topic suffer from short-termism, this choice avoids the error of “judg[ing] contemporary politics by the exceptional standards of mid-20th century Western European politics” (Enyedi and Deegan-Krause 2007, 13). Moreover, the extended temporal scope of this article allows for assessing the long-term evolution of the impact of left power on the reduction of inequalities and its changes over time. The first observation considered is Germany in 1871, and the last one is Ireland in 2020. ⁴

Consistent with this long-term perspective, we consider left parties to be the historical communist, socialist, social democratic, and labor parties that originally emerged to mobilize the working class (Bartolini 2000; Bartolini and Mair 1990). Moreover, we also carefully assess the inclusion of these parties’ direct successors or of new parties emphasizing traditional left issues. This choice has been made to ensure the consistency of the case selection throughout the longitudinal time span of the analysis. This is why we do not consider parties of the so-called “new politics” emphasizing issues like environmentalism, feminism, and civil rights (Müller-Rommel 1989; Poguntke 1987), which have emerged since the 1970s. ⁵ For this classification, we rely on previous studies by Bartolini and Mair (1990) and Emanuele (2023; 2024). The list of 114 left parties resulting from this effort is reported in table A1 in the appendix. ⁶

After clarifying which parties can be considered as part of the left bloc, the subsequent and far-from-straightforward step is to measure the governmental power of the Left. The literature on power resource theory does not always focus directly on governmental power. Indeed, scholars looking at the power of left parties rather than other left organizations (e.g., trade unions, social movements) use various measures with varying degrees of accuracy and sophistication. For instance, some focus on the share of parliamentary seats (Braga, Checchi, and Meschi 2013; Hewitt 1977), which is certainly far from a satisfactory measure, as it tells us nothing about whether left parties are in government and, therefore, can shape policy outcomes. Most scholars, however, do look directly into the governmental arena. Some do so by using Cusack’s (1997) index of cabinet ideological balance, which provides a synthetic score of a cabinet’s overall ideological position (Pontusson, Rueda, and Way 2002; Rueda and Pontusson 2000; for other versions of Cusack’s index, see also Beramendi and Cusack 2009; Iversen and Soskice 2015; Lupu and Pontusson 2011), while others consider whether the Left is in government or not and then weight their power by their share of parliamentary seats vis-à-vis other cabinet partners (Bradley et al. 2003; Huber and Stephens 2000; Kristal 2010). Finally, the most sophisticated measure employed by the literature so far is the share of cabinet seats held by left parties (Allan and Scruggs 2004; Garritzmann and Seng 2016; Jensen 2010). While this is certainly more accurate compared to the previously mentioned alternatives, even this measure does not properly convey either the cabinet’s status in parliament or a party’s status in the cabinet. Indeed, on the one hand, even holding all cabinet seats cannot tell us whether a party governs as a single-party majority or as a minority government. On the other hand, if the left party holds less than 50% of cabinet seats, it is not possible to discern whether the party has a junior or leading status in cabinet because, for instance, a left party with 40% of the seats may be either a leading partner of a multiparty cabinet or a junior partner alongside another party holding the remaining 60% of cabinet seats. Both instances show that even the most advanced among the existing measures miss important information about the effective power the Left can exert over policy outcomes. Overall, we believe that, to different extents, all such operationalizations are unsatisfactory. Hence, to measure the power of the Left, we turn to Bartolini’s (1998) governmental power index. This instrument measures the power of a party (or a bloc of parties) at a given time point (e.g., a given year or legislative term) by considering different characteristics. First, it considers the government status of the party. Here, it distinguishes not only between cabinet and opposition status, but also considers situations in which the party provides an abstention or external support that is necessary for the survival of the minority cabinet. Second, it takes into account the cabinet’s status in parliament. Here, the index considers whether the cabinet has a majority status and whether the party governs alone or in a coalition. Third, it assesses the status of the parties in cabinet. Here, Bartolini’s index looks at the parliamentary seats of cabinet partners to discriminate whether a party leads the cabinet, has an equal standing with others, or is a junior coalition partner. Furthermore, Bartolini considers the additional aspect of cabinet duration. This is relevant only when legislative terms or longer time periods are used as units of analysis in order to weigh the governmental power of each cabinet by its duration within the legislative term. Building on these elements, Bartolini proposes an index ranging from 0 to 13, thus providing a rank-ordering of governmental power across all possible combinations of cabinet characteristics and the role of parties within them.

