Institutional continuity in German collective bargaining: Do employer associations contribute to stability?

Abstract

Employer associations’ organizational capacity to extend collective bargaining coverage (CBC) beyond unionization levels is one important factor contributing to the stability of multi-employer bargaining in Germany. Based on a representative sample of 142 German employer associations, this article carries out an empirical examination of the characteristics influencing this capacity. The major results are that a wider political domain and adherence to social partnership positively contribute to associations’ static capacity to extend CBC, as approximated by membership density. Yet, in a dynamic perspective – approximated by membership growth rates – a paradoxical trend is revealed by which organizational stabilization of associations is achieved at the expense of decoupling firms’ membership status from collective bargaining arrangements, weakening employer associations’ future capacity to extend CBC.

Keywords

collective bargaining employer associations German industrial relations

Introduction

Among European countries, Germany exhibits a unique and relatively stable system of industry-wide collective wage bargaining which contributes to a comparatively high level of collective bargaining coverage (CBC). Together with co-determination at the company and firm level, this bargaining system can justifiably be said to be one of the core institutions associated with German industrial relations (Silvia and Schroeder, 2007; Streeck and Hassel, 2003; Thelen and Van Winbergen, 2003). Apart from trade unions’ strength, employer associations’ organizational capacity to extend industry-wide collective agreements to the wider population of firms might justifiably be said to represent one of the core foundations of this collective bargaining system. And as long as German employer associations can preserve their capacity to extend CBC, regardless of some erosion of the system at the margins, expectations of institutional continuity in multi-employer bargaining are justified (Thelen, 2000).

However, the last two decades saw developments that make it doubtful whether this description of the German collective bargaining system is still valid today. Shrinking coverage of industry-wide, multi-employer collective agreements and membership decline of trade unions and employer associations contribute to concerns about the institutional continuity of the extant system (Ellguth and Kohaut, 2010a, 2010b; Schroeder and Ruppert, 1996). In particular, concerns have been rising about whether employer associations, i.e. collective organizations of firms specialized in industrial relations and wage bargaining, still play their part in stabilizing collective bargaining (Streeck and Hassel, 2003). Responding to declining membership in the late 1990s, employer associations in the metal industry began offering companies a ‘bargaining free’-membership status as an alternative, allowing them to continue to use association services without being bound by industry-wide collective agreements (Haipeter and Schilling, 2006). This was a new development which has since spread to other industries, making the contribution of employer associations to the integration of firms into industry-wide bargaining a debatable affair.

From this starting point, this article examines whether German employer associations still have sufficient room for manoeuvre to extend CBC beyond unionization levels by preserving their current membership densities and even increasing their absolute membership. As a first step, the main characteristics of the German system of collective bargaining are presented and compared to collective bargaining in two other large European economies, i.e. the UK and France. Second, the article develops a member association perspective which is based on considerations from organizational theory to study CBC extension. With reference to (1) an association’s policy domain, (2) the range of the club goods an association provides for its member firms, (3) the adherence to social partnership as an ideology conducive to cooperative relationships with trade unions and (4) a decoupling of firms’ membership status from collective bargaining arrangements, four hypotheses are derived. Third, regression analyses – based on data from a unique survey of 142 German employer associations conducted in the years 2005 and 2006 – are carried out in order to examine the impact of these organizational characteristics on the static and the dynamic aspects of employer associations’ capacity to extend CBC, approximated by membership densities and membership growth respectively. The article concludes with a summary and discussion of these results.

The German collective bargaining system in comparison

To start with, Table 1 compares the French, the German and the British collective bargaining systems. Besides the comparable size of their economies, these three countries have been selected for illustration because in their variety between state intervention (France), laissez-faire (UK) and meso-corporatism (Germany) they represent the range of ideal-typical collective bargaining systems in Europe (Hall and Gingerich, 2009; Traxler et al., 2007; Wallerstein et al., 1997). For comparison, standard indicators of collective bargaining arrangements are given (Traxler et al., 2007). Union density is defined as the share of unionized employees in the workforce. Collective bargaining coverage (CBC) measures the share of employees whose employment conditions are regulated by collective wage agreements. Extension of CBC is defined as the difference between union membership density and collective bargaining coverage. Collective bargaining type is distinguished between bargaining at the level of single employers (‘single employer’) or bargaining at the level of groups of firms (‘multi-employer’). Inter-associational fragmentation refers to the existence of a separate peak association for labour and social policy; in non-fragmented systems only one peak federation for both policy fields exists.

Table 1.

The German collective bargaining system compared

	Collective bargaining coverage (CBC)^a		Union density^a		Extension of CBC^a		Institutional characteristics^b
Year	1996	2005	1996	2005	1996	2005	Collective bargaining type	Peak associations
UK	34.5	35.3	31.7	29.0	+ 2.8	+ 6.3	Single employer	Integrated
Germany	73.4	62.8	28.7	21.2	+ 44.7	+ 41.4	Multi-employer	Fragmented
France	93.7	98.0	8.3	7.7	+ 85.4	+ 90.3	Multi-employer	Integrated

In percentages of waged and salaried employees.

According to Traxler et al. (2007).

Sources: UK: DTI (2007) based on labour force survey. Germany: (a) bargaining coverage: Ellguth and Kohaut (2007) based on establishment panel data (IAB); (b) union density: DGB (2008) based on administrative membership data, BMAS (2008) based on population survey. France: (a) bargaining coverage: DARES (1997), DARES (2006) based on establishment data; (b) membership density: Andolfatto and Labbé (2006a, 2006b) estimation based on administrative membership data.

