Productivity and Wage Effects of Firm‐Level Collective Agreements: Evidence from Belgian Linked Panel Data
| Date | 01 December 2020 |
| DOI | http://doi.org/10.1111/bjir.12525 |
| Author | François Rycx,Andrea Garnero,Isabelle Terraz |
| Published date | 01 December 2020 |
British Journal of Industrial Relations doi: 10.1111/bjir.12525
58:4 December 2020 0007–1080 pp. 936–972
Productivity and Wage Effects of
Firm-Level Collective Agreements:
Evidence from Belgian Linked Panel
Data
Andrea Garnero , Franc¸ois Rycx
and Isabelle Terraz
Abstract
How do firm-level collective agreementsaect firm performance in a multi-level
bargaining system? Using detailed Belgian-linked employer–employee panel
data, our findings show that firm-level agreements increase both wage costs
and labour productivity (with respect to sector-level agreements). Relying on
approaches developed by Bartolucci and Hellerstein et al., they also indicate
that firm-level agreementsexert a stronger impact on wages than on productivity,
so that profitability is hampered. However, this rent-sharing eect mostly holds
in sectors where firms are more concentrated or less exposed to international
competition. Firm agreements are thus mainly found to raise wages beyond
labour productivity when the rents to be shared between workers and firms
are relatively big. Overall, this suggests that firm-level agreements benefit both
employers and employees — through higher productivity and wages — without
being very detrimental to firms’ performance.
1. Introduction
Despite significant interest by academics and policy makers, the role
of unions and collective bargaining systems in firm productivity and
profitability remains a largely open debate. Most reforms to decentralize
collective bargaining have been undertaken with the explicit goal of fostering
productivity growth. However, since the seminal works of Freeman (1976)
Andrea Garnero is at OECD, Universit´
e libre de Bruxelles (CEB and DULBEA) and IZA.
Franc¸ois Rycx is at Universit´
e libre de Bruxelles (CEB and DULBEA), Universit´
e de Mons
(humanOrg), Universit´
e catholique de Louvain (IRES), GLO and IZA. Isabelle Terraz is at
Universit´
e de Strasbourg (BETA, OPEE).
C
2020 John Wiley & Sons Ltd.
Productivity and Wage Eects of Firm-Level Collective Agreements 937
and Freeman and Medo (1984), we know that the theoretical eect of
unions and collective bargaining on productivity is twofold: on the one
hand, unions can behave as ‘rent-seekers’, raise wages, restrict employment
flexibility, aect labour supply by standardizing wages across regions and
industries (Addison and Belfield 2004b) and as a result, negatively aect
firms’ performance. On the other hand, by giving a ‘voice’ to workers,
unions can provide an institutionalized mechanism by which labour and
management can communicate and bargain (Gomez et al. 2008). Unions and
collective bargaining institutions thus enable workers to voice their concerns
and requests, reduce ‘exit’ and, in turn, improve eciency notably through
job and organizational changes. In sum, the eect of unions and collective
bargaining arrangements for firm profitability1remains an empirical issue,
and will depend on their impact on wage costs and labour productivity. If the
wage premium determined by unions or collective bargaining originates from
higher productivity, then both workers and firms can benefit. However. ‘if
raised wagescome at the expense of normal profits, this can damagethe prospects
of firms and employment growth — to the long term detriment of all’(Bryson
2014: 1).
While in the United States, collective agreements between unions and
employers can be negotiated exclusively at the firm level, in many European
countries wages are determined by multi-employer collective agreements (at
the national, sectoral or regional levels).Multi-employer agreements are often
complemented by single-employer agreements (at the firm or establishment
level). The existing literature has shown that firm-level agreements generally
yield higher wages and foster rent sharing, that is, strengthen the wage–
profit elasticity (G¨
urtzgen 2010; OECD 2018; Rusinek and Rycx 2013).
