Developing Institutions. Collective Agreements in the Dutch IT Industry

Developing Institutions Collective Agreements in the Dutch IT Industry

Developing Institutions Collective Agreements in the Dutch IT Industry ACADEMISCH PROEFSCHRIFT ter verkrijging van de graad van doctor aan de Universiteit van Amsterdam op gezag van de Rector Magnificus prof. mr. P.F. van der Heijden ten overstaan van een door het college voor promoties ingestelde commissie, in het openbaar te verdedigen in de Aula der Universiteit op woensdag 17 mei 2006 te 10.00 uur door Adrianus Andreas Godefridus van Liempt Geboren te Waalwijk

Promotiecommissie: Promotores: prof. mr. P.F. van der Heijden prof. dr. A.C.J.M. Wilthagen Overige leden: prof. dr. W. Buitelaar prof. mr. E. Verhulp prof. dr. K.G. Tijdens prof. dr. P.L.M. Leisink dr. R.H. van het Kaar Faculteit der Rechtsgeleerdheid Cover design: Adriaan van Liempt ISBN 10: 907383880 0 ISBN 13: 978907383880 2 2006 by Adriaan van Liempt All rights reserved. Save exceptions stated by the law, no part of this publication may be reproduced, stored in a retrieval system of any nature, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, included a complete or partial transcription, without the prior written permission of the author, application for which should be addressed to author.

For Linda

Contents LIST OF TABLES AND FIGURES LIST OF ABBREVIATIONS ACKNOWLEDGEMENTS III III III 1 INTRODUCTION 3 1.1 Problem Definition 3 1.2 Institutions and Institutional Theory 3 1.3 Origin and Change of Institutions 3 1.4 Explaining the Emergence of Collective Agreements in the Dutch IT Industry 3 1.4.1 Hypothesis 1: The Limited Choice Set Hypothesis 3 1.4.2 Hypothesis 2: The Efficiency Hypothesis 3 1.4.3 Hypothesis 3: The Anachronism Thesis 3 1.5 Methodology and Outline of this Study 3 2 THE HISTORICAL AND LEGAL CONTEXT OF THE DUTCH COLLECTIVE AGREEMENT 3 2.1 The Historical Context of the Dutch Collective Agreement 3 2.1.1 Worker Coalitions: From Prohibition via Toleration to Legalization (c.1400 - c.1870) 3 2.1.2 Development of Worker and Employer Coalitions (c.1870 - c.1918) 3 2.1.3 The Road to Legislation of the Collective Agreement (c.1900 - c.1918) 3 2.1.4 The Inter Bellum Period (c.1918 - c.1939) 3 2.1.5 Collective Agreements as a Means to an End: Wage Control (c.1940 - c.1970) 3 2.1.6 The Retreating Government (c.1970 - c.2004) 3 2.1.7 Conclusion 3 2.2 The Legal Context of the Dutch Collective Agreement 3 2.2.1 International Regulations and Treaties 3 2.2.2 Government Regulations and Legislation 3 2.2.3 The Employment Contract 3

II 2.2.4 Collective Agreements 3 2.2.5 Extension of a Collective Agreement Through the WAVV 3 2.2.6 Works Council Agreements 3 2.2.7 Conclusion 3 3 HISTORY AND DEFINITION OF THE DUTCH IT INDUSTRY 3 3.1 The development of the Dutch IT industry 3 3.1.1 The Roots of Information Technology 3 3.1.2 The Development of the IT Industry Between 1970 and 1990 3 3.1.3 Specialized IT Training and Education During the 1960s - 1980s 3 3.1.4 The IT Industry During the 1990s 3 3.1.5 The Rise of the Dutch IT Industry in the Late 1990s 3 3.1.6 The Post-Millennium Crisis of the Dutch IT Industry 3 3.1.7 The Three Lineages of Software and Service Companies in the Dutch IT Industry 3 3.2 A Definition of the IT Industry 3 3.2.1 The Many Branches of the IT Industry 3 3.2.2 Defining the IT Industry 3 3.2.3 The Dutch IT Industry in Figures 3 3.3 Conclusion 3 4 INDUSTRIAL RELATIONS IN THE DUTCH IT INDUSTRY 3 4.1 Employers Associations and Branch Organizations in the Dutch IT Industry 3 4.1.1 AWVN 3 4.1.2 Branch and Employers Associations Until 2001 3 4.1.3 Nederland-ICT 3 4.1.4 Werkgeversvereniging ICT 3 4.2 Unions and Works Councils in the Dutch IT Industry 3 4.2.1 The Big Sleep: Union Activities in the Dutch IT Industry Until the 1990s 3 4.2.2 A Wake Up Call for the Unions: The Crash of the Dutch IT Industry at the Beginning of the 1990s 3 4.2.3 The Heydays of the Unions in the Dutch IT Industry: 1996-1999 3 4.2.4 The Development of Union Strategies Since the 1990s 3 4.2.5 2001-2004, A New Crisis in the Dutch IT Industry: Hello Works Council, Goodbye Unions? 3 4.2.6 Summary 3 4.3 The ICK Collective Agreement 3 4.3.1 Origins of the ICK Collective Agreement 3 4.3.2 The First ICK Collective Agreement: Conflicts, Compromises, and Disappointments 3

4.3.3 Extending the ICK Collective Agreement: A Mission Impossible? 3 4.3.4 Collective Agreements Connected to the ICK Collective Agreement 3 4.3.5 A Closer Look at the ICK Collective Agreement 3 4.4 Conclusion 3 5 CASE STUDIES 3 5.1 Introduction and Methodological Account 3 5.2 Case Studies of Companies Using a Company Level Collective Agreement 3 5.2.1 Getronics 3 5.2.2 EDS 3 5.2.3 Atos Origin 3 5.2.4 PinkRoccade 3 5.2.5 Torex-Hiscom (isoft since 2004) 3 5.3 Case Studies of Companies Not Using a Collective Agreement 3 5.3.1 LogicaCMG 3 5.3.2 Ordina 3 5.3.3 Capgemini 3 5.3.4 Centric 3 5.3.5 Baan (SSA Global Since 2003) 3 5.4 Flexibility: Collective Agreements vs. Works Council Agreements 3 5.5 Conclusion 3 6 A SURVEY AMONGST IT PROFESSIONALS 3 6.1 Representativeness of the 2003 Berenschot-Automatisering Gids Salary Survey 3 6.1.1 The Problems With Internet Surveys 3 6.1.2 General Characteristics of IT Professionals 3 6.2 IT Professionals in the Internet Survey 3 6.2.1 IT Professionals Inside and Outside the IT Services Industry 3 6.2.2 Job Level and Highest Attained Level of Education 3 6.3 Union Membership Amongst IT Professionals 3 6.3.1 Union Membership in the IT Industry Compared to National Figures 3 6.3.2 Union Membership, Age, Income, and Job Level 3 6.4 The Terms and Conditions of Employment of IT Professionals 3 6.4.1 Company Size and the Type of Agreement 3 6.4.2 Type of Agreement Versus Terms and Conditions of Employment 3 6.4.3 Preferences of IT Professionals in the IT Services Industry 3 6.5 Conclusion 3 III

IV 7 THE PRESENCE AND FUTURE OF COLLECTIVE AGREEMENTS 3 7.1 Explaining the Presence of Collective Agreements in the Dutch IT Industry 3 7.1.1 The Limited Choice Set Hypothesis 3 7.1.2 The Efficiency Hypothesis 3 7.1.3 The Anachronism Hypothesis 3 7.1.4 Linking the Empirical Results Back to Institutional Theory 3 7.2 Explaining the (inter)actions of Actors 3 7.2.1 Unions and Works Councils in the Dutch IT Industry From an Employees Perspective 3 7.2.2 Unions and Works Councils in the Dutch IT Industry: Employers Tools? 3 7.3 Individuality and Flexibility in Collective Agreements: A Paradox? 3 7.4 The Future of Collective Agreements in the Dutch IT Industry. 3 7.5 Some Final Remarks 3 SAMENVATTING (SUMMARY IN DUTCH) 3 LIST OF INTERVIEWED PEOPLE 3 REFERENCES 3 INDEX 3

List of Tables and Figures Table 1.1 Research questions 3 Table 2.1 Clauses and commitments of parties concluding collective agreements 3 Table 2.2 The legal structure of the various terms and conditions of employment available in the Netherlands 3 Table 3.1 Growth of the Internet between 1969 and 2003 3 Table 3.2 Vacancies in the Dutch IT services (SIC 93 class 72) 3 Table 3.3 Services provided in the Dutch IT industry 3 Table 3.4 Combined services branches in the Dutch IT industry 3 Table 3.5 The Dutch IT industry in terms of the Dutch Standard Industrial Classification (SIC 93) 3 Table 3.6 Added SIC 93 subclass 3 Table 3.7 Omitted SIC 93 classes 3 Table 3.8 Number of IT companies in the Dutch IT manufacturing (30) and services (72) industries 3 Table 3.9 Demographics of the Dutch IT industry in terms of company size 3 Table 3.10 Trading results IT services (SIC 93 class 72) 1996-1999 3 Table 3.11 Growth in the number of IT companies compared to the growth in the number of jobs in the Dutch IT industry between 1993 and 2002 (Index 1998 = 100) 3 Table 3.12 Growth of the number of employees in the Dutch IT industry between 1993 and 2002 (SIC 93 class 72) (Index 1998 = 100) 3 Figure 4.1 Three periods of branch and employers associations in the Dutch IT industry 3 Table 4.1 Relative size of unions and union density in the Dutch IT industry 3 Figure 4.2 Actors and the type of agreements in the field of industrial relations in the Dutch IT industry 3 Table 4.2 Types of regulations in the ICK collective agreement 3 Table 4.3 The à la carte terms and conditions of employment 3 Table 4.4 The pay system and job rating system compared 3 Table 4.5 Working hours 3 Table 5.1 Overview of the cases 3 Table 5.2 Interests of the negotiating parties with regard to a company level collective agreement 3