Table 1 depicts the relative stability of collective bargaining patterns. These differ across countries, but only slightly across time: no change in the relative position of countries can be observed, although the UK and France show a slight increase in the extension of CBC, whereas Germany shows a minor decrease. CBC is lowest in the UK, where it reached 35 percent in 2005. In contrast, trade union density is higher than in Germany and France, although declining slightly from 31.7 percent to 29.0 percent between 1996 and 2005. In the UK, CBC does not exceed trade union membership density significantly (6.3 percent), but shows a slight gain over the last decade. In Germany and France, CBC extension is much higher. In France, CBC exceeds union membership density by 90 percent in the year 2005; in Germany, CBC exceeds union membership density by 40 percent. Although membership density in German unions is even lower than in the UK, nearly two-thirds of the workforce (62.8 percent) are covered by collective bargaining agreements – compared to just 35.3 percent in the UK (see Table 1). Regarding the organizational structure of peak-level associations, Germany is different in that there is one federation specialized in trade and competition policy, i.e. the BDI (Bundesverband der Deutschen Industrie), and one specialized in labour policy and industrial relations, i.e. the BDA (Bundesvereinigung der Deutschen Arbeitgeberverbände). In France and the UK, peak-level associations are integrated across policy domains, i.e. represented by the Mouvement des Entreprises de France (MEDEF) and the Confederation of British Industry (CBI) respectively.

Theoretical approaches to collective bargaining coverage

Traditional explanations for the huge differences in CBC across national institutional settings differ according to the theoretical perspective taken to explain the extension of collective agreements. Firm-level explanations highlight voluntary extension by individual firms as the central mechanism. Based on Olson’s (1965) account of the logic of collective action, free-riding of potential member firms can only be overcome if selective goods or regulatory force motivate firms to participate in associations. In the absence of obligatory membership rules, employer associations depend on the services they offer to their members besides collective bargaining, for example legal advice or consulting (Olson, 1965). Seen this way, both the extension of CBC and the propensity of firms to form employer associations can be derived entirely from individual firms’ cost and benefits calculations regarding collective action as revealed by their decision to join an association (Kohaut and Schnabel, 2003; Schnabel et al., 2006). Applied to the country comparison given in the preceding section, the micro-level approach seems to be most appropriate for the situation in the UK. Here, extension of CBC is limited to voluntary extension by individual firms in single-employer bargaining, because union recognition procedures foreclose automatic extension of CBC beyond firms directly organized by unions. The result is a rather low level extension of CBC beyond unionized workplaces (Howell, 2005).

In contrast to that, organizational theory emphasizes that employer associations and their organizational characteristics also matter for extension of CBC. According to this view, associations can be regarded as the structural expression of a general business’s need to organize collectively to take part in the political process, because individual firms lack the capacity to influence political outcomes by unilateral investment decisions or informal coordination (Schmitter and Streeck, 1999). Traditionally, organizational theory has stressed (neo-)corporatist arrangements, i.e. the institutional macro-structure defined by state intervention, as conducive to the extension of CBC. Therefore, empirical studies have focused on macro-level comparisons between countries (e.g. Hall and Gingerich, 2009; Traxler, 2004; Traxler et al., 2007; Wallerstein et al., 1997). According to Traxler et al. (2007), the most important factors explaining CBC extension are state regulation and inter-associational fragmentation, i.e. the separation of employer associations from trade associations. This macro-level approach focusing on institutional features seems to be most suitable for cases such as France where state regulation is mainly responsible for the extension of CBC. In France, the high level of CBC is due to the fact that collective bargaining agreements are extended by law to firms showing only negligible levels of trade union membership and union representation (Andolfatto and Labbé, 2006a, 2006b; Gumbrell-McCormick and Hyman, 2006; Saurugger, 2007).

A member organization based view of collective bargaining extension

Building on the organizational perspective, in this article it is argued that the peculiar German type of CBC extension can be explained best by looking at employer associations directly organizing member firms. By focusing on member associations, this article neither neglects associations’ distinct character as an institutionalized type of organization nor subordinates member associations solely under the features of peak-level associations or state regulation. Such a perspective seems to be most appropriate for the German situation in which neither voluntary extension by firms nor obligatory extension by law can explain the extension of CBC beyond unionization levels (Ellguth and Kohaut, 2007; Traxler et al., 2007). However, this perspective demands a closer look at the organizational characteristics of member associations to explain CBC extension in Germany.

A member organization based view of CBC extension can be summarized as follows. In Germany, in the absence of encompassing extension regulations by the state, employer associations stabilize the collective bargaining system by extending collective agreements to a wider population of firms. In theory, associations’ capacity to extend CBC beyond unionization levels can be derived from organizational processes providing associations with the organizational autonomy typical for formal organizations.

According to such an organizational view, associations’ autonomy rests on the transfer of decision rights from individual members to associations (Coleman, 1982; Thelen, 2000) and the professional administration of political representation (Michels, 1966). From an organizational perspective, associations endowed with these characteristics are generally more reliable, possess a high reproducibility of service quality and can be made accountable for deviations from standard procedures (Hannan and Freeman, 1984). Moreover, by seizing control over interest representation vis-a-vis external political actors on a regular basis, associations achieve the expertise to define new areas of collective action and abandon old ones, to mitigate the centrifugal force of competition among members and to frame ideologies conducive to their role as mediators (Barnett, 2006; Oliver, 1993; Schmitter and Streeck, 1999). Based on this organizational autonomy, employer associations realize their capability by preserving a high membership density, i.e. the relation of the number of actual to the number of potential member firms of a given business community, or by increasing the absolute number of members.

Two important qualifications must be made regarding the notion of ‘organizational autonomy’. First, although employer associations have achieved a status of independence from individual member firms, there is a fundamental tension with regard to their capacity of autonomous decision-making. This tension resides in the contradiction between the interests of an association’s constituency and the interests of external negotiation partners (Child et al., 1973; Schmitter and Streeck, 1999). Consequently, autonomy of associations becomes a relative one which is, on the one hand, restricted by the need to meet the demands of (potential) member firms in order to organize a viable number of firms, and on the other hand, restricted by the demands of external actors like trade unions. Second, the relative autonomy of employer associations is also restricted internally by the differences in the resource endowments of small and large firms (Streeck and Visser, 2006; Traxler, 2007). For example, compared to large corporations, German small and medium-sized firms are more inclined to circumvent CBC extension, since many SMEs face competition carried out predominantly by unit labour costs (Langer, 1994; Schnabel et al., 2006; Thelen and Van Winbergen, 2003; for a critique see Silvia and Schroeder, 2007).

