Delivering the goods? German industrial relations institutions during the COVID-19 crisis.

The COVID-19 pandemic has caused labor market disruptions at an unprecedented scale and is akin to a stress test for industrial relations institutions. Drawing on a large-scale (n = 6111) study of German employees, we empirically investigate whether and how the two institutions comprising Germany's dual system of employee representation-works councils and collective bargaining-have delivered on their protective potential and mitigated the impact of the pandemic on workers. We demonstrate that employees in representative environments fare better on a range of protective outcomes.

INTRODUCTION

Industrial relations (IR) is one of the key institutional domains that distinguishes different “varieties of capitalism” (Hall & Soskice, 2001; Thelen, 2014). Distinctive configurations of IR institutions (e.g., wage‐setting systems, employee voice mechanisms, and employment protections) feature prominently in scholarship on the socio‐economic differences between coordinated and liberal market economies (CMEs and LMEs), typified respectively by Germany and the United States (Jackson & Thelen, 2015). The institutional hallmark of German IR is its “dual system” of employee interest representation through works councils and unions (Gold & Artus, 2015; Jirjahn & Smith, 2006; Müller‐Jentsch, 1979; Thelen, 1991). While works councils provide a mechanism for codetermination at the establishment level, unions and employers' associations negotiate collective agreements at the sectoral level (Keller & Kirsch, 2021). These two IR institutions provide workers with a strong foundation for promoting and protecting their economic and social interests (Behrens et al., 2020). As Addison et al. (2017, p. 195) observe, “this dual system is near‐universally credited with having reduced industrial conflict at establishment level and as having promoted trust and cooperation.” Although both these collective institutions have experienced prolonged decline (Dribbusch & Birke, 2019; Schulten, 2019), and the division of labor between them has changed (Whittall, 2005), they still cover a significant share of employees in Germany: In 2020, the most recent year for which authoritative data are available, about half of all employees were covered by a collective agreement, while 40 percent of the workforce were represented by a works council (Ellguth & Kohaut, 2021).

The COVID‐19 pandemic's effect on the world of work has been described as “cataclysmic” (ILO, 2021a, 2021b). In Germany, as elsewhere, COVID‐19 has caused severe labor market disruptions, including job losses, increased unemployment, reductions in working hours, and a sudden increase in working from home, with these impacts unequal across sectors and jobs (ETUI and ETUC, 2020; ILO, 2020). By confronting workers with these challenges, the pandemic has brought the protective function of IR institutions into sharp relief.

This study investigates whether and how Germany's dual system of IR mitigates the labor market disruptions of COVID‐19 for workers. Our analysis reveals that employees represented by works councils and/or collective bargaining fared better than those who lacked this coverage. We argue that the pandemic has put the protective promise of these IR institutions to the test, foregrounding their capacity to decommodify labor in the face of an unprecedented crisis. We conclude that these collective institutions “deliver the goods” for workers, fulfilling a valuable protective role in supporting employees' interests “on the ground.”

Much prior research has focused on the economic effects of Germany's dual IR system. For example, many studies have examined how works councils impact firm performance (Jirjahn & Smith, 2018). Similarly, scholars have examined the role of Germany's dual system in sustaining the country's highly industrialized export economy (Baccaro & Benassi, 2017; Silvia, 2013). However, at their core these IR institutions serve a social purpose, by protecting workers from arbitrary management decisions and political and market vicissitudes. They not only instantiate and promote fundamental principles such as democracy, dignity, and security at work, but also protect workers from the vulnerabilities that arise from inherent power imbalances in employment and bargaining relationships (Davidov, 2012; Frege & Godard, 2014). The pandemic is akin to a stress test for these IR institutions, enabling us to study their protective effects under conditions of exceptional uncertainty.

Recent IR scholarship drawing on insights from the literature on comparative capitalism has tended to focus mostly on macro‐level institutional developments (Johnstone et al., 2019), calling for greater attention to “the precise nature of firm level practices nurtured by specific institutional configurations” (Wilkinson & Wood, 2017, p. 2505). Recognizing this, here we adopt a more bottom‐up perspective to examine the role of the dual system of IR in Germany's response to the current crisis from the vantage points of workers. Specifically, we examine how works councils, unions, and collective bargaining shape employee perceptions of crisis management at the establishment level. Our quantitative analysis draws on a unique, large‐scale (n = 6111) study of employee experiences during the COVID‐19 pandemic conducted in June 2020. Conceptualizing the protective role of IR institutions through the lens of decommodification (Bosch, 2004), we show that employees in representative environments fare better on a range of protective outcomes than those who are not covered by works councils and/or collective bargaining. Further, our results indicate that COVID‐19 has added new complexities to the division of labor between unions and works councils in Germany's dual system of IR. Finally, our findings also suggest that Germany's key IR institutions have expanded their protective repertoire to address some of the distinctive workplace challenges posed by COVID‐19.

INSTITUTIONS AND LABOR DECOMMODIFICATION: GERMANY'S “DUAL SYSTEM” OF IR

A central tenet of IR is that institutions are needed to protect workers from the adverse effects of market imperfections and inherent power imbalances vis‐à‐vis employers (Kaufman, 2008; Stone, 2004). From this perspective, the logic of capital is to treat labor as any other factor input (Tucker, 2010). Yet, labor's human essence means it is not purely reducible to a commodity; denying this truth inflicts harm on workers. Indeed, according to Kaufman (2010), this proposition constitutes the core theoretical and normative principle for the field of IR. Accordingly, institutional intervention is needed to protect workers from labor market commodification (Kaufman, 2008; Webb & Webb, 1902).