VI Table 5.3 The flexibility regulations contained within the EDS collective agreement and Ordina works council agreement 3 Table 5.4 The flexibility regulations contained within the Atos Origin collective agreement and Capgemini works council agreement 3 Table 5.5 The flexibility regulations contained within the Torex-Hiscom collective agreement and Baan works council agreement 3 Table 5.6 The flexibility regulations contained within the Getronics collective agreement and LogicaCMG works council agreement 3 Table 5.7 The flexibility regulations contained within the PinkRoccade collective agreement and Centric works council agreement 3 Table 6.1 General characteristics of Internet survey respondents compared to other statistical sources (figures are percentages of the total respondents) 3 Table 6.2 General characteristics of Internet survey respondents working inside and outside the IT services industry 3 Table 6.3 Job level versus highest attained level of education and average age 3 Table 6.4 Distribution of main functions 3 Table 6.5 Unions and union membership of active respondents aged 15-64 3 Table 6.6 Union membership amongst IT professionals inside and outside the IT services industry in percentages of the total respondents 3 Table 6.7 Distribution of types of agreement amongst employees in the Netherlands 3 Table 6.8 Company size and type of agreement in the IT services industry 3 Table 6.9 Type of agreement and company size of IT professionals in the IT services industry 3 Table 6.10 Characteristics of IT professionals working with individual agreements in the IT services industry by company size 3 Table 6.11 Type of agreement of IT professionals versus primary and secondary terms and conditions of employment 3 Table 6.12 Preferences of IT professionals regarding the terms and conditions of employment in the Dutch IT services industry 3 Table 6.13 Preferences and opinions of IT professionals in the IT services industry regarding the type of agreement 3 Table 7.1 Original research questions 3 Table 7.2 The empirical data and empirical explanations for the presences of collective agreements versus the hypotheses 3

List of Abbreviations ASP Application service provider ATW Arbeidstijdenwet (Working Hours Act of 1995) AVV Algemeen verbindend en onverbindend verklaring (extension of collective agreements) AWGB Algemene wet gelijke behandeling (the equal treatment Act) AWV Algemene Werkgeversvereniging (General employers association) AWVN Algemene Werkgeversvereniging VNO-NCW (General employers association VNO-NCW) BBA Buitengewoon Besluit Arbeidsverhoudingen (The Extraordinary Decree on Labour Relations) CA Collective agreement CAO Collectieve arbeidsovereenkomst (collective agreement) CBS Centraal Bureau voor Statistiek (Statistics Netherlands) CNV Christelijk Nationaal Vakverbond (National Federation of Christian Trade Unions) CRM Customer relationship management EAI Enterprise application integration ERP Enterprise resource planning ESC European Social Charter FTU Full time unit ISIC International Standard Industrial Classification ITOP IT Ondernemingsraden Platform (Platform for IT Works Councils) FENIT Federatie Nederlandse IT-industrie (Federation of Dutch IT firms) FHI Federatie van Technologiebranches (Federation of Technology Branches) FNV Federatie Nederlandse Vakvereniging (Netherlands Trade Union Confederation) HIO Hogere Informatica Opleiding (school for higher computer science) HTML Hypertext Mark-up Language HTTP Hypertext Transfer Protocol ICT Information and communication technology ILO International Labour Organisation IOP Individueel ontwikkelingsplan (Individual development plan) IPO Initial public offering ISP Internet service provider ISV Independent software vendor

VIII IT Information technology OEM Original equipment manufacturers PC Personal computer SER Sociaal Economische Raad (Social and Economic Council) SOI Sectoraal Overleg Ondernemingsraden in de IT-branche (Sector level Consultation for Works Councils in the IT industry) SOC Standard occupation classification system SOP Software Ondernemingsraden Platform (Software Works Council Platform) STAR Stichting van de Arbeid (Labour Foundation) VIFKA Vereniging van Importeurs en Fabrikanten van Kantoormachines (Association of Importers and Manufacturers of Office Machines) VNO-NCW Vereniging van Nederlandse Ondernemingen-Nederlands Christelijk Werkgeversverbond (Confederation of Netherlands Industry and Employers) VAR Value added reseller VHP Vereniging voor Hoger Personeel (Union for Higher Personnel) V-ICTN Vereniging ICT Nederland (Dutch ICT Association) VONIT Vereniging van Onafhankelijke IT-specialisten (Association of Independent IT Specialists) WAVV Wet algemeen verbindend/onverbindend verklaren van CAO s Act on extending and un-extending collective agreements of 1937) WAZ Wet Arbeid en Zorg (labour and care Act of 2001) WCAO Wet op de collectieve arbeidsovereenkomst (Collective Agreement Act of 1927) WLV Wet op Loonvorming (Wage Formation Act of 1969) WMM Wet minimumloon en minimumvakantiebijslag (minimum wage and minimum holiday allowance Act of 1968) WOR Wet op de ondernemingsraden (Works Council Act of 1998) WSOI Werkgroep Sectoraal overleg Ondernemingsraden in de ITindustrie (Study group Sector level Consultation for Works Councils in the IT industry)

Acknowledgements There IT is! It is 1982, I am fourteen years old, and my dad finally arrives with a big box under his arm. I have to control myself to an extent unbearable for a boy that age, but I manage, and let my dad open the box. Out comes lots and lots of stuff amongst which the user manual... Uh oh, I hear myself thinking, and yes, my dad starts carefully sifting through the pages, but stops quite suddenly and gives my mum a wink. IT, a brand new Commodore 64 home computer, is quickly installed, hooked up to a small black and white TV, and turned on. After a few seconds the magic line appears that would forever change my life: READY. I have heaps of friends who think my passion for computers is just plain strange. To them it is a machine that does things. They did not understand that to me computers were instruments I can control. Something unimaginable in real life at that time with constricting factors such as parents and school. It seems strange that in those days, and also later when the time arrives to decide what study to take and what to become, I never considered the Information Technology industry. I wanted to become a movie director instead and tell stories to people. Obviously, the most sensible path to pursue a career in the movies was to study sociology, learn about society, and have something to talk about. During my studies I became infected with theories about how societies probably worked and even attempted to do my own bit of theorizing. During my studies I met two wonderful people with an even greater passion for theory and society: Frans Kerstholt and Jan van Wezel. They took me under their wings and introduced me to a range of theories from various disciplines and kept me in blissful ignorance of the real life research topics of a social scientist trying to make a living. After having worked at Tilburg University for a year, I started working for the Hugo Sinzheimer Institute (HSI), a brilliant and dynamic research institute for labour and law, associated with the University of Amsterdam. During my first years at the HSI, I researched various fields such as labour, training and education, most notably together with Harm van Lieshout, who gave me a lot of trust and taught me a great many things during the time we worked together. After some time, I wanted to master a field of my own, which became industrial relations in the Dutch Information Technology industry.

X It was predominantly through the efforts and faith of Paul van de Heijden in the research topic that working on this thesis was possible. I would also like to express my warm thanks to a lot of other people important to me during this study. First and foremost the superpowers Ton Wilthagen and Robbert van het Kaar, who were my direct supervisors throughout the whole project. In hindsight, and after having supervised a number of students myself since, my respect for them and the difficulty of their job has increased significantly. Also Martijn van Velzen and his seemingly bottomless pit of altruism, expertise and humor was of great value to the successful completion of this study. Other former HSI colleagues, in particular Frank Tros, Dlanor Reztleb, and Jan Popma, should be thanked for their time and comments. I am also grateful to Astrid Ornstein for making working life at the HSI a living heaven. The expertise of people such as Harrie Gooskens and Gerard Alberts have proven invaluable to me for writing a pioneering chapter on the Dutch history of computing. Something which simply did not exist at the time I met them. Both Bas Linders and Berenschot should be thanked for allowing me to include a number of research questions on their Internet survey. Finally I would like to express my heartfelt thanks for your continuing support throughout the years to all those close to me and the closest of them all, my wife and love, Linda.

