Access the full text.
Sign up today, get DeepDyve free for 14 days.
Loading next page...
/lp/ou_press/extreme-work-hours-in-western-europe-and-north-america-diverging-RqEaFK4J2F
- Publisher
- Oxford University Press
- Copyright
- © The Author(s) 2018. Published by Oxford University Press and the Society for the Advancement of Socio-Economics. All rights reserved. For permissions, please email: journals.permissions@oup.com
- ISSN
- 1475-1461
- eISSN
- 1475-147X
- DOI
- 10.1093/ser/mwy020
- Publisher site
- See Article on Publisher Site
Abstract
Abstract This article presents a political economy analysis of extreme work hours in 18 advanced Western economies since the 1970s. Empirically, it shows that the culture of long work hours has gained significance not only in the Anglo-Saxon but also in most Continental European welfare states. Theoretically, it provides an institutionalist argument against the neoclassical, or supply-side, point of view on the drivers of long work hours in post-industrial labour markets. It demonstrates that the choice to work long hours is not entirely, or even mainly, left to the preference of the individual. Instead, individual choices are constrained by labour market policies, collective bargaining institutions and new labour market structures, the pattern and trends of which do not necessarily follow the contours of the regime typology. Data on extreme work hours was compiled from the Luxembourg Income Study and the Multinational Time Use Study micro-data collections. 1. Introduction In the second half of the 19th century, scholars were increasingly concerned about the issue of work time. The legal limitation of the work day to 8 hours was one of the most important demands of the early social-democratic and labour movements in Europe. The movement for the 40-hour work week was an answer to the dramatically changing conditions of work in the period of transformation from agricultural production to a predominantly industrial market structure. Before the first labour regulations were enacted, work days had often been extended to 12 or 14 hours, 6 days a week, at the discretion of the employer. By the first decades of the 20th century, trade unions were organized, and strict work time regulation was successfully enacted in most Western European countries. Therefore, the topic seemed less relevant and received less focus in social science research throughout the middle and second half of the 20th century. Then in 1991, when Juliet Schor (1991) presented evidence that US-Americans were spending significantly more time at paid work in the late 1980s than the 1960s, the topic of work time received renewed interest. Schor’s revelation was surprising, and, at the same time, disappointing, as it suggested that during the transformation to post-industrialism, the fruits of technological advancements were, again, not used in a labour-friendly manner. Since then, Schor’s main finding was repeatedly corroborated by labour market research that analysed time-use data and population surveys. In addition to an increase in average work hours, the proportion of people working extremely long hours has also increased in the USA since the 1970s (Coleman and Pencavel, 1993a,b; Leete and Schor, 1994; Clarkberg and Moen, 2001). Regarding the trend in Western Europe, existing comparative work suggests that most Western European societies followed a qualitatively different path (Jacobs and Gerson, 1998; Bosch et al., 1993; Ausubel and Grübler, 1995; Bosch and Lehndorff, 2001; Alesina et al., 2005). As national average hours of work declined or stagnated in all Western European countries, the literature is dominated by accounts of the success of the ‘short work week movement’ in Europe (Bosch et al., 1993; Bosch and Lehndorff, 2001; Berg et al., 2004). In contrast, the present study demonstrates that a deeper look at the higher end of European countries’ work time distribution suggests a different narrative. In most European labour markets (including not only the Anglo-Saxon but most Continental European labour markets), declining or stagnating average hours mask a new and adverse work time tendency: the increasing prevalence of extreme work hours. In line with earlier literature (Jacobs and Gerson, 2004; OECD, 1998, 2015), very long hours, or extreme work hours, or extreme work, all used as synonyms throughout the article, are conceptualized and operationalized as individuals' weekly work hours of 50 or more. With an increasing number of full-timers working more than 50 hours per week and an increasing number of involuntary part-timers (OECD stats), the reconciliation of work and life might not be as achievable for many Europeans as it is suggested by existing studies of work time. The negative repercussions of long work hours on individuals’ health status, family and community life, as well as social cohesion are well-known. Devoting long hours to work increases the risk of burnout (Spurgeon et al., 1997) and has a negative impact on sleeping habits (Virtanen et al., 2009). Most people who regularly work long hours feel that their job not only undermines their health but also their spousal relationship (Hewlett and Luce, 2006). Furthermore, long work hours might result in the neglect of children (Folbre, 1994; Jacobs and Gerson, 1998) and a reduction in fertility rates (Bettio and Villa, 1993). As women are still the primary caregivers in Western societies (Esping-Andersen, 2009), a rat race type of competition for long work hours creates a work environment in which women are less able to compete than men (Landers et al., 1996). Thus, the long work hours culture is an impalpable hindrance to gender equality, particularly in high-skilled labour markets (Bertrand et al., 2010; Hewlett and Luce, 2006; Burke, 2009; Gerson, 2009). Yet, despite its far-reaching repercussions, we know surprisingly little about the comparative patterns and institutional foundations of this re-emerging phenomenon. Most studies on very long work hours concentrate on the USA (Coleman and Pencavel, 1993a,b; Figart and Golden, 1998; Jacobs and Gerson, 1998) and explain the proliferation of extreme work hours by supple-side preferences, that is, by individuals’ voluntary choices for longer work weeks (e.g. Bowles and Park, 2005; Hochschild, 1997). Rather surprisingly, the scholarship taking a demand side point of view is still in its infancy. Micro-level sociological studies show that long work hours are oftentimes not in line with individuals’ self-reported preferences (Clarkberg and Moen, 2001), but they do not investigate the structural forces behind these incongruences. The few comparative studies that emphasize the role of institutions in shaping work time rely extensively on the premises of Esping-Andersen’s (1990) welfare regime approach, thus, they are unable to explain polarizing work time trends in Continental Europe. Thus, to fill the gap in comparative work time scholarship, this article provides an empirical and a theoretical contribution. Empirically, it analyses the patterns and trends of extreme work hours in Western Europe and North America since the 1970s. Using data from the Luxembourg Income Study (LIS) and the Multinational Time Use Study (MTUS) data collections, it shows that the proportion of extreme work hours has increased not only in the Anglo-Saxon but also in most Continental European labour markets. As the latter group is commonly celebrated as the champion of the short work week movement (Bosch 2000; Berg et al., 2014; Burgoon and Baxandall, 2004), our empirical results present a puzzle for the literature on national work time regimes. Theoretically, the article provides an institutions-based argument against the neoclassical, or supply-side, view on the drivers of extreme work in post-industrial labour markets. Drawing on the rich political economy literature that links the capacity of macro-institutions to differences in distributional outcomes (Rueda, 2006; Martin and Thelen, 2007; Baccaro and Howell, 2011; Thelen, 2014) and the literature on the complementarity and coherence of institutions (Hall and Soskice, 2001; Hancké, 2009; Witt and Jackson, 2016), this article demonstrates that the choice whether to work long hours is not entirely, or even mainly, left to the preference of the individual but is guided by policy and collective socio-economic institutions. Contrary to conventional wisdom, the most relevant work time tendencies of the past decades are shaped by liberalizing trends in labour market policies, industrial relations arrangements and labour market structures not only in the Anglo-Saxon world but also on most parts of Continental Europe, rather than by regime-conform developments. The theoretical expectations are tested on 18 Western political economies at the macro- and the broad skills-based meso-level. The article identifies not only the cross-national variety in institutional drivers but also the impact of political inaction in the face of changing market structures, especially at the bottom and top of the broad skills scale. The remainder of this article is structured as follows. The next section situates the theoretical argument within debates on the drivers of long work hours in post-industrial labour markets. Section 3 explicates in detail how institutions shape extreme work time. Subsection 4.1 discusses the data and methodology used for the empirical analysis. Subsections 4.2 and 4.3 present the main empirical patterns and trends in extreme work since the 1970s. Section 5 supports the theoretical proposition empirically. Finally, Section 6 concludes the analysis and suggests implications for comparative political economy. 