The protective role of IR institutions is often conceptualized through the lens of decommodification (Bosch, 2004; Greer, 2016; Herman et al., 2021; Rubery et al., 2018). Following Esping‐Andersen's (1990, 2000, p. 353) influential use of this concept in his analysis of welfare states, decommodification “signals a citizen's relative independence from pure market forces; it captures one important dimension of freedom and constraint in the everyday life of advanced capitalism.” However, as other IR scholars have noted, Esping‐Andersen (1990) emphasizes that under capitalism, decommodification is never absolute—it always remains partial and cannot completely protect labor from commodification (Bosch, 2004; Herman et al., 2021). For example, Rubery et al. (2018, p.514) argue that decommodification requires that individuals have “some degree of empowerment in their engagement in market‐based work,” while Bosch (2004, p.619) suggests the need for “a tolerable level of individual welfare and security” underpinned by an adequate, secure income and an ongoing, stable job with a regular work schedule. IR research has long shown that systems of worker representation are a central institutional mechanism for establishing protective norms that decommodify labor (Budd, 2004; Heery, 2010; Kaufman, 2015).

In Germany, IR is structured around the so‐called “dual system” of employee interest representation comprised of two key institutions: works councils and unions (Markovits & Allen, 1984; Müller‐Jentsch, 2003; Thelen, 1991; Turner, 1991). While works councils provide a mechanism for codetermination at the establishment level, unions and employers' associations negotiate collective agreements predominantly at sectoral level (Keller & Kirsch, 2021).

Legally, there is a strict separation between the responsibilities and rights of works councils and unions, although in practice they cooperate closely (Keller & Kirsch, 2021; Nienhüser, 2020). With collective bargaining over wages, working hours, and other core conditions the exclusive preserve of unions, works councils enjoys statutory information, consultation, and co‐determination rights over different aspects of HRM and work organization in negotiations with management. Workers can initiate the creation of works councils in establishments with more than five employees, and its members are elected by the entire workforce (not just by union members). The logic of this “division of labor” is that distributive issues and their conflict potential are mostly removed from the establishment level, leaving works councils to represent workers' day‐to‐day interests and explore the potential for joint problem‐solving with management in areas such as hiring, dismissal, disciplinary proceedings, safety, working time arrangements, and the introduction of new technology (Behrens, 2020). As Müller‐Jentsch (1979, p. 202) emphasized in his seminal analysis of the dual system, this functional differentiation distinguishes Germany from other European countries with enterprise‐level mechanisms for employee representation. Nonetheless, in practice unions and works councils have had a “symbiotic relationship” (Müller‐Jentsch, 2003). While unions provide works councils with training and support, works councils recruit the lion's share of members for unions (Müller‐Jentsch, 1997, p. 276; Behrens, 2009).

In terms of the protective role of unions and works councils, studies show that the wages and conditions in German workplaces covered by a collective agreement are generally much better than in those devoid of collective bargaining coverage (Addison et al., 2016; Bossler, 2019; Lübker & Schulten, 2021). And while much of the research on works councils has tended to focus on economic performance outcomes (Jirjahn & Smith, 2018; Nienhüser, 2020), there is evidence to show that works councils help protect employee interests by monitoring employer compliance with labor standards and promoting workplace health (Jirjahn et al., 2021; Nienhüser, 2020). In short, the two pillars of the dual system of IR create representative environments where employees can fare better. Germany is therefore a strong case for examining whether and how these IR institutions mitigate the impact of the COVID‐19 crisis on workers.

COVID‐19: A STRESS TEST FOR DECLINING IR INSTITUTIONS

The dual system of interest representation was long considered the bedrock of stability for German IR. Far from a “fair weather” phenomenon (Thelen, 1991, p. 227), it proved robust in the face of major shocks such as the mass unemployment and associated union militancy of the 1970s (Müller‐Jentsch, 1979). As Thelen (1991) observed in her seminal work on the central role of the dual system in Germany's “negotiated adjustments” to economic changes in the 1970 and 80s, any remaining conflicts between labor and capital were “largely conflicts over the balance between the two levels of the organization in the dual system” (Thelen, 1991, p. 21).

In recent decades, however, scholars have highlighted different challenges to the stability and functioning of the dual system of IR. Part of this debate has focused on various processes of decentralization, which have challenged the balance within this system. Rogers and Streeck (1995), for example, suggest that parochial works councils may engage in “wildcat cooperation” with management to make local concessions that undermine collectively agreed standards at the industry level.

One major mechanism allowing for a controlled variety of decentralization are so‐called opening clauses whereby the parties to a collective agreement empower plant management and works councils to adjust standards to the specific need of their companies (Ellguth & Kohaut, 2014; Whittall & Trinczek, 2019). In a wildcat variety of decentralization, however, actors at the establishment level deviate from collectively agreed standards without union permission, quasi against the law (Bispinck & Schulten, 1999; Traxler, 1995). Pressures for decentralization escalated when in 1998 the federal government initiated an “alliance for jobs and competitiveness” and—in search for increased flexibility—encouraged establishment‐level actors to establish such alliances on the ground, a strategy which some observers (Deppe, 2013; Schulten, 2001) have titled “competitive corporatism.” In his famous parliamentary speech from March 14, 2003, the then Chancellor Schröder threatened to empower works councils and plant management to deviate from binding agreements by introducing statutory opening clauses.