1 Introduction During 1997 and 1998 three collective agreements emerged in the Dutch Information Technology (IT) industry. 1 They seemed to come from nowhere, because collective agreements, especially in the Dutch IT software and services industry, 2 had been a rarity until then. 3 Taking into consideration the fact that collective agreements have been around in the Netherlands since 1927 and that in 2003 85 percent of all employees in the Netherlands fell under the scope of a collective agreement (SZW 2004), 4 the relative absence of collective agreements in the Dutch IT industry is quite conspicuous. Obviously this has something to do with the age of the sector, which is, compared to other industries, quite young. This, however, still does not explain why collective agreements have never become as popular in the Dutch IT industry as they have been in other industries. It also raises the question as to why a number of them suddenly did appear between 1997 and 1998. These basic questions underlying this study gain importance if one also takes the following facts into consideration: The IT industry is generally associated with highly skilled and highly trained employees who are expected to be more than capable of negotiating their own terms and conditions of employment; The legitimacy of collective agreements has been under discussion even before they were introduced in the Netherlands in 1927. Collective agreements are sometimes criticized as unwieldy and outdated institutions; Because the Dutch IT industry is a relatively young sector, it does not share the same historical background and traditions regarding industrial relations which have come to be associated with other industries such as the metal industry or the construction industry. So the basic question of this study is: Why are there collective agreements in the Dutch IT industry? 1 Information technology, in the Netherlands, is often referred to as Information and Communication Technology (ICT). Throughout this study however, the term Information Technology (IT) will be used. Despite the obvious relations between Information technology and communication technology, they are still different enough in nature to be identified as separate industries. 2 See chapter three for an elaborate definition of the Dutch IT industry. 3 See chapters four and five for a detailed historical discussion of industrial relations in the Dutch IT industry. 4 See chapter two for a more elaborate discussion of the origin of the collective agreement in the Netherlands.

2 Developing Institutions 1.1 Problem Definition This study aims at a detailed and distinguished case study of the developing governance regime for labour relations in general, and collective agreements in particular, in the Dutch IT industry. The first goal of the study is to map the emergence of collective bargaining in the Dutch IT industry (See table 1.1, 1); and the second goal is to understand and explain its emergence (See table 1.1, 4). In order to understand the position of the collective agreement in the Dutch IT industry it is also important to look at the socio-political context of the negotiation process (See table 1.1, 2), as well as analyse of the contents of the collective agreements in comparison to other types of arrangements (See table 1.1, 3). Which actors have stimulated the emergence of collective agreements, and why? Why have some IT firms introduced collective agreements and others not? Is this a matter of union power and/or management preference? Finally, as the emergence of collective agreements in this sector is a relatively new phenomenon, the future developments that actors in this sector expect will be examined (See table 1.1, 5). Will the number of collective agreements continue to grow? And, is firm-level bargaining the end of the line, or is an industry wide collective agreement conceivable at some point in the future? Table 1.1 Research questions 1 To what extent and in which particular ways do collective agreements emerge in the Dutch IT industry? 2 Which actors (firms, unions, employers associations and others) are involved in this process? What are their opinions on (various types of) collective agreements? What role have they played in enabling (or hampering) the emergence of collective agreements in this industry? What are the reasons for their preferences and actions? 3 What are the contents of these collective agreements in terms of flexibility and individuality? How do the stipulations in these agreements compare to prevailing arrangements on similar topics in similar IT firms that lack a collective agreement? 4 How can the extent and nature of these collective agreements in the Dutch IT industry be accounted for? 5 What are the prospects for collective agreements in the Dutch IT industry? This chapter will start by discussing the theoretical starting point of this study, which is based on the institutional theories provided by Douglass North and Jack Knight. North is an economist and uses institutions to explain the differences in economic performance between countries. Knight is a political scientist and uses institutions to

Introduction 3 explain social conflict, which he understands as distributional conflicts over resources. Sections 1.2 and 1.3 discuss what institutions are, how they come about and how they change. Collective agreements, works council agreements, and individual employment contracts - the three major ways to regulate the terms and conditions of employment in the Netherlands - are all considered institutions in this study. This is so because they regulate the relation between two parties, the employer and the employees, and describe what both parties can expect from each other. This chapter then continues to formulate three hypotheses, which are based on insights from sections 1.2 until 1.4, and may offer explanations for the presence of collective agreements in the Dutch IT industry. Having or not having a collective agreement, at least in the Dutch IT industry, is a matter of choice. Finally, this chapter concludes with a discussion of the methodologies used in this study. 1.2 Institutions and Institutional Theory The goal of this chapter is to formulate a number of hypotheses which offer theory embedded explanations for the presence or absence of collective agreements in the Dutch IT industry. The theoretical starting point will be institutional theory as defined by North and Knight. The basic assumption underlying this study is that actors (i.e. Dutch IT firms) make rational choices within a given institutional setting that restricts their decisions. 5 Before continuing to discuss whether the collective agreement or its alternatives can be regarded as institutions, it is first necessary to shed some light on institutions, institutional theory, and what it is able to explain. At this point it is important to understand that within traditional sociology, institutions have been identified with group survival and group complexity. Within economics, institutions have come to be associated with economic growth and efficiency, or the absence thereof. In this study institutions are defined as: shared concepts between humans in situations of repeated interaction that are prescriptions for behaviour that are enforced by external agency (rules) or by the individuals themselves (norms), or derive from strategies that individuals adopt within the accepted rules, norms and their expectations about how everybody else will behave (Kay 2003, 3). According to contemporary institutional theory institutions are embedded in an institutional matrix. The institutional matrix consists of all institutional systems. An institutional system itself is a system of functionally interrelated institutions (Bush 1987, 1076). 6 In order to bring systematization into the multitude of institutions, institutions are often categorized into three different levels: the constitutional level, the collective choice or 5 The concept of rationality will be explained later on. 6 In chapter two the institutional system, in which the collective agreement resides, will be discussed in greater detail.

4 Developing Institutions policy decision level, and the operational level of individual actions. 7 Because of the interrelatedness of institutions, there is no strict hierarchical division between the institutional levels. Institutions are imperative for the social structuring of our lives; at least that is what sociologists like us to believe, such as for instance Durkheim, 8 who even went as far as defining sociology as the science of institutions. 9 In sociology institutions are commonly defined as webs of interrelated norms and rules (both formal and informal), governing social relationships. This so-called governing of social relationships implies that institutions can be considered as solutions (though on occasion far from optimal or efficient) to recurrent social problems. Institutions thus structure human behaviour. Saying that institutions are solutions though, immediately puts institutional theory into a functional context. It is, however, not a functional statement in the traditional sense, because the function of institutions is not ascribed afterwards by a theorist, but beforehand by those who created the institutions. Institutions are a product of intentional design. Institutions are created for dealing with recurrent social problems. Recurrent social problems take place in situations where people are living together and where problems with coordinating the interactions between these people emerge. If interactions are repeated and the coordination problems of these interactions are not effectively dealt with by means of rules or agreements, these problems will recur. If people are unable to form such solutions when they are essential for the groups functioning and cohesion, these groups will find it very hard to live together or produce those goods that enable the society to continue. An important factor in the discussion of why institutions perform such a significant social function is group size. As groups grow larger the probability of social problems tends to occur more frequently as there are more potential and complex interactions in larger groups. 10 This latter point has also been noted in economics by some authors, notably North. North defines institutions in two ways. In North (1991a, 97), North defines institutions as the humanly devised constraints that structure political, economic and social interaction. In North (1991b, 3), North continues that institutions are the incentive structure of an economy and, therefore, fundamentally influence individual choices. These humanly devised constraints consist of formal and self imposed informal rules. A noteworthy observation made by North is that these constraints reduce uncertainty and create order. In the end institutions define the choice set and although North restricts this 7 With the introduction of international laws and regulations often influencing the constitutional level of individual states, it would make sense to introduce a fourth institutional level of international legislations. 8 Contemporary sociologists also value institutional analysis highly. See for example Brinton and Nee (1998) and Smelser and Swedberg (1994). Both volumes are characterized by their multidisciplinary approach, which incorporates both sociology and economics. 9 Emile Durkheim (1895) Règles de la méthode sociologique. 10 More on this when Mancur Olson s theory of collective action is discussed below.

Introduction 5 set to economics by combining this choice set with the constraints coming from economic laws, this choice set can be defined as the set of contingent social actions. The concept of incentive structures is especially interesting as it relates institutions and institutional systems to economic outcomes. North (1991b, 8) states that economic growth is dependent on institutions that are able to lower the cost of transactions. 11 Incentive structures are the reflected opportunities provided by the institutional matrix. North (1994) gives an example of what he means by these reflected opportunities: if the institutional framework rewards piracy then piratical organizations will come into existence; and if the institutional framework rewards productive activities then organizations firms will come into existence to engage in productive activities. In this study the collective agreement is regarded as an institution that is embedded in an institutional matrix defining the Dutch system of industrial relations. Figure 2.1 in chapter two gives a schematic presentation of the institutional matrix for the Netherlands and illustrates how actors, institutions, and legislation are related in this system. It would go too far to discuss the entire figure at this point as the Dutch system of industrial relations will be discussed in the next chapter, but it is at this point important to understand the following: Collective agreements and Works council agreements can be a part of individual agreements. Collective agreements are not contracts of employment, but rather a set of agreements on the terms and conditions of employment made by an employer or a body of employers with one or more unions. These agreements are imposed upon each individual contract of employment covered by the collective agreement; Collective agreements can only be imposed upon employers through extension of a collective agreement to an industry or industrial branch by the Minister for Social Affairs and Employment. When a collective agreement is extended to an industry, it is called an industry wide collective agreement; Employers organizations and unions can also create an industry wide collective agreement without it being extended by the Minister for Social Affairs and Employment. These types of collective agreements cannot be imposed upon employers and are only accepted voluntarily; Like collective agreements, also works council agreements are not contracts of employment, but rather a set of agreements on the terms and conditions of employment made by an employer and its works council. In the Netherlands a works council's degree of involvement is limited by law. It is up to the employer to extend the level of participation of its works council beyond this minimum level of participation. 11 Transaction costs are costs involved by bringing about a transaction, which do not include the costs of that which is bought or sold itself. Transaction costs can take many forms, such as e.g. the costs of looking for a buyer or seller, et cetera.