To the best of our knowledge, besides Bartolini’s (2000) initial descriptive account of the governmental power of the class bloc, this index has never been applied empirically in subsequent works, probably because of the very time-consuming effort required for the data collection process. Despite this hurdle, we believe that this endeavor is worthwhile to give a more accurate account of left power in Western Europe.

We provide a refinement of Bartolini’s original measure. Our index ranges from 0 (when the Left is in opposition) to 10 (when the Left holds a single-party majority cabinet) as we eliminate the original measure’s scenario in which the left party has an equal standing vis-à-vis other governing partners. Indeed, we look at parties’ share of cabinet seats—which we believe is a more precise assessment of a party’s status in cabinet than parliamentary seats—and, in cases where two parties hold the same share of seats, rather than attributing an equal status to both, we consider the following criteria to break the tie and distinguish between leading and junior partners: if the prime minister belongs to one of these parties, we deem that party to be the leading partner; otherwise, if the prime minister does not belong to either party, we deem the leading partner to be the one with the larger share of parliamentary seats. ⁷ Our governmental power index (GPI) is calculated at the level of legislative terms and can be applied to both single parties or blocs of parties: in this case, the left bloc. ⁸

Table 1 reports the GPI value for the different combinations of the three aforementioned aspects (government status of the party, cabinet status in parliament, and party status in cabinet). Therefore, the GPI of a given cabinet is 0 if the Left is in the opposition; 1 or 2 if the Left provides, respectively, necessary abstention or necessary external support to a minority cabinet; 3 to 10 if the Left is in government, according to the different combinations of cabinet status (i.e., multiparty minority, single-party minority, surplus coalition, minimum winning coalition, and single-party majority) and party status (i.e., junior or leading), with higher values attributed to majority governments and leading statuses, and 10 corresponding to a single-party majority government. The index is first calculated for a given cabinet and then directly attributed to the respective legislative term if the latter coincides with the former. Conversely, if there are multiple cabinets in a legislative term, the GPI is obtained by weighting the score of each cabinet by its duration (in number of days) within the term. As a result, the GPI is a continuous measure ranging from 0 to 10.

Table 1 Governmental Power Index: Dimensions and Values

Source: Adapted and revised from Bartolini (1998, 48–49).

The striking piece of evidence we draw from figure 1—reporting the evolution over time of the GPI—is that, since the initial politicization of the class cleavage, the overall majority of legislative terms in Western Europe (330 out of 616) are characterized by the absence of left governmental power, as the left bloc was in opposition for most of the period considered. The first instance of a GPI that is different from 0 is recorded in France in 1906, but only in the interwar period do we see a noticeable improvement of the index. The march of social democracy toward power accelerated just after World War II and reached its heyday between the 1960s and the 1990s, as shown by the locally weighed estimation fit of the GPI. Under the programmatic formula of the “Third Way” (Giddens 1998), social democratic parties achieved unprecedented success, and the index recorded its highest peak precisely in the 1990s (4.73 on average), before declining in the following two decades. In the 2010s, the power of the Left decreased to the lowest point since the 1930s (3.45).