A second strand of this debate has argued that works councils have lost power to new competitors in workplace representation. For example, the diffusion of new management concepts such as high‐performance work systems has seen more firms adopt practices such as group work and semi‐autonomous teams, and these in turn have assumed some of the traditional responsibilities of works council (Keller, 2008; Schumann, 2003). Decisions on vacation times and the distribution of working time, to take just two examples of functions usually performed by works councils, are now entrusted to working groups and their elected speakers. Similarly, the emergence of so‐called “other representative bodies” such as elected or appointed workforce spokespersons, ombudspersons, and committees serve either to take powers from statutory works councils or to replace them altogether. (Ellguth, 2009; Hertwig, 2011).

Finally, a third challenge to this system is connected to the debate on the erosion of German labor relations (Hassel, 1999; Schulten, 2019). In quantitative terms, coverage by both institutions has been in decline since the 1990s. In 1995, 72 percent of western German employees were covered by collectively agreed standards for wages, hours, and working conditions (multi‐employer agreement), however this share declined to 45 per cent by 2020 (and in eastern Germany from 56 percent in 1996 to 32 percent in 2020). (Ellguth & Kohaut, 2021, p. 307). Similarly, the share of the western German workforce represented by a works council declined from 51 percent (1993) to 41 percent (2020) and in eastern Germany from 45 percent (1996) to 36 percent (2020) (Ellguth & Kohaut, 2021, p. 314). There is ongoing debate on whether—as the term erosion suggests—declining coverage by works councils and collective bargaining is an indication of across‐the‐board weakening of these institution (see also Baccaro & Howell, 2017) or rather the emergence of a new equilibrium within a dualized system whereby the core of the system stabilizes itself at the expense of the periphery (Hassel, 2014; Palier & Thelen, 2010; Thelen, 2014).

Against this backdrop of challenges to the stability and functioning of the dual system of IR, we ask whether and how works councils and trade unions still deliver on their protective promise to decommodify labor in the face of the COVID‐19 crisis. As we noted above, labor decommodification means that workers have some degree of protection from rampant market forces that would otherwise render them “replaceable, easily redundant, and atomized” (Esping‐Andersen, 1990, p. 37). Workers require protection from commodification across different dimensions of the employment relationship, ranging from guarantees of adequate income to opportunities for ongoing skill development (Sengenberger, 2005). These protective needs may evolve as socio‐economic conditions change (Bosch, 2004).

Thus, and for the purposes of our analysis, decommodification is conceptualized as a process that spans three different dimensions of employees' work‐related interests. The first dimension captures the very essence of the concept of commodification (replaceable and easily redundant): employees' fears about losing their jobs. Job and income security has been a fundamental concern for workers since the early days of industrialization and is a central objective of labor decommodification (Bosch, 2004). As the employment relationship is not simply a major source of income but also of recognition and self‐esteem, job protection is probably the most encompassing aspect of our analysis. Our second dimension concerns the reconciliation of work and private life. With the weakening of the male breadwinner model and the increase in female labor force participation in recent decades, this dimension has gained importance as more and more employees find it difficult to combine employment with family duties such as caring for children or other family members (Bosch, 2004; Rubery et al., 2018). As different spheres of human life are involved, we call this dimension boundary management. Finally, we recognize a third dimension of decommodification: further training and professional development to enable workers to maintain or even improve their skills and remain employable (Bosch, 2004; Rubery et al., 2018). The shift from the Fordist mass production model to the fast‐changing knowledge economy and the digitalized manufacturing landscape of “Industry 4.0” has meant that workers' skills are at risk of becoming obsolete more quickly (Hirsch‐Kreinsen, 2016; Stone, 2004). Employees who are supported to develop their skills and knowledge might find it easier to keep their current job, or transfer to a new (and potentially better paid) position.

DATA

Our quantitative analysis draws on a unique, large‐scale panel study of employee experiences during the corona pandemic (HBS Befragung “Die Situation von Erwerbstätigen in Zeiten von Corona”; Hövermann, 2020a). Commissioned by the Hans Böckler Foundation, the survey was conducted by the KANTAR polling institute. Surveys were conducted in the form of a Computer Assisted Web Interview (CAWI) and included respondents aged 16 and older who are either employed or self‐employed. Our analysis excludes self‐employed persons, as well as lifetime civil servants (Beamte) who are excluded from collective bargaining. The interviews used for our analysis were conducted between June 18 and June 29, 2020. The survey took advantage of a structured online access panel provided by KANTAR. The sample used for setting up the access panel aimed to represent the German working population with respect to the key variables age, sex, industry, location (state), and level of formal education. Data were corrected by calculating post‐stratification weights. Survey respondents were recruited from participants in the popular “Payback” card program, which is used by German retailers as a tool to improve customer relationships.1 It is estimated that more than 50 percent of all German households carry a Payback card (Hövermann, 2020b). All cardholders serve as the population from which participants in the online access panel were selected. Individuals selected for participation were contacted in writing and invited to take part in the access panel. As compensation for their participation in each survey, respondents receive a certain amount of Payback points.

The major focus of the panel survey “The situation of employees during the corona pandemic“ is on working conditions during the pandemic (including income, training and working time), the distribution of care work within a household, and public policies with respect to the pandemic. In addition to these panel questions, the survey also includes a set of questions about persons' savings, debt, and consumption behavior.

Responses were checked for accuracy and consistency. The survey drew 6111 valid responses, which is a response rate of 82 percent in relation to the number of individuals who were invited to participate.

