Labor market dualization or cultural divide?: A comparative analysis of the declining support for center-left parties in Western Europe and Japan

Abstract

Center-left parties in Western Europe and Japan have experienced a persistent decline in voter turnout. This study examines the factors contributing to this weakening support. Using data from the 2016 European Social Survey across 13 countries and the 2017–2018 Japanese General Social Surveys, we tested three hypotheses: labor market dualization between insiders and outsiders; divergence in policy preferences between social investment and social compensation; and a cultural divide between liberalism and authoritarianism. Our findings provide little support for the labor market dualization and policy preference divergence hypotheses in both Western Europe and Japan. The third hypothesis, concerning a cultural divide in attitudes toward immigration, is supported only in Western Europe. Conversely, Japan is an anomaly because of the absence of a substantial left-liberal constituency. The scarcity of liberal values among Japan's sociocultural professionals suggests that center-left parties have underdeveloped core constituencies.

Introduction

Over the past three decades, left-wing parties in advanced industrialized countries have experienced significant erosion of their power resources. With globalization and post-industrialization, the number of manufacturing workers has declined and unionization rates have fallen (Oesch, 2006; Wallerstein and Western, 2000). Some researchers suggest that service and white-collar workers constitute the new backbone of center-left parties (Kitschelt, 1994). Indeed, there was a temporary “revival” of center-left parties in the late 1990s in countries such as the UK, Germany, and France (Manwaring and Kennedy, 2022: 7). However, since the 2000s, the share of votes for these parties has seen a steady decline. Data from the Political Data Yearbook indicate a decrease in the share of votes for center-left parties in lower house elections: from 25.3% in 2002 to 7.5% in 2017 in France; from 38.5% in 2002 to 20.5% in 2017 in Germany; from 39.9% in 2002 to 28.3% in 2018 in Sweden; and from 40.7% in 2001 to 32.2% in 2019 in the UK.

Japan has largely followed a trajectory similar to that of Western Europe. With the end of the Cold War, the Japan Socialist Party, rooted in the working class, was dissolved in 1996. Subsequently, the Democratic Party of Japan, a center-left party targeting the urban middle class, gained 42.4% of the votes in 2009. Nevertheless, after losing power in 2012, the party has experienced steady stagnation. Its successor, the Constitutional Democratic Party of Japan, won only 19.1% of the votes in 2017.

A number of studies have examined the factors contributing to the decline of center-left parties, focusing on their strategies and organizational aspects (Arndt, 2013; Evans and Tilley, 2017; Kats and Mair, 1995). However, given the universal trend of this decline, it is equally important to analyze the demand side, namely, the changes in voter preferences. This study examines three theories that aim to explain the decline in center-left party support: the labor market dualization between insiders and outsiders; the divergence in policy preferences between social investment and social compensation; and the cultural divide between liberalism and authoritarianism. Using data from the 2016 European Social Survey (ESS) and the 2017–2018 Japanese General Social Surveys (JGSS), this study systematically compares the factors contributing to the weakening of support for center-left parties in Western Europe and Japan.

The remainder of this paper is organized as follows. The next section briefly reviews the literature on weakening support for center-left parties and formulates three hypotheses. The third section explains the data and methods used in this study, and the fourth section presents the results of the data analysis. Finally, the fifth section summarizes the main findings and discusses their implications.

Literature review and hypotheses development

In this section, we summarize the literature on the weakening of support for center-left parties, with a focus on Western Europe. As for Japan, the existing literature finds that partisan conflict between left- and right-wing parties has centered on constitutional and security issues, with relatively less contention regarding economic and welfare policies (Jou and Endo, 2016; Taniguchi, 2020). This can be attributed to the prevalence of clientelism of the conservative party (Lipscy and Scheiner, 2012; Scheiner, 2005). The literature also emphasizes the lack of trust among voters in the governing capabilities of center-left parties (Funabashi and Nakano, 2016; Jou and Endo, 2016). Recent research by Berlucchi and Hino (2022) reveals that the rise of authoritarian values in Japanese society has contributed to an increasing support for the ruling party, the Liberal Democratic Party (LDP). However, few studies have systematically compared the attributes and preferences of center-left supporters in Japan with those in Western Europe.

