Innovation, industrial relations and employee outcomes: evidence from Italy

Abstract

Purpose – This paper seeks to examine the relationships between working conditions, innovation activities and industrial relations in two local production systems located in the northern Italy, exploiting data collected through ad-hoc surveys in 2004 and 2006. Design/methodology/approach – The paper follows a recent stream of literature, which is still quite scanty, that addresses the issue of the implication for workers derived from the introduction of organizational changes. This topic is addressed and extended, taking into consideration both the role of firm level industrial relations and the role of other innovation activities that may influence working conditions. Findings – The results seem to support the position maintained by the advocates of organizational changes. Job empowerment is spurred by the form of organizational changes usually defined as high performance workplace practices. However, such typology of changes in the organization does not seem to be linked with positive trends in safety/security and stress. At the same time the positive role of cooperative industrial relations on the working condition emerges, also as complementary elements to innovation activities. Originality/value – This work confirms some relevant empirical results obtained in international literature about the linkages between organizational changes and working conditions. At the same time it provides an original perspective of analysis taking into account other influencing factors of workers' well being: good quality industrial relations at firm level and innovation activities such as technological innovation and ICT.

Full text

Page 1

Paper to be presented at the Summer Conference 2009
on
CBS - Copenhagen Business School
Solbjerg Plads 3
DK2000 Frederiksberg
DENMARK, June 17 - 19, 2009
INNOVATION, INDUSTRIAL RELATIONS AND EMPLOYEE OUTCOMES
EVIDENCE FROM ITALY
Davide Antonioli
University of Ferrara
ntndvd@unife.it
Massimiliano Mazzanti
University of Ferrara
ma.maz@iol.it
Paolo Pini
University of Ferrara
pnipla@unife.it
Abstract:
The present work aims to provide further evidence on the relationship between firms innovation activities and
working conditions. The literature on such a topic, although quite scanty, has proved to be lively recently, and
it is mainly focused on the linkages between organizational innovation and workers well being, given the
increasing concern about the effects of new work practices or high performance work practices on working
condition. During the 1990s several economists and managerial scholars pointed to the positive effects on
workers from the introduction of new forms of work and production organization: however, some more recent
studies highlight its potential negative effects. On this basis we mainly investigate the effects of organizational
changes and firm level industrial relations climate on workers well being for two Italian local production
systems, Reggio Emilia and Modena, with the further aim of not to overlook other aspects of firms innovation
activity, often neglected in previous studies. The information used come from two unique datasets.
The empirical results point to a general positive impact of organizational changes on working conditions,
although with some exception. Other innovation activities in training, ICT and technological areas are positively
related, when significant, with the indexes of working conditions as well. Finally, we confirm the relationship
between cooperative industrial relations at firm level and workers well being for both the local production
systems
JEL - codes: L60, O33, J81

Page 2

INNOVATION, INDUSTRIAL RELATIONS AND EMPLOYEE OUTCOMES
EVIDENCE FROM ITALY

This version February 2009


Abstract


The present work aims to provide further evidence on the relationship between firms innovation activities
and working conditions. The literature on such a topic, although quite scanty, has proved to be lively
recently, and it is mainly focused on the linkages between organizational innovation and workers’ well
being, given the increasing concern about the effects of ‘new work practices’ or ‘high performance work
practices’ on working condition. During the 1990s several economists and managerial scholars pointed to the
positive effects on workers from the introduction of new forms of work and production organization;
however, some more recent studies highlight its potential negative effects. On this basis we mainly
investigate the effects of organizational changes and firm level industrial relations climate on workers’ well
being for two Italian local production systems, Reggio Emilia and Modena, with the further aim of not to
overlook other aspects of firms’ innovation activity, often neglected in previous studies. The information
used come from two unique datasets.
The empirical results point to a general positive impact of organizational changes on working conditions,
although with some exception. Other innovation activities in training, ICT and technological areas are
positively related, when significant, with the indexes of working conditions as well. Finally, we confirm the
relationship between cooperative industrial relations at firm level and workers well being for both the local
production systems.