Figure 1 Evolution over Time of the GPI (Left)

However, compared to recent accounts about the electoral fall of left parties in Western Europe (Benedetto, Hix, and Mastrorocco 2020; Delwit 2021; Polacko 2022), at the governmental level the retrenchment is less pronounced so far. This is because, as the trajectory of political parties usually follows a sequence of phases with specific thresholds, from electoral contestation to governmental power (Rokkan 1970), the access to the executive for left parties represented the last step of a long developmental process. At the same time, once they achieved power and consolidated as part of the “core” of their respective party systems, left parties (particularly social democratic parties) started to benefit from a favorable position that allowed them to hold more power than they deserved based on their recent electoral fortunes. Indeed, even in the 2010s, left parties have been in power in all Western European countries. ⁹

4. The Dependent Variable(s): Measuring Varieties of Equality

In the operationalization of the dependent variable, our approach entails both a multidimensional conception of equality and a comparative longitudinal perspective. As a result, we need to look at equality from different viewpoints and employ data covering a long temporal span, which is crucial to address our research question. However, on the one hand, the literature has mostly focused on economic inequality by specifically looking at different measures of income inequality (e.g., the Gini index and 90:10 or 90:50 income distribution ratios; the share of national income held by the top decile or percentile; annual labor income share as a percentage of the gross domestic product; poverty reduction after tax and transfers), thus overlooking political and social forms of inequality. On the other hand, so far, this literature has often relied on the same source, the Luxembourg Income Study (LIS), which, despite being accurate, provides data starting only from 1968 and lacks information for Cyprus, Malta, and Portugal (e.g., Bradley et al. 2003; Iversen and Soskice 2006; Mahler 2010).

On these bases, we contend that there is a need for long-term measures of political, social, and economic inequalities. Despite some clear empirical limitations, the Varieties of Democracy (V-Dem) dataset (Coppedge et al. 2019) is, to the best of our knowledge, the only available source that provides this kind of data. ¹⁰ Therefore, we employ three indicators for political inequality, pointing to how political power is distributed in a given society across different categories, namely socioeconomic position, social group, and gender (see Coppedge et al. 2019, 190–91). These are interval variables coded by country experts with a specific assessment scheme and range from 0 to 4. The first, power equality by socioeconomic position, measures the extent to which wealthy people monopolize political power. In more detail, the variable ranges between wealthy people enjoying a virtual monopoly on political power and wealthy people having no more political power than those whose economic status is average or poor. The second, power equality by social group, measures the extent to which one social group (ethnic, racial, linguistic, or religious) monopolizes political power. Here, at one extreme, we find a situation in which one social group comprising a minority of the population monopolizes political power, and at the other, a situation in which all social groups have roughly equal political power. The third, power equality by gender, measures the extent to which men monopolize political power. It ranges between men having a near monopoly of political power and men and women having roughly equal political power.

Furthermore, we also look at two indicators of social inequality by considering the level of educational and health equality in a given society (Coppedge et al. 2019, 192–93). Once again, both are country-experts-based interval variables ranging from 0 to 4. Educational equality considers the percentage of children receiving low-quality education and ranges from a situation in which at least 75% of children receive such low-quality education that it undermines their ability to exercise their basic rights as adult citizens, to a situation in which only less than 5% of children are in this scenario. Health equality considers the percentage of citizens excluded from quality healthcare. Here, at one extreme, at least 75% of citizens receive poor-quality healthcare, which undermines their ability to exercise their basic rights, and at the other, less than 5% of citizens are in this scenario.

Finally, for economic inequality, we take into account two indicators. The first, welfare universalism, also derived from V-Dem (Coppedge et al. 2019, 150), measures the extent to which welfare provisions tend to be universalistic. It is a country-experts-based interval variable ranging from 0 to 5, namely from a situation in which there are no or extremely limited welfare state policies to an opposite situation in which almost all welfare state policies are universal in character. The second indicator is a widespread measure of income inequality, namely the Gini index after taxes and transfers. The data collection related to this variable was rather challenging, given that there is no single source covering the whole analyzed time span, and there are also some inconsistencies between the values attributed to a given observation among different sources. Therefore, we collected data from six different publicly available databases (World Income Inequality Database, OECD, LIS, World Bank, World Inequality Database, and Clio Infra). When more than one source provided data for a given observation, we simply took the average of the different sources. ¹¹