The first hypothesis used to explain the decline in center-left support is labor market dualization. This theory argues that an increase in precarious employment resulting from post-industrialization has created a division of interests between two labor market groups known as “insiders” and “outsiders” (Goldthorpe, 1984; Lindbeck and Snower, 1989, 2001). Recent scholars have highlighted the impact of labor market dualization on individual policy preferences and party competition (Emmeneger et al., 2012; Häusermann et al., 2020; Rueda, 2007). For example, using data from the 1996 Eurobarometer, Rueda (2007) underscores the contrasting preferences of insiders and outsiders. Insiders, comprising full-time workers and voluntary part-time workers with robust employment protection, tend to prefer job security while remaining indifferent to labor market policies. Conversely, outsiders, who are involuntary and non-standard workers such as short-term contractors and part-time workers as well as the unemployed, tend to favor active and passive labor market policies. This divergence in policy preferences poses a strategic dilemma for center-left parties (Lindvall and Rueda, 2013; Rueda, 2007).

This theory has provoked debates regarding the definition and voting behavior of insiders and outsiders. While some scholars argue that dualization is a universal trend (Emmeneger et al., 2012; Lindvall and Rueda, 2013), others contend that it is specific to conservative-corporatist regimes (Thelen, 2014). Rueda (2007) and Emmenegger (2014) characterized insiders and outsiders based on the status of the labor market and employment protection. Nevertheless, this definition has been criticized owing to the diversity of the two groups. Rovny and Rovny (2017) claim that outsiders are heterogeneous, comprising high- and low-skilled workers and the unemployed. Rehm (2009) and Iversen and Soskice (2015) point out that unemployment risk varies according to occupation. Considering the different prospects of unemployment risk, Schwander and Häusermann (2013) propose a classification of 17 social groups by combining factors such as occupation, sex, and age.

To date, researchers have not reached a consensus on the definition and policy preferences of insiders and outsiders (Häusermann et al., 2020; Rovny and Rovny, 2017). Owing to the statistical difficulty in distinguishing between voluntary and involuntary employment, this study defines insiders and outsiders based on the most common understanding—the employment status—while excluding the unemployed. Therefore, the first hypothesis is as follows.

H1: Insiders with stable employment prefer to maintain employment protection and are indifferent to both active and passive labor market policies, whereas outsiders with unstable employment prefer labor market policies.

The second hypothesis focuses on the divergence in policy preferences between the working and middle classes. Some scholars posit that left-wing parties lost the support from the working class owing to their adoption of centrist “third way” policies (Arndt, 2013; Evans and Tilley, 2017; Karreth et al., 2013). However, this argument is disputed by others who contend that center-left parties’ core constituencies have shifted from the working class to the educated middle class, comprising service and white-collar workers (Gingrich and Häusermann, 2015). The educated middle class prefers social investment policies aimed at enhancing individual human capital, while the working class tends to favor passive income maintenance policies.

For instance, Gingrich and Häusermann (2015) conducted an analysis using data from the Eurobarometer (1972–2002) and the ESS (2002–2012). They found that sociocultural professionals and skilled service workers show a strong preference for social investment policies over traditional redistributive policies. These findings were supported by Busemeyer and Garritzmann (2017) and Garritzmann et al. (2018). Busemeyer and Garritzmann (2017) conducted the “Investing in Education in Europe” (INVEDUC) survey across eight Western European countries, revealing that individuals with higher educational backgrounds tended to support social investment policies, while those with lower income and education levels were more inclined toward income maintenance policies. Bremer (2022) examined data from the International Social Survey Program (ISSP) conducted in 2012 across Europe, Asia, and Latin America and concluded that sociocultural professionals exhibit a strong preference for social investment policies, whereas low-skilled service and manufacturing workers demonstrate less support for these policies.

Previous research has yet to confirm whether there is a conflict between the preference for social investment and social compensation among center-left supporters. This study uses the occupational classifications outlined by Kitschelt and Rehm (2014) and Oesch and Rennwald (2018) and analyzes data from the ESS in 2016 to investigate the existence of such conflicts. The ESS (2016) is particularly well suited for this purpose, because it includes newly added questions addressing preferences for social investment and social compensation. Thus, we propose the following hypothesis.