VARIABLES AND RESULTS

As outlined in Section 3, job and income security, skill development, and boundary management can be considered key dimensions of labor decommodification. In the following analysis, we employ binary logistic regression analyses to test whether and how key IR institutions deliver on their protective potential across these dimensions to mitigate the impact of the pandemic on workers in Germany. Rather than measuring the intensity of decommodification, our multivariate analysis investigates the likelihood of labor decommodification in enterprises covered by works councils and/or collective agreements compared to those enterprises devoid of such coverage.

The dimensions of decommodification serve as the dependent variables in four independent regression models. Job and income security, our first dimension, is represented by the subjective perceptions of job security by employees. Respondents to our survey were asked whether they are concerned about losing their jobs due to the COVID‐19 crisis.2 This variable takes the value of “1” when respondents agree or somewhat agree to this statement.

As it is indicated in the literature, the COVID‐19 pandemic affects not only job security but also income security (Dörre, 2020). As Kohlrausch et al. (2020) highlight, low‐wage earners are particularly prone to suffering wage cuts due to the pandemic. While special compensation schemes such as short‐time work (Kurzarbeit) help to avoid mass dismissals, they still cut into workers' incomes and mostly hurt employees at the lower end of the income distribution. In several cases, compensation payments by the federal employment agency (Bundesagentur für Arbeit) are topped up by additional payments by the employer. Some collective agreements also include provisions for topping up pay. Our second model examines whether such additional payments are available to employees. Our dependent variable takes the value of “1” when employers provide for co‐payments on top of subsidies provided by the federal employment agency. As inclusion into the federal short‐time work program is a precondition for topping up payments, our regression only includes respondents who have identified as working short‐time at the time of our survey.

Boundary management, our second dimension of decommodification, is captured by employees' access to provisions for remote work/home office arrangements. Prior to the COVID‐19 pandemic, provisions for home office arrangements gave employees greater choice in where to work—or in the terminology of our decommodification framework—to manage work–life boundaries. Most commonly, provisions for remote work serve to enable employees to better reconcile work with caring responsibilities, such as raising children or caring for other family members in need of support. With the COVID‐19 pandemic accelerating, provisions for remote work have also been used to protect employees from being infected at the workplace. By way of avoiding potentially hazardous commutes as well as crowded office spaces, providing employees with remote work options has served as one of the most powerful weapons for protecting employees during the pandemic (Gabler et al., 2021; Neumann et al., 2020). In our estimate, the dependent variable takes the value of “1” when there are provisions available at the establishment level which allow employees to take advantage of remote work/home office arrangements. Since this question concerns provisions applying to the establishment‐level only, arrangements which might be the result of collective bargaining are not included.

To capture skill development as the third dimension of decommodification, we asked respondents whether they have participated in training measures since the start of the pandemic. Our dependent variable takes the value of “1” for an affirmative response.

All estimates statistically control for establishment size, including six size categories with “2000 and more employees” serving as the reference category. As we know from previous research on German labor relations, establishment or company size respectively are often powerful explanatory variables.3 As establishment size increases, the structures of employee interest representation tend to become more complex, and often more powerful. For example, works councils in establishments with at least 200 employees are entitled to one full‐time member on release from regular work duties (“Freistellung”), with additional full‐time positions guaranteed in larger enterprises. And in companies with 200 or more employees, works councils have the right to establish an economics committee to gain additional information on the economic performance of their firm.

In addition, all our models control for industry (with “other services” serving as the reference category) as we find close, social partnership‐like labor relations in specific industries such as construction, where unions and employers jointly govern protective institutions established to guarantee workers additional pension entitlements and other benefits.

Further, the pandemic has affected industries differently, with healthcare services in high demand while hotels and restaurants had to close during lockdowns. Accordingly, our industry dummies also seek to account for the extent to which workers felt the brunt of the pandemic.

All models also control for the location of an establishment in east Germany (“1” if yes). It was only with Germany's unification in 1990 that its traditional IR institutions were transferred to its eastern states, with some scholars suggesting this might have affected their functioning (Artus, 2001; Schroeder, 2000). In addition, our regression models on skill development also statistically control for five different categories of formal school education (with the lower level, 9‐year school certificate serving as the reference category), as previous research has established that it is mostly higher skilled employees who benefit from training measures at the establishment level (Keller & Seifert, 2020).

Key IR institutions in the form of a works council at the establishment level or coverage by a collective agreement are the major independent variables included in the model. The close relationship between collective bargaining and works council coverage raises the possibility of multicollinearity.4 Accordingly, separate models are estimated for each independent variable including collective bargaining only, works council only, and both institutions. As our variable on the existence of provisions for home office/remote work is restricted to establishment‐level measures, this model considers works councils only. Descriptive statistics are provided in Table 1.