Hypotheses

In order to examine whether German employer associations continue to be an independent source of institutional stability of collective bargaining, four organizational characteristics of associations are tested for their impact on associations’ capacity to extend CBC: (1) an association’s policy domain (H1), (2) the range of club goods an association provides for its member firms (H2), (3) the adherence to social partnership as an ideology conducive to cooperative relationships with trade unions (H3) and (4) a decoupling of firms’ membership status from collective bargaining arrangements (H4).

By enlarging its policy domain, an association increases the number of firms within its reach, thereby attaching more firms to collective bargaining (Traxler, 2007). If already organized members are represented by their association in more policy arenas, members identify more strongly with their association and could develop more similarity in their political demands (Kim and Bearman, 1997). A broader policy domain also serves a broader range of potential members’ political interests. All else being equal, a wider definition of an association’s political activities should lead to an increase in an association’s capacity for CBC extension.

Hypothesis 1: The wider an employer association defines its policy domain, the greater is its organizational capacity to extend CBC.

From an organizational perspective, the most interesting point of the Olsonian tradition is the insight that an association must attract enough clients for its service to stay in the business of interest representation (Olson, 1965). A wider definition of the range of services, i.e. an increase in the number of services provided as club goods like legal advice in labour relations or vocational training, makes an association more attractive for potential members to actually join an association and provides incentives for the already organized firms to remain members (Schnabel and Wagner, 1996). Ceteris paribus, the supply of a higher number of club goods should increase employer associations’ capacity to extend CBC.

Hypothesis 2: The more club goods an employer association offers, the greater is its organizational capacity to extend CBC.

Intermediating firms’ goals with unions’ goals is one of the central tasks of employer associations in collective bargaining. Associations can achieve such an intermediating function more easily if they use framing ideologies that define a common ground for negotiations (Benford and Snow, 2000). In the case of German collective bargaining, employer associations can entertain an ideology of ‘social partnership’ in order to define more cooperative relations with the unions (Behrens and Jacoby, 2004; Streeck and Hassel, 2003). Ceteris paribus, the stronger an association’s adherence to the ideology of social partnership is, the greater its capacity to extend CBC should be.

Hypothesis 3: The stronger the adherence to social partnership of an employers’ association, the greater is its organizational capacity to extend CBC.

As outlined in the introduction, the rise in ‘bargaining free’-membership status has become a highly debated topic in the literature on German collective bargaining (Haipeter and Schilling, 2006; Streeck and Hassel, 2003; Völkl, 2002). A decoupling of the membership status from collective undermines the link between an association’s membership and CBC, which has serious implications for the argument that associations’ organizational autonomy is closely connected to the extension of CBC. By allowing firms to join without the duty to formally accept collective bargaining agreements, employer associations might stabilize themselves as organizations by simultaneously contributing to the erosion of the German model of collective bargaining. As a consequence, ‘bargaining free’-membership status can be interpreted as a proxy for organizational change which is detrimental to employer associations’ capacity to extend CBC.

Hypothesis 4: If an employer association uses decoupling by having a ‘bargaining free’-membership status, its organizational capacity to extend CBC is reduced.

Data and findings

Data description and measurement

The data come from a survey among 1053 top managers, i.e. chief executive directors or presidents of all German business associations, conducted in the years 2005–6. The postal survey was responded to by 244 business associations, which is a response rate of 23 percent. The survey questionnaire covered topics ranging from associations’ domains, membership structure, membership growth and associations’ business services to characteristics of associations’ executives. However, not all of the responding associations are employer associations, and not all responding employer associations are subject to the present analysis. With respect to the first selection problem, it was necessary also to include trade associations and craft associations – the so-called Innungen (roughly trade guilds) (Thelen, 2004) – into the original address data collection, which was done by using different sources like administrative registries of public addresses (Oeckl, 2006) and internet pages of (peak) associations. This approach is due to the rather complex structure of business interest intermediation in Germany, making it impossible to judge from the outside whether an association is de facto an employer association. The second selection problem is a conceptual one, because the theoretical hypotheses only apply to those associations that directly organize member firms. Therefore, associations of professions, pure trade associations as well as peak-level employer associations – organizing associations themselves – have all been excluded from the present analysis.

Based on respondents’ statements, 142 of all associations in the sample classify as employer associations directly organizing member firms. This provides an adequate representation of the entire population of German employer associations, estimated to consist of around 650 member associations (Behrens, 2010). In sum, these employer associations influence the employment conditions of around 4.9 million employees. Most of these employer associations operate within their respective industry domain at the regional state level (90 percent), and in a small minority of cases at the federal or the local level. The average membership size is 571 member firms. On average, they have 11 employees and nearly 40 percent of them command a yearly budget in the range between €750,000 and €2.5 million. Another 48 percent of associations have a budget below €750,000; the remaining 12 percent dispose of more than €2.5 million annually. The majority of these employer associations operate in manufacturing industries (53 percent), the metal industry and the chemical industry having the lead. The other associations operate in services with retailing as the most important segment, and in the construction trades.

Membership density and membership growth of employer associations

To test the hypotheses on employer associations’ capacity to extend CBC, associations’ membership density and membership growth rates are used as proxy indicators (for an overview on variables definition and description, see Table 2). A parallel analysis of both dependent variables has important advantages, because they capture the static and dynamic aspect of CBC extension respectively. In the absence of obligatory membership rules, an employer association’s membership density, i.e. the relationship of the number of member firms to the total number of firms in its business community, approximates the static level of CBC extension at a given point in time. An association’s membership growth, i.e. the average annual growth rate of absolute membership figures, reveals how an association can preserve a given CBC level over time.