H2: Sociocultural professionals are inclined to favor social investment over social compensation, while manufacturing workers show a stronger preference for social compensation.

The third hypothesis explores the cultural divide between liberalism and authoritarianism. According to this theory, the shift in industrial structure has resulted in an increase in the number of educated white-collar workers. They tend to hold “postmaterialist” and liberal values that emphasize self-determination, individual autonomy, and equality (Flanagan and Lee, 2003; Inglehart, 1989; Kitschelt, 1994). ¹ The spread of these liberal values has triggered a cultural backlash (Norris and Inglehart, 2019). Analyzing data from the ESS conducted between 2002 and 2014 across 30 European countries, Norris and Inglehart (2019) demonstrate that older, less educated, and rural residents are more likely to hold authoritarian values and chauvinistic attitudes toward immigrants.

With the advance of globalization, this cultural divide has reshaped the landscape of political competition in Western democracies (Bornschier, 2011; Kriesi et al., 2008). According to Kriesi et al. (2008) and Bornschier (2011), individuals with traditional-authoritarian values, often referred to as the “losers” of globalization, tend to be hostile to immigration and European integration while supporting right-wing populist parties. Conversely, those with liberal, universalist values, known as the “winners” of globalization, are generally more supportive of these issues and tend to align with center-left parties. Using data from the Chapel Hill Expert Survey, Hooghe et al. (2002) argues that party competition in Western European countries now revolves around a new cultural axis between green-alternative-libertarian and traditional-authoritarian-nationalist orientations (Hooghe et al., 2002; Hooghe and Marks, 2018).

Previous studies have identified various factors that contribute to the cultural divide, such as age, sex, education level, and occupation (Ford and Jennings, 2021). We specifically examine the influence of occupation using the occupational classifications outlined by Kitschelt and Relm (2014) and Oesch and Rennwald (2018). Our focus is on attitudes toward immigration among left-wing supporters. Therefore, the third hypothesis is as follows.

H3: Among left-wing supporters, a cultural divide exists between those with liberal values and those with authoritarian values in terms of their attitudes toward immigration.

Data and method

We use data from the ESS (2016), focusing on individuals in paid jobs across 13 Western European countries: Austria, Belgium, Switzerland, Germany, Spain, Finland, France, the UK, Ireland, Italy, the Netherlands, Norway, and Sweden (N  =  23,748). The ESS (2016) is recognized as one of the most comprehensive social surveys, covering occupational attributes, attitudes toward welfare, and preferences for social investment and social compensation. However, as this survey does not include questions related to employment policy preferences, we supplement our dataset with data from the ESS (2008) across 12 countries (excluding Austria, as its data are missing in this edition of the survey).

For Japan, we use data from the JGSS (2017–2018), specifically focusing on individuals in paid jobs (N  =  1,643). The JGSS is a comprehensive survey covering occupational attributes and welfare preferences. Importantly, it includes questions comparable with those in the ESS. To further investigate employment policy preferences, we supplement our dataset with data from the JGSS (2006).

To define insiders and outsiders, we use the type of employment contract as a criterion. In the employment contract section of the ESS, labeled as “wrkctra,” insiders are classified as individuals with an unlimited contract. Similarly, in the employment status section of the JGSS, labeled as “TPJOB,” insiders are classified as regular employees. Outsiders are categorized as individuals with limited-contract employment in the ESS and temporary workers, dispatched workers, contract employees, and non-regular employees in the JGSS. Unemployed individuals are excluded from the outsiders’ category.

We use the classification of seven occupational groups proposed by Oesch and Rennwald (2018). However, it is important to note one deviation. Oesch and Rennwald classify lawyers and dentists as liberal professionals, whereas we categorize them as sociocultural professionals. This classification is consistent with the understanding that these professionals are primarily involved in interpersonal communication with clients (Kitschelt and Rehm, 2014). To implement this classification, we rely on the “isco08” variable from ESS (2016) and the “XXJOB” variable from JGSS (2017–2018) data (see Appendix 1 for more details).