2. The debate on the drivers of long work hours Existing theories on the drivers of long work hours use either of two opposing lines of arguments. The first approach takes a supply-side point of view assuming that long hours of work represent workers’ preferences. That is, based on their preferences, workers decide voluntarily whether to work long hours on a daily and weekly basis. The most widely articulated supply-side driven argument draws on the consumption theory of Thorsten Veblen, a nineteenth century economist–sociologist. In a 21st century interpretation of the theory, the ‘Veblen effect’ is evoked to explain that people choose to work more hours to earn enough so that they can emulate the consumption standards of the very rich (Schor, 1998). Bowles and Park (2005) place this argument in a historical perspective by arguing that rising income inequality contributes to the manifestation of the Veblen effect because the pulling away of top incomes further increases the gap in consumption standards between the wealthy and the rest of the population. In a similar vein, neoclassical economic theory provides a supply-side driven explanation. According to the standard textbook argument (Varian, 2014), high earners choose to work longer hours because it is too expensive for them not to do so: due to the ‘substitution effect’, the price of substituting an extra hour of work with leisure is more expensive for high earners than it is for those earning less. Thus, high earners make a free rational choice when opting for more hours of work. Though approached from a different angle, the sociological theory of Hochschild (1997) provides a corroboration to neoclassical arguments about workers’ preferences for long work hours. She suggests that with the increase of the dual earner household model, it has become more difficult for workers to reconcile family and work responsibilities. Many couples live under constant time pressure which often leads to having a neglected home and problems with their children. It is thus precisely difficulties at home, as Hochschild maintains, that drives people back to work—a sphere of life where they are more likely to be able to keep things under control and eventually reap success. The supply-side approach has been criticized by scant literature taking a demand-side point of view. Micro-level sociological studies compare individuals’ stated preferences with their actual hours of work from surveys to show that much of the overtime of full-timers is involuntary (Clarkberg and Moen, 2001; Reynolds, 2004; Baslevent and Kirmanoglu, 2014; Crompton and Lyonette, 2006;). From a methodological point of view, this approach has the disadvantage of making cross-national comparisons difficult as preferences for work hours is a subjective measure that systematically varies across cultures. It is thus at least partly for methodological reasons that these works either concentrate on the individual level drivers (age, gender, family status, etc.) of long daily and weekly work schedules rather than the effect of the national institutional environment, or they analyse separate micro-level models and then compare a few cases qualitatively. Approaching the question from a macro-political frame of reference, the literature on national work time regimes provides a methodologically and conceptually important demand-side contribution to the debate (Rubery et al., 1998; Berg et al., 2004, 2014; Burgoon and Baxandall, 2004; Mutari and Figart, 2010). In particular, it introduces power resources logic (Stephens, 1979; Korpi, 1983) into work time analysis by explicating how policy institutions constrain employers in their ability to control workers’ actual work time, as well as individual workers’ actual palette of choices for work hours. However, by relying as a reference on the welfare regime paradigm, and focusing on the coexistence, or configuration, of a range of work time practices, its contribution to the debate on long work hours remains to be mainly conceptual rather than substantive. To gain a better understanding of the type of contribution that the literature on national work time regimes provides, it is worth highlighting the relevant results of two key articles. Berg et al. (2014) identify three work time configurations in Western democracies which they define by a range of work time practices. They postulate that the configurations emerge from dissimilar power relations between the state, employers and worker representatives. They characterize the ‘unilateral’ configuration (with the country example of the USA) by employers’ control over work time, the ‘negotiated’ configuration (with the country example of Sweden) by extensive tripartite negotiations, while the ‘mandated’ configuration (with the country example of France) by the strong role of the state in shaping work time practices. The question in which category other Continental European countries best fit is left unanswered. Burgoon and Baxandall (2004) explicitly argue that Esping-Andersen’s three regime types generate three worlds of work time. They propose that regime-specific policies and welfare institutions shape three distinct constellations of work time practices, which are discernible on two work time indicators: annual work hours per persons employed and per total working-age population. From the analysis of these two indicators, they suggest that Continental welfare states are the ‘champions’ of the short work week movement (Burgoon and Baxandall, 2004, p. 447). Building on the power resources logic introduced into work time analysis by the literature on work time regimes, but focusing on the prevalence of extreme work, the present study demonstrates that the literature was too quick in enunciating the group of Continental European welfare states as the forerunners in the race toward shorter work weeks. Critical changes in labour market structures, labour market policies and collective bargaining institutions opened the way to polarizing work time tendencies not only in the Anglo-Saxon world but also in most Continental European labour markets. 3. Theory and hypotheses on the role of macro-institutions This article proposes two main arguments on how the national political economy environment influences extreme work patterns. First, individuals’ choices for allocating more than 50 hours per week to work are encouraged or constrained by a range of direct macro-institutions. The strength of employment protection legislation, unions and whether unions are encompassing—that is, whether they represent all workers from low-skilled services through manufacturing to high-skilled services—play a crucial role in shaping the legal and practical framework within which individual choices for work hours take place. Second, two recently revealed changes in labour market structures have similarly important impacts. One is the widening division of labour markets into an insider and an outsider segment (Rueda, 2006; King and Rueda, 2008; Palier and Thelen, 2010; Allmendinger et al., 2015). The other one is the growing size of the non-shielded high-skilled services sector (Wren, 2013), in which professionals are often exempt from all stipulations related to the regulation of work time (Gerson, 2009; Hermann, 2014). These tendencies create new labour market conditions which pave the way for the spreading of the long work hours culture at the bottom and the top of the skills scale. Following Hacker and Pierson (2010), the present study maintains that political inaction in the face of new structural circumstances creates a policy ‘drift’ in itself. Therefore, a proper political economy analysis of extreme work must examine not only the direct impact of political economy institutions but also the indirect impact of political inaction in the face of structural changes. As much of the changes in the institutional setup and labour market structures are not in alignment with the regime typology, this article moves beyond the broad clustering approach and examines the institutional influencers in more detail. 3.1 The role of institutions: labour market policies and collective bargaining institutions Labour market policies The most important policy pillar of extreme work is employment protection legislation rather than the regulation of work time. This argument contrasts with conventional wisdom, as work time practices are believed to be first and foremost shaped by the regulation of work time—e.g. the stipulation of and collective agreements on standard weekly hours, paid leave and maximum work time (Rubery et al., 1998; Bosch, 2000; Lehndorff, 2000; Burgoon and Baxandall, 2004; Mutari and Figart, 2010). Considering that 50 hours of work is above any standards (standard weeks typically range between 37 and 40 hours), and even above the 48-hour maximum, as stipulated by the EU Working Time Directive, there is a compelling case for the notion that relative differences in the stringency of these standards do not directly influence extreme work patterns. Moreover, with the increasing use of the ‘opt-out’ clause—Article 11 (1)—of the EU Working Time directive across countries and industries,1 the number of workers who are exempt from regulation has been augmenting (Hermann, 2014), rendering the policy instrument of direct work time regulation even less effective. In contrast, employment protection legislation (henceforth, EPL) plays a key role in guiding the work time norms of full-time employment in post-industrial labour markets. Through defining the basic power relations between employers and employees in the case of a disagreement, EPL acts as an important policy constraint against employer encroachment. If workers are well protected against individual dismissal, they are more likely to reject undesired overtime requests from their superiors. Employers are also less likely to pressure workers into staying late if employment protection puts a clear boundary on their power. In general, strong EPL contributes to the creation and maintenance of an organizational culture in which the conditions of work are not exclusively controlled by employers. Placing