Keywords: organizational innovations, working conditions, industrial relations, local production systems.


JEL: L60, M54, O33, J51, J81

Page 3

1. Introduction

In the present work we intend to complement the widespread literature that investigates the
effects of innovation on firms’ economic outcomes providing evidences on the effects of innovation
activities on employee outcomes.
The subject is not new, but is rather under-researched. However, the increasing diffusion of
so called High Performance Workplace Practices (HPWP henceforth) and the contraction of the
traditional Fordist-Taylorist organization of production have spurred a renewed interest in the
consequences of ‘new’ forms of production organization for workers’ conditions. A widespread
endorsement of the HPWP benefits on workers emerged during the ‘90s, especially in the US
context: wider discretion, greater opportunity for using their skills thorough job redesign, decreased
level of control by management and greater worker involvement in decision-making processes were
perceived as aspects that increased workers’ well being (see Handel, Levine, 2004 for a review).
However, some scholars (Gallie, 2005; Green, 2004; Brenner, Fairris, Ruser, 2004; Askenazy,
Caroli, 2006) have highlighted that there are costs to workers associated with organizational
changes: intensification of the working activity, reduction in working dead-times, psychological and
physical pressures.
Firm’s activities potentially influencing the working conditions cannot be bounded to the
organizational changes, but they surely involve other forms of innovations and the specificities of
the labour/industrial relations, such as cooperative relations between management and workers or
their delegates. The latter issue has not received the deserved attention by a stream of literature that
mainly analysed organizational changes/working condition relations in liberal market economies
(Kalmi, Kauhanen, 2008). The present work shows results for local production systems, included in
a market economy that shares characteristics belonging both to a coordinated and to a liberal market
economy (Italy) and defined as Mediterranean economy by some scholars, with historical traditions
of strong and active unionism.
The empirical work is grounded on the above considerations and it aims to verify the
existence of robust linkages between working conditions, as dependent variable, innovation
activities and cooperative industrial relations at firm level, as covariates. The data utilized come
from two unique datasets, each of them for a single Local Production System (LPS henceforth)
located in Northern Italy: Reggio Emilia (RE henceforth) and Modena (MO henceforth).

Page 4

The paper is structured as follows. Section 2 reviews the existing literature in order to
contextualize the present work. Section 3 outlines RE and MO LPSs in addition to data description
and methodological issues. Section 4 provides the results of the analysis and section 5 concludes.

2. Related Literature

Notwithstanding the importance of the subject here investigated, the relations between
working conditions and the innovation activities of the firm, it is relatively under-researched1
(Askenazy, Caroli, 2006; Kalmi, Kauhanen, 2008), when compared to a widespread literature
concerning the relation between the introduction of HPWP and firm outcomes (see among others
Black, Lynch, 2001; Caroli, Van Reenen, 2001; Janod, Saint-Martin, 2004; Huselid, 1995; Huselid,
Becker, 1996; Cappelli, Neumark, 2001; Ichniowski, Shaw, Prennushi, 1997; Ichniowski, 1990;
Zwick, 2004; BLINDED; Leoni, 2008).
In this work we focus on the issue regarding the linkages between innovation activities, with
a special attention on organizational innovations usually intended as enhancers of firm’s economic
outcome, and workers’ well being, without neglecting to analyse the nexus between cooperative
industrial relations and working conditions, which can be thought of as being complementary to the
literature on the effects of innovation activities.
In the attempt of providing a brief systematization of the existing literature we can move the
first step following a stream of works that recognizes in the introduction of HPWP a way for the
management to strengthen control over workers’ efforts and to intensify the pace of work (see
Ramsay, Scholarios, Harley, 2000 for a review). According to this negative position (intensification
thesis) self-managing teams, for example, are considered as instruments that substitute the
hierarchical control with the peers control. In the same vein, worker participation is perceived as a
method for co-opting workers into a managerial perspective in order to preserve hierarchical
authority without bureaucratic control (Vidal, 2007). The work intensification that several
economies have experienced during the 90s and in more recent years is explained, by some scholars
(Green, McIntosh, 2001; Green, 2004; Askenazy, 2004; Fairris, 2004), as a result of the diffusion of
HPWP.
Parallel to the critical view a positive position maintains that the adoption of HPWP
increases workers’ well being (empowerment thesis) through a multiplicity of channels (Handel,
Levine, 2004; Askenazy, Caroli, 2006). In this perspective both management and workers gain from
the introduction of organizational changes: the former obtain higher levels of productivity; the latter