Regarding employment policy preferences, we rely on specific questions from the ESS (2008) and the JGSS (2006). In the ESS (2008), we examine a question on the government's responsibility to ensure employment for all, labeled as “gvjbevn.” Respondents who answered “no” to this question (score 0–4) are classified as having market-oriented preferences, while those who answered “yes” (score 6–10) are classified as having government-oriented preferences. In the JGSS (2006), we utilize an item that asked respondents about their perceptions of government expenditure related to employment measures, labeled as “BD3EMPLY.” Respondents who indicated “too much” (score 1) are classified as having market-oriented preferences, while those who indicated “too little” (score 3) are classified as having government-oriented preferences.

To investigate preferences for social investment, we rely on specific questions from the ESS (2016) and the JGSS (2017–2018). In the ESS (2016), we use a question that asked whether the government should spend more on education at the cost of unemployment benefit, indicated as “eduunmp.” Respondents who answered “in favor” to this question (score 3–4) are considered to favor social investment, while those who answered “against” (score 1–2) are considered to favor social compensation.

In the JGSS (2017–2018), there are no directly comparable questions. However, as social investment policies often include children's education (Bremer, 2022; Garritzmann et al., 2018; Hemerijck, 2017), we use a question about the responsibility for children's education, labeled as “OP5CCED.” Respondents who supported the government's assumption of responsibility (score 3–4) are considered to favor public responsibility for social investment, while those who believed that “individuals and families” should assume this role (score 1–2) are considered to favor private responsibility.

To examine redistribution preferences, we rely on specific questions from the ESS (2016) and the JGSS (2017–2018). In the ESS, we use a question that asked whether the government should reduce differences in income levels, labeled as “gincdif.” Respondents who answered “agree” to this question (score 1–2) are classified as having government-oriented preferences, while those who answered “disagree” (score 4–5) are classified as having market-oriented preferences. In the JGSS, we use a question that asked whether it is the government's responsibility to reduce income disparities between high- and low-income families, labeled as “Q5GVEQAA”. Respondents who answered “agree” to this question are classified as having government-oriented preferences, while those who answered “disagree” are classified as having market-oriented preferences.

To operationalize sociocultural values, we use a specific question from the ESS (2016) and the JGSS (2017–2018). In the ESS, we use a question that asked whether immigrants make the country a better or worse place to live, indicated as “imwbcnt.” Respondents who answered “worse” to this question (score 0–4) are classified as having authoritarian values, while those who answered “better” (score 6–10) are classified as having liberal values. There are no directly comparable questions in the JGSS because the Japanese government does not officially accept immigrants. As an alternative, we use an item that asked about opinions on increasing the number of foreigners in your community, labeled as “QFNRINCR.” Respondents who answered “for” to this question are categorized as having liberal values, while those who answered “against” are categorized as having authoritarian values.

Results and discussion

First, we investigated the impact of labor market dualization on employment policy preferences. As shown in Figure 1, although outsiders are slightly more inclined to support government responsibility, most insiders also hold government-oriented preferences. There are no significant differences in their preferences.

Figure 1.

Preferences for employment policy in Western Europe.

To examine the differences in preferences for active labor market policies between insiders and outsiders, we analyzed data from the ESS (2016) concerning social investment for the unemployed. Among insiders, 66% expressed a preference for social investment, whereas 30% favored social compensation. Similarly, 62% of the outsiders expressed a preference for social investment, whereas 32% favored social compensation. Thus, our findings do not support the assertion that insiders are indifferent to labor market policies, whereas outsiders express a strong preference for them.

For Japan, data from the JGSS (2006) reveal that among insiders, 26% held government-oriented preferences, whereas only 2% demonstrated market-oriented preferences. Similarly, among outsiders, 22% favored government responsibility, whereas only 1% showed market-oriented preferences. Although most responses are classified as “not applicable,” there is a stronger inclination among insiders toward the government's responsibility for employment policy than among outsiders.

Based on our examination of preferences for employment and active labor market policies, we conclude that the labor market dualization hypothesis lacks sufficient supporting evidence in both Western Europe and Japan.