TABLE 1

Descriptive statistics

Variable	Percent
Concerned about job loss (1 = yes)	10.4
Co‐payments for short‐time work (1 = yes)	46.2
Provisions for remote work/home office available (1 = yes)	60.9
Taken part in qualification measures (1 = yes)	21.8
Works council (1 = yes)	60.2
Collective agreement (1 = yes)	63.4
East Germany	19.8
Education: Hauptschule certificate (9 years of schooling)	22.5
Education: Mittlere Reife certificate (10 years of schooling)	36.2
Education: Abitur (university entrance qualification)	39.9
Education: other degree	1.0
Education: no degree	0.3
Industry: Other Services	10.0
Industry: Agriculture and Forestry	1.1
Industry: Public Sector	10.9
Industry: Utilities and Mining	4.2
Industry: Manufacturing	20.2
Industry: Construction	7.4
Industry: Retail	12.0
Industry: Transportation	6.0
Industry: Hotels and Restaurants	6.8
Industry: Communication, Media, and Arts	3.4
Industry: Financial Services, Insurance	3.5
Industry: Housing	5.1
Industry: Healthcare and social services	9.5
Establishment size: 2000 employees and more	13.2
Establishment size: less than 5 employees	6.4
Establishment size: 5–19 employees	17.6
Establishment size: 20–199 employees	34.6
Establishment size: 200–499 employees	14.5
Establishment size: 500 bis 1999 employees	13.7

Note: Source: Survey “Die Situation von Erwerbstätigen in Zeiten von Corona,” June 2020.

Results

Our first regression model uses employees' concerns about job loss due to the COVID crisis as the dependent variable. As our estimates show (Table 2), there is a significant (p < 5%) and negative association between worries about potential job loss and coverage by a collective agreement in both models which include this variable. When an establishment is covered by collective bargaining, employees (controlling for the other variables in the model) perceive their job to be more secure when compared to employees not covered. The variable identifying the existence of an elected works council fails to reach conventional levels of significance. A number of significant associations of our industry variables deserve mentioning as they confirm what we might expect in those of parts of the economy most severely hit by the crisis. Hotels and restaurants, most severely affected by the lockdown in Spring, come along with an increased threat of job loss while the public sector (p < 1 in all models), utilities and mining (p < 1% in model 1, p < 5% in models 2 and 3), and healthcare and social services (p < 1 in all models) are associated with lower concern about job loss.

TABLE 2

Binary logistic regression, DV: “Are you concerned that in the near future you might lose your job due to the corona crisis?” (1 = “yes” or “rather yes”)

Variable	Coefficient B (standard error)
	Model with works council	Model with collective agreement	Model with works council and collective agreement
Works council (1 = yes)	0.003 (0.131)		0.162 (0.145)
Collective agreement (1 = yes)		−0.264** (0.116)	−0.325** (0.128)
East Germany	0.076 (0.129)	0.060 (0.129)	0.054 (0.129)
Industry: other services	(Reference)	(Reference)	(Reference)
Industry: agriculture and forestry	−0.158 (1.064)	−0.147 (1.064)	−0.132 (1.064)
Industry: public sector	−1.596*** (0.286)	−1.471*** (0.288)	−1.499*** (0.289)
Industry: utilities and mining	−0.996*** (0.370)	−0.911** (0.370)	−0.929** (0.371)
Industry: manufacturing	0.195 (0.170)	0.219 (0.170)	0.203 (0.170)
Industry: construction	−0.534 (0.354)	−0.481 (0.355)	−0.465 (0.355)
Industry: retail	0.189 (0.192)	0.232 (0.193)	0.225 (0.193)
Industry: transportation	0.364* (0.212)	0.408* (0.212)	0.398* (0.212)
Industry: hotels and restaurants	0.675*** (0.207)	0.729*** (0.209)	0.754*** (0.210)
Industry: communication, media, and arts	−0.095 (0.292)	−0.088 (0.291)	−0.106 (0.292)
Industry: financial services, insurance	−0.614* (0.330)	−0.550* (0.330)	−0.575* (0.330)
Industry: housing	−1.176 (0.736)	−1.149 (0.736)	−1.143 (0.736)
Industry: healthcare and social services	−1.399*** (0.281)	−1.325*** (0.282)	−1.337*** (0.282)
Establishment size: 2000 employees and more	(Reference)	(Reference)	(Reference)
Establishment size: less than 5 employees	0.347 (0.270)	0.196 (0.263)	0.273 (0.272)
Establishment size: 5–19 employees	−0.047 (0.221)	−0.188 (0.212)	−0.109 (0.223)
Establishment size: 20–199 employees	0.047 (0.179)	−0.064 (0.176)	−0.017 (0.181)
Establishment size: 200–499 employees	0.122 (0.193)	0.052 (0.194)	0.062 (0.195)
Establishment size: 500 bis 1999 employees	0.123 (0.196)	0.086 (0.197)	0.088 (0.197)
Constant	−2.097*** (0.226)	−1.892*** (0.216)	−1.980*** (0.230)
Valid cases, pseudo‐R 2 (Nagelkerke)	4318/0.077	4318/0.080	4318/0.080

Note: Source: Survey “Die Situation von Erwerbstätigen in Zeiten von Corona,” June 2020.

***/**/* significant at 1/5/10 percent level.

Our second model (Table 3) focuses on income security and includes only those individuals who indicate that they have taken advantage of the federal short‐time work program. The dependent variable takes the value of “1” when respondents indicate that their employer provides for co‐payments on top of subsidies provided by the federal employment agency. The likelihood that workers benefit from co‐payments increases where they represented by an establishment‐level works council and covered by a collective agreement. This association is significant and robust in all three models.