This is an expected outcome, since in many countries, firm-level agreements
can only improve wages and working conditions. In contrast, very little is
known regarding the consequences of firm-level agreements on productivity
(Doucouliagos et al. 2017; G¨
urtzgen 2010; Magda et al. 2012). This is
somehow surprising since the decentralization of collective bargaining was
mainly promoted in order to better align wages with productivity and to
enhance productivity by promoting incentive schemes that can be better
implemented at the firm level, or in a multi-level bargainingsystem (i.e. where
firm-level bargaining complements or tailors the conditions set in collective
agreements) such as performance-related pay schemes (Bechter et al. 2019;
OECD 2018).
Our paper provides one of the first attempts at measuring the impact of
firm-level collective agreements on both wagesand productivity in the context
of a two-tier bargaining system, which is typical of continental European
countries. More precisely, we rely on detailed matched employer–employee
panel data for Belgium, covering all years from 1999 to 2010, to investigate
the consequences of firm-level collective agreements (with respect to multi-
employer agreements) on wage costs and productivity–wage gaps, that is,
profits. Our data oer several advantages. The panel covers a large part of
the private sector, provides accurate information on average productivity and
C
2020 John Wiley& Sons Ltd.
938 British Journal of Industrial Relations
wages within firms and allows us to control for a wide range of worker, job
and firm characteristics. It also enables us to address important econometric
issues that are often not accounted for in this literature, such as firm-level
fixed eects (FE) and endogeneity. While each national bargaining system
has its own specificities, the Belgian bargaining system shares many features
with the systems in Finland, France, Iceland, Italy, Portugal, Slovenia, Spain
and Switzerland (see OECD 2017 for a detailed discussion), where multi-
employer bargaining is complemented by firm-level bargaining. Debates on
the (re-)introduction of some forms of sector-level bargaining or regulations
are also present in the United States, United Kingdom and New Zealand.
Therefore,while keeping the institutional dierences in mind, the conclusions
of this analysis for Belgium can be of interest for several other countries.
Toassess the impact of fir m-levelcollective agreements on wages, empirical
studies typically relyon Mincerian wage equations and/or Oaxaca’s (1973) and
Blinder’s (1973) type decomposition techniques using detailed information
at worker le vel (Ba rt et al. 1994; Cardoso and Portugal 2003; Card and
de la Rica 2006; Daouli et al. 2013; Dell’Aringa and Pagani 2007; Magda
et al. 2012, 2016; Plasman et al. 2007; Rycx 2003). This approach, also
known as the residual method, decomposes the dierence in mean wages
between bargaining regimes into explained and unexplainedcomponents. The
share of the wage gap that cannot be explained by dierences in observable
characteristics is then interpreted as the wage premium earned by workers
covered by a firm-level collective agreement. As highlighted by Bartolucci
(2014: 1167), ‘the residual approach can be understood as a comparison of
wages and productivity where the latter is approximated using a function
of observable characteristics’. However, if unobservable characteristics vary
across bargaining regimes and arecorrelated with productivity, then it is likely
that the wage premium associated to firm-level collective bargaining will be
biased.
The availability of linked employer–employee data oers a response to this
potential weakness of the residual approach (Bartolucci 2014). If worker-level
data are not good enough to account for dierences in productivity between
bargaining regimes, one possibility is to directly estimate a production
function and a wage cost equation at the firm level. The production function
yields estimates for the average marginal product of workers covered by
dierent bargaining regimes, whilethe wage equation estimates the impact of
each bargaining regime on the average wage bill paid by the firm. Estimating
both equations with the same samples and identical control variables allows
comparing the parameters for marginal productsand wage costs, and drawing
conclusions on how firm-level collective agreementsaect productivity, wages
and productivity-wage gaps. This method has been developed by Hellerstein
et al. (1999) and refined by Van Ours and Stoeldraijer (2011) among others.
It is now standard in the literature on the productivity and wage eects
of labour/firm heterogeneity (Cardoso et al. 2011; Dong and Zhang 2009;
Garnero et al. 2014a,b; Hellerstein and Neumark 2004; Mahlberg et al. 2013;
Nielen and Schiersch 2014; Rickne 2012; Vandenberghe 2018).
C
2020 John Wiley& Sons Ltd.
To continue reading
Request your trial