2. Theoretical Background and Hypotheses

The mission of the political Left is described by Norberto Bobbio (Reference Bobbio1997, 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 Reference Anderson1998; Bartolini Reference Bartolini2000; Lukes Reference Lukes, Ball and Bellamy2003; Noël and Thérien Reference Noël and Thérien2008). This is because of the Left’s conception of human nature, whereby what makes people similar is greater than what sets them apart (Bobbio Reference Bobbio1997; White Reference White2011). 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 Reference Bircan, Brück and Vothknecht2017; Fernández-i-Marín et al. Reference Fernández-i-Marín, Knill, Steinbacher and Steinebach2023; Jahn Reference Jahn2018; Mair Reference Mair2009). 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 Reference Garrett1998; Pierson Reference Pierson and Pierson2001).

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 Reference Brady, Blome, Kleider, Brady and Burton2016). However, this literature has evident shortcomings.Footnote 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 Reference Garritzmann and Seng2016; Kauder and Potrafke Reference Kauder and Potrafke2013) or as part of it (Busemeyer Reference Busemeyer2009; Hewitt Reference Hewitt1977; Jensen Reference Jensen2010). 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 Reference McKnight, Loureiro and Vizard2019), 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 Reference Kenworthy and Malami1999), hence substantively overlapping with the existing contributions on gender inequality at large (e.g., Avdeyeva Reference Avdeyeva2009; Esping-Andersen Reference Esping-Andersen1993). 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 Reference Ansell2008; Braga, Checchi, and Meschi Reference Braga, Checchi and Meschi2013; Busemeyer Reference Busemeyer2009; Hewitt Reference Hewitt1977; Kauder and Potrafke Reference Kauder and Potrafke2013). 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 Reference Busemeyer2009; Schmidt Reference Schmidt and Castles2007), others contend that the impact of left power on education spending is unclear (Ansell Reference Ansell2008; Jensen Reference Jensen2011) or absent (Garritzmann and Seng Reference Garritzmann and Seng2016). Beyond this form of inequality, little research exists on the effect of left power on healthcare, with available works (e.g., Huber and Stephens Reference Huber and Stephens2000) 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 Reference Allan and Scruggs2004; Huber and Stephens Reference Huber and Stephens2000; Iversen and Soskice Reference Iversen and Soskice2015; Jensen Reference Jensen2010), while on income inequality evidence from the literature is more mixed. Several studies (e.g., Beramendi and Cusack Reference Beramendi and Cusack2009; Bradley et al. Reference Bradley, Huber, Moller, Nielsen and Stephens2003; Huber and Stephens Reference Huber and Stephens2014; Iversen and Soskice Reference Iversen and Soskice2006; Pontusson, Rueda, and Way Reference Pontusson, Rueda and Way2002; Rueda and Pontusson Reference Rueda and Pontusson2000) highlight a direct and positive effect of left power on the reduction of income inequality. Yet, many contributions (e.g., Hicks and Kenworthy Reference Hicks and Kenworthy2003; Lupu and Pontusson Reference Lupu and Pontusson2011; Mahler Reference Mahler2010; Morlino Reference Morlino2020; Rueda Reference Rueda2008; Yi Reference Yi2013) 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 (Reference Kristal2010), 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 Reference Bobbio1997), is positively associated with the reduction of political, social, and economic inequalities. Hence, we can derive the following hypothesis:

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 Reference Lipset, Rokkan, Lipset and Rokkan1967), 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 Reference Benedetto, Hix and Mastrorocco2020; Delwit Reference Delwit2021; Loxbo et al. Reference Loxbo, Hinnfors, Hagevi, Blombäck and Demker2019). 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 (Reference Alvaredo, Chancel, Piketty, Saez and Zucman2018, 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 Reference Mair2009). 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 Reference Evans and Tilley2017; Karreth, Polk, and Allen Reference Karreth, Polk and Allen2013; Mair Reference Mair2008; Polacko Reference Polacko2022).