this argument in a macro context, it is suggested that EPL is the most important stand-alone policy institution that effectively diminishes the pressures coming from globalization and post-industrial structural change toward longer and atypical work hours. As the level of EPL is historically entrenched in advanced economies, it explains an important part of the cross-national variation in extreme hours but only a marginal part of the longitudinal trends. Anglo-Saxon countries have historically weak EPL and high proportions of extreme work while Continental European and Scandinavian countries have relatively strong EPL and lower proportions of extreme hours, though with immense within-group variation. Collective bargaining institutions A large part of the longitudinal trend in extreme work can be traced back to changes in industrial relations arrangements which, to a large extent, do not follow the contours of the welfare regime typology. The literature on the new politics of social solidarity extensively analyses these changes and draws attention to some of their adverse distributional effects (Streeck and Thelen, 2005; Martin and Thelen, 2007; Baccaro and Howell, 2011; Thelen, 2014). Scholars of the new politics of social solidarity identify either a uniform trend in the weakening of worker representation across Western Europe (Baccaro and Howell, 2011) or distinct varieties of liberalization (Martin and Thelen, 2007). Beyond drawing attention to profound declines in union density and bargaining centralization in large parts of Continental Europe, the most original contribution of these works lies in shedding light on unions’ deteriorating capacity to represent workers in an encompassing manner. More specifically, Thelen (2014) postulates that the gradual employment shift to services resulted in the erosion of traditional collective bargaining arrangements because union membership remained to be concentrated in manufacturing. In some countries (e.g. in Germany), cooperation between labour and capital even intensified while the growing services sectors remained underrepresented. Consequently, formal institutional stability in terms of union density and bargaining levels might mask profound changes in unions’ capacity and interest to represent all groups of workers. The work time effect of these changes is relatively straightforward. With the growing importance of the 24/7 service economy, the proportion of service workers who are underrepresented in collective negotiations is increasing. The deteriorating representativeness of unions undermines their capacity and interest to counterbalance post-industrial pressures toward longer and atypical work hours in the low-skilled and the high-skilled services sectors. At the bottom of the skills scale, low-skilled workers provide services, such as serviced food and child care; high-skilled professionals provide high-end services, such as legal and business consulting. State capacity Besides labour market policies and collective bargaining institutions, extreme work is indirectly influenced by a range of policy areas, including educational, industrial policies and the rules of corporate governance. As a differentiated analysis of these indirect factors is beyond the scope of the article, we resort to theories on the coordinating capacity of the state (Acemoglu and Robinson, 2008; Martin and Thelen, 2007; Bustikova and Corduneanu-Huci, 2017) to emphasize and account for the influence of the state beyond the well-discernible institutional channels. Differences in state capacity, broadly defined as bureaucratic authority, efficacy and penetration within a national territory, are believed to explain paths of political development. More specifically, Martin and Thelen (2007) suggest that a large public sector enhances a government’s political capacity to sustain macro-corporatist institutions that lead to more equal distributional outcomes and higher levels of social solidarity. Linking this argument to the study of long hours, this suggests that extreme work is less likely to be present in societies with a large and capable state apparatus than in others where state bureaucracy is confined to a minimalist role. Finally, the size of the public sector might even have a ‘composition’ effect. This effect could work through two channels. First, through maintaining good working conditions in the public sector (direct effect), and, secondly, through influencing private sector practices by the conditions set for the public sector (setting standards). 3.2 The role of political inaction in the face of new labour market structures Beyond the examination of the direct effect of policy institutions, a proper political economy analysis of extreme work must also examine the impact of political inaction in the face of new labour market structures. The long work hours culture have found a fertile soil in new tendencies that are reshaping the bottom and the top of the labour markets of Western societies. At the bottom of the skills scale, an increasing number of workers are becoming labour market outsiders who are in atypical, or precarious, employment or unemployment (Lindbeck and Snower, 2001; Rueda, 2006; King and Rueda, 2008; Palier and Thelen, 2010; Allmendinger et al., 2015). The practice of very long hours is particularly wide-spread among outsiders for two reasons. First, due to a lack of regulatory protection and high replaceability, outsiders are in a vulnerable position vis-à-vis their employers. Not complying with an employer’s request for overtime might result in an outsider’s immediate dismissal and replacement. Secondly, in many cases, outsiders consent to, sometimes even initiate, working very long hours in order for their income to reach subsistence level. In today’s increasingly unequal economies, an ever-larger number of low-skilled workers must compensate for their relatively low hourly pay by allocating more time to work. While this decision is formally voluntary, in substance it is not because the choice is strongly shaped by the restrictive political economy environment. At the top of the skills ladder, extreme hours are imposed on many high-skilled professionals by a different set of structural pressures. With the internationalization of professional labour markets (Rodrik, 1997, Krings et al., 2009; Wren, 2013) and the increasing use of regulatory exemptions (Hermann, 2014), a race-to-the-bottom type of competition is prevailing on professional labour markets (Landers et al., 1996). The advancements in information technology and the increased interconnectedness of post-industrial labour markets created a new labour market structure in which much of the barriers that once protected professionals from fierce competition with each other have been removed (Rodrik, 1997). In effect, many high-skilled professionals are now employed in non-shielded services (Wren, 2013), in which the conditions of work are profoundly shaped by competitive pressures coming from economic globalization. Because of this, high-end service jobs in post-industrial labour markets do not provide the employment security and work-life balance that was once guaranteed by industrial white-collar jobs. Existing policy institutions do not adequately target the adverse distributional effects of these new market structures. Labour market dualism and employment growth in high-end services provide a fertile soil for new labour market vulnerabilities, and deteriorating conditions of work, including the increasing proportion of extreme hours. 4. Trends and patterns of extreme work 4.1 Data and methodology Data on the proportion of workers with extreme work hours in 16 Western European countries (Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, the Netherlands, Norway, Spain, Sweden, Switzerland, the UK), as well as Canada and the USA is sourced from two micro data collections: the Luxembourg Income Study Database (henceforth, LIS) and the Multinational Time Use Study (henceforth, MTUS). Both microdata collections contain numerous harmonized country-level surveys from various years, starting as early as the 1970s. The macro-level indicator—the share of workers working extreme hours—was calculated from 104 nationally representative surveys for 24 different socio-economic subgroups. These subgroups are formed by all combinations of three gender categories (female, male, all); four educational categories (ISCED 0-2, ISCED 3-4, ISCED 5-6, all); and two employment statuses (full time worker, all in employment). The share of extreme workers was calculated from all LIS and MTUS surveys conducted between 1970 and 2010 in which individual respondents’ age, gender, highest educational level, employment status and weekly work hours were reported. In line with earlier literature (Jacobs and Gerson, 2004; OECD, 1998, 2015), very long hours, or extreme work hours, or extreme work, all used as synonyms throughout the article—is operationalized as weekly work hours of 50 or more. Descriptive and inferential statistics is conducted using indicators of extreme work over three different educational pools. This makes it possible to empirically identify not only the effect of macro-institutions but also the cross-cutting impact of changing market structures and education on the prevalence of extreme work. Macro-indicators on EPL, union strength, the size of the public sector, labour market dualization and economic globalization are sourced from publicly available data from OECD, Amsterdam Institute for Advanced labour Studies, Armingeon Comparative Political Dataset 2013, Fraser Institute and UNCTAD. 