1 Part of the literature is focused on the relation between changes in organizational practices and workers’ wages (Black,
Lynch, Krivelyova, 2004; Handel, Gittleman, 2004; Forth, Milward, 2004)

Page 5

receive economic benefits (e.g. higher wages), enjoy higher levels of job satisfaction and are
capable of fully exploiting their skills. The advocates of HPWP are used to stress the accent on the
role of complementarities between such practices. A wide endorsement about the positive impact of
bundles of practices on firm performances (see Lynch, 2007 for recent evidence), rather than single
practices adoption, leads us to expect the same when the outcome variables are measures of
working conditions. Although, it is not necessarily true that the introduction of HPWP implies high
level of workers involvement (Vidal, 2007); the adoption of bundles of HPWP may indicate the
willingness of managers to deeply reconsider the production and labour organization structures in a
way that may also alter the hierarchical structure and the equilibria among the firms constituencies.
The latter represent a key point to understand whether or not the implementation of
organizational changes affects the workers well being. High commitment workplace practices may
be used in the attempt to align management and workers’ objectives, but in this respect the role that
can play a cooperative and participative industrial relations climate cannot be neglected.
Participatory industrial relations may be legitimately thought of as complementary aspects of
innovation activities in influencing working conditions2 and not only firm economic performance
(Menezes-Filho, Van Reenen, 2003; Metcalf, 2003). In fact, unions are clearly concerned about the
well being of their members, while management may be less sensible towards workers’ well being
and more focused on workers’ efforts. If industrial relations are not poor, but driven by
union/management cooperation3, then the union presence may contribute to non-conflictual
resolution of the partially divergent objectives of the two parties, ending up with advantages both
for the firm and for the workers. International institutions have also recognized the importance of
social dialogue in improving the workers’ well being (Eurofound, 2007).
In addition to and interrelated with organizational changes and workers representation we
also have to enumerate other innovation activities, which revolve around training policies (here
intended as a separate element with respect to HPWP), technological changes and ICT adoption,
that are potential sources of changes in working conditions (Antonioli, Mazzanti, Pini, 2009). For
all the innovation activities, but for training, which is usually thought to have an unambiguous
positive role on workers, may hold the dichotomous view that roughly synthesises the scholars
positions about the effects of HPWP on working conditions. On the one hand, the increasing

2 Within this line of empirical research the issue of reverse causality emerges (Renaud, 2002): the good quality of the
dialogue between union delegates and management spurs workers well being, because management is more receptive to
the union voice and consequently to the workers’ needs, or instead, better quality working conditions promote less
adversarial industrial relations?
3 If unions are conservative (and likely operating in a poor industrial relations environment) they may be prone to adopt
behaviours that hinder the introduction of innovations, both technological and organizational, independently by the
interest of their members. Some empirical studies have pointed out that union power, usually “proxied” by union
density variable, may exert a hindering effect on R&D expenditure and on profitability (Menezes-Filho, Ulph, Van
Reenen, 1998; Betts, Odgers and Wilson, 2001)

Page 6

diffusion of ICT (Brynjolfsson, Hitt, 2000; Bresnahan, Brynjolfsson, Hitt, 2002) has been identified
as a cause of stress among workers because of the possibility they provide to management to
monitor and control workers more intensively (Green, 2004). On the other hand, ICT help in
improving information sharing, in spreading information about best safety practices, and providing
workers with greater degrees of autonomy in their jobs (Askenazy, Caroli, 2006). The same
reasoning may hold for innovation in processes or quality control. On the one hand, their
introduction can increase mental strain for workers and reduce safety because the ‘ever’ changing
production environment reduces the possibility of setting and learning safety procedures; on the
other hand, the focus on quality, and especially in the processes implemented, can be thought of as
improving occupational safety (Askenazy, Caroli, 2006).