6 Developing Institutions Works council agreements cannot be imposed upon employers, nor can they be imposed upon employees. An employee can refuse to accept the terms and conditions of employment set in the works council agreement. If a contract of employment is neither based on a collective agreement nor works council agreement, it is based on an individual agreement between the employer and employee. The contract of employment contains the terms and conditions of employment without referring to a(n industry wide) collective agreement or a works council agreement. Because of the absence of an extended industry wide collective agreement, employers in the Dutch IT industry have the choice to select by what means they wish to regulate the terms and conditions of employment for their employees. The same holds true for employees who do not fall under the scope of a collective agreement. 1.3 Origin and Change of Institutions Describing the functions of institutions and the way they are embedded in society does not say anything about the origin and change of institutions. Institutions are often criticized for their rigorousness and their limited ability of adapting to general changes in society. Historically, collective agreements have had to deal with similar criticisms. Collective agreements for instance have generally been criticized because they restrict the freedom of employers to implement changes, which criticasters of collective agreements argue is required to be internationally competitive. 12 These criticisms have not dissipated over the years and can still be heard today. The rise of newly industrialized countries and the globalization of markets has changed the global economy and increased international competition. 13 These changes call for adapting to a cheaper, more flexible system of production, which collective agreements prevent, or at least make difficult because of a multitude of regulations and because unions are unwilling to help bring these changes about. Creating a more flexible production system, at least in terms of labour, is limited to two factors: price and quantity. Whilst there are a number of ways imaginable to manipulate these factors, 14 these are usually restricted by laws and institutions, which limit the extent to which these flexibility measures can be executed. This often leads employers to perceive institutions such as the collective agreement as rigid and inefficient. 12 In chapter two the position of the collective agreement in the Netherlands will be discussed in greater detail. 13 See for instance De Korte (2000), Steijn (2001), Tros (2000a), Van der Meer & Smit (2000), and Nagelkerke & Wilthagen (2000). 14 For instance by introducing a pay for performance system, or a merit based system. One can also think of hiring and firing personnel based on demand, without having laws penalizing the breaking of a contract of employment.

Introduction 7 Increased international competition has also affected traditional Taylor inspired modes of production and require a so-called post-taylorist mode of production (De Korte 2000). The institutional matrix used in the Netherlands is attuned to a classical, 15 Taylorist notion of work and income, which is based on fixed income, set working times, a transparent job description, authority, collective social securities, and life long employment. This classical mode of production clashes with today s demands for product differentiation and flexibility. This is especially apparent in companies that are driven by fast technological developments, a strong focus on knowledge, and fierce competition, such as IT and telecommunication companies. These types of companies need to be able to flexibly deploy their means of production, which are often employees and not machines (De Korte 2000). De Korte s criticism, as well as the historical criticisms against the collective agreement, offer an empirical and theoretical challenge. It is challenging to determine empirically whether collective agreements do not fit in well with IT companies. Theoretically the challenge is to put forward ideas on how institutions are created, whether they are able to change, and if so, how they are able to do so. For this the insights of Knight and Olson will be added to those of North. Knight discusses the origin, functioning, and change of institutions as a product of differences in power. 16 Mancur Olson discusses the problems of collective action and socially beneficial outcomes, 17 that is, outcomes that will be beneficial to the greater good, outcomes which supersede the individual. Knight defines institutions as networks of social conventions, norms, rules, and rights that form the basis of society. These networks structure social interactions in particular ways and provide information on how people are expected to act in given situations. Similar to North, Knight recognizes the structuring qualities of institutions. Institutions create the incentive structure for actors by determining which actions are more favourable in social and economic terms. These institutions need not be efficient though as they are an expression of the power relations in the society. Institutions reflect the conflicts over the allocation of resources between actors. The allocation of resources is inherently unequal because of differences in power. Power stems from distributional advantages, which is defined as the ability to affect by some means the alternatives available to a person or group. Distributional advantages are thus created by limiting the preferred strategic actions of others. Knight explains the origin of institutions out of micro interactions between individuals or groups of individuals. This is a two stage process. Firstly, behavioural expectations need to stabilize into conventions, and secondly, these conventions again have to be generalized into institutions that hold for all members of a group or 15 De Korte (2000) uses a brief historical account of the origin of the collective agreement in the Netherlands. A more elaborate historical account of the Dutch system of industrial relations will be presented in chapter two of this study. 16 The discussion of Knight s theory of institutions and institutional change is based on Knight (1992). 17 The discussion of Mancur Olson s insights on collective action draws on Todd Sandler (1992), who has systemized Olson's (1965) The Logic of Collective Action.

8 Developing Institutions society. Conventions may come about through repeated interactions, which can be regarded as coordination games. Conventions only tend to come about in those situations where there is a unique equilibrium. In the case of multiple equilibriums matters tend to become more difficult and conventions may or may not come about. In the case of multiple equilibriums that are valued differently by those involved, the chances of stable conventions coming about are even more improbable. The generalization of these conventions into institutions is a process of social recognition of strategic rules of action. Actors are confronted with the stabilized behavioural expectations and realize that compliance with them is their best response to the actions of others. Knight regards institutional change as a very difficult process, which may only come about under certain conditions that involve changes in either the distributional consequences of institutions or relative bargaining power. Even if one of these conditions is fulfilled, the process of institutional change will be very difficult due to the problems involved in the process of collective action. Knight (1992, 147) writes that it is a problem inherent in any attempt to alter existing institutional arrangements, and it explains why neither a preference for new institutions nor shifts in power may lead to immediate change through a decentralized process. Knight thereby implies that there is no ground for the belief that institutions are efficient for a collective. Institutions mainly represent the interests of those with more resource holding power. Only under rare conditions do institutions arise that are collectively efficient. Olson has pioneered the exploration of the problems of collective action that Knight refers to as a major obstacle to institutional change. Collective action arises in situations where a problem creates costs and or benefits for more than one person, so that some coordination of efforts is required. Problems with collective action are typified by those situations in which the actions of the participants are interdependent. This situation looks remarkably like an interdependent strategic situation commonly found in game theory textbooks. 18 Olson's main argument rests on the assumption that individual rationality is not sufficient for collective rationality. By this Olson means that there are differences with regard to the rationality of collective outcomes. Olson's Logic of collective action is mainly concerned with situations in which the pursuit of self-interest leads to sub-optimal outcomes. This does not tally with one of the core assumptions of neoclassical economics, which states that rational self interested actions lead to the best outcome for the whole group. 18 Elster (1986, 6-9) discerns three types of interdependency in strategic situations. The first type of interdependency is that in a strategic situation the rewards of each depend on the rewards of all. Rewards are also dependent on the choices everyone makes. This is the second interdependency. The third interdependency relates to choices in the sense that in a strategic situation the choices of each depend on the choices of all. It is most notably this third interdependency, which characterizes game theory.

Introduction 9 Problems of collective action leading to suboptimal outcomes arise only in those situations in which goods are public and or markets are not competitive. Olson mentions several factors why collective action may not come about or is suboptimal. One factor is group size. As group size increases, individual contributions are limited. This causes actual provision levels to decline as the group grows. An explanation for this is that an individual may think that his share of the gain of collective contribution is so small that he refrains from making a contribution to the public good. A second possible explanatory factor is group asymmetry. Individuals in groups may have different tastes and abilities. This is mostly the case in heterogeneous groups. Here the individuals and collectives are confronted by an exploitation problem, where the 'rich' carry the burden of the 'poor'. Whether or not public goods are provided depends on the notion of privilege. A privileged group is a group in which the benefits of collective action for all the individuals of that group exceed the costs of not having access to the public good, even if these costs are solely borne by that individual or coalition. If benefits decline when the size of the group increases, then the likelihood of contribution depends on whether the costs decrease or remain static, as the decrease in costs may offset the decline in benefits. Olson stresses the importance of institutions for the sharing of the costs of providing for public goods, since if costs were shared through institutions, individual costs would indeed decrease with an increase in group size. Olson addresses two ways of overcoming problems of collective action. The first one is that private benefits should be tied up with the provision of collective goods. Olson calls this selective incentive. Selective incentives can be both social sanctions and social rewards. They are restricted to smaller groups however. A second approach, which is more effective on a societal level, is the design of institutional structures. Olson writes that large groups can foster cooperation by organizing themselves in a federal structure with small subgroups. This, however, may affect collective benefits, as interactions over the entire collective may be limited. But institutions can be designed to create an underlying game structure that promotes cooperation instead of defection. The theoretical insights of North, Knight, and Olson have been brought together here to demonstrate that though this criticism may be true, modern institutional theory is able to connect micro interactions to macro outcomes, 19 is able to explain why not all institutions are efficient, what their relevance is to societies, and why institutions are difficult to change. 19 Classical institutional theory is often associated with Adam Smith who considered all institutions to be efficient. Ronald Coase (1937), however, radically changed this belief by introducing the concept of transaction costs. Transaction costs have a big impact on economic outcomes, and are always present and significant. In a world of zero transaction costs, institutions would not matter (Yeager 1999, 33).