Figure 2 plots the evolution over time of the different measures of equality introduced above. They are reproduced in three separate charts due to the different range of the variables. ¹² This visual representation tells a very clear story: the last 150 years have represented a long march toward the reduction of inequalities in Western Europe. To different extents, all indicators move toward the more egalitarian end of their respective scales. Of course, despite the common path, the reduction of inequalities has followed different timings and tempos according to various factors. The increase in educational and health equality has been more rapid and intense compared to the other types of equality, so much so that today both measures have almost reached the upper limit of their respective indices. Conversely, there is still scope for improvement in other varieties of equality, such as gender, socioeconomic position, and income (as illustrated by the Gini index). ¹³

Figure 2 Evolution over Time of Different Measures of Equality

Overall, this picture seems to reject the idea that the pursuit of equality is a mission that has been accomplished in Western Europe and that, therefore, the Left is unable to significantly reduce inequalities today simply because there is no room for maneuver. Indeed, only in the case of educational and health equality would it be acceptable to talk about the Left having achieved a very satisfying outcome, as all other indicators are still far from the upper end of their indices; as such, the Left still has substantial opportunity to intervene via government policy to make national societies more equal.

5. Controls and Method

The relationship between left governmental power and the reduction of inequalities in Western Europe will be tested by controlling for some factors that are either deemed relevant by the literature on the topic or may constitute relevant economic, social, and institutional constraints. For a start, periods of economic shocks may hinder the ability of left parties in government to rectify inequalities, as economic downturns reduce the opportunity for redistribution (Jahn 2018). Therefore, we include a dichotomous variable, economic shocks, that assumes a value of 1 if at least one year in the legislative term recorded a GDP growth rate of −5% or below. Data are taken from the latest available version of the Maddison Project Database, which is the only available source for longitudinal GDP data that covers our spatial-temporal framework. ¹⁴ Armed conflicts are also an important potential confounder to account for, as they have been shown to negatively impact inequalities (Bircan, Brück, and Vothknecht 2017). This is why we include a dichotomous variable, periods of war, which assumes a value of 1 if, in at least one year of the legislative term, the country is directly engaged in an armed conflict. ¹⁵

Additionally, we include two controls concerning the relationship between political and societal actors: corruption and clientelism. We deem that political corruption may exert a powerful constraint on the ability of governments to be responsive toward citizens’ demands and to deliver effective public policies (see also Houle 2018). More generally, high levels of political corruption may curtail the capacity of governments to change the status quo, a crucial prerequisite for left parties aiming to accomplish their flagship goal. In a similar vein, a context of pervasive clientelism is conducive to personalistic, client-based policy outcomes rather than an approach oriented toward the collective benefit. With this in mind, we control for the political corruption index and the clientelism index, both provided by V-Dem (Coppedge et al. 2019). Both are continuous measures: in the former, high values indicate more corrupt societies by considering corruption in the public sector, executive, legislature, and judiciary; in the latter, high values indicate more pervasive clientelistic relationships by means of targeted and contingent distribution of resources (e.g., goods, services, jobs, or money) in exchange for political support.