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 Reference Best2011; Goldberg Reference Goldberg2020) 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 Reference Abou-Chadi and Wagner2019; Oesch and Rennwald Reference Oesch and Rennwald2018).Footnote ² This transformation may have also been reinforced by the decline in size and political relevance of unions (e.g., Rennwald and Pontusson Reference Rennwald and Pontusson2021), which traditionally compel left parties to pursue a leftist policy agenda (Garrett Reference Garrett1998; Piazza Reference Piazza2001). 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:

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.Footnote 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 Reference Enyedi and Deegan-Krause2007, 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.Footnote 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 Reference Bartolini2000; Bartolini and Mair Reference Bartolini and Mair1990). 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 Reference Müller-Rommel1989; Poguntke Reference Poguntke1987), which have emerged since the 1970s.Footnote ⁵ For this classification, we rely on previous studies by Bartolini and Mair (Reference Bartolini and Mair1990) and Emanuele (Reference Emanuele2023; Reference Emanuele2024). The list of 114 left parties resulting from this effort is reported in table A1 in the appendix.Footnote 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 Reference Braga, Checchi and Meschi2013; Hewitt Reference Hewitt1977), 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 (Reference Cusack1997) index of cabinet ideological balance, which provides a synthetic score of a cabinet’s overall ideological position (Pontusson, Rueda, and Way Reference Pontusson, Rueda and Way2002; Rueda and Pontusson Reference Rueda and Pontusson2000; for other versions of Cusack’s index, see also Beramendi and Cusack Reference Beramendi and Cusack2009; Iversen and Soskice Reference Iversen and Soskice2015; Lupu and Pontusson Reference Lupu and Pontusson2011), 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. Reference Bradley, Huber, Moller, Nielsen and Stephens2003; Huber and Stephens Reference Huber and Stephens2000; Kristal Reference Kristal2010). Finally, the most sophisticated measure employed by the literature so far is the share of cabinet seats held by left parties (Allan and Scruggs Reference Allan and Scruggs2004; Garritzmann and Seng Reference Garritzmann and Seng2016; Jensen Reference Jensen2010). 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 (Reference Bartolini1998) 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 (Reference Bartolini2000) 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.Footnote 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.Footnote 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 (Reference Bartolini1998, 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 Reference Giddens1998), 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 Reference Benedetto, Hix and Mastrorocco2020; Delwit Reference Delwit2021; Polacko Reference Polacko2022), 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 Reference Rokkan1970), 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.Footnote 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. Reference Bradley, Huber, Moller, Nielsen and Stephens2003; Iversen and Soskice Reference Iversen and Soskice2006; Mahler Reference Mahler2010).

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. Reference Coppedge, Gerring, Knutsen, Lindberg, Teorell, Altman and Bernhard2019) is, to the best of our knowledge, the only available source that provides this kind of data.Footnote 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. Reference Coppedge, Gerring, Knutsen, Lindberg, Teorell, Altman and Bernhard2019, 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. Reference Coppedge, Gerring, Knutsen, Lindberg, Teorell, Altman and Bernhard2019, 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. Reference Coppedge, Gerring, Knutsen, Lindberg, Teorell, Altman and Bernhard2019, 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.Footnote 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.Footnote 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).Footnote 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 Reference Jahn2018). 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.Footnote 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 Reference Bircan, Brück and Vothknecht2017). 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.Footnote 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 Reference Houle2018). 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. Reference Coppedge, Gerring, Knutsen, Lindberg, Teorell, Altman and Bernhard2019). 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 (Reference Mahler2010) 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 Reference Iversen and Soskice2006; Jensen Reference Jensen2010). As a result, we control for Gallagher’s (Reference Gallagher1991) 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.Footnote 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 Reference Bartolini2000). 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 (Reference Bartolini2000). 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 Reference Mair2009), 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 (Reference Fernández-i-Marín, Knill, Steinbacher and Steinebach2023), 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. Reference Coppedge, Gerring, Knutsen, Lindberg, Teorell, Altman and Bernhard2019, 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 Reference Roberts and Wibbels1999). 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 Reference Stimson1985), 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.Footnote 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 Reference Beck and Katz1995). This method has been by far the most widely used in the literature dealing with our topic (see, among many, Allan and Scruggs Reference Allan and Scruggs2004; Kristal Reference Kristal2010; Lupu and Pontusson Reference Lupu and Pontusson2011; Rueda Reference Rueda2008). 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 Reference Emanuele and Trastulli2024).Footnote 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.Footnote 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 Reference Mair2009).Footnote 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 Reference Allan and Scruggs2004; Garritzmann and Seng Reference Garritzmann and Seng2016; Jensen Reference Jensen2010). 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.Footnote 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.Footnote ²² This result further confirms the underlying multidimensionality of equality, where its components are each explained by a different set of determinants.Footnote 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.Footnote 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 (Reference Berry, Golder and Milton2012), 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.Footnote 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.Footnote 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.