Table 2.

Definition and description of variables

Variable name	Indicator (range of values)	Relative frequency or mean
		Sample (N = 142)	Density estimation (n = 96)	Growth rate estimation (n = 100)
Dependent: Membership density	Actually organized member firms as percentage of potential member firms (5-100%)	55.60	55.72	–
Dependent: Membership growth	Average rate of annual change in membership in percent (-15.9 to 104.5%)	−0.32	–	−0.33
Political domain	Association active in policy domain (0 = industrial relations only, 1 = industrial relations and trade policy)	0.33	0.33	0.31
Club goods	No. of club goods provided to member firms (from 1 up to 8 services provided)	5.80	6.00	6.00
Social partnership	‘German social partnership is an advantage in global competition’ (1 = absolutely disagree to 5 = absolutely agree)	2.64	2.66	2.66
Decoupling	‘Bargaining free’-membership status (0 = no, 1 = yes)	0.67	0.65	0.65
Age	Age of association in years since founding (6-181 years)	55.36	55.55	55.61
East Germany	Association active in East Germany (0 = West, 1 = East)	0.13	0.15	0.14
Member group (size)	Most important member group by firm size (1 = up to 20, 2 = 20 up to 100, 3 = 100 up to 250, 4 = 250 up to 500, 5 = 500 employees and more)	2.27	2.33	2.28
Member group (firm age)	Most important member group by firm age (1 = up to 5, 2 = 5 up to 15, 3 = 15 up to 25, 4 = 25 up to 35, 5 = 35 years and more)	4.00	3.98	3.95
Trade unions	Chemical industry union (IG BCE)	0.13	0.17	0.15
	Metal industry union (IG Metall)	0.26	0.18	0.19
	Services union (Ver.di)	0.25	0.23	0.23
	Small unions	0.35	0.42	0.43

For the following analysis, both dependent variables are measured subjectively as they have been self-reported by employer associations’ executive directors. The average membership density in the selected sample has a mean value of 56 percent for 2005–6: 64 percent of the employer associations in the sample report a density rate of 50 percent or higher, while only 14 percent report a membership density below 25 percent. To capture the dynamics of CBC extension, the average annual rate of membership growth between 2001 and 2005 is used. On average, the yearly growth rate is slightly negative (–0.32 percent), indicating that German employer associations have lost members in the years between 2001 and 2005.

Associations’ domain definition, club goods, ideology and decoupling

Independent variables are measured as follows. An association’s political domain is approximated by a dummy variable indicating whether an association reports to engage only in industrial relations or in both, trade policy and industrial relations (H1). Associations that operate within both policy fields are said to have a wider political domain than those associations only operating within the field of industrial relations. Those associations with a wider political domain, i.e. mixed employer and trade associations, account for 33 percent of the associations in the sample; 67 percent are pure employer associations.

The scope of the service range is measured by the number of services an association reports to provide as club goods (H2), i.e. services exclusively offered to its member firms. The most important services are legal advice in labour law-suits, vocational training and consultancy in personnel management. Other services include provision of industry information, auditing in the implementation of standards, working groups, general business consulting and publication of professional periodicals. The average number of club goods provided by associations is 6 different services.

As argued earlier, employer associations adhering to a more collaborative ideological orientation (H3) towards unions should be capable of bridging contradicting demands of unions and firms more easily. As a proxy for such an ideological orientation, executive managers’ assessments – due to their role as key informants about an association’s policy – are used based on the statement ‘German social partnership is an advantage in global competition’. This statement was measured on a five-point scale, ranging from ‘absolutely disagree’ to ‘absolutely agree’; 24.4 percent of associations’ executive managers agree with this statement, 45.0 percent strongly disagree or disagree and 30.6 percent are undecided on the issue.

A major limitation of employer associations’ capacity to extend CBC is decoupling membership status from collective bargaining (H4) by using a ‘bargaining free’-membership option for firms. As the data used here indicate, roughly two-thirds of employer associations use ‘bargaining free’-membership status. These associations, on average, report that 32.7 percent of their members actually make use of the ‘bargaining free’-option, lending some support to include this variable as an indicator for organizational change reducing employer associations’ capacity to extend CBC.

Control variables

A number of factors intervene in the relationship between associations’ organizational characteristics and their capacity to extend CBC. In particular, an association’s membership structure, its age, its operating region and the industry-specific culture of collective bargaining seem to be important in this regard.

The main results of micro-level studies are that larger and older firms are more often members in associations and more likely to be covered by collective bargaining (Ellguth and Kohaut, 2007; Kohaut and Schnabel, 2003; Schnabel and Wagner, 1996; Schnabel et al., 2006). To capture these age and size effects, the size and age classes of associations’ most important membership groups are used as control variables indicating membership structure. For the majority of associations, small and medium-sized firms, i.e. firms up to 250 employees, are the most important member group (78.6 percent), whereas associations dominated by member firms with 250 employees and more account for 21.4 percent of the sample. With respect to firms’ age, associations’ membership is dominated on average by firms between 25 and 35 years old.

With respect to an association’s age, it can be assumed that older associations are the backbone of the traditional German model of collective bargaining. As this model has come under pressure by defections of individual member firms in the last years, one would expect older associations to reveal a membership decline, although on a comparatively high density level. Most of German employer associations were founded directly after the Second World War, the time the German model started to evolve gradually into its present form (Schroeder, 2000; Streeck, 1997). On average, the associations in the sample are 55 years old.

Similarly, associations should be distinguished by region, because industrial relations of the West German kind have still not been extended completely to the former East Germany and it is widely disputed as to whether collective bargaining institutions can be regarded as having become fully functional in the east in a manner comparative to the western part of Germany (Artus, 2001; Schroeder, 2000). Consequently, East German associations should have a lower membership density than West German associations; 13 percent of the associations in the sample are located in the eastern part of Germany.