Furthermore, we also control for some institutional constraints. First, many scholars state that the electoral system plays a role for two main reasons. On the one hand, Mahler (2010) shows that postelectoral coalition governments, the typical outcome under nonmajoritarian systems, have a higher level of efficacy compared to pre-electoral coalition agreements, typical of majoritarian or majority-assuring systems. On the other, the electoral system affects the likelihood of a left government, which decreases as the system becomes more disproportional (Iversen and Soskice 2006; Jensen 2010). As a result, we control for Gallagher’s (1991) least-squares index of disproportionality under the assumption that the larger the proportion of the dispersed and unrepresented vote, the lower the chances of the Left being in government. ¹⁶ Second, we take into account the process of enfranchisement, namely the progressive enlargement of suffrage. The latter can be considered as an institutional precondition for left parties’ governmental power: this is because, as the working class is traditionally the core constituency of the Left, the electoral success of the latter is severely limited if the former is not entitled to vote (Bartolini 2000). Therefore, we include an ordinal variable (suffrage) that is 0 for restricted suffrage, 1 for universal male suffrage, and 2 for universal suffrage. Data comes from Bartolini (2000). Third, as external constraints deriving from membership in supranational institutions such as the European Union (EU) allegedly force governing parties to prefer responsibility even at the expense of responsiveness (Mair 2009), we include an ordinal variable (EU constraints) assuming a value of 0 if the country is not a member of the EU in a given legislative term, 1 if it is a member of the EU but did not adopt the Euro, and 2 if it also adopted the Euro. Fourth, a further institutional constraint comes into play when the Left’s egalitarian policies are to be implemented: the extent to which public administration is rigorous and impartial in the process of law implementation. The capacity of public administration to implement and enforce the law is a crucial condition in the passage from policy output to effective social transformation. Following Fernández-i-Marín and colleagues (2023), we take the rigorous and impartial public administration measure from V-Dem. This is a continuous variable where low values indicate that the law is not respected by public officials and arbitrary or biased administration of the law is the norm, and high values indicate that the law is generally fully respected by public officials and arbitrary or biased administration of the law is very limited (Coppedge et al. 2019, 162–63).

Finally, we include a trend variable, operationalized as the starting year of the legislative term (year), as typically done in time-series analyses (Roberts and Wibbels 1999). This variable will also allow us to test our H1 in relation to the change over time in the impact of left governmental power on the reduction of inequalities. Table A3 in the appendix reports the descriptive statistics of our dependent and (unlogged) independent variables.

We test our models by taking into account that we have a time-series cross-section dataset with repeated observations over time (legislative terms) on the same fixed units (countries). More specifically, we have a time-serially dominant pool (Stimson 1985), where temporal units (on average, 29.8 legislative terms per country) are more numerous than cross-section units (20 countries). In this context, three kinds of issues may arise, namely heteroskedasticity, autocorrelation, and unobserved heterogeneity. Our diagnostic tests confirmed the presence of all such issues. ¹⁷ We tackle these by running our models through panel-corrected standard errors (PCSE) regression with lagged dependent variables and country-fixed effects (Beck and Katz 1995). This method has been by far the most widely used in the literature dealing with our topic (see, among many, Allan and Scruggs 2004; Kristal 2010; Lupu and Pontusson 2011; Rueda 2008). In our data, we test the impact of the GPI of the left bloc in a given observation, say the 2013–18 Italian legislative term, on the various indicators of equality measured at the beginning of the subsequent term, say Italy in 2018 (Emanuele and Trastulli 2024). ¹⁸

6. Analysis and Results

To test our first hypothesis about the impact of left power on equality, we ran seven separate models where our indicators of political, social, and economic equality were regressed on the GPI for the left bloc. The results of these “baseline” models, controlling only for the lagged dependent variable and country-fixed effects, are displayed in table A4 in the appendix. The effect of GPI is always positive and statistically significant for the V-Dem variables, while it is negative and significant for the Gini index, where high values indicate higher inequality. Therefore, the preliminary piece of evidence we draw from this analysis is that, regardless of the specific indicator used, the governmental power of the Left has been positively associated with the pursuit of equality. In the aggregate, since the end of the nineteenth century, the story of left parties and socialism appears to be inextricably linked with the historical reduction of inequalities in Western Europe. Moreover, the presence of country-fixed effects ensures that the effect is robust across national specificities: in each country, the higher the governmental power of the Left, the higher the political, social, and economic equality.