Second, we examined the divergence in policy preferences between social investment and social compensation by occupation. As shown in Figure 2, managers and technical professionals in Western Europe have a stronger preference for social investment, whereas production workers exhibit a slightly weaker preference. However, the overwhelming majority across all occupations consistently show a robust preference for social investment. Our findings do not provide evidence of a conflict between the preference for social investment among sociocultural professionals and that for social compensation among service and production workers.

Figure 2.

Preferences for social investment and social compensation by occupation in Western Europe.

Figure 3 illustrates preferences for social investment in Japan, particularly concerning the allocation of public and private responsibilities for children's education. While this data is not directly comparable with Western European data that focuses on social investment for the unemployed, it can be used to analyze social investment preferences based on occupation in Japan, as social investment policies often include children's education (Garritzmann et al., 2018; Hemerijck, 2017). As shown in Figure 3, sociocultural professionals and clerks are more inclined to support public responsibility, while managers and small business owners are more likely to support private responsibility. A distinctive feature of the Japanese case is that a significant proportion of respondents across all occupations emphasize private responsibility. This is in stark contrast to Western Europe, where approximately 80% of citizens support increasing public spending for children's education (Busemeyer et al., 2018).

Figure 3.

Preference for social investigation (children's education) by occupation in Japan.

In summary, our findings indicate that individuals across all occupations in Western Europe have significantly higher levels of support for social investment, and there is no observable conflict of preference in this regard. In Japan, the equilibrium between the approval and disapproval of public social investment is balanced without any clear inclination among sociocultural professionals to favor social investment. Therefore, our second hypothesis, which suggests conflicting policy preferences between sociocultural professionals and manufacturing workers, has little empirical support in both Western Europe and Japan.

The third hypothesis posits a cultural divide between sociocultural professionals and low-skilled workers. Figure 4 illustrates the varied attitudes toward immigration by occupation in Western Europe. We find a clear tendency for managers, sociocultural and technical professionals to hold liberal values. Conversely, individuals in other occupations tend to have less liberal attitudes. Notably, there is a contrasting tendency between sociocultural professionals and production workers.

Figure 4.

Attitudes toward immigration in Western Europe.

In the Japanese context, as shown in Figure 5, individuals across all occupations, except managers, generally exhibit authoritarian attitudes toward foreigners. Contrary to the patterns observed in Western Europe, there is no indication that sociocultural professionals in Japan are inclined to hold liberal values.

Figure 5.

Attitudes toward foreigners in Japan.

It should be noted that the ESS surveys attitudes toward immigrants, while the JGSS focuses on attitudes toward foreigners. Additionally, the ESS conducted in 2016 may have reflected heightened exclusionary sentiments toward immigrants in response to the European refugee crisis in 2015. However, these factors do not significantly impact the results of our analysis. Even if there was a surge in anti-immigrant sentiments in Western Europe during this period, Japan exhibited notably stronger anti-foreigner sentiments than Western Europe. Furthermore, the attitudinal divide between sociocultural professionals and production workers is a prominent feature unique to Western Europe, in contrast to the situation in Japan.

To examine the relationship between occupation and policy preferences among left-wing supporters, we combine the axis of distribution preferences with cultural attitudes. In Figure 6, the vertical axis represents public attitudes toward immigration, calculated as the percentage of respondents with a positive attitude minus those with a negative attitude. The horizontal axis represents the level of the respondents’ support for the government's responsibility to reduce income inequality, with the midpoint set at 55%. As shown in Figure 6, managers and technical professionals tend to be market-oriented on redistribution, and liberal in their attitudes toward immigration. Notably, among those who support government redistribution, there is a cultural divide between liberalism and authoritarianism. Sociocultural professionals and production workers exhibit contrasting cultural attitudes.

Figure 6.

Occupation and political preference in Western Europe.

Similarly, in Figure 7, we combine the redistribution preferences axis with cultural attitudes in Japan. The vertical axis represents public attitudes toward foreigners, calculated as the percentage of respondents with a positive attitude minus those with a negative attitude, with the midpoint set at −10%. The horizontal axis represents the level of support for the government's responsibility to reduce income inequality, with the midpoint set at 40%.