4.2 Cross-national patterns and trends To provide a first look at extreme work patterns across Western democracies, Figure 1 shows the ratio of extreme workers to all full-time workers in Western European and North American countries in two periods: in (i) 1970–1989 (grey bars) and (ii) 1990–2010 (black bars). Each bar shows a period average which was calculated from all available observations for the respective period. Countries are ranked by increasing order in the ratio of extreme work in 1990–2010. The graph offers two major insights. Figure 1 View largeDownload slide The share of workers with extreme work hours in Western democracies. Notes: The graph depicts period averages of the share of workers who work extreme hours in seventeen Western countries. The averages by country and period were calculated based on all available LIS and MTUS observations. Figure 1 View largeDownload slide The share of workers with extreme work hours in Western democracies. Notes: The graph depicts period averages of the share of workers who work extreme hours in seventeen Western countries. The averages by country and period were calculated based on all available LIS and MTUS observations. First, countries with the lowest proportion of extreme work after 1990 include the Scandinavian states (Sweden, Denmark, Finland), and France. Countries with the highest incidence of extreme work in the same period include the North American countries (USA, Canada) and European Anglo-Saxon countries (UK and Ireland). These are followed directly by some of the Continental European welfare states (Austria, the Netherlands, Germany and Spain). Secondly, apart from France, we see a significant rise in extreme work after 1990 in all countries for which data were available for both periods. In fact, the biggest increases in extreme work between the two periods are detected in the labour markets of Continental European countries (Austria, the Netherlands, Germany, Luxembourg, Belgium and Italy). In the Netherlands, Germany, Luxembourg and Belgium, the share of extreme workers has doubled. In Austria and Italy, the same ratio has tripled. In contrast, the share of extreme workers in France dropped to one of the lowest levels. For the Scandinavian states, no pre-1990 observations were available. However, the low proportion of extreme work after 1990 suggests that the rise in long hours, if at all, was marginal in Scandinavia. Figure 2 illustrates the cross-national variation and changes over time in extreme work in Western European countries only. The graph depicts one observation for each country for two periods: the observation closest to 1985 for the first period and one closest to 2000 for the second period. Figure 2 View largeDownload slide The share of workers with extreme work hours in Western European countries in two periods. Notes: The graph depicts one observation for each country for two periods: the observation closest to 1985 for the first period and one closest to 2000 for the second period. The graph includes observations that were sourced from LIS surveys, except for Denmark, for which the single available observation, sourced from the MTUS dataset, is shown. For both periods, all countries are included for which at least one LIS observation was available for the given period. Figure 2 View largeDownload slide The share of workers with extreme work hours in Western European countries in two periods. Notes: The graph depicts one observation for each country for two periods: the observation closest to 1985 for the first period and one closest to 2000 for the second period. The graph includes observations that were sourced from LIS surveys, except for Denmark, for which the single available observation, sourced from the MTUS dataset, is shown. For both periods, all countries are included for which at least one LIS observation was available for the given period. As suggested by the graph, extreme work was a marginal phenomenon in Western European labour markets around 1985. By the end of the 1990s, however, European labour markets became more diverse. In a small group of countries, including France and the Nordic countries, the incidence of extreme hours remained low. Meanwhile, other European labour markets shifted away from a balanced work time pattern to a more polarized one, with a sizeable proportion of their population working more than 50 hours per week. 4.3 Variation across broad educational categories Extreme work hours are not distributed equally across broad skills groups. Figure 3 shows the ratio of extreme work in Western economies for three different subgroups: dark shaded bars show the overall ratio across all educational categories; empty bars show the same share among low-skilled workers; while light shaded bars show the same ratio among high-skilled workers. The figure depicts averages over the period between 1990 and 2010, calculated based on all available observations. Figure 3 View largeDownload slide Extreme work across educational categories. Notes: The graph depicts averages of the share of workers who work extreme hours for three skills-groups in seventeen Western countries between 1990 and 2010. The averages were calculated based on all available LIS and MTUS observations for the respective skills-group and period. Figure 3 View largeDownload slide Extreme work across educational categories. Notes: The graph depicts averages of the share of workers who work extreme hours for three skills-groups in seventeen Western countries between 1990 and 2010. The averages were calculated based on all available LIS and MTUS observations for the respective skills-group and period. As suggested by the graph, extreme work is more prevalent at the low and high ends of the educational scale than in the middle. Apart from the USA (where structural and political changes essentially normalized the long work hours culture in all segments of the labour market during the transition to post-industrialism), and Sweden (where the phenomenon is practically unknown), extreme work is concentrated among low-skilled workers, high-skilled professionals, or both. Countries with a high concentration of extreme work among the low-skilled include Denmark, Greece, Spain and Ireland. In contrast, extreme work concentrates among the highly educated in Luxembourg, Germany, Austria and the UK. Finally, in Belgium, both the low- and the high-skilled are much more likely to work extreme hours than those with an intermediate level of skills. Medium-skilled workers, in general, are less likely to work long hours in the Western world than those with low and high qualifications. Figure 4 shows the change over time in the share of extreme workers across educational categories. Since the 1970s, the most radical increase in extreme hours occurred among high-skilled professionals. While in the 1970s this group had the most balanced work time profile, to the 1990s professionals became the most time-deprived strata of the workforce. Figure 4 View largeDownload slide Trends over time in extreme work in three educational categories. Notes: The graph depicts averages in the share of workers with extreme hours in four decades across three skills-groups in eleven Western countries. Countries with at least one observation from the pre-1990s are included. These are Austria, Belgium, France, Germany, Italy, Luxembourg, the Netherlands, Norway, the UK, Canada and the USA. The averages were calculated based on all available LIS and MTUS observations for the respective period. Figure 4 View largeDownload slide Trends over time in extreme work in three educational categories. Notes: The graph depicts averages in the share of workers with extreme hours in four decades across three skills-groups in eleven Western countries. Countries with at least one observation from the pre-1990s are included. These are Austria, Belgium, France, Germany, Italy, Luxembourg, the Netherlands, Norway, the UK, Canada and the USA. The averages were calculated based on all available LIS and MTUS observations for the respective period. Even though the latest observations in our dataset suggest some decreases between 2004 and 2010 after an increasing trend that lasted a quarter of a century, a number of recently published statistics report that the increasing trend has returned in many Continental European and Anglo-Saxon countries after 2010. An OECD report on well-being (OECD, 2015) finds that the percentage of all employees usually working fifty hours or more per week has doubled in Switzerland and Portugal and moderately increased in Ireland, the UK, the USA, Greece and Belgium between 2009 and 2013. During the same period, these percentages remained very low in all Scandinavian states and France. In a similar vein, a recent analysis of the British Trade Union Congress (TUC, 2015) shows that the number of people working more than 48 hours a week in the UK has risen by 15% to nearly three million people between 2010 and 2015. Overall, the increasing trend in long hours that preceded the crises years seem to have resumed in most Western European countries after 2010, with the apparent exception of France and the Scandinavian states. The systematic concentration of extreme work in groups of countries, skills groups, and time periods suggests that it can neither be sufficiently well explained by workers’ individual socio-economic characteristics nor by supply-side theories. On the contrary: workers’ decision about the length of their work week seems to be guided by country- and skills-specific institutional constraints. 