3. Empirical framework and methodology

Two Northern Italy provinces in Emilia-Romagna, Reggio Emilia (RE) and Modena (MO),
are the geographical locations of the manufacturing firms analysed in the present work. The two
LPSs share some common features both in terms of industrial system structures and of institutional
settings.
Both the LPSs are characterized by a predominant presence of Small and Medium
Enterprises (SME)4 as we can see by the population distributions in tables A1a and A1b in
Appendix. A second characteristic of the RE and MO LPSs, which is linked to the prevalence of
SME, is the existence of districts (Brusco, 1982; Becattini, 2004): non-electrical machinery and
equipment and ceramic tiles for RE; ceramic tiles, wood processing machinery and biomedical for
MO. The presence of districts that are usually constituted by networks of small firms and the
institutional settings characterized by a strong social cohesion and by a traditional deep rooted
unionism, make the two LPSs paradigmatic versions of the so called ‘Emilian model’ (Amin, 1999;
Brusco, 1982). The latter, indeed, is characterized by a well marked entrepreneurship spirit and an
equally strong unionism, which coexists with a productive apparatus characterized by the presence
of a district-like production system5.

4 According to the European Commission (Official Journal of the European Union, L 124, May 2003) a SME is defined
as a firm with less than 250 employees. Table A.1 shows that in our local production system about 85% of the firms
with union representatives are below this threshold.
5 We especially refer to the role of CGIL, the left wing union. For an overview of the union history and the linkages
with political parties we refer the interested reader to Baglioni (1998).

Page 7

The empirical analysis is conducted using two different datasets. The first data source is a
firm level survey conducted on the manufacturing firms located in RE6. The criteria we adopted for
the identification of the population of 376 firms were: (a) firms with at least 20 employees; (b)
firms belonging to manufacturing sectors according to the ISTAT ATECO 20027 classification; (c)
presence of union representatives to be interviewed. The data for 2004 were provided by union
representatives, through face-to-face interviews. The interviews led to 192 respondents, which
constituted 51% of firms with union representation (UR) within the firm.
The second survey has been conducted on manufacturing firms located in MO province. The
criteria for the identification of the population slightly differ from those used on RE firms. In fact,
only the firms with at least 50 employees were used to select those having union representatives to
be interviewed: 246 firms is the firm population identified in MO province and the usable
respondents were 176. The data collected refer to the period going from 2004 to 20068.
As we can see in tables A.1a and A.1b only some minor distortions of the two samples of
interviewed firms, with respect to the populations, emerge. A version of the Cochran Test (Cochran,
1977) for sample distortions shows acceptable results9.
Both the two surveys are unique sources of information about firms’ structural
characteristics, workforce composition, innovation activities, working conditions and industrial
relations. The information stemming from the two questionnaires has been selected in order to get
identical data for both the LPSs.
The empirical model is based on the following regression function:

(1) Working Condition (WC_EMP; WC_SAFE) = β0i + β1i[structural variables] + β2i[innovation
activities] + β3i[industrial relations] + εi

The variables in the above reduced form may be shortly described as follow, while for
detailed information on their construction we remind the reader to the Tab.A2 in Appendix.
Dependent variables. The working conditions indexes are two. The index WC_EMP
concerns the trend in items more akin to job content and it synthesises aspects related to the concept

6 Several official sources were used to construct the firm population of RE: Reggio Emilia Chamber of Commerce, Istat
Census, Aida data bank, “Impero” data bank. For reasons of homogeneity and information availability the population
refers to the year 2001.
7 The sectors are: food, textiles, wood, chemicals, non-metallic mineral products, machinery, other industries. The
ISTAT ATECO classification coincides with the NACE Rev1.1 and thus with ISIC Rev3.1.
8 For MO database we gratefully acknowledge the work of IRES-ER. A special thank goes to Loris Lugli and Stefano
Tugnoli for providing us the database which we worked on. We further refer the interested reader to the report by Lugli
and Tugnoli ‘Innovazione e Qualità del Lavoro a Modena’ www.er.cgil.it/ireser.nsf/Documenti.
9 For details about the data see BLINDED and Lugli, Tugnoli (2008).