10 Developing Institutions 1.4 Explaining the Emergence of Collective Agreements in the Dutch IT Industry At the end of section 1.2 the collective agreement was set against the works council agreement and the individual contract of employment, to explain that the research question of this study is based on the empirical state that companies in the Dutch IT industry are free to choose a way to regulate the terms and conditions with their employees. In the Dutch IT industry no extended industry wide collective agreement exists. This enables IT companies to employ either a works council agreement, a collective agreement, a combination of the previous two, or to simply choose to arrange the terms and conditions of employment on an individual basis by means of a contract of employment solely. In order to formulate hypotheses that have empirical relevance, factors and conditions need to be addressed that influence the potential choices of Dutch IT companies. The hypotheses thus will reflect just how free actors really are in their choices. It is important to identify the decision making power of individual companies in the Netherlands. In the Dutch system of industrial and labour relations, decision making power resides on three levels. Shifts in decision making power between these levels are usually identified as processes of centralization or decentralization. The highest level of decision making power resides on the macro level, which is defined by the government, as well as two government advising bodies, the Social and Economic Counsel (Sociaal Economische Raad, SER) and the Foundation of Labour (Stichting van de Arbeid, STAR). 20 The second, meso level is defined by the employers and employees organizations. The micro level itself can be subdivided into the company level and the level of individuals. Centralization and decentralization can be understood as shifts in decision making power from one hierarchical level to another. 21 For instance, from the government to local authorities. In the case of centralization power is shifted from a lower to a higher hierarchical level and in the case of decentralization from a higher to a lower hierarchical level. Processes of centralization and decentralization can occur between various levels: from the individual level to a company wide level, and vice versa, and from a company level to an industry level and vice versa. Special forms of decentralization are sometimes referred to as flexibilization or individualization. Flexibilization can take place at different levels and implies that actors at a lower hierarchical level are given options instead of rules. Individualization, which takes place at the level of the individual, is a special case of flexibilization where options are given to individuals directly, instead of to a group of individuals. The hypotheses mentioned below will predominantly operate on the level of the individual firm, or employees, and not take into account higher levels of decision making. This study claims that IT companies are relatively free in choosing the type 20 See chapter two for a more detailed account of these latter two organizations. 21 The definitions of the processes of centralization and decentralization draw on Tros (2000a, 25-32).

Introduction 11 of arrangement they prefer to regulate their terms and conditions of employment. None of the institutions in question have undergone serious changes throughout the historic scope of this study, these institutions are meaningful to the actors selecting them. Having alternatives available may even tone down potential forces that could bring about institutional change. De Korte s critique of the collective agreement mentioned above, if correct, thus may induce actors to choose less rigorous alternatives to the collective agreement instead of bringing about institutional change. Moreover, as shall be demonstrated in chapter two, the contents of works council agreements, collective agreements, and contracts of employment, are not fully determined by law. Actors working with these institutions have a certain degree of freedom with regard to what they agree upon. In this sense the institutions that are studied here are less rigorous than most institutions. In this study it is thus not institutional change per se, but rather institutional selection that needs to be accounted for. 1.4.1 Hypothesis 1: The Limited Choice Set Hypothesis Path dependency often refers to a sequence of events whereby later events depend on earlier events. Path dependency is a process, which constrains future choice sets. To put this in the words of North (1990, 98-99): At every step along the way there are choices political and economic that provide real alternatives. Path dependence is a way to narrow the choice set conceptually and link decision-making through time. It is not a story of inevitability in which the past neatly predicts the future. Note that in North s definition of path dependency, the word constrains is used, rather than determines. Path dependency in North s definition does not exclude the possibility of institutional change, or the creation of institutions. The choice set, however, is restricted and not infinite. Though path dependency is commonly used to explain the difficulty of institutional change, it can also be used to explain why employers and employees in this case may not prefer an alternative institution to regulate the terms and conditions of employment. Factors within companies that contribute to this path dependency effect are amongst others the corporate culture and the background and history of the firm, but they may also include exogenous factors such as the dominance of the system of industrial relations in which the company operates. These factors may limit the choice set to those choices that appear most natural, recognizable, familiar, and attractive. This is especially the case when a certain institution poses no serious problems and actors are satisfied with it. Herbert Simon (1961) introduced the term satisficing to describe that actors have no intention in gathering additional information about alternatives. Simon states that people are by nature bounded in their rationality. This is so because they have no perfect foresight, nor do they possess the computational skills required for such an endeavour. Moreover the information required for rational behaviour is usually asymmetrically distributed. In order to cut decision costs (the time it costs to make a decision), a solution is chosen that is good enough.

12 Developing Institutions Emphasizing the power of corporate culture in this respect is important. As culture is everything non-material which is passed on from generation to generation. Kreps (1990) defines corporate culture as a set of collective templates for action. Corporate cultures can, according to Kreps, be an efficient supplement to traditional monitoring and enforcement mechanisms for organizations. Especially when the organization cannot readily specify appropriate behaviour for all contingencies. North, though somewhat later in his career, understood the power of culture. Yet instead of using it to explain possible efficiencies, North used it to explain the existence of inefficient institutions. Denzau and North (1994) elaborate on the topic of ideology and corporate culture. 22 Ideologies and mental models are formed from common experiences, which in turn influence the efficiency of institutions. So mental models and ideologies can explain inefficient institutions. Cognitive processes create path-dependence in ideas and ideologies, and then in institutions. This is a circular process though, since institutions again influence choice by limiting available choices to individuals. North and Denzau (1994) conclude that: Ideas matter and the way by which ideas evolve and are communicated is the key to developing useful theory which will expand our understanding of the performance of societies both at a moment of time and over time. Reasons for not choosing or establishing a different institution may thus be that the costs of gathering additional information are too high, or alternatives may bring about uncertainties, or introduce risks. In the case of firms, the use of a particular institution may be part of a particular corporate culture. Propositions 1.1 and 1.2 summarize the above. 1.1 If a dominant majority of employees within the company is familiar and satisfied with a certain type of institution that directly or indirectly regulates the terms and conditions of employment and gives preference to this type of regulation, they will choose this institution over other institutions if they have the choice. 1.2 If employers are accustomed to working with a certain type of institution that directly or indirectly regulates the terms and conditions of employment for their employees and they are satisfied with this type of regulation, they will choose this institution over other institutions if they have the choice. The hypothesis takes into consideration the background of a company, its culture and history, and exogenous factors, which can limit a company s choice set. Whatever the choice actors at 22 North (1992, 1993a, 1993b) was dissatisfied with his earlier, neoclassical, treatment of human behaviour. He was sure that, although the neoclassical assumptions about human behaviour are the right tool in such situations where low-cost transacting exists, these assumptions were not able to explain ideologies and the institutions based on them. In order to find those answers, North turned to cognitive psychology. He admitted that by doing so, he was opening up the black box of rationality - a complex and still very incomplete part of science.

Introduction 13 one point made, it is a long term choice and will only be re-evaluated if one of the previously mentioned factors changes considerably. 1.4.2 Hypothesis 2: The Efficiency Hypothesis The conclusion from section 1.3 was that institutions can never be fully efficient and will always lead to transaction costs. This does not imply that different institutions generate the same (amount of) transaction costs in similar conditions. The general idea behind the efficiency hypothesis is that actors, i.e. employers, seek to reduce transaction costs of the institutions they are working with. Not using a collective agreement implies, at least in theory, that a company has to negotiate with each and every individual over the terms and conditions of employment. This may be very costly, especially in those cases where companies employ a lot of people. Harmonization, a process whereby various existing regulations within a company are centralized into one general regulation, of these various regulations in the company may thus be a goal in itself as it brings about a reduction in the total transaction costs involved in individual and collective bargaining. Whilst different companies may wish to harmonize their individual regulations, their ulterior goals may be quite opposite. One company may use harmonization in order to offer tailor-made terms and conditions of employment. This is quite paradoxical as the employer wants to create tailor-made solutions by reducing individual solutions. It is however easier to offer individual arrangements from a common platform. Other companies arguments for harmonization may be less ambiguous as they simply want a single solution that allows them to negotiate with a single party. The efficiency thesis predicts that companies seek to reduce transaction costs involved in processes of collective and individual bargaining. Depending on their own estimations of the costs involved, companies will choose to negotiate over the terms and conditions of employment with either individuals, the works council, or the unions. If such alternatives do not exist, existing institutions will be changed, given that the group with the greatest bargaining power seeks to maximize efficiency, is strong enough to bring about these changes, and is able to come up with a more efficient alternative. 1.4.3 Hypothesis 3: The Anachronism Thesis According to the notion of post-taylorism described in section 1.3, collective agreements may very well be an anachronism, something out of harmony with the time. Post-Taylorism suggests that the employee is empowered within this new system of production. The negotiation powers of the employee are greater in this new system of production than they used to be and he can focus on his individual demands, rather than those of the collective. Empowerment of the employee may very well be one of the factors Knight recognizes as being able to shift the balance in power towards the individual employee, and thus, bring about institutional change, or in this case the selection of an alternative institution. Individual empowerment may