The union-specific culture of collective bargaining is also of importance for employer associations’ capacity to extend CBC, in particular because the bargaining style of the union might have an impact on individual employers to join or leave an association due to wage pressure. On the level of member associations, the union-specific culture can be approximated by recoding associations’ industry classification according to the domains of the relevant trade unions. This reclassification was done by using the registry of bargaining agreements provided by the WSI Tarifarchiv (WSI, 2005). In the analysis, the employer associations dealing with the three largest German trade unions are compared, i.e. IG Metall (metal industry, paper, wood, plastics and textiles) with 2.3 million members; Ver.di (printing, retailing and financial services) with 2.1 million members; and IG BCE (chemical industry, mining, glass, rubber) with 0.7 million members. Whereas IG BCE is a very partnership-oriented trade union (Behrens and Jacoby, 2004), both IG Metall and Ver.di have a reputation for a rather resolute bargaining style which also includes occasional strike action. Employer associations dealing with these larger unions are compared with those associations either confronted with smaller unions like the construction union (IG Bau) or the food processing, catering and hotels union (NGG), or those associations operating in industry segments at the margins of established union jurisdictions like temporary work, industry-related services or the craft sector. This heterogeneous group is characterized by a fragmentation of industrial relations cultures in which the unions have often a rather weak presence, and sometimes, are even non-existent at the workplace level (Bosch and Kalina, 2007; Völkl, 2002).

In Table 2, the descriptive statistics for the subsamples of the separate linear regressions for membership density and membership growth rates are also reported. Compared to the whole sample of 142 cases, the drop-out due to single missing values, leading to a listwise deletion of 46 and 42 cases respectively, does not distort the variables as they enter into the estimations. The only exception is the decrease in the share of associations operating within the jurisdiction of IG Metall. Checking for relations among the regressors, Table 3 shows the partial correlation matrix of the independent variables. Only 21 percent of the entries in Table 3 show a considerable and highly significant (at the .01-level) partial correlation coefficient, indicating that there are no serious problems with collinearity among the regressors. The most relevant correlations can be observed between the control variables of association age, the dummy for East Germany and membership structure by firm age. This is due to the founding process of member associations after German reunification in 1990.

Table 3.

Partial correlation matrix of independent variables

	Political domain	Club goods	Social partnership	Decoupling	Age	East Germany	Member group (size)	Member group (age)	IG BCE	IG Metall	Ver.di
Club goods	−.02
Social partnership	.19**	−.03
Decoupling	−.08	.09	.02
Age	.12	.11	.21**	.10
East Germany	−.10	−.10	−.09	.05	−.45***
Member group (size)	.07	.01	.03	.14	.25***	−.23**
Member group (age)	.08	.02	.07	.03	.43***	−.51**	.18**
IG BCE	.16*	−.10	.16*	−.07	.23***	−.09	.18**	.27***
IG Metall	−.14*	.21**	−.01	.16*	.05	−.09	.08	.03	−.23***
Ver.di	−.03	−.18**	−.09	.28***	−.09	.06	−.07	.13	−.23***	−.35***
Small unions	.05	.04	−.02	−.33***	.13	.10	.14*	−.33***	−.29***	−.44***	−.43***

Notes: sig. at ***= .01, **= .05, *= .01 level. N =142.

Findings

Table 4 contains the results of four regression models of employer associations’ capacity to extend CBC beyond unionization levels, as measured by membership density and membership growth. After testing for control variables, the four organizational characteristics of associations are introduced into the regression. All estimations reported assume a standard linear model of the OLS type. In the case of membership density, it has been considered that density is a fraction what implies technical limitations to performing an OLS regression (Papke and Wooldridge, 1996). However, estimation of a general linear model (GLM) produces redundant results not reported here.

Table 4.

Membership density and membership growth of employer associations

	Membership density Base model (OLS)		Membership density Full model (OLS)		Membership growth Base model (OLS)		Membership growth Full model (OLS)
	B	t-value	B	t-value	B	t-value	B	t-value
Political domain	–	–	9.52*	1.94	–	–	−2.37	−0.91
Political domain	–	–	(4.90)		–	–	(2.61)
Club goods	–	–	−0.54	−0.36	–	–	−1.45*	−1.83
	–	–	(1.50)		–	–	(0.79)
Social partnership	–	–	4.03**	2.13	–	–	−0.57	−0.58
Social partnership	–	–	(1.89)		–	–	(0.99)
Decoupling	–	–	2.27	0.47	–	–	5.45**	2.08
	–	–	(4.85)		–	–	(2.62)
Association age	−0.07	−0.78	−0.13	−1.40	−0.13***	−2.77	−0.13***	−2.73
Association age	(0.09)		(0.09)		(0.05)		(0.05)
East Germany	−12.72	−1.54	−12.14	−1.51	−12.05***	−2.82	−12.45***	−2.95
East Germany	(8.24)		(8.05)		(4.27)		(4.22)
Member group (size)	6.17***	3.50	5.88***	3.39	0.65	0.71	0.71	0.77
Member group (size)	(1.76)		(1.74)		(0.92)		(0.92)
Member group (age)	1.39	0.66	1.56	0.75	−2.37**	−2.14	−2.03*	−1.85
Member group (age)	(2.12)		(2.07)		(1.11)		(1.10)
IG BCE	21.50***	3.08	18.91***	2.73	0.63	0.17	−0.88	−0.23
	(6.98)		(6.92)		(3.74)	–	(3.75)
IG Metall	7.60	1.20	10.42	1.65	−3.78	−1.19	−4.54	−1.38
	(6.31)		(6.33)	–	(3.17)		(3.29)
Ver.di	4.69	0.83	6.19	0.97	−2.45	−0.79	−6.33*	−1.85
	(5.99)		(6.36)		(3.11)		(3.42)
Constant	35.66***	3.50	26.26*	1.89	17.82***	3.40	24.92***	3.41
	(10.19)		(13.90)		(5.24)		(7.31)
No. of observations	96		96		100		100
Adj. R²	0.27		0.32		0.12		0.16
F-value	5.90***		5.02***		2.97***		2.75***
χ² (d.f. = 4)	−		11.52**		−		9.15*
d.f.	7		11		7		11

Notes: sig. level: *** .01, ** .05, * .10. Standard error in parentheses. d.f. degrees of freedom.