Page 8

of job empowerment: the higher (lower) the index, the higher (lower) is the intensity of job
enrichment and empowerment. The index WC_SAFE was constructed using the variables stress and
safety/security: the higher (lower) the value of the index the higher (lower) the level of
safety/security and lower (higher) the level of stress10. Because WC_SAFE assumes only few values
in the interval 1-3 we have rescaled the variable using integers on a range going from 1 to 5; then,
we have used ordered probit regressions in the estimation phase, instead of OLS11.
The separate utilization of the two dependent variables is aimed at verifying the impact of
innovations and industrial relations variables on different items enclosed in the concept of working
conditions: some related to the concept of job empowerment, and others related to ‘critical’ aspects
that can be used to “measure” psychological and physical strain.
It is important to note that there may be a lack of simultaneity between the dependent
variables and the innovation variables, which derives from how the question was framed for both
the questionnaires: “…subsequently to the introduction of changes by the management, how did the
working conditions change?”. The generic term ‘changes’ without other specifications, in the
context of the questionnaires, refers to changes in innovation areas (technology, organization,
training and ICT).
Structural variables. The vector of structural variables captures as much as possible the firm
specific heterogeneity in our cross sectional environment in order to reduce, to some extent, the
likelihood of relevant variables being omitted.
Innovation variables. Innovation indexes encompass several spheres. In fact, the
administered questionnaire allowed collecting information not only on product and process
innovations (with a wealth of information on R&D activities), but also on organizational changes
(including those practices that international literature defines as HPWP), training activities and ICT
adoption.
Industrial Relations variables. The indexes of industrial relations takes into account both
formal (e.g. presence of bilateral technical commissions or existence of firm level bargaining) and
informal aspect of the firm level relations between management/union representative and
management/workers.

10 The respondents to the questionnaire (union representatives) were asked to indicate the trend for ten items (effort,
security and job stability, employees competences, information disposable to the employees, autonomy in accomplish
the job tasks, influence over the managerial decisions, monetary incentives, non-monetary incentives, stress,
safety/security) concerning the working conditions on a scale from 1 (decreased) to 3 (increased). Only for the stress
item the number 1 means increased and 3 decreased, so that the higher is WC_SAFE the better is for workers.
11 We have also run OLS estimations and the results, which are available from the authors upon request, rarely differ.

Page 9

Given the literature review provided above it is possible to expect the list of “signs” as
reported in table 1 associated to the linkages between working conditions, on the one hand,
innovations and good quality industrial relations on the other hand.




TABLE 1- Hypothesized signs of influencing variables on the working conditions: WC_EMP, WC_SAFE
Influencing factors
Hypothesized causality directions:
Innovations à Working Conditions
Organizational innovation
Technological innovation
Training
ICT
Hypothesized relations12:
Industrial relations ß à Working Conditions
(Cooperative) Industrial relations


4. Results of the econometric analysis
Empowerment Thesis

Intensification Thesis

(+)
(+)
(+)
(+)

(-)
(-)
(+)
(-)

(+)
(+)

The results of the econometric exercises are reported in tables 2a.1, 2a.2, 2b.1 and 2b.213.
At first, it should be stressed that because of the importance of organizational changes in
influencing working conditions, we estimate both a specification with only the synthetic index of
organizational innovation (INNO_ORG) and a specification with the disaggregated variables used
to construct INNO_ORG. The same has been done with the industrial relation index (IND_REL).
Second, we have settled up different specifications for each synthetic innovation index because of
their quite high bivariate correlations (Tab.A.3 in Appendix). Third, we have also run regressions
with multiplicative interaction terms between innovation indexes and the industrial relations one, in
order to verify if the interaction of industrial relations and innovations reinforce the impact of the
latter on working conditions (tables 2a.2 and 2b.2). Finally, it must be stressed that for Reggio
Emilia sample of firm we decided to drop those firms having less than 50 employees, in order to
make the comparison between the two LPSs more direct14.
Looking at table 2a.1 we note a positive and significant impact of the innovation activities
on WC_EMP for both the LPSs, with some specificities. When we disaggregate the synthetic index