TABLE 3

Binary logistic regression, DV: Employers provide for co‐payments on top of subsidies provided by the federal employment agency (1 = yes)

Variable	Coefficient B (standard error)
	Model with works council	Model with collective agreement	Model with works council and collective agreement
Works council (1 = yes)	0.817*** (0.224)		0.566** (0.242)
Collective agreement (1 = yes)		0.839*** (0.209)	0.648*** (0.225)
East Germany	−0.059 (0.260)	0.021 (0.263)	0.034 (0.265)
Industry: other services	(Reference)	(Reference)	(Reference)
Industry: public sector	1.539** (0.717)	1.364* (0.720)	1.250* (0.727)
Industry: utilities and mining	−0.121 (0.871)	−0.327 (0.866)	−0.352 (0.874)
Industry: manufacturing	0.223 (0.313)	0.210 (0.314)	0.152 (0.317)
Industry: construction	0.544 (0.892)	0.301 (0.896)	0.291 (0.902)
Industry: retail	0.322 (0.381)	0.232 (0.383)	0.214 (0.387)
Industry: transportation	0.537 (0.414)	0.508 (0.414)	0.467 (0.418)
Industry: hotels and restaurants	0.383 (0.321)	0.089 (0.329)	0.162 (0.332)
Industry: communication, media, and arts	1.036* (0.537)	1.001* (0.537)	0.987* (0.543)
Industry: financial services, insurance	1.990* (1.157)	1.820 (1.161)	1.837 (1.176)
Industry: healthcare and social services	0.680 (0.494)	0.582 (0.495)	0.575 (0.496)
Establishment size: 2000 employees and more	(Reference)	(Reference)	(Reference)
Establishment size: less than 5 employees	−2.232*** (0.557)	−2.239*** (0.554)	−2.016*** (0.565)
Establishment size: 5–19 employees	−2.180*** (0.504)	−2.350*** (0.486)	−2.015*** (0.508)
Establishment size: 20–199 employees	−2.044*** (0.452)	−2.037*** (0.450)	−1.835*** (0.459)
Establishment size: 200–499 employees	−1.870*** (0.477)	−1.752*** (0.481)	−1.668*** (0.484)
Establishment size: 500 bis 1999 employees	−2.087*** (0.472)	−2.075*** (0.474)	−2.040*** (0.474)
Constant	1.017** (0.515)	1.092** (0.500)	0.736 (0.526)
Valid cases, pseudo‐R 2 (Nagelkerke)	554/0.215	554/0.221	554/0.232

Note: Source: Survey “Die Situation von Erwerbstätigen in Zeiten von Corona,” June 2020.

***/**/* significant at 1/5/10 percent level.

Our third regression model focuses on home office arrangements to capture decommodification through boundary management. The dependent variable takes the value of “1” where respondents indicate that provisions for home office are available at their establishment. As noted above, because our dependent variable is restricted to provisions at the establishment level, our regression only considers associations with our works council variable. As our estimate indicates (Table 4), chances that provisions for remote work/home office arrangements are available increase where workers are represented by a works council. Compared to large firms (2000 employees and more), employees in medium‐sized and small firms to a lesser degree benefit from provisions for home office.

TABLE 4

Binary logistic regression, DV: Provisions for remote work/home office available at the establishment level (1 = yes; excluding cases where respondents indicate that “this does not apply to my working situation”)

Variable	Coefficient B (standard error)
	Model with works council
Works council (1 = yes)	0.605*** (0.108)
Collective agreement (1 = yes)
East Germany (1 = yes)	−0.157 (0.106)
Industry: other services	(Reference)
Industry: agriculture and forestry	−1.035 (0.835)
Industry: public sector	−0.437*** (0.163)
Industry: utilities and mining	0.557** (0.260)
Industry: manufacturing	−0.291* (0.150)
Industry: construction	−0.607** (0.246)
Industry: retail	−0.956*** (0.170)
Industry: transportation	−0.943*** (0.202)
Industry: hotels and restaurants	−1.510*** (0.230)
Industry: communication, media, and arts	0.297 (0.245)
Industry: financial services, insurance	0.515** (0.250)
Industry: housing	−0.078 (0.401)
Industry: healthcare, social services	−1.470*** (0.175)
Establishment size: 2000 employees and more	(Reference)
Establishment size: less than 5 employees	−1.496*** (0.240)
Establishment size: 5–19 employees	−1.350*** (0.177)
Establishment size: 20–199 employees	−0.787*** (0.140)
Establishment size: 200–499 employees	−0.298* (0.154)
Establishment size: 500 bis 1999 employees	0.003 (0.163)
Constant	1.245*** (0.182)
Valid cases, pseudo‐R 2 (Nagelkerke)	2984/0.233

Note: Source: Survey “Die Situation von Erwerbstätigen in Zeiten von Corona,” June 2020.

***/**/* significant at 1/5/10 percent level.

Finally, our fourth model focuses on skill development. Our dependent variable has the value of “1” when respondents indicate that they have taken part in skill development measures such as presentations, seminars, training, e‐learning, or workshops since the start of the pandemic. Our three regression models show a significant association with works councils in both models which include this variable (p < 1%), while collective bargaining turned out to be significant (p < 5%) only when not controlled for the works council influence (Table 5). As predicted by the relevant literature, higher levels of formal school education (when compared to our reference category, the 9th‐grade Hauptschule certificate) are associated with a higher likelihood that employees benefit from qualification measures.