Table 2 reports the full regression models with all the control variables introduced before, while figure 3 shows the coefficient plot that compares the effect of the GPI across the different measures of equality. We find that most of the indicators of equality (five out of seven) are still significantly affected by the GPI of the left bloc. The full models show that left power has a positive impact on the extent to which political power is equally distributed among socioeconomic positions and social (i.e., ethnic, racial, linguistic, and religious) groups, on the share of the population that receives high-quality education and health, and the extent to which the welfare state is based on universal measures. More specifically, the coefficient of the GPI is significant at p < 0.001 for educational equality and at p < 0.01 across all the other models except the one regarding power equality by social group, where the GPI is significant at p < 0.05. By looking at the coefficient plot of figure 3—all dependent variables are standardized, so the effects of the GPI across models can be compared—we can see that left power exerts a slightly stronger effect on power equality by socioeconomic position and educational equality. Conversely, after the inclusion of control variables, the effect of left power on power equality by gender and income distribution disappears. ¹⁹ More specifically, in light of these results, it might be argued that leveling the playing field in terms of political power, education, health, and welfare is politically less contentious than redistributing income, which instead is a divisive and politically adversarial measure likely to be perceived as a zero-sum game and also subject to stronger external constraints from financial and economic actors at the supranational and international levels (e.g., Mair 2009). ²⁰ These results are fully confirmed by replacing the GPI with a measure reporting, for each legislative term, the weighed share of cabinet seats of the left bloc (table A12). As discussed above, the latter can be considered the most accurate measure of left power employed by the current literature (Allan and Scruggs 2004; Garritzmann and Seng 2016; Jensen 2010). The fact that this additional robustness test corroborates our results further confirms that they are not driven by the specific measure of left power used.

Table 2 Governmental Power Index (Left) and Measures of Equality: Full Models

Notes: PCSE regressions with lagged dependent variable and country-fixed effects; PCSEs are reported. +p < 0.10; *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 3 Coefficient Plot of the Regressions between Measures of Equality and GPI (Left)

Notes: For the sake of comparability, all dependent variables have been standardized, and income inequality (Gini index) has been inverted so that high values mean higher income equality, consistent with the other dependent variables. Control variables, lagged dependent variables, and country-fixed effects not shown.

Overall, H1 is mostly confirmed, as the higher the governmental power of the Left, the lower the (political, social, and economic) inequalities. Our multidimensional operationalization of inequalities has allowed us to specify the exceptions to this finding and has confirmed the importance of providing a multifaceted assessment of the investigated relationship. ²¹

It is also interesting to notice that, among the control variables, none has a consistent effect across models, as different groups of factors exert significant effects on the various measures of equality. ²² This result further confirms the underlying multidimensionality of equality, where its components are each explained by a different set of determinants. ²³ Moreover, for all the measures of equality for which a positive trend over time could be detected at the descriptive level (see figure 2), such an increase over time seems to be absorbed by the impact of left governmental power, as the control for time (year) appears not significant.

Assessing that left governmental power has a significant effect on the reduction of inequalities does not necessarily mean that the Left really makes a difference when it comes to power. Indeed, given that levels of inequality have been massively reduced over the course of the last 150 years, as shown in figure 2, it might be the case that left and right parties have similarly contributed to this trend. To control for this potential occurrence, we have calculated the GPI of the center-right bloc, and we have rerun the regression models with the GPI for this bloc as the focal predictor. ²⁴ The results are displayed in table A17 in the appendix. We find that the GPI of the center-right bloc always shows a consistently negative association with the various measures of equality, although the coefficients are never significant at p < 0.05. This means that the Center-Right has not significantly increased inequalities in Western Europe, but at the same time has not contributed to reducing them either. Therefore, this analysis clarifies that, in the aggregate period including the last 150 years, there has been a sharp difference in how left and right parties in government have shaped inequality.

Besides the overall assessment of the effect of left governmental power on political, social, and economic inequalities, we are also interested in testing how this relationship varies over time, under the expectation that the impact of the Left’s governmental power on inequalities has decreased in recent decades (H2). As discussed in the theoretical section, different factors may have contributed to the lower capacity of the Left to pursue its flagship goal. To test H2, we rerun all regression models in table 2 with the addition of the interaction between the GPI and our trend variable (year). The full regression table is reported in table A18 in the appendix. The analysis shows a negative and significant interaction for all measures of equality that were significantly influenced by the GPI of the Left in table 2 except for power by social group, which means that, overall, the impact of left power on the reduction of inequalities has decreased over time. This temporal change is particularly prominent in the case of power by socioeconomic position and welfare state universalism (in both cases, the interaction coefficient is significant at p < 0.001).