14 Developing Institutions reduce the power of trade unions that reflect and defend the power and interests of the collective employees. Traditionally, union power stemmed from the capacity to raise a strike fund which enabled the threat of a strike. With an increase in the individual bargaining power of employees, they may opt to cancel their union membership, or not become members at all, which then will reduce the relative strength of unions. Firstly, because less contributions are made to the strike fund, and secondly, the overall number of employees willing to strike will diminish. In the Netherlands strikes are relatively uncommon historically and the power of the unions is usually measured through the degree of representation. 23 This holds particularly for industries lacking an industry wide collective agreement, where representation is measured per company and where numbers may vary to a greater extent. Reduced union power may indirectly affect the choice for regulating the terms and conditions of employment through a collective agreement. Collective agreements typically represent the terms and conditions of employment for employees as a collective. Empowered employees may not wish to regulate their terms and conditions of employment through a collective agreement. In terms of bargaining power, which it is assumed every actor seeks to maximize, the employer may think it is in his interest to negotiate individually with each employee, or to opt for a works council agreement, where he negotiates with the works council on a few terms and conditions of employment, rather than with the unions on all terms and conditions of employment. 3.1 Employees seek to increase to maximize their bargaining power. If an employee s bargaining power is increased through the level of education he has received, or by some other means, in such a way that enables him to individually negotiate over the terms and conditions of employment, he will prefer to negotiate individually rather than in a collective via a union. The same holds vice versa. If their individual bargaining power is limited, they will seek the help of the unions. 3.2 Employers seek to increase their bargaining power and will favour individual negotiations with employees over collective negotiations with unions. According to the anachronism hypothesis the collective agreements present in the Dutch IT industry represent the interests of an outdated bargaining power and as such are anachronisms. These collective agreements will disappear and be replaced by agreements that express the interest of the dominant bargaining power. 23 Rojer (2002, 69) uses Eurostat data to indicate that in the Netherlands the lost number of labour days per 1,000 employees between 1974 and 1994 was 5. Germany, who had the second lowest number of lost days, had 9. The United Kingdom had 51, Finland 187, Spain 365, and Italy had lost 565 days on average.

Introduction 15 1.5 Methodology and Outline of this Study Industrial relations in the Dutch IT industry have, until this study, not been studied extensively in the Netherlands. 24 One can wonder why this has been the case. Though the IT industry is relatively young compared to other industries, it has been around for more than half a century now, so the age of the industry does not seem a very good excuse. When people are referring to the Dutch IT industry, they are often referring to a young and dynamic industry that is host to highly skilled and trained employees who work with cutting edge technology. In this study one of the goals is to refine this image, to give a more detailed account of the many stereotypical assumptions that are associated with the (Dutch) IT industry. Because there is very little written about industrial relations in the Dutch IT industry, this study has a strong focus gathering empirical information. In order to obtain all the information relevant to this study and required to answer the main research question about the presence of collective agreements in the Dutch IT industry, a number of methods have been used. Though each chapter will give a more detailed account of the methodologies used, a brief account will be given here. Chapter two explains the development of the Dutch model of industrial relations, and gives an account of the Dutch legal system in order to present an image of how industrial relations in the Netherlands have been shaped throughout history. This is necessary because, in order to understand the unique position of the IT industry in the Netherlands in terms of its model of industrial relations, the context should be presented as well. A historical account of the Dutch system of industrial relations is not enough to give a proper account of what is and what is not possible within this system in the Netherlands. Things are thought to be different in the Dutch IT industry and therefore there must be some freedom residing in the institutional system. Chapters three through six are about the Dutch IT industry as well as the actors that are relevant to a system of industrial relations. In the Netherlands these are: employers, employers associations, employees, works councils and trade unions. Each chapter uses different methodologies to obtain information. Chapter three introduces the Dutch IT industry by giving both an historical and statistical account. The historical part is predominantly based on articles and interviews with experts. The definition of the IT industry and its statistical description stem predominantly from the Dutch statistical services, the Central Bureau for Statistics (CBS). 24 A few studies exist, which give a brief description of industrial relations in the Dutch IT industry. See for instance Smit 1995; Tijdens, et al. 2000; Van Liempt and Van Uffelen, 2000; Van Liempt 2002a; Leisink 2002; and Van Velzen 2004. Internationally the image is similar and only a few studies can be found, such as for instance Boes and Baukrowitz (2002) who study workers participation in the German IT industry, and Kahmann (2003), who studied possible constraints from the growing IT industry on trade union action.

16 Developing Institutions Though the most important actors in the Dutch system of industrial relations have already been introduced in chapter two, in chapter four they are examined more closely in terms of their presence and actions in the Dutch IT industry. This chapter draws heavily on various articles in IT magazines and interviews with representatives of employers associations and unions. In what may be considered the core of this study, chapter five, ten case studies are presented of companies that use, or do not use a collective agreement. For each of these cases a detailed account is given of the type of company, their history, and how they came to use a collective agreement, or why they have never started to use one. These cases are based on IT magazine articles, publications provided by the various companies, and interviews with both representatives of the works council and the human resources departments of these companies. Also every relevant union was consulted in each case. The Dutch IT industry itself has been discussed in chapter three, its system of industrial relations in chapter four, and the employers in chapter five. This leaves room for one final actor in chapter six: the employee. In order to obtain up to date information on a large scale, an Internet survey was used, which asked employees about a number of relevant indicators for this study, such as the type of regulation that covered their terms and conditions of employment, whether they are members of a union, et cetera. Finally, in chapter seven, all of the empirical results are brought together and used to answer the questions this study has and see if the theory embedded hypotheses have any relevance in explaining the presence of collective agreements in the Dutch IT industry.

2 The Historical and Legal Context of the Dutch Collective Agreement The purpose of this chapter is twofold. Firstly, it explains why and how collective agreements have been created in the Netherlands. This historical analysis may be needed in order to put into perspective the yet to be determined causes and reasons underlying the presence of collective agreements in the Dutch IT software and services industry. In the historical analysis five periods are highlighted. The first period starts somewhere in the 15 th century and focuses on the gap between employers and employees, and the way in which the government intervened between these two parties through prohibition of coalitions. The second period, which is split up in two parts, starts at the beginning of the 20 th century and ends in 1918. The first part of this period focuses at the development of workers and employers coalitions, whilst the second part discusses the legislation of the collective agreement during this period. The third period centres on the preparation and realization of various labour regulations between the two World Wars, the inter bellum period. The fourth period focuses on the Wage Formation Act of 1969 (Wet op Loonvorming, WLV) and deals with developments up until 1970. The fifth and final period highlights the phenomenon of the retreating government. This final period starts in 1996 and discusses developments that are difficult to grasp into a central theme, but concentrate problems associated with the expanding role of employees representation. The second aim of this chapter is to present the legal context for regulation of the terms and conditions of employment in the Netherlands. This section demonstrates how the Dutch system of industrial relations operates and how the rules are hierarchically organized. I also discuss the possible means of formalizing the terms and conditions of employment. This discussion should help in understanding the decisions of IT firms when choosing the way in which they formalize and regulate the terms and conditions of employment.

18 Developing Institutions 2.1 The Historical Context of the Dutch Collective Agreement The collective agreement as we know it today in the Netherlands has a stirring history. This institution, which may be regarded as the culmination of collective worker empowerment, did not come about as easily as current figures indicate. Today 85 percent (SZW 2004) of Dutch workers is covered by a collective agreement. Before 1927 there were hardly any at all. This section highlights the historical context in which the collective agreement evolved in the Netherlands. 2.1.1 Worker Coalitions: From Prohibition via Toleration to Legalization (c.1400 - c.1870) The early phases of the development of collective agreements are typified by power imbalances between the various parties (employers, employees, and governments). These power imbalances were often created by the state, which protected interests of employers by banning employee coalitions. Two different periods of prohibition on coalitions can be discerned. The first period covers the late-middle ages to approximately the 19 th century, when various European states introduced a ban on coalition in response to various forms of social unrest. The combination of poor socio-economic conditions and the lack of voice or exit options proved to be a breeding ground for violent confrontations between workers and those who opposed them (Jacobs 1986, 10-12). A second period of coalition prohibition started at the advent of the 19 th century. New technologies that could replace workers, food shortages, taxes, et cetera, proved to be new sources for social unrest. Governments saw no other means of ending the social unrest other than to act upon existing laws that had gradually been relaxed over time. As social unrest returned it was simply a matter of falling back on these reinstated laws. Governments not only used laws that addressed collectives in their fight against social unrest. Laws were also used that could address individuals. Examples of this are laws against conspiracy and work interruptions (Jacobs 1986, 12-15). Governments, however, were not only apprehensive about worker coalitions, but also about employer coalitions. Although these employer coalitions never met the same opposition as worker coalitions did, they became a focus of attention whenever employer strategies were focused at decreasing wages (Jacobs 1986, 15-16). Great Britain was the first nation to abolish prohibition of coalitions. The country was far more industrialized than other nations during the first half of the 19 th century. A philosophy arose there that was a mixture of utilitarianism and the classical economic philosophy of laissez-faire that was becoming popular during the 19 th century. The belief was that if workers were free enough to do what they want, they would not need to coalesce. It was now possible for worker coalitions and employer coalitions to meet and discuss the terms of employment on a collective level. Thus the first steps towards a regulating institution were taken. The freedom that was