12 When we take into consideration the linkages between working conditions and industrial relations we do not have the
potential causality direction implied by the question posed to the respondents. Thus, we can strictly refer to correlations
rather than to weak causality relations.
13 Because the aim of the empirical analysis is not to calculate elasticities we avoid presenting the coefficients, rather we
adopt a more qualitative way of results displaying.
14 As a matter of fact we also conducted the estimation with all the sample of Reggio Emilia firms and the results do not
differ. The results are available from the authors upon request.

Page 10

INNO_ORG, using its components we observe that the sign of the synthetic index is mainly driven
by the two components that capture changes in production and labour organization (ORG_PROD
and ORG_LAB). This means that workers empowerment is spurred by such typologies of
organizational changes, in accordance with the advocates of HPWP implementation.

Page 11

TABLE 2a.1 – Results of the econometric exercise with WC_EMP as dependent variable (OLS)
Dependent variable
WC_EMP

1 2 3

0.932***
(3.30)
0.311*
(2.48)
0.506**
(1.76)
0.052
(0.54)

0.594***
(3.51)
RE
1

MO
4 5 6 7 2 3 4 5 6 7
Organizational Innovation Organizational Innovation

INNO_ORG


0.623**
(2.21)


INNO _LAB_ ORG


0.367**
(2.22)
0.506***
(4.55)
0.159*
(1.67)


INNO_PROD_ ORG




REW


Industrial Relations Industrial Relations
0.482**
(2.53)
0.025
(0.31)
0.096
(0.98)
0.052
(0.93)
0.109***
(2.80)
Innovation Activities

COOP_INDREL
-0.028
(-0.59)
0.014
(0.20)
0.172***
(2.86)
0.122**
(2.39)
Innovation Activities












BTC
BARG


REL_ITEMS


INDREL_TREND





INNO_TRAIN
0.290***
(3.72)



0.353***
(5.12)
0.132
(1.11)

INNO_TECH


0.362***
(3.28)
0581***
(3.00)
0.28
0.00
134
1.82



0.354**
(2.45)
0.15
0.00
176
1.94
INNO_ICT
R2
F test prob.
N
VIF
Notes: levels of significance 10% *, 5%**, 1%***; t-statistics in parenthesis; robust standard errors have been used; empty cells mean the variable is not included in the
specification; additional controls include sector and size dummies, index of perceived firm performance, strategy dummies, belonging to a group, delocalization dummy,
international turnover, diffusion of short term contracts and their conversion in long lasting ones, in/out-sourcing activities, relations with client/suppliers; VIF is the Variance
Inflation Factor and it represents a ‘test’ to recognize the existence of multicollinearity when the threshold of 10 is exceeded.

0.26
0.00
134
1.55


0.29
0.00
134
1.50

0.24
0.00
134
1.58

0.16
0.00
176
1.9


0.23
0.00
176
1.93

0.12
0.00
176
1.97
0.34
0.00
134
1.35
0.25
0.00
134
1.25
0.28
0.00
134
1.19
0.28
0.00
176
1.46
0.16
0.00
176
2.08
0.18
0.00
176
1.99

Page 12

The same strong and positive relations emerge when the index capturing the intensity of
good quality (cooperative) industrial relations at firm level is considered. Also in this case when we
disaggregate the index in its components we find that the sign of the synthetic index is driven by
specific components, which can be considered the less formal aspects of good quality industrial
relations: the index capturing the management attitude to inform, consult or negotiate with union
representatives on several items (REL_ITEMS) and the index providing the union representative
perception of the industrial relations trend over time (INDREL_TREND). The result does not come
unexpected (BLINDED; BLINDED), but despite the proximity of the two LPSs an interesting
difference concerning the role of good quality industrial relations on working conditions emerges.
For RE firms the variable REL_ITEMS is strongly significant, while it is not for MO firms. We
may argue that due to historical and idiosyncratic conditions, local unions in RE have a stronger
power than those in MO. Such power translates into the capacity to make the union voice listened
within the firm. At the same time it is possible that RE firm’s management is more inclined to
discuss union instances. This finding support a literature that recognizes in the institutional setting a
factor influencing the sign of the impact of organizational changes implementation on working
conditions (Godard, 2004).