In general, regression results support the view that organizational characteristics of employer associations matter for employer associations’ capacity to extend CBC, although in a different way for the static and the dynamic dimension respectively. The estimation model for the static capacity of employer associations to extend CBC, as approximated by membership density, improves significantly when the four organizational characteristics are added. Defining the model with only the control variables as the base model, the estimation including the indicators for organizational characteristics reveals a higher explanation power (χ² = 11.52, significance at the .05-level). However, for the dynamic capacity of associations to extend CBC, the model improvement through inclusion of the four organizational variables is a weak one (χ² = 9.15, .10-significance level).

Findings for membership density

Regarding the static dimension in terms of employer associations’ membership density, results in Table 4 indicate that an enlarged policy domain (H1) has a positive, albeit weakly significant impact on associations’ capacity to extend CBC. Associations that are active in two policy areas reveal a membership density that is 9 percent higher than the membership density of their more specialized counterparts. If members can be attracted by interest representation in several fields, they can also be attached more strongly to collective bargaining. Similarly, adherence to social partnership seems to have a positive impact for CBC extension in terms of membership density (H3). For every step on the agreement scale, the membership density increases by 4 percent. However, the number of club goods provided and the decoupling of the membership status from participation in collective bargaining does not affect the static capacity of employer associations to extend CBC.

For the control variables, results indicate that an association with larger firms has a higher static capacity for CBC extension, as approximated by membership density (see Table 4). In both models, one can observe a positive relationship between the most important member group’s size class and membership density. This can be explained by referring to Traxler (2007) and Streeck and Visser (2006), who argue that, in particular, larger firms provide resources to an association that secure high membership density levels. With respect to union relationships, one can conclude from Table 4 that the overall bargaining culture among social partners can have a profound effect on employer associations’ capacity to extend CBC. In particular, associations dealing with the IG BCE have a larger capacity to extend CBC than those dealing with other unions, even if simultaneously controlling for adherence to social partnership and the other characteristics in the estimation model.

Findings for membership growth

With respect to the dynamic dimension of employer associations’ capacity to extend CBC, as approximated by membership growth rates, the estimations reveal paradoxical results. As opposed to the estimation for membership density, results on the policy domain (H1) and the social partnership orientation (H3) are insignificant for membership growth. In addition, employer associations’ dynamic capacity for CBC extension seems to be negatively affected by the range of club goods provided (H2) and positively influenced by decoupling (H4). Both results contradict the hypotheses.

Hypothesis 2 predicts associations with a broader scope of services to have a greater capacity for CBC extension. In fact, the opposite seems to be true if one looks at the dynamic side of CBC extension, which is somewhat puzzling. This unexpected result might be reinterpreted as an indication that firms expect their associations to stick with their core competencies and not to offer additional services members can buy from law firms or private business consultancies.

Hypothesis 4 predicts a negative impact of decoupling membership status from collective bargaining extension by allowing a ‘bargaining free’-option. However, a strong positive association is observed between decoupling and membership growth. As such, this result contradicts the expectation that decoupling reduces the capacity of employers’ associations to extend CBC. Rather, one might say that German employer associations increase this capacity in the short run by choosing the decoupling option to increase their absolute membership, although this might undermine their future capacity to extend multi-employer bargaining, since it is doubtful whether ‘bargaining free’-members will adopt collective agreements later. As it is, this paradoxical result contradicts the stability hypothesis on German collective bargaining.

The most important control variables associated with membership growth are East Germany, followed by association age and the firm age of the most important member groups. Older associations, those with an older membership and those operating in eastern Germany are less capable of attracting additional firms, undermining their organizational capacity to extend CBC. Additionally, declining capacity of CBC extension seems to be most pronounced in services (Ver.di), even more so than in the control group of fragmented industrial relations.

Discussion and conclusions

Summing up, the findings based on a data set representative for German employer associations reveal that German employer associations contribute significantly to the stability of the typical collective bargaining pattern in Germany by extending CBC far beyond unionization levels. Therefore, Germany still provides a unique case of an organization-based extension of CBC that differs markedly from low, voluntary CBC extension in the UK or high, state-regulated CBC extension in France. Furthermore, the results show that it is useful to treat associations as relatively independent actors and to look at employer associations from a perspective inspired by organizational theory. Such a member organization based perspective allows us to study the interplay between organizational characteristics of employer associations and CBC extension in detail. Moreover, the distinction between the static and the dynamic aspects of CBC extension reveals important insights for the evaluation of employer associations’ political strategies towards their existing and potential members, but also towards their union counterparts.

In more detail, this article’s findings highlight a wider policy domain and ideological adherence of employer associations to social partnership as conducive to an extension of CBC. Particularly by widening the scope of their political domain, associations can increase the propensity of firms to stay a member of an association. In this way, employer associations stabilize their extant membership by representing their constituency in more policy arenas and, indirectly, make firms stick with collective bargaining. Additionally, a strong case for the positive impact of cooperative relationships with the unions can be made, as indicated by the ideological adherence to social partnership. Employer associations’ capacity to stabilize collective bargaining rests on their ability to bridge ideological conflicts between their member firms and trade unions by adherence to the ideology of social partnership. This holds true even if the collective bargaining culture of different trade unions is accounted for. Therefore, it seems reasonable to conclude that cooperative relations with unions can help associations to keep members within the association by opening a corridor for compromise.