The Historical and Legal Context of the Dutch Collective Agreement 19 granted regarding coalitions was ambiguous though, since there still existed some minor clauses in the law that could be used to declare a coalition illegal (Jacobs 1986, 17-20). Meanwhile in the rest of Europe prohibitions on coalitions remained in force and close to the middle of the 19 th century even expanded. Nevertheless Great Britain turned out to be the example for the rest of Europe regarding industrial relations. The New Model trade unions that were introduced in Great Britain in the middle of the 19 th century turned out to be more than the local and temporal coalitions of employees that had existed until then. These unions were the model for the unions of the present day. During the second half of the 19 th century prohibitions on coalition were also abandoned in the rest of Europe (Jacobs 1986, 21-25). 2.1.2 Development of Worker and Employer Coalitions (c.1870 - c.1918) Abandoning the prohibition of coalitions was a major precondition for developing a legal basis for collective agreements. Allowing unions to represent workers, however, did not immediately imply that they were accepted as a negotiation party. The first trade unions and federations of labour arose in the Netherlands around 1870 and they expanded rapidly for a century. Examples of these are for instance the Nationaal Werkliedenverbond (1869), the Sociaal Democratische Bond (1881), and the Nationaal Arbeids Secretariaat (1893). These first unions were predominantly small-scale unions and had an ideological or religious background. Most of them originated directly as a result of a labour conflict. Employers associations, which shared to a large extent the same types of religion or ideological background as the unions did, can according to Van Peijpe (1985, 94-97) also be typified by other characteristics besides religion or ideology. Some employers associations had a distinct sense of social responsibility, whilst others focused on the commercial and economic interests of their members. A more important characteristic with regard to the subject of collective agreements was the scope of an employers association. These branch type associations concentrated primarily on their members interests with regard to labour conditions. Their goal was to form the employers into a united front. Some of these organizations even went as far as fining individual members who negotiated with their employees on an individual basis. Allowing coalitions was not enough by itself to establish a legal basis for collective employment contracts. Van Peijpe (1985, 97-99) distinguishes several other factors one of which is the rise of modern industrial capitalism. Industrial mass-production had an important impact on both employers and employees because of the scale of production. A second factor, the consciousness of workers, also proved crucial for the development of collective employment contracts. Workers were not interested in abstract matters such as a collective agreement. They were more interested in shortterm benefits and rather chose to go on strike first and negotiate later. A third factor was the fragile position of workers. Most of the workers were untrained and had no

20 Developing Institutions permanent contract in the first place and one can wonder whether collective agreements would really have benefited workers. A fourth factor was the lack of power of the workers unions. If collective agreements were the reflection of the actual power relations, workers would not have had any advantages and would be in exactly the same spot as they were without a collective agreement. In this light it is not surprising that those instances where forms of collective agreements did arise were not on the lower end of workforce, 1 but rather on the higher end of trained and skilled professionals such as e.g. diamond cutters, who had considerably more power than the low skilled workers. 2 For higher skilled workers the balance between going on strike and negotiating was biased towards the latter. Strikes only occurred when there were no other options left. During this period important steps were taken towards the balancing of interests between employers and employees which reduced the need for hostility. The first collective agreements that arose were predominantly one-sided employer determined regulations. Though there have also been instances where for instance diamond cutters were able to agree with jewellers not to work below a certain wage. Also more institutionalized forms of collective agreements had started to materialize, e.g. the establishment of the Permanente Commissie (Permanent Committee) in the Amsterdam printing business. The first collective agreement was realized in 1894 in the Amsterdam construction business. The first collective agreements that had a more national character followed in 1895 (diamond industry) and 1899 (the Rhine trade). In 1910 the number of collective agreements had risen to 46 and by 1918 to 923. The number of organized workers rose from 23,000 to 96,000 in the same period. In 1914 a relatively extensive collective agreement was concluded in the typography industry. This collective agreement consisted of a number of new regulations regarding compulsory membership and industrial jurisdiction. This collective agreement was promoted by employers who wanted to reorganize the branch and strived for the determination of binding price and wage regulations (OSA 1995, 11). 2.1.3 The Road to Legislation of the Collective Agreement (c.1900 - c.1918) The first careful steps towards legal regulation of collective agreements were taken at the beginning of the 20 th century. Afraid that the unions would be too weak to compete against the employers and still unsure whether a collective agreement was 1 It is important to note at this point that while a legal scholar may discern simply one type of collective agreement, which is statutory and can therefore only exist after the 1927 Collective Agreement Act (Wet op de collectieve arbeidsovereenkomst, WCAO), historical examples exist of collective agreements, which are not statutory, but nevertheless are contracts mentioning terms and conditions of employment and which are concluded collectively, rather than individually. In this study, these latter types of collective agreements are regarded as less institutionalized forms in comparison to the statutory collective agreement since the WCAO of 1927. 2 This will be interesting later on in this study as IT professionals, especially in the software and services niche of the Dutch IT industry this study is interested in, are typically high skilled professionals.

The Historical and Legal Context of the Dutch Collective Agreement 21 something desirable, a majority in the Dutch Lower Chamber opted for a moderated statutory regulation. This first Act regulating contracts of employment in the Netherlands, the Employment Contract Act (Wet op de arbeidsovereenkomst) dates from 1907 and contained, amongst other things, regulations regarding the determination of wages, although it said nothing about the level of these wages. Fase (1980, 17-18) mentions that this early Act was too meagre to be called a genuine law governing collective employment contracts. In order to be called a genuine contract Act it should not be possible to legally divert from the collective contract in an individual contract. Moreover, the collective contract should coercively complete those parts of an individual contract that have been omitted (OSA 1995, 11-12). The collective agreement in general met with some criticisms. Collective agreements supposedly endangered international competitiveness by restricting the freedom of employers to implement changes in the terms and conditions of employment. Collective agreements would also drive up the general level of wages, which could hurt international competitiveness. Higher wages on the other hand are only one of the effects collective agreements can have on the terms of employment. Proponents of the collective agreement would stress that these are positive effects and would indirectly benefit employers as well. Besides having effects such as ensuring industrial peace, higher wages could also lead to a general increase in living standards. An additional advantage of contracts is that they specifically mention what is asked of and demanded from both parties, which reduces uncertainty between the employer and the employee (OSA 1995, 12-14). While some employers accepted the argument that collective agreements could offer industrial peace, they were less enchanted by the prospect of the coercive characteristics of collective agreements. Particularly the discussion of the extension of collective agreements (in Dutch algemeen verbindend verklaren, or avv for short), was not welcomed with open arms. The debate on the extension of collective agreements started almost as soon as the Employment Contract Act was introduced. Arguments for extension of collective agreements were predominantly based on the view that collective agreements should hold for the entire branch or sector. Most notably, religious factions argued for an organic society, which is neither a hotchpotch of individuals nor a mechanism, but rather an organism consisting of interlinked parts. They argued that if a majority benefited from certain regulations, why not compel others who work in the same branch or sector to fall under the same regulations (OSA 1995, 14-18). Opponents of extension stressed that forcing rules and regulations upon a minority implied that the minority should give way to the majority. Collective agreements are a product of individual companies, yet they are forced upon all companies in the same branch. While collective agreements come about in relative freedom through the individual actions of companies, this freedom would be taken away from them through the avv. Proponents of the avv dismissed these arguments and ironically referred to them as the freedom to uphold poor terms and conditions of employment. Besides these arguments of principle, economically grounded arguments against the avv were also formulated, which again, as in the case of the

22 Developing Institutions collective agreement, concentrated on the rise of labour costs and the increased level of unemployment caused by the rise in wages (OSA 1995, 14-18). 2.1.4 The Inter Bellum Period (c.1918 - c.1939) Shortly after World War I, when the Netherlands found itself in the midst of a crisis, opposition increased against state influence through an act governing extension of collective agreements. New arguments against the avv were brought into the discussion. Besides the obvious detrimental economic effects the avv would supposedly have in these days of crises, the avv would in principal imply a minimum wage. A minimum wage could have two effects, it was argued. Firstly, it could mean that the general level of wages rose above the economic value of those wages. Secondly, it could mean that some companies would go bankrupt, because they would not be able to afford to pay those wages. The reason for this is that with an avv collective agreements are connected to the wages of those who are organized. With an avv it would not be possible to lower the wages below the minimum set in the avv. An avv should therefore only be made possible in those branches that only operate on a local level and hardly have to deal with international competition. In these branches an avv could even have a positive effect, as it would reduce competition. Opponents of the avv were predominantly employers in the export industries who noted that even these so-called local niches would indirectly effect the exporting business as prices on the local market would rise for everyone and therefore also the costs of the exporting industry. Whilst those operating on local markets can easily compensate for these higher costs (as everyone has to deal with those firms), the internationally oriented businesses cannot. Some argued that industrial peace might come at too high a price with the avv (OSA 1995, 18-21). The post-war turmoil gave rise to new incentives for collective bargaining over terms and conditions of employment. The density of unions increased and boosted closer organizational connections between employers and employees. 3 Between 1918 and 1920 the number of union members had risen considerably. Nor were employers far behind in organizing themselves. Besides the already existing employers confederations trade union confederations for employees were also established. The impulse in collective bargaining also contributed to the discussion of extension of the Collective Agreement Act by enshrining it in public law. During this period (1918-1923) discussions regarding nullification of certain collective agreement regulations and the possibility of withdrawal of an avv first started appearing. In particular, the conditions in some surrounding countries of the Netherlands, where wages had risen considerably in artificially protected branches, were giving rise to concern (OSA 1995, 18-23). 3 This is interesting as apparently the conflicts between employers and employees did not increase despite the poor conditions.