Following Berry, Golder, and Milton (2012), to provide a correct interpretation of the results in the case of interaction models, it is always better to look at the marginal effect plots. The marginal effects of the GPI for the left bloc on various measures of equality across time are graphically reported in figure 4. What strikingly emerges from the chart is that not only does the impact of the GPI decrease over time (H2 is confirmed), but also its marginal effect crosses the zero line in all plots, thus indicating that the relationship becomes not significant at a certain time point. More specifically, since the 1980s, left governmental power has ceased to be statistically associated with the reduction of inequalities in Western Europe. The figure shows that the decline over time is common to all indicators of equality that displayed a significant GPI impact, even for power equality by social group, although the latter shows a less steep negative slope compared to the other measures, especially in the case of power equality by socioeconomic position and welfare state universalism. ²⁵

Figure 4 Marginal Effects of GPI (Left) on Various Measures of Equality across Time

Note: For the sake of comparability, all dependent variables have been standardized, and income inequality (Gini index) has been inverted so that high values mean higher income equality, consistent with the other dependent variables.

In general, the ability of the Left to exploit its governmental power to pursue egalitarian policies was very high at the beginning of the period under analysis, namely at the time of the first left governments in the initial decades of the twentieth century. Then, it progressively decreased until it completely faded at the time of the fall of the Berlin Wall.

As a final point, in light of this finding, we have replicated the same test for center-right parties to see whether, in recent decades, left and right parties have behaved similarly, meaning that both have not contributed to reducing inequalities. As table A19 in the appendix shows, the marginal effect of the GPI of the center-right bloc is never significant. ²⁶ This demonstrates that, since the 1980s, left and right parties have become indistinguishable from one another as far as the politics of inequality is concerned.

7. Discussion and Conclusion

This article has investigated the relationship between left power and inequalities in Western Europe. Despite the considerable attention it has received in the literature, existing studies have suffered from a limited temporal span, the use of varied and often inadequate measures of left power, and a bias toward economic forms of inequality at the expense of political and social ones. We have addressed these limitations by introducing three innovations. First, to provide a comprehensive account of this relationship and these changes over time, we have put forward a comparative longitudinal analysis by examining the effect of left power on various forms of inequality in 20 Western European countries over the last 150 years. This broad temporal span embraces the whole political history of the Left in the region and allows us to test whether and how the hypothesized relationship changes over time. Second, by refining Bartolini’s (1998) original proposal, we have presented an index of left governmental power that returns, in a given legislative term, the power exercised by left parties within their respective national governments. The index captures the “effective” power of the Left more adequately than the existing alternatives presented by the literature, as it considers not only the presence or absence of left parties within the executive but also the cabinet’s status in parliament and the status of the party or parties in cabinet, as well as the duration of each cabinet within any given legislative term. Third, we have accounted for the multidimensional nature of inequality by employing different measures of this phenomenon from political, social, and economic viewpoints.

Our time-series cross-section analysis has shown that, overall, the Left in government has historically reduced inequality, in line with prior expectations. However, while such a relationship holds for all measures of inequality in our baseline models, when we also control for potential confounders, the picture becomes more nuanced. Indeed, the positive effect of left governmental power on the reduction of inequalities is confirmed for most aspects (political power of socioeconomic classes and social groups, education, health, and welfare), but not for all of them (women’s political power and income distribution).

Moreover, the most interesting and original finding of our empirical analysis concerns the change over time in the hypothesized relationship between the power of the Left and the reduction of inequalities. That is, we find a consistent trend across the various measures of inequality that, albeit with different rates of change, all point to the same conclusion: the equalizing effect of the Left in government was very high in the initial decades of the twentieth century and then progressively decreased over time, until it became not significant in the 1980s and beyond. Additionally, our empirical tests show that, in the last 40 years, left and right parties have become indistinguishable from one another as far as the politics of inequality is concerned.