In a dynamic perspective that looks at the rate of change of absolute membership levels, it can be concluded that other organizational characteristics contribute to an explanation of the change in membership levels than those contributing to an explanation of the absolute level of membership density. As the results suggest, German employer associations do decouple membership status from collective bargaining in order to increase their membership, thereby undermining their capacity for CBC extension in the future. This paradoxical result needs additional and more detailed qualitative research on whether decoupling by introducing a ‘bargaining free’- membership status is an employer association strategy directed predominantly at undermining multi-employer bargaining, or a special recruitment strategy for attracting new firms, or whether this sort of decoupling is better understood as a pressuring device to bring down labour union demands in wage negotiations.

Likewise, it seems surprising that an increase in the number of services provided as club goods is not conducive to increasing associations’ dynamic capacity for CBC extension. This result is even more striking in a comparative perspective, since in other countries the dissolution of multi-employer bargaining has been accompanied by a transformation of employer associations into pure service organizations (Traxler, 2007). In Germany, such transformations are very rare. However, also regarding this finding more research needs to be done to find out whether this finding is a result of members’ expectations of employer organizations to focus exclusively on collective bargaining instead of services, or whether the increase of the service level is a reaction to organizational stress caused by a shrinking membership.

Finally, it should be noted that additional questions for the stability of German collective bargaining remain. For example, the analysis in this article is biased towards extant employer associations. This bias neglects those industries populated by younger, small and medium-sized firms not covered by any form of collective interest representation at all, i.e. operating in an ‘employer association free zone’, for example in the IT industry (Lehrer, 2000). However, based on the results of this article, one may doubt whether this development is more than a transitory phenomenon. Moreover, the current debates in Germany about the minimum wage extension rules and the introduction of a general, nationwide minimum wage show that there is a certain willingness to stabilize wage setting in exactly those parts of the economy (Bosch and Kalina, 2007). Moreover, employers in these industries have already formed their own associations, for example in temporary agency work or postal services. Another topic neglected in this article is the rise in single-employer bargaining by Germany’s larger corporations. Such a rise in voluntaristic collective bargaining policy might undermine employer associations’ capacity to extend CBC (Streeck and Hassel, 2003; Streeck and Visser, 2006). As both of these issues indicate, more research on collective bargaining in settings dominated by voluntary employer associations is necessary.

Footnotes

Acknowledgements

The author would like to express his most sincere gratitude to Martin Behrens, Michael Fichter and Jörg Sydow for commenting on earlier drafts of the article.

Funding

Part of this research was funded by the Hans Boeckler Foundation (grant number S-2006-878-2-1), for which the author expresses his thanks.

Markus Helfen is a PostDoc research fellow working as the chair of Inter-Firm Cooperation associated with the Department of Management at Freie Universität Berlin, School of Business and Economics. His research areas include international labour and employment relations, human resource management, collective action as well as organization theory.

References

Andolfatto

Labbé

(2006a) Histoire des syndicats (1906–2006). Paris: Éditions du Seuil.

Andolfatto

Labbé

(2006b) La transformation des syndicats français. Vers un nouveau ‘modèle social’? Revue Française de Science Politique 56(2): 281–297.

Artus

(2001) Krise des deutschen Tarifsystems. Die Erosion des Flächentarifvertrags in Ost und West. Wiesbaden: Westdeutscher Verlag.

Barnett

(2006) Finding a working balance between competitive and communal strategies. Journal of Management Studies 43(8): 1753–1773.

Behrens

(2010) Strukturen der Interessenvertretung in den Verbänden der Wirtschaft. In: Schroeder

Weßels

(eds) Handbuch Arbeitgeber- und Wirtschaftsverbände in Deutschland. Wiesbaden: VS Verlag, 148–168.

Behrens

Jacoby

(2004) The rise of experimentalism in German collective bargaining. British Journal of Industrial Relations 42(1): 95–123.

Benford

Snow

(2000) Framing processes and social movements: An overview and assessment. Annual Review of Sociology 26: 611–639.

BMAS (Bundesministerium für Arbeit und Sozialordnung). (2008) Arbeitnehmer in Vollzeit- und Teilzeitbeschäftigung. Available at: www.bmas.de/coremedia/generator/18112/statistisches__taschenbuch__2006.html.

Bosch

Kalina

(2007) Niedriglöhne in Deutschland – Zahlen, Fakten, Ursachen. In: Bosch

Weinkopf

(eds) Arbeiten für wenig Geld, Niedriglohnbeschäftigung in Deutschland. Frankfurt and New York: Campus, 20–105.

10.

Child

Loveridge

Warner

(1973) Towards an organizational study of trade unions. Sociology 7(1): 71–91.

11.

Coleman

(1982) The Asymmetric Society. New York: Syracuse University Press.

12.

DARES (Direction de l’animation de la recherche, des études et des statistiques). (2006) La couverture conventionelle a fortement progressé entre 1997 et 2004. Première Informations 46(2).

13.

DARES (Direction de l’animation de la recherche, des études et des statistiques). (1997) La couverture conventionelle à la fin 1997. Première Informations 29(2).

14.

DGB (Deutscher Gewerkschaftsbund). (2008) Mitglieder in den DGB-Gewerkschaften. Available at: www.dgb.de/dgb/mitgliederzahlen/mitglieder.htm.

15.

DTI (Department of Trade and Industry). (2007) Trade Union Membership 2006. London: DTI.

16.

Ellguth

Kohaut

(2007) Tarifbindung und betriebliche Interessenvertretung: Aktuelle Ergebnisse aus dem IAB-Betriebspanel. WSI Mitteilungen 60(9): 511–514.

17.

Ellguth

Kohaut

(2010a) Tarifbindung und betriebliche Interessenvertretung: Aktuelle Ergebnisse aus dem Betriebspanel 2009. WSI-Mitteilungen 63(4): 204–209.

18.