5. Quantitative evidence 5.1 Macro-correlations The rise of extreme work in most Continental European countries and the Anglo-Saxon world occurred simultaneously to liberalizing trends in labour market policies and industrial relations arrangements over the past decades. In most Continental European and Anglo-Saxon countries, the stringency of EPL stagnated or weakened, while union density rates and the level of bargaining centralization decreased. Above all the institutional factors, cross-national differences in extreme work are driven by differences in the strength of employment protection legislation. As Figure 5 illustrates, there is a strong negative association between the two. Countries with strong EPL exhibit systematically lower levels of extreme work. The figure plots the ratio of extreme work hours (y-axis) against the stringency of EPL on regular contracts, sourced from the OECD (x-axis). Extreme work shares shown in the figure were calculated as averages over the period between 1990 and 2010 for each country. Figure 5 View largeDownload slide The association between employment protection legislation and extreme work. Notes: For each country, the graph plots the average of all available observations on the ratio of extreme work over the period 1990–2010, against the simple average of the EPL indicator (sourced from OECD) in the given country from the same years. Figure 5 View largeDownload slide The association between employment protection legislation and extreme work. Notes: For each country, the graph plots the average of all available observations on the ratio of extreme work over the period 1990–2010, against the simple average of the EPL indicator (sourced from OECD) in the given country from the same years. Bivariate macro-correlations confirm that extreme work ratios are negatively correlated with EPL (r2 = 0.54, t-statistic = −3.16). They also indicate that extreme work is negatively correlated with the regulation of work time measured by Fraser Institute’s composite indicator (r2 = 0.54, t-statistic = −2.97). This composite indicator incorporates nine aspects of the work time regulation for workers in three different labour market positions. In line with our theoretical expectation, multivariate regression analysis presented in the next Subsection suggests that the association between the composite index of work time regulation and extreme work is partly due to common confounders. The link between work time standards and extreme work diminishes once we control for a range of other political economy factors. Pairwise macro-correlations between extreme work ratios and indicators of the strength of worker representation within collective bargaining broadly fit expectations. Extreme work ratios have a negative, but weaker, correlation with union density (r2 = 0.50, t-statistic = −1.09) or with the extent to which collective bargaining is centralized (r2 = 0.51, t-statistic = −1.53). This provides a first indication on the validity of our theoretical argument on unions’ lack of capability in fighting for policy measures that target the adverse work time effects of the increasing importance of the 24/7 service economy. Finally, the size of the public sector is strongly correlated with extreme work (r2 = 0.52, t-statistic = −2.77), suggesting that the state has the capacity to shape work time through various direct and indirect channels. 5.2 Regression analysis A series of pooled cross-section ordinary least squares (OLS) regressions estimate the underlying causal dynamics between the theoretical key institutions, labour market structures and extreme work. Variables - key institutions (i) Employment protection legislation is measured by the OECD indicator: EPL on regular contracts. (ii) The Composite index of union strength is calculated with equal weights from Union density, measured as the net union membership in the proportion of wage and salary earners in employment (sourced from the Amsterdam Institute for Advanced labour Studies’ ICTWSS database); Collective bargaining centralization [sourced from Fraser Institute’s 2013 Economic Freedom of the World Indicators Report (henceforth, EFW)], and Collective bargaining coverage (sourced from ICTWSS). The (iii) Size of public sector is measured by general government consumption as a percentage of GDP (sourced from EFW). To proxy the (iv) Size of the outsider labour market segment, a composite index is calculated from (a) the share of involuntary part-time workers (OECD), (b) the ratio of temporary employment in total employment (OECD), and (c) the unemployment rate (Armingeon Comparative Political Dataset 2013), with equal weights. Finally, the openness of the economy is measured by (v) FDI inward stock aspercentageof GDP (UNCTAD). Controls To control for the potential effect of broad political ideological, an indicator of Left party strength is included: Left parties inpercentageof total cabinet posts (Armingeon Comparative Political Dataset 2013). To control for the possibility that standard stipulations of work time influence extreme work as suggested by the literature on national work time regimes and the simple pairwise correlations, a composite indicator of Work time regulation, sourced from EFW, is included. To control for differences in labour market conditions, GDP per capita growth, sourced from OECD statistics, is used. This measure is included to ensure that the effects of institutions are not confounded by the effect of economic cycles. Finally, a Datasource dummy is introduced because extreme work hours are systematically higher in surveys that are originally sourced from an MTUS dataset than those sourced from an LIS dataset. Estimation technique A series of pooled cross-section OLS regressions are estimated, which use both the longitudinal and cross-section aspects of the data for identification. As the unbalanced panel data set on extreme work hour ratios consists of 104 observations from 18 countries (and 27 country-datasource combinations), it is inappropriate for panel estimation methods. Panel methods with a stronger focus on the longitudinal aspect work best when there are many observations for the same unit, which is not the case in our dataset. Furthermore, the objective of this study in comparing various institutional systems implies that the cross-country aspect should also be in the focus of interest, rather than be blended out. Consequently, the most appropriate estimation method is pooled cross-section estimation with robust standard errors to correct for possible heteroskedasticity. The nine columns in Table 1 show the regression outputs for three different estimations: the dependent variable in columns (1)–(3) is Extreme work: the overall ratio of extreme work hours among full time employees; in columns (4)–(6), Low-skilled extreme work: the ratio of extreme work hours among low-skilled full time workers; and in columns (7)–(9), High-skilled extreme work: the ratio of extreme work hours among high-skilled full time workers in a given country in a given year. Table 1. Institutional drivers of extreme work hours (1) (2) (3) (4) (5) (6) (7) (8) (9) Variables Extreme work Extreme work Extreme work Low-skilled extreme work Low-skilled extreme work Low-skilled extreme work High-skilled extreme work High-skilled extreme work High-skilled extreme work Employment protection on regular contracts −1.850** −1.187* −1.651*** −2.980** −1.374 −1.769** −2.276* −1.578* −1.598** (0.855) (0.625) (0.609) (1.286) (0.831) (0.871) (1.168) (0.867) (0.777) Composite index of union strength 0.565 1.409* 0.564 (0.636) (0.789) (0.731) Union density 0.031 0.110** 0.105** −0.001 (0.039) (0.048) (0.048) (0.044) Size of public sector −0.408*** −0.406*** −0.265** −0.512*** −0.522*** −0.474*** −0.360** −0.337** −0.429*** (0.140) (0.145) (0.118) (0.176) (0.183) (0.154) (0.159) (0.166) (0.153) Size of the outsider labour market segment 0.440 0.701* 0.870** 1.047** 0.777* 0.308 0.583 (0.365) (0.418) (0.420) (0.497) (0.432) (0.393) (0.449) FDI inward stock as % of GDP 0.021 0.033 0.016 0.028 0.027* 0.040* 0.039*** (0.016) (0.025) (0.017) (0.031) (0.015) (0.022) (0.015) Left parties in % of total cabinet posts −0.002 −0.005 0.023 0.015 0.014 0.016 (0.015) (0.016) (0.019) (0.021) (0.018) (0.020) Work time regulation −0.583 −0.746 −0.438 −0.527 −0.446 −0.508 (0.477) (0.459) (0.497) (0.514) (0.462) (0.452) GDP per capita growth (annual %) −0.145 −0.232 0.131 0.125 0.333 0.180 (0.359) (0.406) (0.429) (0.471) (0.336) (0.377) Datasource dummy 14.038*** 13.692*** 13.577*** 5.495*** 5.473*** 5.189*** 7.994*** 7.476*** 7.949*** (1.749) (1.578) (1.739) (1.747) (1.639) (1.650) (1.781) (1.662) (1.710) Constant 4.415 5.421 2.682 14.204*** 14.463*** 14.469*** 12.132*** 13.291*** 14.747*** (3.797) (4.059) (3.509) (4.548) (4.806) (4.336) (4.163) (4.414) (3.848) Observations 100 95 104 95 90 94 95 90 99 R2 0.604 0.618 0.562 0.209 0.231 0.182 0.347 0.356 0.323 (1) (2) (3) (4) (5) (6) (7) (8) (9) Variables Extreme work Extreme work Extreme work Low-skilled extreme work Low-skilled extreme work Low-skilled extreme work High-skilled extreme work High-skilled extreme work High-skilled extreme work Employment protection on regular contracts −1.850** −1.187* −1.651*** −2.980** −1.374 −1.769** −2.276* −1.578* −1.598** (0.855) (0.625) (0.609) (1.286) (0.831) (0.871) (1.168) (0.867) (0.777) Composite index of union strength 0.565 1.409* 0.564 (0.636) (0.789) (0.731) Union density 0.031 0.110** 0.105** −0.001 (0.039) (0.048) (0.048) (0.044) Size of public sector −0.408*** −0.406*** −0.265** −0.512*** −0.522*** −0.474*** −0.360** −0.337** −0.429*** (0.140) (0.145) (0.118) (0.176) (0.183) (0.154) (0.159) (0.166) (0.153) Size of the outsider labour market segment 0.440 0.701* 0.870** 1.047** 0.777* 0.308 0.583 (0.365) (0.418) (0.420) (0.497) (0.432) (0.393) (0.449) FDI inward stock as % of GDP 0.021 0.033 0.016 0.028 0.027* 0.040* 0.039*** (0.016) (0.025) (0.017) (0.031) (0.015) (0.022) (0.015) Left parties in % of total cabinet posts −0.002 −0.005 0.023 0.015 0.014 0.016 (0.015) (0.016) (0.019) (0.021) (0.018) (0.020) Work time regulation −0.583 −0.746 −0.438 −0.527 −0.446 −0.508 (0.477) (0.459) (0.497) (0.514) (0.462) (0.452) GDP per capita growth (annual %) −0.145 −0.232 0.131 0.125 0.333 0.180 (0.359) (0.406) (0.429) (0.471) (0.336) (0.377) Datasource dummy 14.038*** 13.692*** 13.577*** 5.495*** 5.473*** 5.189*** 7.994*** 7.476*** 7.949*** (1.749) (1.578) (1.739) (1.747) (1.639) (1.650) (1.781) (1.662) (1.710) Constant 4.415 5.421 2.682 14.204*** 14.463*** 