Our findings suggest that, although left parties entered into government less frequently in the initial decades of the twentieth century than in post-World War II years, their impact on the reduction of inequalities was particularly strong in the earlier period precisely because the inequalities to be tackled at that time were far greater than in later decades. Then, as inequalities gradually lowered and the external pressures for responsibility on left governments simultaneously increased, the ability of the Left to pursue egalitarian outcomes progressively decreased until it completely faded at the time of the fall of the Berlin Wall. This indicates that, despite what is commonly argued, the Left’s retreat from egalitarianism did not come about abruptly with the “Third Way” but rather was the result of a long-term process whose deep roots could only be unveiled through a comparative longitudinal design.

In light of these findings, and taking account of the possible underlying factors discussed in the theoretical section, our interpretation points in a clear direction while discounting potential alternatives. Indeed, on the one hand, the descriptive analyses provided in the previous sections have shown that, despite declining in the last few decades, the Left’s loss of governmental power has not been as great as its electoral decline (Benedetto, Hix, and Mastrorocco 2020; Delwit 2021; Emanuele and Trastulli 2023) and in any case not so great as to undermine left parties’ ability to deliver egalitarian policies once in government. On the other hand, notwithstanding the remarkable advancements made by Western European societies in reducing inequalities over the past century, it cannot be said that the Left is unable to pursue further egalitarian policies because it has achieved its goals: according to our data, only in the areas of educational and health equality may it be argued that the Left has accomplished its egalitarian mission. Our data clearly shows that the Left has become increasingly unable to reduce inequalities even in areas where there is still scope for improvement, as in the case of power equality by socioeconomic position. In this regard, it is also hard to imagine that left parties have interrupted their egalitarian efforts due to a shift in the preferences of a public that, according to the thermostatic model (e.g., Wlezien 1995), have readjusted their issue priorities following the achievement of high levels of equality. Indeed, not only are highly egalitarian outcomes still largely unaccomplished, but—as public opinion data shows—public preferences are still oriented toward the reduction of inequalities. ²⁷

Conversely, what is plausibly the most likely interpretation of governing left parties’ recent inability to reduce inequality is that they have become increasingly constrained by external institutions (e.g., the EU, global markets). The progressive liberalization of financial markets and the setting of binding fiscal and economic parameters for the European member states meant that “an autonomous monetary policy became extremely hard to pursue” (Boix 1998, 70). In this context, governing parties have faced limited domestic autonomy as globalization has forced them to assume common positions (Mair 2008, 218), especially within the area of welfare policy (Huber and Stephens 2001). As a result of these long-term irreversible transformations, governing parties’ room for maneuvering is, de facto, much more limited today than it was in the past. This limited autonomy particularly impacts the agenda of social democratic and socialist parties, which emerged in the nineteenth century with the purpose of realizing a radical transformation of the status quo, pointing foremost to the reduction of political, social, and economic inequalities. Over the course of the twentieth century, their actions in government brought about important accomplishments—public education and healthcare systems, universal welfare states—and contributed to transforming Western Europe as a whole into the most egalitarian region of the world (Alvaredo et al. 2018). However, their transformative capacity is now behind them. Since the 1980s, as shown by our analysis, reducing political, social, and economic inequalities has proved to be beyond the governmental power of the Left. In recent decades, the pressure of external constraints and the related responsibility have outweighed left parties’ political willingness to be responsive toward their flagship goals. And most importantly, left parties with governing responsibility do not seem to have the strength and perhaps lack even the will to challenge the dominant paradigm and change the status quo. This may also be because the progressive shift of their reference constituency from industrial workers to sociocultural professionals and the decline in the size and political relevance of unions reduce the willingness of left parties to engage with traditional left goals.

Overall, based on the presented evidence, we can conclude that the Left in Western Europe is today incapable of accomplishing its historical mission to reduce inequalities. Further research is needed to delve deeper into the implications of this finding with regard to the future perspectives of the Left vis-à-vis their very raison d’être.