TABLE 5

Binary logistic regression, DV: “Since the start of the corona pandemic, have you taken part in qualification measures such as presentations, seminars, training, e‐learning or workshops?” (estimate excludes self‐employed and unemployed)

Variable	Coefficient B (standard error)
	Model with works council	Model with collective agreement	Model with works council and collective agreement
Works council (1 = yes)	0.447*** (0.103)		0.434*** (0.115)
Collective agreement (1 = yes)		0.202** (0.092)	0.025 (0.103)
East Germany (1 = yes)	−0.023 (0.095)	−0.012 (0.095)	−0.022 (0.095)
Education: hauptschule certificate (9 years of schooling)	(Reference)	(Reference)	(Reference)
Education: mittlere reife certificate (10 years of schooling)	0.623*** (0.126)	0.631*** (0.125)	0.623*** (0.126)
Education: Abitur (university entrance qualification)	1.322*** (0.120)	1.324*** (0.120)	1.324*** (0.120)
Education: other degree	0.989** (0.434)	0.929** (0.432)	0.989** (0.434)
Education: no degree	0.331 (0.787)	0.333 (0.785)	0.332 (0.787)
Industry: other services	(Reference)	(Reference)	(Reference)
Industry: agriculture and forestry	−0.734 (1.082)	−0.775 (1.080)	−0.736 (1.082)
Industry: public sector	−0.249* (0.150)	−0.190 (0.153)	−0.257* (0.154)
Industry: utilities and mining	0.233 (0.193)	0.266 (0.194)	0.227 (0.195)
Industry: manufacturing	−0.406*** (0.141)	−0.367*** (0.140)	−0.407*** (0.141)
Industry: construction	−0.444* (0.269)	−0.489* (0.269)	−0.450* (0.270)
Industry: retail	−0.020 (0.161)	−0.007 (0.161)	−0.023 (0.162)
Industry: transportation	−0.125 (0.190)	−0.104 (0.190)	−0.129 (0.191)
Industry: hotels & restaurants	−0.500** (0.218)	−0.592*** (0.219)	−0.506** (0.220)
Industry: communication, media, and arts	0.493** (0.202)	0.540*** (0.201)	0.493** (0.202)
Industry: financial services, insurance	0.857*** (0.191)	0.916*** (0.190)	0.853*** (0.192)
Industry: housing	−0.923* (0.542)	−0.926* (0.541)	−0.925* (0.542)
Industry: healthcare, social services	0.183 (0.148)	0.203 (0.149)	0.178 (0.149)
Establishment size: 2000 employees and more	(Reference)	(Reference)	(Reference)
Establishment size: less than 5 employees	−0.407* (0.220)	−0.613*** (0.213)	−0.404* (0.220)
Establishment size: 5–19 employees	−0.533*** (0.158)	−0.730*** (0.150)	−0.531*** (0.159)
Establishment size: 20–199 employees	−0.252** (0.117)	−0.343*** (0.116)	−0.249** (0.118)
Establishment size: 200–499 employees	−0.322** (0.129)	−0.335** (0.130)	−0.318** (0.130)
Establishment size: 500 bis 1999 employees	−0.233* (0.130)	−0.234* (0.130)	−0.231* (0.130)
Constant	−1.992*** (0.193)	−1.781*** (0.186)	−0.2000*** (0.195)
Valid cases, pseudo‐R 2 (Nagelkerke)	4297/0.121	4297/0.116	4297/0.121

Note: Source: Survey “Die Situation von Erwerbstätigen in Zeiten von Corona,” June 2020.

***/**/* significant at 1/5/10 percent level.

DISCUSSION AND CONCLUSION

The COVID‐19 pandemic is an unprecedented public health crisis, killing more than 5.9 million people to date globally (World Health Organisation, 2022). Unsurprisingly, there has been intense scrutiny of governmental efforts to curb the spread of the virus and prevent deaths, bringing the performance of national healthcare systems into sharp focus (e.g., Desson et al., 2020). While safeguarding public health is the highest priority, the pandemic has also caused tremendous disruptions to the world of work. This raises an important question: How do IR institutions protect workers during a crisis of this magnitude?

We addressed this question with a focus on Germany, which has a distinctive “dual system” of employee interest representation through works councils and collective bargaining by unions. Previous research has focused on the economic performance of these IR institutions, paying little attention to their effects on employee interests (Jirjahn & Smith, 2018; Nienhüser, 2020). To help address this gap in the current knowledge, we applied a decommodification lens to conceptualize how these two IR institutions have served to protect different employee interests to date during the pandemic.

We found that employees in representative environments fare better on a range of protective outcomes. Table 6 summarizes the novel empirical insights of our analysis:

TABLE 6

Summary of regression results

Dimension of decommodification	Dependent variables	Industrial relations institution
Job and income security	Perception of job security	Works council: n.s.
		Collective agreement: +
	Payment to top up short‐time work compensation	Works council: +
		Collective agreement: +
Skill development	Participated in training during pandemic	Works council: +
		Collective agreement: n.s. (when works council is included)
Boundary management (reconciling work and family life)	Provisions for home office available	Works council: +
		Collective agreement: not included

Looking first at collective bargaining, we found that it gives workers stability in uncertain times by providing job and income security. Not only do workers covered by collective agreements feel less worried about losing their jobs due to the pandemic than their non‐union counterparts, but they also benefit from top‐up payments that supplement limited government subsidies during periods of short‐time work. Our findings thus add weight to prior research highlighting the benefits of collective bargaining for employees (Bossler, 2019; Lübker & Schulten, 2021).

Turning to works councils, our results show they play a critical role in decommodifying non‐monetary aspects of work by promoting workforce skill development and work–life balance. Notably, employees represented by works councils are also more likely than unrepresented workers to benefit from top‐up payments during short‐time work even when controlling for collective bargaining. This finding is remarkable, as collective bargaining over wages is usually the prerogative of trade unions, rather than works councils at the establishment level. Our analysis therefore extends the limited empirical evidence on the actual protections afforded to employees by works councils (Jirjahn & Smith, 2018; Nienhüser, 2020). Overall, our findings indicate that the dual system of IR “delivered the goods” for German workers during the pandemic. This is a success story for both unions and works councils, and a potentially useful insight to help frame their ongoing campaigns to recruit and mobilize workers for collective action (Carstensen et al., 2022; Gahan & Pekarek, 2013).