14.469*** 12.132*** 13.291*** 14.747*** (3.797) (4.059) (3.509) (4.548) (4.806) (4.336) (4.163) (4.414) (3.848) Observations 100 95 104 95 90 94 95 90 99 R2 0.604 0.618 0.562 0.209 0.231 0.182 0.347 0.356 0.323 Robust standard errors in parentheses. *** P < 0.01, ** P < 0.05, * P < 0.1. Table 1. Institutional drivers of extreme work hours (1) (2) (3) (4) (5) (6) (7) (8) (9) Variables Extreme work Extreme work Extreme work Low-skilled extreme work Low-skilled extreme work Low-skilled extreme work High-skilled extreme work High-skilled extreme work High-skilled extreme work Employment protection on regular contracts −1.850** −1.187* −1.651*** −2.980** −1.374 −1.769** −2.276* −1.578* −1.598** (0.855) (0.625) (0.609) (1.286) (0.831) (0.871) (1.168) (0.867) (0.777) Composite index of union strength 0.565 1.409* 0.564 (0.636) (0.789) (0.731) Union density 0.031 0.110** 0.105** −0.001 (0.039) (0.048) (0.048) (0.044) Size of public sector −0.408*** −0.406*** −0.265** −0.512*** −0.522*** −0.474*** −0.360** −0.337** −0.429*** (0.140) (0.145) (0.118) (0.176) (0.183) (0.154) (0.159) (0.166) (0.153) Size of the outsider labour market segment 0.440 0.701* 0.870** 1.047** 0.777* 0.308 0.583 (0.365) (0.418) (0.420) (0.497) (0.432) (0.393) (0.449) FDI inward stock as % of GDP 0.021 0.033 0.016 0.028 0.027* 0.040* 0.039*** (0.016) (0.025) (0.017) (0.031) (0.015) (0.022) (0.015) Left parties in % of total cabinet posts −0.002 −0.005 0.023 0.015 0.014 0.016 (0.015) (0.016) (0.019) (0.021) (0.018) (0.020) Work time regulation −0.583 −0.746 −0.438 −0.527 −0.446 −0.508 (0.477) (0.459) (0.497) (0.514) (0.462) (0.452) GDP per capita growth (annual %) −0.145 −0.232 0.131 0.125 0.333 0.180 (0.359) (0.406) (0.429) (0.471) (0.336) (0.377) Datasource dummy 14.038*** 13.692*** 13.577*** 5.495*** 5.473*** 5.189*** 7.994*** 7.476*** 7.949*** (1.749) (1.578) (1.739) (1.747) (1.639) (1.650) (1.781) (1.662) (1.710) Constant 4.415 5.421 2.682 14.204*** 14.463*** 14.469*** 12.132*** 13.291*** 14.747*** (3.797) (4.059) (3.509) (4.548) (4.806) (4.336) (4.163) (4.414) (3.848) Observations 100 95 104 95 90 94 95 90 99 R2 0.604 0.618 0.562 0.209 0.231 0.182 0.347 0.356 0.323 (1) (2) (3) (4) (5) (6) (7) (8) (9) Variables Extreme work Extreme work Extreme work Low-skilled extreme work Low-skilled extreme work Low-skilled extreme work High-skilled extreme work High-skilled extreme work High-skilled extreme work Employment protection on regular contracts −1.850** −1.187* −1.651*** −2.980** −1.374 −1.769** −2.276* −1.578* −1.598** (0.855) (0.625) (0.609) (1.286) (0.831) (0.871) (1.168) (0.867) (0.777) Composite index of union strength 0.565 1.409* 0.564 (0.636) (0.789) (0.731) Union density 0.031 0.110** 0.105** −0.001 (0.039) (0.048) (0.048) (0.044) Size of public sector −0.408*** −0.406*** −0.265** −0.512*** −0.522*** −0.474*** −0.360** −0.337** −0.429*** (0.140) (0.145) (0.118) (0.176) (0.183) (0.154) (0.159) (0.166) (0.153) Size of the outsider labour market segment 0.440 0.701* 0.870** 1.047** 0.777* 0.308 0.583 (0.365) (0.418) (0.420) (0.497) (0.432) (0.393) (0.449) FDI inward stock as % of GDP 0.021 0.033 0.016 0.028 0.027* 0.040* 0.039*** (0.016) (0.025) (0.017) (0.031) (0.015) (0.022) (0.015) Left parties in % of total cabinet posts −0.002 −0.005 0.023 0.015 0.014 0.016 (0.015) (0.016) (0.019) (0.021) (0.018) (0.020) Work time regulation −0.583 −0.746 −0.438 −0.527 −0.446 −0.508 (0.477) (0.459) (0.497) (0.514) (0.462) (0.452) GDP per capita growth (annual %) −0.145 −0.232 0.131 0.125 0.333 0.180 (0.359) (0.406) (0.429) (0.471) (0.336) (0.377) Datasource dummy 14.038*** 13.692*** 13.577*** 5.495*** 5.473*** 5.189*** 7.994*** 7.476*** 7.949*** (1.749) (1.578) (1.739) (1.747) (1.639) (1.650) (1.781) (1.662) (1.710) Constant 4.415 5.421 2.682 14.204*** 14.463*** 14.469*** 12.132*** 13.291*** 14.747*** (3.797) (4.059) (3.509) (4.548) (4.806) (4.336) (4.163) (4.414) (3.848) Observations 100 95 104 95 90 94 95 90 99 R2 0.604 0.618 0.562 0.209 0.231 0.182 0.347 0.356 0.323 Robust standard errors in parentheses. *** P < 0.01, ** P < 0.05, * P < 0.1. In sum, the general form of the estimators is as follows: Extreme workit = ΣβaKey institutionsit + ΣβbControls + βcDatasource dummyit + µit.; (1)–(3) Low-skilled extreme workit = ΣβaKey institutionsit + ΣβbControls + βcDatasource dummyit + µit.; (4)–(6) High-skilled extreme workit = ΣβaKey institutionsit + ΣβbControls + βcDatasource dummyit + µit.; (7)–(9) where the βs are parameter estimates for the main independent, control, and datasource dummy variables. The subscripts i and t represent the country and year of the observations, respectively. For each of the three dependent variables, three specifications are reported. The first ones, columns (1), (4) and (7), include all main explanatory and control variables. The second specifications, columns (2), (5) and (8), include all explanatory and control variables, with one change to the previous specification: Union density is plugged-in instead of the composite index of Union strength. In the third specifications, columns (3), (6) and (9) the significant variables from the previous specifications are included. Results The results reported in Table 1 provide strong quantitative support for our argument on the coercive effect of macro-institutions on extreme work. One stand-alone policy institution, Employment protection legislation, and the general measure of state capacity, the Size of public sector, have strong and significant reducing effects on very long hours across all specifications. These results support our arguments on the prominent role of EPL and, more generally, the capacity of the state, in hindering the spreading of the long work hours culture. Union strength has ambiguous effects on extreme work: neither the composite index of Union strength nor Union density has a significant effect in the expected direction. Moreover, they have a positive effect in the specifications in columns (4) through (6), implying that union strength might even increase the proportion of very long hours among the low-skilled. This result is in line with earlier evidence on the offsetting work time effect of unions (Blanchflower, 1996; Burgoon and Baxandall, 2004), and provides another layer of evidence for theories on the hollowing out of the formal institutional arrangements in collective bargaining in Western Europe (Streeck and Thelen, 2005; Thelen, 2014). Finally, the hypothesis on the effect of political inaction in the face of new market structures is overwhelmingly supported by the results, presented in Table 1. The Size of the outsider labour market segment has a statistically significant positive effect in columns (2), (4), (5) and (6), implying that dualism indeed increases the proportion of very long hours, especially among the low-skilled. Both the magnitude and the significance of the effect are more stable in columns (4) through (6), which suggests that the causal relationship is most pronounced among the low-skilled. In countries, where a sizeable proportion of the workforce falls in the outsider category, very long hours among the low-skilled are significantly more prevalent. The impact of economic openness, measured as FDI inward stock aspercentageof GDP, is positive and significant in columns (7) through (9), whereas it is insignificant in the first six columns. This suggests that economic globalization indeed has a significant increasing effect on the ratio of extreme work among the highly educated but not among those with lower skills. Economic openness indeed introduces a new layer of competition among professionals on several occupational labour markets. The notion that the effects of the two structural changes are discernible even after controlling for all theoretically relevant policy institutions implies that existing policy institutions do not provide adequate counterweight to the adverse work time effect of changing labour market structures. The results related to the theoretically important control variables broadly fit expectations inasmuch as most coefficients have the expected signs and their influence is not statistically significant. The explanatory power of left partisanship, Left parties inpercentageof total cabinet posts, is largely taken away by the concrete policy institutions, implying that much of the influence that broad ideological setups exert can be captured by direct policy instruments. Work time regulation has a negative but insignificant effect in all specifications, suggesting that differences in the regulated or collectively agreed standards for work hours and paid vacation days do not have a significant effect on whether people work more than fifty hours per week. 6. Conclusion This article provides an empirical and a theoretical contribution to comparative work time scholarship. Empirically, it provides a comprehensive analysis of extreme work hours in Western Europe and North America. Theoretically, it presents an institutionalist perspective against supply-side, or neoclassical, approaches to the analysis of long work time patterns in post-industrial labour markets. The supply-side position explains the existence of extreme work by individuals’ preferences. It suggests that workers freely opt to work long hours to reach their financial and social goals, or to find shelter from their problem-loaded (dual earner) households. In contrast to supply-side approaches, this article demonstrates that the choice whether to work more than 50 hours per week is not entirely, or even mainly, left to the preference of individuals but is guided by collective institutions. Individual choices are constrained by labour market policies, collective bargaining institutions, and changes in labour market structures that the state fails to act upon. As the most important policy institution, employment protection legislation contributes to the formation and maintenance of an organizational culture in which it