Figure 7.

Occupation and policy preferences in Japan.

In Japan, as in Western Europe, managers and technical professionals tend to exhibit market-oriented preferences. However, only managers have a liberal attitude toward foreigners. A distinctive feature of the Japanese context is the absence of a social group that combines liberal attitudes with support for government redistribution. Small business owners and manufacturing workers strongly endorse government redistribution while maintaining authoritarian attitudes toward foreigners. The position of sociocultural professionals in Japan is particularly noteworthy because of its contrast to Western Europe. Japanese sociocultural professionals do not adhere to liberal values, neither in their attitudes toward foreigners nor in their perspectives on the government's responsibility for social investment, as previously discussed.

In conclusion, our findings support the third hypothesis within the context of Western Europe, where a cultural divide between liberalism and authoritarianism contributes to declining support for center-left parties. However, this hypothesis does not hold true in Japan because those who favor government redistribution are predominantly authoritarian and there is no cultural divide within this group. Instead, we posit that the challenges faced by center-left parties in Japan stem from the absence of core constituencies, specifically social groups with a left-liberal orientation. Since individuals across all occupations, except managers, tend to hold authoritarian values, conservative parties are more likely to attract public support than center-left parties.

Conclusion

In most advanced industrialized countries, the voting rate for left-wing parties has substantially decreased over the past two decades. Scholars have examined the causes of this decline, focusing primarily on the strategies and organization of these parties. This study analyzes the demand side, namely, changes in voter preferences. By examining data mainly from the ESS (2016) and the JGSS (2017–2018), this study systematically compares the factors contributing to the weakening support for center-left parties in Western Europe and Japan.

Our findings reveal that the labor market dualization hypothesis does not hold to empirical evidence in both Western Europe and Japan. In Western Europe, more than 60% of insiders and outsiders support passive and active labor market policies. There is no conflict of preferences in which insiders are indifferent to labor market policies and outsiders favor them. Similarly, in Japan, we find no significant differences in employment policy preferences between insiders and outsiders.

We also tested the hypothesis of divergence in policy preferences between social investment and social compensation by occupation. Again, our findings provide little empirical support for this hypothesis in both Western Europe and Japan. While there are slight variations in preferences, the overwhelming majority across all occupations in Western Europe consistently show a robust preference for social investment. In Japan, the level of support for public social investment is roughly balanced by the level of non-support. Specifically, there is no clear inclination among sociocultural professionals to favor public social investment.

Finally, we tested the third hypothesis regarding the cultural divide between liberalism and authoritarianism among left-wing supporters. Our results reveal that only this hypothesis explains the weakening of support for center-left parties in Western Europe. While both sociocultural professionals and service and manufacturing workers express support for government redistribution, there is a significant divergence between the two groups in terms of their attitudes toward immigration.

This conclusion is supported by additional data from the ESS (2016). When examining attitudes toward environmental protection, we observe a higher percentage of both liberals among sociocultural professionals and authoritarians among manufacturing workers (Figure 8). These findings suggest that the primary obstacle for center-left parties in realigning their constituencies lies in the cultural divide between liberalism and authoritarianism.

Figure 8.

Preferences for environmental protection by occupation in Western Europe.

However, this hypothesis does not apply consistently to Japan owing to the absence of a significant social group with a left-liberal orientation. Specifically, Japanese sociocultural professionals tend not to hold liberal values, either in their attitudes toward foreigners or in their perspectives on the government's responsibility for social investment. Our findings suggest that a core constituency for center-left parties, which could potentially compensate for the decline in the number of manufacturing workers, is yet to fully materialize in Japan. This underdevelopment may partly explain the lower support for center-left parties compared with the LDP, the dominant conservative party.

This study systematically compared the attributes and preferences of center-left supporters in Western Europe and Japan. Despite using two different datasets, the ESS and the JGSS, this study suggests that the causes of the declining support for center-left parties differ between the two regions. However, further research is needed to comprehend the underlying reasons for the scarcity of social groups with a left-liberal orientation in Japan.