Works Councils and Labour Productivity: Looking beyond the Mean
| DOI | http://doi.org/10.1111/bjir.12037 |
| Published date | 01 June 2015 |
| Date | 01 June 2015 |
Works Councils and Labour Productivity:
Looking beyond the Mean
Steffen Mueller
Abstract
After more than 25 years of empirical investigation, the analysis of works
council effects on establishment productivity is still an active field of economic
research. I present first evidence for works council regression coefficients along
the unconditional log labour productivity distribution in Germany. The highest
(lowest) coefficients were found at the bottom (top) of the distribution, which
suggests that works councils have the most beneficial effects at low levels of
productivity. Importantly, coefficients are positive at each productivity level.
The results shed light on a classical theoretical argument concerning moderat-
ing factors for the productivity effect of employee participation.
1. Introduction
A growing body of research documents the effects of German works councils
on establishment productivity, wages, profits and investment behaviour. The
rising interest in works councils as a form of employee participation other
than union representation is likely driven by de-unionization in many indus-
trialized countries. Although in recent years the literature converges towards
the view that works councils are, on average, positively related to productiv-
ity, it is silent with respect to the effects that works councils have at different
quantiles of the unconditional labour productivity distribution.1
Knowledge about the ceteris paribus effects of works councils along the
labour productivity distribution allows a richer description of how worker
co-determination shapes the economy. It discloses whether works councils
have adverse effects on certain parts of the productivity distribution other
than the mean. This is a reasonable presumption as rent-seeking is one of
several possible motivations for workers to set up a works council, and also
because of the essentially inflexible one-works-council-fits-all-firms frame-
work as provided by the German works council legislation, probably ham-
pering tailor-made, firm-specific solutions.
Steffen Mueller is at the University of Erlangen-Nuremberg.
© John Wiley & Sons Ltd/London School of Economics 2013. Published by John Wiley & Sons Ltd,
9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA.
British Journal of Industrial Relations
53:2 June 2015 0007–1080 pp. 308–325 doi: 10.1111/bjir.12037
A related point to be explored is the ongoing puzzle of why works councils
have to be mandated if they have beneficial effects on firms. For example,
large positive effects for some highly productive firms and negative effects for
the majority could explain the resistance of the majority of employers against
works council introduction despite, on average, positive effects of councils.2
Works council effects at different points of the unconditional productivity
distribution also shed light on a classical hypothesis put forward by the
pioneers of the empirical analysis of German works councils more than
25 years ago. FitzRoy and Kraft (1985) argued that works councils might
be beneficial for badly managed firms but should make things worse in the
presence of competent managers, thereby suggesting a more positive works
council effect for the least efficiently organized firms.
I address the above questions by applying unconditional quantile regres-
sions as proposed by Firpo et al. (2009) to the rich data of the Institute for
Employment Research (IAB) establishment panel. A further contribution is
to provide a detailed subgroup analysis by splitting the sample by time
period, region, sector and firm size. To make this point clear from the
beginning, unconditional quantile regression does not solve identification
problems also faced by conventional estimators. The purpose of this article
is to present for the first time ceteris paribus effects of works councils along
the unconditional labour productivity distribution. As with standard ordi-
nary least square (OLS) estimation, ceteris paribus means nothing more than
holding all observables fixed, and causality can be achieved if works council
existence is exogenous conditional on the rich set of control variables
included in this study. Section 5 provides additional endogeneity checks.
2. German works councils
Addison (2009) gives an excellent and up-to-date treatment of the history and
the functioning of German works councils, and I will therefore focus on
the most important aspects in the following. The Works Constitution Act
(Betriebsverfassungsgesetz) of 1972 and revised in 2001 constitutes the legal
base for German works councils. Given that there are at least five permanent
employees in the establishment, the law entitles (any small group of at least
three) workers to initiate a works council election, however without imposing
any automatism to do so. As a consequence, there are eligible establishments
with and without works councils, and works council incidence is rising with
establishment size.
The law affords works councils a bunch of legal rights, including rights on
information, consultation and even co-determination, whereas the number of
rights increases with the number of employees in the plant. Employer and
works council meet frequently in order to discuss a broad range of topics, for
example working conditions. However, workers are also entitled to be heard
by the employer or to formally complain with the help of the works council
if they feel treated unfairly.
© John Wiley & Sons Ltd/London School of Economics 2013.
Works Councils and the Labour Productivity Distribution 309
To continue reading
Request your trial