TABLE 2a.2 – Results of the econometric exercise with WC_EMP as dependent variable with interaction terms (OLS)
Dependent
variable
WC_EMP RE

1 2 3 4

Interaction Terms
INNO_ORG*
COOP_INDREL (4.30)
INNO_TRAIN*
COOP_INDREL (4.11)
INNO_TECH*
COOP_INDREL (3.54)
INNO_ICT*
COOP_INDREL (3.97)
R2 0.27 0.31 0.28 0.30
F test prob. 0.00 0.00 0.00 0.00
N 134 134 134 134
VIF 1.12 1.66 1.98 1.25
Notes: Notes: levels of significance 10% *, 5%**, 1%***; t-statistics in parenthesis; robust standard errors have been
used; empty cells mean the variable is not included in the specification; additional controls include sector and size
dummies, index of perceived firm performance, strategy dummies, belonging to a group, delocalization dummy,
international turnover, diffusion of short term contracts and their conversion in long lasting ones, in/out-sourcing
activities, relations with client/suppliers; VIF is the Variance Inflation Factor and it represents a ‘test’ to recognize the
existence of multicollinearity when the threshold of 10 is exceeded.

In order to verify the importance of industrial relations we construct several multiplicative



MO
1 2
Interaction Terms
3 4
1.898***

1.185***
(2.78)
0.609***
(5.46)

0.658***



0.360**
(2.28)

0.750***


0.624***
(3.10)
0.18
0.00
176
2.10
0.857***






0.18
0.00
176
2.11

0.25
0.00
176
2.14

0.14
0.00
176
2.17
interaction terms between the synthetic industrial relations index and the innovation ones. As
expected (table 2a.2) the relations between the interaction terms and WC_EMP are reinforced

Page 13

versions of those obtained without interactions. As a relevant example the INNO_TECH variable,
which is not significant for MO firms, when it is interacted with the industrial relations index turns
out to be significant. This result seems to support the idea that good quality industrial relations are
complementary elements to innovations. We might infer, although we do not test it here, that
cooperative industrial relations at work is a mediating factor that allows to obtain stronger effect of
innovations on workers’ well being15.
If the results are quite clear cut when we consider the index WC_EMP, the same cannot be
said when the index capturing the trend in workers stress and safety/security is used as dependent
variable of our regressions (tables 2b.1 and 2b.2).
At first, we notice that only the training index is significant and positive for the firms
belonging to both local production systems. This result does not come unexpected since a better
trained workforce is likely to better know and implement safety/security norms16. The other
innovation indexes do not increase the probability of having high values of the dependent variable,
that is a positive trend in the aspects enclosed in the dependent. However, when the INNO_ORG
variable is disaggregated we find that for RE the changes in production organization likely spur
better working conditions while the presence and intensity of diffusion of reward schemes is
negatively related with the dependent variable. The latter result may be interpreted as a
consequence of the intensification in working pace due to the linkages between the reward system
and performance indicators. Things are slightly different for MO context: here we have a negative
sign associated to ORG_LAB variable, while positive is the association of the dependent with
ORG_PROD and REW. The absence of significance of the synthetic index of organizational
changes is given by the fact that the negative sign associated to ORG_LAB seems to be strong
enough to offset the positive impact of the other two main variables used to construct the synthetic
index. The result of the ORG_LAB variable is in line with the recent, critical literature on the
relations between organizational changes and working conditions (Green, 2004; Brenner, Fairris,
Ruser, 2004; Askenazy, Caroli, 2006). Although those organizational changes clustered in
ORG_LAB variable are usually interpreted as ameliorative elements of the job content, the results
we obtained underlines the possibility that the workers are subject to higher level of stress when
changes in labour organization are introduced, notwithstanding the fact that such changes positively
affect the job contents. A potential explanation calls in question the synergic nature of
organizational innovations and good quality industrial relations. As we can see in table 2b.1
industrial relations indexes are less robustly linked with working conditions in MO local system

15 In our previous researches we have found that industrial relations exert an indirect role on labour productivity,
spurring innovation activities: BLINDED
16 We also cannot forget that part of training activities on safety/security norms is compulsory.