is relatively difficult for employers to pressure people into working late. Beyond setting labour market policies, the state influences work time patterns through various indirect channels. In general, a large and capable public sector can shore up a state’s capacity in implementing policies that target adverse distributional patterns, including that of the proliferation of the long work hours culture. The proportion of extreme work is further influenced by political drifts which result from political inactivity in the face of new economic structures. One such structural change is the widening division of labour markets into an insider and an outsider segment. The other is the employment growth in non-shielded high-end service sectors. Both changes contribute to the emergence of new labour market vulnerabilities and worsening conditions of work (including longer hours). Beyond providing an important contribution to comparative work time scholarship, this article has implications on three ongoing debates in comparative political economy. First, the finding about the ambiguous effect of unions on work hours provides new impetus into the debate on unions’ capabilities to represent workers in an inclusive manner in post-industrial economies. The notion that unions are incapable to fight the proliferation of the long work hours culture in several Western European countries introduces a new angle in the discussion on the hollowing out of the formal institutional arrangements in collective bargaining (Streeck and Thelen, 2005; Baccaro and Howell, 2011; Thelen, 2014). Secondly, the result relating to the cross-cutting effect of the political economy environment and education on the share of extreme work has implications on the study of the new labour market divide between outsiders and insiders. Labour market vulnerability has been widely equated in the literature with the precarious position of low-skilled, low-wage labour (King and Rueda, 2008; Tomlinson and Walker, 2012; Emmenegger et al., 2012). It is only in most recent studies that the emergence of labour market vulnerability among the highly educated is recognized (Häusermann et al., 2015). The notion that very long work hours have become most prevalent among the high-skilled population indicates that vulnerability is very much present at the higher strata of the labour market. As this article suggests, it is precisely the highly educated part of the workforce that is most pressured in terms of work time by the additional layer of competition introduced by economic globalization. Economic globalization, with the frequent reorganization of global value chains and the internationalization of professional labour markets, has reshaped professional labour markets in an unprecedented way. Fierce competition, insecurity about medium term employment prospects, along with the increasing use of regulatory exemptions, boost a race-to-the-bottom type of competition in professional labour markets. This manifests itself, among others, in an increasing precedence of extreme work hours. The effect of changing labour markets structures at the bottom and the top of the skills scale is partially counterbalanced by policy institutions, such as employment protection legislation, but overall there is more room for targeted policy measures. Thirdly, from a broader perspective, the study of extreme hours might provide a new impetus in the debate on European political economies’ convergence toward the US-American pattern (Baccaro and Howell, 2011) or their divergence along distinct varieties of capitalism (Hall and Soskice, 2001). Extreme work might be an instrument that sheds light on some of the undiscovered converging tendencies in the labour markets of Continental Europe that no other indicator, e.g. income inequality or temporary work, can properly capture. Though the article presents a multifaceted political economy analysis of extreme work, not every aspect of the institutional palette could be directly considered. Further research could complement this study by concentrating on other institutional factors, including those incorporating the core of the gendered work time regime approach (e.g. Rubery et al., 1998). These, mostly welfare-based or social policy related, factors could include the quality and availability of childcare facilities, access to parental leave schemes, social norms and policies that incentivise men and women to distribute family care responsibilities equally. Perhaps a more suitable level of analysis for such an inquiry would be at the micro-, firm-, or multi-level. Furthermore, it is an inherent shortcoming of quantitative comparative studies that they do not provide abundant qualitative and conceptual discussion around country cases. Therefore, to reveal the role of country-specific institutional developments in shaping extreme work, a qualitative historical analysis of institutional reforms would be needed. This could be a fertile new terrain for future research. Finally, a comparative analysis of the cross-reinforcing incentives incorporated in the tax and social security system could provide fresh insights into the topic. Jacobs and Gerson (2004) show that the US-American regulatory environment induces labour practices that trigger the growing bifurcation of work into overwork and underwork through creating a legal environment in which it is the most cost-effective strategy for employers to pay the overtime hours of full-time workers while employing a large part of the workforce in atypical forms (e.g. part-time or project-based employment), as the latter do not require employers to pay social security contributions. Hiring workers on a full-time basis and not pressuring them into overwork is the most ‘expensive’ form of employment in the US-American regulatory environment. Whether similar incentives are present in the tax and social contributions system of other Western political economies could be revealed by a series of qualitative country case studies. Acknowledgements The author is grateful for comments from Achim Kemmerling, Bob Hancké, Evelyne Hübscher, Martin Kahanec, Dorothee Bohle, Gerhard Bosch, Jon Messenger, Ignace Glorieux, as well as the members of the Political Economy Research Group at Central European University, the participants of the yearly conference of the Society for the Advancement of Socio-Economics (SASE) in 2017 and 2015, the RDW conference in Geneva in 2015, a PhD workshop in Amsterdam in 2014, the Annual Doctoral Conference in Budapest in the years 2016, 2015, 2014 and the Graduate Network Conference in London in 2013. The author would like to express her gratitude to members of the Awards Committee of the 2015 conference of SASE for selecting this paper as an outstanding submission meritorious of the Graduate Student Prize. Footnotes 1 The opt-out clause effectively gives European countries the possibility to allow their employers to require more than 48 hours of work from their employees, provided that these consent to the exemption in advance. The opt-out is used in all sectors in the UK since 1993, in Malta, Cyprus and Estonia since 2004, and in Bulgaria since 2007. Germany uses the opt-out in public health services, local police and firefighter services since 2003, and among federal civil service since 2008. Spain, France, the Netherlands and Belgium introduced the opt-out only in public health services. References Acemoglu D. , Robinson J. ( 2008 ) ‘Paths of Economic and Political Development’ In Wittman D. , Weingast B. (eds) The Oxford Handbook of Political Economy . Oxford University Press . Alesina A. , Glaeser E. , Sacerdote B. ( 2005 ) Work and Leisure in the U.S. and Europe: Why so Different? Working Paper Series No. 11278 . Cambridge, MA : National Bureau of Economic Research (NBER ). Allmendinger J. , Bernhardt J. , Hipp L. ( 2015 ) ‘ Institutions and the Prevalence of Nonstandard Employment ’, Socio-Economic Review , 13 , 351 – 377 . Google Scholar CrossRef Search ADS Armingeon Comparative Political Dataset 2013. Accessed at http://www.cpds-data.org/ on August 2014. Ausubel J. , Grübler A. ( 1995 ) ‘ Working Less and Living Longer: Long-Term Trends in Working Time and Time Budgets ’, Technological Forecasting and Social Change , 50 , 113 – 131 . Google Scholar CrossRef Search ADS Baccaro L. , Howell C. ( 2011 ) ‘ A Common Neoliberal Trajectory: The Transformation of Industrial Relations in Advanced Capitalism ’, Politics & Society , 39 , 521 – 563 . Google Scholar CrossRef Search ADS Başlevent C. , Kirmanoğlu H. ( 2014 ) ‘ The Impact of Deviations from Desired Hours of Work on the Life Satisfaction of Employees ’, Social Indicators Research , 118 , 33 . Google Scholar CrossRef Search ADS Berg P. , Appelbaum E. , Bailey T. , Kalleberg A. L. ( 2004 ) ‘ Contesting Time: International Comparisons of Employee Control of Working Time ’, Industrial and Labor Relations Review , 57 , 331. Google Scholar CrossRef Search ADS Berg P. , Bosch G. , Charest J. ( 2014 ) ‘ Working-Time Configurations: A Framework for Analyzing Diversity across Countries ’, Industrial and Labor Relations Review , 2014 67 , 805 . Google Scholar CrossRef Search ADS Bertrand M. , Goldin C. , Katz L. ( 2010 ) Dynamics of the Gender Gap for Young Professionals in the Financial and Corporate Sectors. NBER WORKING PAPER SERIES, Working Paper 14681. Bettio F. , Villa P. ( 1993 ) ‘ Strutture Familiari e Mercati Del Lavaro Nei Paesi Svilupatti. L’emergere Di Un Percorso Mediterraneo per L’integrazione Delle Donne Nel Mercato Del Lavaro ’, Economica e Lavoro , XXVII , 3 – 30 . Blanchflower D. ( 1996 ) ‘The Role and Influence of Trade Unions in the OECD’. In ILO Report , Geneva , International Labour Organization . Bosch G. ( 2000 ) ‘Working time reductions, employment consequences and lessons from Europe: defusing a quasi-religious controversy’. In Golden L. , Figart D. (eds) Working Time: International Trends, Theory