We Want Them All Covered! Collective Bargaining and Firm Heterogeneity: Theory and Evidence from Germany
| DOI | http://doi.org/10.1111/bjir.12239 |
| Author | Florian Baumann,Tobias Brändle |
| Date | 01 September 2017 |
| Published date | 01 September 2017 |
British Journal of Industrial Relations doi: 10.1111/bjir.12239
55:3 September 2017 0007–1080 pp. 463–499
We Want Them All Covered! Collective
Bargaining and Firm Heterogeneity:
Theory and Evidence from Germany
Florian Baumann and Tobias Br¨
andle
Abstract
This articleestablishes a link between the degree of productivity dispersion within
an industry and collective bargaining coverage of the firms in the industry. In
a stylized unionized oligopoly model, we show that dierences in productivity
levels can aect the design of collective wage contracts a sector-union oers to
heterogeneous firms. Using German linked employer–employee data, we test a
range of our theoretical hypotheses and find empirical support for them. The
dispersion of sector-level labour productivity decreases the likelihood of firms
being covered by a collective bargaining agreement on the industry level, but
increases the likelihood of firms being covered by firm-level agreements. The
results hold for dierent subsamples and (panel) estimation techniques.
1. Introduction
Negotiations on wages can take on various forms, with dierent actors
involved. Apart from individual employers and employees, labour unions or
employers’ associations may take part in the bargaining process, forming
so-called collective bargaining agreements (CBAs) between firms and their
employees. In fact, the question of how dierent bargaining modes arise
in dierent countries, or even within industries, has not yet been answered
convincingly in the economic literature, both theoretically and empirically.
Furthermore, we observe that dierent kinds of bargaining modes exist at
the same time and even within the same industry. For example, in Germany
we observe some firms being covered by an industry-level agreement, some
being coveredby a firm-level agreement with the industry union and others not
being bound by union wages at all coexisting at the same time, all within the
Florian Baumann is at the University of Bonn. Tobias Br¨
andle is at the Institute for Applied
Economic Research in T¨
ubingen.
C
2017 John Wiley& Sons Ltd.
464 British Journal of Industrial Relations
same economic sector (see, e.g. Addison et al. 2013). In this article, we argue
that, in addition to firm-level characteristics, industry heterogeneity, that is,
the productivity dispersion between firms, explains the extent of collective
coverage as well as the parallel existence of bargaining regimes within an
industry. Moreover, we show that the dierences in productivity dispersion
between industries are a relevantand until now unexplored factor in empirical
estimations determining the bargaining status of firms.
This article contributes to the explanation of the occurrence of dierent
wage bargaining regimes and, especially, their co-existence. The argument
for the observed dierences in wage bargaining that we develop in this
analysis is based (a) on the fact that firms are heterogeneous with regard to
productivity and (b) on the idea that collective bargaining can be a means
to save on the transaction costs of individual negotiations. The fact that
transaction costs determine the extent of collective bargaining is described in,
for example, Arrowsmith et al. (2003) or Dinlersoz and Greenwood (2012).
Card et al. (2013), among others, find that heterogeneous firms also dier
in their wage structure, and Hirsch et al. (2014) establish, for Germany, that
more productive firms are more likely to be under collective coverage (for a
detailed literature review, see Section 2). The argument in Hirsch et al. (2014)
is that high-productivity firms like to pool with less productive firms in wage
bargaining as this allows them to pay lower wages. Indeed, G¨
urtzgen (2009)
establishes that collectively bargained wages at the industry level should fall
with productivity dispersion within the industry.
Building on these insights, webuild up a stylized unioniz ed oligopolymodel
in which dierent bargaining regimes in the same industry, that is, collective
bargaining at the industry level, firm-level bargainingwith the industry union
and bargaining outside union coverage, may co-exist at the same time. In
doing so, we establish that productivity dispersion within industries is an
important determinant of the mode of collective bargaining; more dispersed
productivity levels among firms lead to lower collective coverage and greater
importance of firm-level agreements. We develop the argument theoretically
and provide empirical tests with German linked employer–employee data
using a measure of within-industry productivity dispersion to explain firms’
bargaining regimes.
The empirical data illustrated in Figure 1 provide a starting point: the
scatter plot shows within-industry labour productivity dispersion on the
abscissa and the percentage of plants1covered by collective and firm-specific
union wage contracts on the ordinate, both averaged overtime (plants without
union coverage constitute the remaining third group not displayed). It can be
seen that there is a negative association between within-industry productivity
dispersion and collective bargaining coverage at the industry level, and a
positive relationship between productivity dispersion and coverage by firm-
level contracts.
To help explain this relationship, we use a stylized unionized oligopoly
model. The model is held by simple abstracting frommany real-world features
(e.g. the restriction on the number of firms). Nevertheless, the model is
C
2017 John Wiley& Sons Ltd.
Collective Bargaining and Firm Heterogeneity 465
FIGURE 1
WageBargaining and Productivity Dispersion within German Industries.
Source: Own calculations based on LIAB QM2 9310, years 1996–2010 using controlledremote
data access via FDZ.
Notes: The means are calculated across all waves and using employee-representative sample
weights. As industry classificationwe use the WZ-93 2-digit level further explained in section
‘Data’. Some industries do not havefirms with fir m-levelagreements and are therefore only
represented for collectivebargaining agreements.
suciently complex to convey the basic insights we want to illustrate. We
consider an industry in which heterogeneous firms (one low-productivity
firm, two medium-productivity firms, one high-productivity firm) compete
in quantities. Principally, the workforce is represented by an industry union.
Firms pay wages resulting either from a collective wage agreement covering
the industry, from firm-level negotiations with the industry union or wages
bargained by its ownemployee representatives.The industry wage contract set
by the union depends on the productivity distribution of firms covered.Firms
agreeing to wagesaccording to the collective agreement do not have to bargain
individually with their workforce and therefore save on additional negotiation
costs.2However, as firms dier with respect to productivity, the uniform
wage imposes a greater burden on relatively unproductive firms, making
coverage less attractive. These firms might want to abstain from coverage by
the industry contract and instead bargain directly with their own workforce.
In this case, first, the lower productivity level is recognized within the
negotiations, and, second, wages are lower due to the ensuing competition in
wage setting. Accordingly, the firm has to trade-o the increase in transaction
costs and a likely worseningof industrial relations on the one hand with lower
wages on the other. At the same time, highly productive firms in the industry
reap extra profits as the collective wage is relatively low for them. Therefore,
the union has an incentive to conclude separate firm-level agreements
involving higher wages for the employees in these firms. This is a costly
C
2017 John Wiley& Sons Ltd.
To continue reading
Request your trial