Page 14

than in RE one. Moreover, the simple bivariate correlations between innovation indexes and
industrial relations ones (Tab.A.3 in Appendix) clearly show how innovations and good quality
industrial relations are more related in RE than in MO. Thus, in the latter LPS there are less
opportunities to exploit synergies between organizational changes and cooperative industrial
relations. In synthesis, the results for the two LPSs partially support both the empowerment thesis
and the intensification thesis.
Another interesting result emerges when the interaction terms are considered. Here the
evidence does not seem to robustly support the existence of synergies, although for ICT innovations
in Reggio Emilia we register a significant result when interacted with cooperative industrial
relations. In firms having high intensity of ICT adoption and high level of cooperation between
management and unions/employees, that is to say high values in the interaction term
INNO_ICT*COOP_INDREL, then the level of stress is lower and the level of safety/security is
higher than in firms having low levels of the interaction term, because of a low degree of ICT
adoption or because of poor industrial relations or both. In addition to the result for ICT innovation
we cans see that the role of training is also reinforced when we interact the training index with that
of cooperative industrial relations in the case of Modena firms. In conclusion, we may say that if
firms have both high intensity in training activities or in ICT adoption and in cooperation between
management and union delegates or workers then, they are likely to have better working conditions
than those firms whose ICT and training innovating activities and industrial relations are both poor.

Page 15

TABLE 2b.1 – Results of the econometric exercise with WC_SAFE as dependent variable^ (Ordered Probit)
Dependent variable
WC_SAFE

1 2 3

Organizational Innovation
-0.713
(-0.67)
0.381
(0.44)
1.299*
(1.81)
-0.749**
(-2.20)

Industrial Relations
1.467**
(1.97)
RE
1

MO
4 4 5 6 7 2 3
Organizational Innovation
5 6 7
INNO_ORG
0.468
(0.55)
-1.119**
(-2.13)
1.242***
(2.92)
0.649**
(2.11)

INNO _LAB_ ORG
INNO_PROD_ ORG
REW


Industrial Relations
1.052*
(1.81)
-0.249
(-1.04)
0.038
(0.14)
0.018
(0.11)
0.662***
(4.85)
Innovation Activities
COOP_INDREL
-0.241
(-1.18)
0.019
(-0.09)
0.437**
(2.11)
0.492***
(2.76)
Innovation Activities

BTC
BARG
REL_ITEMS
INDREL_TREND




INNO_TRAIN
0.907***
(2.67)
0.471
(0.93)

0.413*
(1.69)
0.285
(0.69)

INNO_TECH
0.099
(0.21)
0.07
0.00
176
INNO_ICT
Pseudo R2
Prob. Wald Chi2
N
Notes: Notes: levels of significance 10% *, 5%**, 1%***; z-statistics in parenthesis; robust standard errors have been used; empty cells mean the variable is not included in the
specification; additional controls include sector and size dummies, index of perceived firm performance, strategy dummies, belonging to a group, delocalization dummy,
international turnover, diffusion of short term contracts and their conversion in long lasting, ones in/out-sourcing activities, relations with client/suppliers; VIF is the Variance
Inflation Factor and it represents a ‘test’ to recognize the existence of multicollinearity when the threshold of 10 is exceeded.



0.11
0.00
134

0.14
0.00
134

0.11
0.00
134

0.12
0.00
134

0.13
0.00
134

0.11
0.00
134
0.289
(0.44)
0.11
0.00
134

0.08
0.00
176

0.12
0.00
176

0.08
0.00
176

0.13
0.00
176

0.08
0.00
176

0.08
0.00
176