and Policy Perspectives , Routledge , pp. 177 – 195 . Bosch G. , Dawkins P. , Michon F. ( 1993 ) Working Time in Fourteen Industrialized Countries: An Overview , Geneva , International Institute for Labour Studies , pp. 1 – 45 . Bosch G. , Lehndorff S. ( 2001 ) ‘ Working-Time Reduction and Employment: Experiences in Europe and Economic Policy Recommendations ’, Cambridge Journal of Economics , 25 , 209 – 243 . Google Scholar CrossRef Search ADS Bowles S. , Park Y. ( 2005 ) ‘ Emulation, Inequality, and Work Hours: Was Thorsten Veblen Right? ’ The Economic Journal , 115 , F397 . Google Scholar CrossRef Search ADS Burgoon B. , Baxandall P. ( 2004 ) ‘ Three Worlds of Working Time: The Partisan and Welfare Politics of Work Hours in Industrialized Countries ’, Politics and Society , 32 , 439 – 473 . Google Scholar CrossRef Search ADS Burke R. ( 2009 ) ‘ Cultural Values and Women’s Work and Career Experiences ’, Culture, Organizations, and Work , 442 – 461 . Bustikova L. , Corduneanu-Huci C. ( 2017 ) ‘ Patronage, Trust, and State Capacity: The Historical Trajectories of Clientelism ’, World Politics , 69 , 277 – 326 . Google Scholar CrossRef Search ADS Clarkberg M. , Moen P. ( 2001 ) ‘ Understanding the Time-Squeeze: Married Couples' Preferred and Actual Work-Hour Strategies ’, American Behavioral Scientist , 44 , 1115 – 1135 . Coleman M. , Pencavel J. ( 1993a ) ‘ Changes in Work Hours of Male Employees, 1940-1988 ’, Industrial and Labor Relations Review , 46 , 262 – 283 . Coleman M. , Pencavel J. ( 1993b ) ‘ Trends in Market Work Behavior of Women since 1940 ’, Industrial and Labor Relations Review , 46 , 653 – 676 . Google Scholar CrossRef Search ADS Crompton R. , Lyonette C. ( 2006 ) ‘ Work-Life “Balance” in Europe ’, Acta Sociologica , 49 , 379 – 393 . Google Scholar CrossRef Search ADS Emmenegger P. , Häusermann S. , Palier B. , Seeleib-Kaiser M. ( 2012 ) ‘How We Grow Unequal’. In Emmenegger P. , Häusermann S. , Palier B. , Seeleib-Kaiser M. (eds) The Age of Dualization. The Changing Face of Inequality in Deindustrializing Societies , Oxford University Press , pp. 3 – 26 . Google Scholar CrossRef Search ADS Esping-Andersen G. ( 1990 ) The Three Worlds of Welfare Capitalism , Princeton , Princeton University Press . Esping-Andersen G. ( 2009 ) The Incomplete Revolution. Adapting to Women’s New Roles . Cambridge Polity Press . Figart D. , Golden L. ( 1998 ) ‘ The Social Economics of Work Time: Introduction ’, Review of Social Economy , 56 , 411 – 424 . Google Scholar CrossRef Search ADS Folbre N. ( 1994 ) Who Pays for the Kids ? London , Routledge . Google Scholar CrossRef Search ADS Fraser Institute’s 2013 Economic Freedom of the World Indicators Report (EFW). Accessed at https://www.fraserinstitute.org/research/economic-freedom-of-the-world-2013-annual-report on August 2014. Gerson K. ( 2009 ) The Unfinished Revolution: How a New Generation Is Reshaping the Family, Work, and Gender in America . Oxford University Press . Hacker J. , Pierson P. ( 2010 ) ‘ Winner-Take-All Politics: Public Policy, Political Organization, and the Precipitous Rise of Top Incomes in the United States ’, Politics & Society , 38 , 152 – 204 . Google Scholar CrossRef Search ADS Hall P. , Soskice D. ( 2001 ): Varieties of Capitalism: The Institutional Foundations of Comparative Advantage . Oxford University Press . Google Scholar CrossRef Search ADS Hancké B. [Edited by] ( 2009 ): Debating Varieties of Capitalism: A Reader . Oxford University Press . Häusermann S. , Kurer T. , Schwander H. ( 2015 ) ‘ High-Skilled Outsiders? Labor Market Vulnerability, Education and Welfare State Preferences ’, Socio-Economic Review , 13 , 235 – 258 . Google Scholar CrossRef Search ADS Hermann C. ( 2014 ): Capitalism and the Political Economy of Work Time . Routledge . Hewlett S. A. , Luce C. B. ( 2006 ) ‘ Extreme jobs: The dangerous allure of the 70-hour workweek ’. Harvard Business Review , 84 , 49 – 59 . Google Scholar PubMed Hochschild A. ( 1997 ) The Time Bind: When Work Becomes Home, and Home Becomes Work , New York , Metropolitan Books . ICTWSS database of the Amsterdam Institute for Advanced labour Studies. Accessed at http://www.uva-aias.net/en/ictwss on September 2015. Jacobs J. , Gerson K. ( 1998 ) ‘ Who Are the Overworked Americans? ’ Review of Social Economy , 56 , 442 – 459 . Google Scholar CrossRef Search ADS Jacobs J. , Gerson K. ( 2004 ) The Time Divide. Work, Family, and Gender Inequality , Harvard University Press . King D. , Rueda D. ( 2008 ) ‘ Cheap Labor: The New Politics of “Bread and Roses” in Industrial Democracies ’, Perspective on Politics , 6 , 279 – 297 . Korpi W. ( 1983 ) The Democratic Class Struggle , Routledge . Krings B. J. , Nierling L. , Pedaci M. , Piersanti M. ( 2009 ) ‘Working time, gender and work-life balance’. Katholieke Universiteit Leuven, HIVA. Landers R. , Rebitzer J. , Taylor L. ( 1996 ) ‘ Rat Race Redux: Adverse Selection in the Determination of Work Hours in Law Firms ’, The American Economic Review , 86 , 329 – 348 . Leete L. , Schor J. ( 1994 ) ‘ Assessing the Time-Squeeze Hypothesis: Hours Worked in the United States 1969–89 ’, Industrial Relations , 33 , 25 – 43 . Lehndorff ( 2000 ) ‘Working Time Reductions in the European Union. A Diversity of Trends and Approaches’. In Golden L. , Figart D. (eds), Working Time: International Trends, Theory and Policy Perspectives , Routledge . Lindbeck A. , Snower D. ( 2001 ) ‘ Insiders versus Outsiders ’, The Journal of Economic Perspectives , 15 , 165 – 188 . Google Scholar CrossRef Search ADS Luxembourg Income Study Database. Accessed at http://www.lisdatacenter.org/our-data/lis-database/ on August 2014. Martin C. J. , Thelen K. ( 2007 ) ‘ The State and Coordinated Capitalism: Contributions of the Public Sector to Social Solidarity in Postindustrial Societies ’, World Politics , 60 , 1 – 36 . Google Scholar CrossRef Search ADS Multinational Time Use Study . Accessed at https://www.timeuse.org/mtus on August 2014. Mutari E. , Figart D. ( 2010 ): ‘The Social implications of European Work Time Policies: Promoting Gender Equality?’ In Golden L. , Figart D. (eds), Working Time. International Trends, Theory and Policy Perspectives , Routledge . OECD statistics. Accessed at http://stats.oecd.org/ on July 2017. OECD . ( 2015 ) How Is Life? 2015: Measuring Well-Being , Paris , OECD . Palier B. , Thelen K. ( 2010 ) ‘ Institutionalizing Dualism: Complementarities and Change in France and Germany ’, Politics & Society , 38 , 119 – 148 . Google Scholar CrossRef Search ADS Reynolds J. ( 2004 ) ‘ When Too Much is not Enough: Actual and Preferred Work Hours in the United States and Abroad ’, Sociological Forum , 19 , 89 – 120 . Google Scholar CrossRef Search ADS Rodrik D. ( 1997 ) ‘Sense and Nonsense in the Globalization Debate’ , Foreign Policy , 107 , 19 – 37 . Google Scholar CrossRef Search ADS Rubery J. , Smith M. , Fagan C. ( 1998 ) ‘ National Working-Time Regimes and Equal Opportunities ’, Feminist Economics , 4 , 71 – 101 . Google Scholar CrossRef Search ADS Rueda D. ( 2006 ) ‘ Social Democracy and Active Labour-Market Policies: Insiders, Outsiders and the Politics of Employment Promotion ’, British Journal of Political Science , 36 , 385 – 406 . Google Scholar CrossRef Search ADS Schor J. ( 1991 ). The Overworked American: The Unexpected Decline of Leisure . New York : Basic Books . Schor J. ( 1998 ) ‘ In Defense of Consumer Critique: Revisiting the Consumption Debates of the Twentieth Century ’, The Annals of the American Academy of Political and Social Science , 611 , 16 – 30 . Google Scholar CrossRef Search ADS Spurgeon A. , Harrington J. M. , Cooper C. L. ( 1997 ) ‘ Health and Safety Problems Associated with Long Working Hours: A Review of the Current Position ’, Occupational and Environmental Medicine , 54 , 367 – 375 . Google Scholar CrossRef Search ADS PubMed Stephens J. ( 1979 ) The Transition from Capitalism to Socialism , Springer . Google Scholar CrossRef Search ADS Streeck W. , Thelen K. ( 2005 ) ‘Institutional Change in Advanced Political Economies’. In Streeck W. , Thelen K. (eds) Beyond Continuity , Oxford University Press . Thelen K. ( 2014 ) Varieties of Liberalization and the New Politics of Social Solidarity , New York , Cambridge University Press . Google Scholar CrossRef Search ADS Tomlinson M. , Walker R. ( 2012 ) ‘Labor Market Disadvantage and the Experience of Recurrent Poverty’. In Emmenegger P. , Häusermann S. , Palier B. , Seeleib-Kaiser M. (eds) The Age of Dualization. The Changing Face of Inequality in Deindustrializing Societies , Oxford University Press , pp. 52 – 70 . Google Scholar CrossRef Search ADS TUC analysis. Accessed at https://www.tuc.org.uk/international-issues/europe/workplace-issues/work-life-balance/15-cent-increase-people-working-more on September 2015. United Nations Conference on Trade and Development (UNCTAD). Accessed at http://unctadstat.unctad.org/wds/ReportFolders/reportFolders.aspx? sCS_ChosenLangCS_n on August 2017. Varian H. ( 2014 ) Intermediate Microeconomics: A Modern Approach: Ninth International Student Edition , WW Norton & Company . Virtanen M. , J. E. et al. ( 2009 ) ‘ Long Working Hours and Sloop Disturbances: The Whitehall II Prospective Cohort Study ’, SLEEP , 32 , 737 – 745 . Google Scholar CrossRef Search ADS PubMed Witt M. A. , Jackson G. ( 2016 ) ‘ Varieties of Capitalism and Institutional Comparative Advantage: A Test and Reinterpretation ’, Journal of International Business Studies , 47 , 778 – 806 . Google Scholar CrossRef Search ADS Wren A. (ed.) ( 2013 ) The Political Economy of Service Transition , Oxford University Press . Google Scholar CrossRef Search ADS © The Author(s) 2018. Published by Oxford University Press and the Society for the Advancement of Socio-Economics. All rights reserved. For permissions, please email: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)
Journal
Socio-Economic Review – Oxford University Press
Published: Apr 19, 2018