Ellguth

Kohaut

(2010b) Auf der Flucht? Tarifaustritte und die Rolle von Öffnungsklauseln. Industrielle Beziehungen 17(4): 345–371.

19.

Gumbrell-McCormick

Hyman

(2006) Embedded collectivism? Workplace relations in France and Germany. Industrial Relations Journal 37(5): 473–491.

20.

Haipeter

Schilling

(2006) Arbeitgeberverbände in der Metall- und Elektroindustrie. Tarifbindung, Organisationsentwicklung und Strategiebildung. Hamburg: VSA.

21.

Hall

Gingerich

(2009) Varieties of capitalism and institutional complementarities in the political economy: An empirical analysis. British Journal of Political Science 39(3): 449–482.

22.

Hannan

Freeman

(1984) Structural inertia and organizational change. American Sociological Review 49(2): 149–164.

23.

Howell

(2005) The Unions and the State: The Construction of Industrial Relations Institutions in Britain, 1890–2000. Princeton, NJ: Princeton University Press.

24.

Kim

Bearman

(1997) The structure and dynamics of movement participation. American Sociological Review 62(1): 70–93.

25.

Kohaut

Schnabel

(2003) Tarifverträge – nein danke!? Ausmaß und Bestimmungsgründe der Tarifbindung west- und ostdeutscher Betriebe. Jahrbücher für Nationalökonomie und Statistik 223(3): 312–331.

26.

Langer

(1994) Arbeitgeberverbandsaustritte – Motive, Abläufe und Konsequenzen. Industrielle Beziehungen 1(2): 132–154.

27.

Lehrer

(2000) From factor of production to autonomous industry: The transformation of Germany’s software sector. Vierteljahreshefte zur Wirtschaftsforschung 69(4): 587–600.

28.

Michels

(1966) Political Parties: A Sociological Study of Oligarchical Tendencies of Modern Democracy. New York: Free Press.

29.

Oeckl. (2006) Taschenbuch des Öffentlichen Lebens – Deutschland 2006. Bonn: Festland Verlag.

30.

Oliver

(1993) Formal models of collective action. Annual Review of Sociology 19: 271–300.

31.

Olson

(1965) The Logic of Collective Action, Public Goods and the Theory of Groups. Cambridge, MA: Harvard University Press.

32.

Papke

Wooldridge

(1996) Econometric methods for fractional response variables with an application to 401(k) Plan participation rates. Journal of Applied Econometrics 11: 619–632.

33.

Saurugger

(2007) France. In: Traxler

Huemer

(eds) Handbook of Business Interest Associations, Firm Size and Governance. London: Routledge, 122–140.

34.

Schmitter

Streeck

(1999) The Organization of Business Interests: Studying the Associative Action of Business in Advanced Industrial Societies. Cologne: MPIfG.

35.

Schnabel

Wagner

(1996) Ausmaß und Bestimmungsgründe der Mitgliedschaft in Arbeitgeberverbänden: Eine empirische Untersuchung mit Firmendaten. Industrielle Beziehungen 3(4): 293–306.

36.

Schnabel

Zagelmeyer

Kohaut

(2006) Collective bargaining structure and its determinants: An empirical assessment with British and German establishment data. European Journal of Industrial Relations 12(2): 165–188.

37.

Schroeder

(2000) Das Modell Deutschland auf dem Prüfstand. Zur Entwicklung der industriellen Beziehungen in Ostdeutschland. Wiesbaden: Westdeutscher Verlag.

38.

Schroeder

Ruppert

(1996) Austritte aus Arbeitgeberverbänden: Motive-Ursachen-Ausmaß. WSI-Mitteilungen 49(5): 316–328.

39.

Silvia

Schroeder

(2007) Why are German employer associations declining? Arguments and evidence. Comparative Political Studies 40(12): 1433–1459.

40.

Streeck

(1997) German capitalism: Does it exist? Can it survive? New Political Economy 2(2): 237–257.

41.

Streeck

Hassel

(2003) The crumbling pillars of social partnership. West European Politics 26(4): 101–124.

42.

Streeck

Visser

(2006) Organized business facing internationalization. In: Streeck

. (eds) Governing Interests: Business Associations Facing Internationalization. London: Routledge, 242–272.

43.

Thelen

(2000) Why German employers cannot bring themselves to dismantle the German model. In: Iversen

Pontusson

Soskice

(eds) Unions, Employers, and Central Banks: Macroeconomic Coordination and Institutional Change in Social Market Economies. Cambridge: Cambridge University Press, 138–169.

44.

Thelen

(2004) How Institutions Evolve: The Political Economy of Skills in Germany, Britain, the United States, and Japan. New York: Cambridge University Press.

45.

Thelen

Van Winbergen

(2003) The paradox of globalization: Labor relations in Germany and beyond. Comparative Political Studies 36(8): 859–880.

46.

Traxler

(2004) Employer associations, institutions and economic change: A crossnational comparison. Industrielle Beziehungen 11(1): 42–60.

47.

Traxler

(2007) The theoretical and methodological framework of analysis. In: Traxler

Huemer

(eds) Handbook of Business Interest Associations, Firm Size and Governance. London: Routledge, 10–36.

48.

Traxler

Brandl

Pernicka

(2007) Business associability, activities and governance: Cross-national findings. In: Traxler

Huemer

(eds) Handbook of Business Interest Associations, Firm Size and Governance. London: Routledge, 351–406.

49.

Völkl

(2002) Der Mittelstand und die Tarifautonomie. Arbeitgeberverbände zwischen Sozialpartnerschaft und Dienstleistung. München: Hampp.

50.

Wallerstein

Golden

Lange

(1997) Unions, employers’ associations, and wage setting institutions in Northern and Central Europe, 1950–1992. Industrial and Labor Relations Review 50(3): 379–401.

51.

WSI (Wirtschafts- und Sozialwissenschaftliches Institut der HBS). (2005) WSI-Tarifhandbuch 2005. Düsseldorf: Bund-Verlag.