The different ways in which collective bargaining and works councils have helped to decommodify work during the pandemic—in the main, by improving the monetary or more qualitative conditions of work, respectively—are consistent with traditional accounts of the institutional “division of labor” in Germany's dual system of IR (Keller, 2008). However, as the pandemic has unfolded, alongside the question of monetary conditions, the qualitative aspects of work have attracted growing concern. For example, lockdowns—involving the closure of schools and childcare services to help drive down new infections—have amplified the challenge of balancing work and caring commitments for many workers (Kohlrausch & Zucco, 2020). Further, with the pandemic transforming work practices and increasing labor market uncertainty, employees' ability to access workplace training has become a more urgent consideration (ILO, 2021a, 2021b). And while not an explicit concern of our analysis, questions of workplace health and safety to reduce infection risks have become paramount for those employees who must physically report to work (Franklin et al., 2020).

As these examples suggest, the pandemic has underscored the importance of works councils as the primary channel of representation on these qualitative concerns at the enterprise level. A fruitful avenue for future comparative analysis would be to examine how workers in countries with different systems of employee representation have fared during the pandemic. For example, workers in LMEs such as the UK and the United States lack statutory mechanisms to create the works councils found in LMEs such as Germany (Wilkinson et al., 2020). This raises questions about the relative capacity of unions and other employee voice mechanisms to mitigate the pandemic's impact on workers in those countries.

While we found that collective bargaining and works councils have so far “delivered” for German workers during the pandemic, we do not wish to imply that they are immune to change or, indeed, erosion. The continued decline in both works council and collective bargaining coverage is well documented, as is their uneven distribution across the Germany economy (Ellguth & Kohaut, 2021; Schroeder & Keudel, 2008). Our findings broadly reflect the core–periphery landscape portrayed in dualization accounts (Hassel, 2014; Thelen, 2014), in that we see relatively strong IR institutions covering workers in large manufacturing firms and the public sector (the core), while these safeguards tend to be weak or absent in private services, a sector that is dominated by small‐ to medium‐sized firms (the periphery). This means workers in the core and the periphery are likely to experience the potential dislocations of the pandemic very differently.

For workers on the periphery within Germany and beyond, the pandemic could either further reinforce their precarity or prove to be a turning point toward institutional reintegration and decommodification. For example, Herman et al. (2021) show that some employers in the UK will “never let a good crisis go to waste,” and will use the pandemic as a catalyst to lower the employment conditions of precarious workers. At the same time, COVID‐19 has also laid bare the essential but undervalued nature of many jobs in areas such as retail, health care, and transport (Hennekam et al., 2020). The pandemic has fueled political discussion of the need to raise wages and improve working conditions for these unsung heroes, with unions campaigning hard to capitalize on the newfound public goodwill toward frontline workers (UNI Global Union, 2020). If these campaigns succeed, the gap between the core and the periphery may change for the better.

Our findings also have some important implications for understanding how IR institutions change over time. One prominent account holds that slow and incremental changes can add up gradually to produce institutional transformation (Streeck & Thelen, 2005; Thelen, 2009). These accrued changes have resulted in institutional “liberalization” in CMEs, which can be understood as “the steady expansion of market relations in areas that under the postwar settlement of democratic capitalism were reserved to collective political decision making” (Streeck & Thelen, 2005, p. 30). However, as our findings indicate, in areas where key IR institutions still enjoy substantial coverage, they have the capacity to take on new tasks and functions. The pandemic has elevated qualitative aspects of work (e.g., work–life balance, skill development, OHS) as key concerns for worker representatives. This is not to suggest these dimensions of decommodification have displaced collective bargaining over wages as the traditional focus of German unions. Indeed, institutions for safeguarding job and income security are of critical importance for protecting workers during the pandemic. However, in Germany's dual system of IR, our analysis implies that the novel emphasis on non‐monetary aspects of decommodification is not a zero‐sum game at the expense of traditional union concerns. Our findings suggest that rather than ceding ground to market forces, as per liberalization accounts (Baccaro & Howell, 2017; Streeck, 2009; Thelen, 2014), Germany's key IR institutions have expanded their protective repertoire to address the novel workplace challenges posed by COVID‐19. Whether and how these key IR institutions can maintain this more expansive role in the long run is an important topic for future research.

Finally, while much recent IR research has focused on macro‐level institutional developments (Johnstone et al., 2019), our results underscore the value of examining how institutions shape IR practices “on the ground” to gain a fuller picture of what and how they deliver for workers (Wilkinson & Wood, 2017). As noted, the pandemic saw many workers face new or intensified challenges in balancing work and caring responsibilities. Yet, mainstream IR research has long neglected the close relationship between production in the workplace and social reproduction in the home (Rubery & Hebson, 2018). While macro‐level analyses provide an important lens for understanding IR dynamics, it is at the granular level of the workplace that employees must navigate the work–life interface, and they look to their representative institutions for support and protection. Indeed, our findings point to the critical role of works councils in establishing home office provisions that help employees manage work–life demands. Complementing recent debates over macro‐level institutional continuity and change in IR, our study affirms that adopting a bottom‐up perspective can yield valuable insights on how IR institutions address the evolving protective needs of workers in practice.