The Historical and Legal Context of the Dutch Collective Agreement 23 In 1926 a bill was introduced for a Collective Agreement Act. Members of the government at the time did not have the avv on the agenda. During the parliamentary discussion several amendments were introduced. One of them refers to article 10 of the bill, which stated that members of an organization can withdraw themselves from the collective agreement if they do so within two weeks after its conclusion. Another amendment refers to articles 12 and 13 of the bill, which gave legal superiority to the regulations of the collective agreement. In the case that no regulations were made, or if they were different, the regulations of the collective agreement would hold. This amendment tried to expand articles 12 and 13 to article 14, which stated that a collective agreement also holds for non-organized employees. This amendment was not accepted because liberal factions criticized article 14 for its links to a possible avv. In 1927, the new Collective Agreement Act (Wet op de collectieve arbeidsovereenkomst, WCAO) was accepted (OSA 1995, 22-23). This 1927 Act did not bind those employers and employees who did not themselves participate in the negotiations. The collective agreement, in this Act, merely sufficed as a company level regulation, not as a sector, regional or national regulation. This had the consequence that collective agreements could change the competitive relation between those firms working within a collective agreement and those working outside it. This was so because a non-committed employer could produce at a lower cost because he was not bound by the collective agreement to pay the minimum wages. Therefore an uncommitted employer could pay his employees lower wages. Because other firms were not bound, nor obliged to be bound, by a collective agreement, they could offer worse terms and conditions of employment than they could have if all firms were bound to the same collective agreement. These examples of these potential effects of non-extended collective agreements only held in a loose labour market though and held more for homogeneous branches rather than heterogeneous ones (Fase 1980, 23-24). 4 It would take another ten years after the introduction of the Collective Agreement Act before the law that allowed collective agreements to be extended, the avv Act (Wet algemeen verbindend/onverbindend verklaren van CAO s, WAVV) was enacted. During these ten years more attention was paid to other matters such as a bill for joint industrial councils, which was enacted in 1933 as the Joint Industrial Councils Act (Bedrijfsradenwet). In 1936 a bill for the avv was submitted. The bill had several goals. Firstly, it attempted to protect the collective agreement and to promote collective bargaining. Secondly, its goal was to decentralize legislation since it was in essence the branch or sector itself that would ask for an avv within the branch or sector. When the avv Act was passed in 1937, it was accepted on the condition that the factual implementation of an avv should be restricted to those situations where the conditions of a collective agreement were applicable to a majority of those who were working in a certain branch, the so-called majority requirement (meerderheidseis) (OSA 1995, 23-24). 4 Branches characterized by a uniform type of firm sharing a similar structure of costs.

24 Developing Institutions At the time, not only the extension of collective agreements was discussed, but also the nullifying of collective agreements. Nullification was considered a means to tackle the abuse of power, a card that could be played in case an avv did not have the desired effect. The general drawback of the possibility of these nullifications was that they would put pressure on collective agreements. The unions were not pleased with the existence of these nullification possibilities, as these made their job more difficult (OSA 1995, 24-27). 2.1.5 Collective Agreements as a Means to an End: Wage Control (c.1940 - c.1970) It is hard to tell whether the Second World War has had a great impact on the development of the process of collective bargaining. A pay freeze and price freeze were introduced in 1940 by the German occupier and in 1942 both the Collective Agreement Act and avv Act were rendered inoperative. 5 After the war both laws were reinstated. A new private organization was founded in 1945, the Labour Foundation (Stichting van de Arbeid, STAR), which comprised representatives of the federations of employers associations and federations of unions. During World War II employers and employees had been secretly making preparations for its founding. The belief that both employers and employees had to work together to rebuild the country after the war had ended, had been an important consideration. For that reason the STAR was soon officially accepted as the advisory body for social-economic matters and functioned as a consultation platform for the top of the organized business community (OSA 1995, 27). The Extraordinary Decree on Labour Relations (Buitengewoon Besluit Arbeidsverhoudingen, BBA) was passed in 1945. The basic principle of the BBA was that terms and conditions of employment would be formed through collective consultation in industries and enterprises. At the same time the BBA gave the Boards of Government Mediators a number of powers which limited the freedom of collective formation of terms and conditions of labour. The Boards of Government Mediators were re-instated boards that already existed before the Second World War. The goal of these boards was to mediate in labour conflicts. Firstly, article 14 of the BBA stated that the Board of Government Mediators must approve new collective agreements to begin with, and then focus on alterations to existing collective agreements. Secondly, article 17 stated that employers cannot deviate from terms and conditions stated in the collective agreement or another binding pay regulation. This binding pay regulation was based on article 12, which stated that the Board was empowered to determine regulations with regard to wages and other terms and conditions of labour. A binding wage regulation had the same status as an extended collective agreement. Finally, article 15 transferred the power of extending collective agreements or declaring them non-binding to the Board. The Board, however, had to follow guidelines set by the Minister of Social Affairs and Health. During the 5 These freezes meant that wages and prices could not be raised or lowered.

The Historical and Legal Context of the Dutch Collective Agreement 25 following years collective agreements were only approved if they had standard features rather than minimal characteristics (OSA 1995, 27-28). The post-war consensus cloaked the underlying differences, creating a situation in which different parties with different agendas shared the same objective of wage moderation. Employers regarded the centrally directed wage policy as a means of reducing production costs, which was necessary to protect their export interests. The liberals did not agree with the lack of wage differentiation and with the uniformity of the pay rounds. In their view the pay policy should be abandoned as soon as the social-economic conditions would allow so. The social democrats though had different reasons to support the centralized pay policy. They pursued the structuring of a planned economy through state influence. Other parties such as unions also agreed with the pay policy, but urged that it should be linked to a centralized price policy (OSA 1995, 28-29). The post-war inspired consensus did not last very long. Within five years after the end of WWII, when the economy was booming, wages rose beyond the agreed wage limits. The Korea crisis of 1950, which led to price increases, did nothing to help the consensus, which dissipated as all parties, except the government, demanded the pay policy to be abandoned. 1950 was also the year in which the Social and Economic Counsil (Sociaal Economische Raad, SER) was founded. The SER comprised representatives of employers associations and unions, as well as independent members (so-called crown members ). The SER had no government members, nor representatives. The SER resided in public law rather than private law. Ministers were obliged to ask the SER for advice when they wanted to make changes to important regulations (such as legislation) in the social or economical field. The SER urged a return to a branch level organization of wage formation (preferably through collective agreements). Due to the economic boom more and more employers tried to avoid the centralized pay policy by paying employees wages above the agreed rate, especially in those instances where qualified personnel was hard to find. Even the prosperity rounds of 1954 and 1956 proved insufficient to stop these so-called black wages. 6 The STAR urged a more decentralized wage policy and recommended that wage formation was a matter for the business community and that the government only had a supplementary role. They also advised that the SER should approve and extend collective agreements in the future. The SER took up this proposal in 1955. The government, who at that time justified its policy by referring to the employment level, was also willing to accept this proposal. At the same time, however, the government urged that the final responsibility regarding the wage policy should rest in its own hands. In 1959 the system of uniform wage rounds was abandoned and replaced by systematic wage raises. In practice this new policy s effect proved hard to measure and the credibility of this policy was soon discredited (OSA 1995, 29-31). 6 The prosperity rounds allowed for small wage raises.

26 Developing Institutions During the 1960s, the discussion regarding the rate of wages was connected to productivity. If productivity was rising, so should wages. In practice, however, it was usually the other way around; first the level of wage rises was decided upon and the rise in productivity was calculated afterwards. The rigorous governmental control on wages imposed serious tensions on an already strained labour market. This made matters quite complex for the government, as they had to manoeuvre between inflation on the one hand, and labour peace on the other hand. With the wage rounds not producing the desired effect of releasing some of the tensions, pressure was still upon the government to pull back its influence on the matter. Therefore a minor emergency Act was introduced, which allowed the STAR to decide collective agreement approvals relatively independently. The STAR had to consult both the government and the SER in order to be able to make a final decision. The idea behind this emergency Act was to allow wage developments to be co-ordinated internally. The effect of the Act was that the government had to intervene at least once a year because wages had risen explosively (about seven times higher than the norm set). These interferences included nullifying some collective agreements and introducing a new wage pause (OSA 1995, 31-34). The government was gradually loosening the reins on its Wage Control Policy and the increasingly independent process of wage formation was eventually formalized in the Wage Base Determination Act of 1969 (Wet Loonvorming, WLV). Negotiating parties were free to determine whether a collective agreement is a minimum collective agreement, or a standard collective agreement. The government did not want to completely deregulate wage policy and retained some degree of control. This was made possible through article 10, which effectively allowed the government to control wages even those that were not determined through collective agreements. The criterion that applied to this article was whether it is in the interest of the national economy to take measures regarding the level of wage costs in connection to the other required measures. Prior to using this measure, advice had to be asked of both the STAR and the SER (OSA 1995, 34-35). 2.1.6 The Retreating Government (c.1970 - c.2004) During the seventies and the beginning of the eighties the government used the pay policy measure nine times. The first occasion was in 1970 when after a strike in the Rotterdam port, wages were raised nationwide. Each time the pay policy measure was used it was backed up by the same argument: the parties within the Labour Foundation were not able to reach an agreement. In 1974 the control options of the government were increased through a temporary Enabling Act (Machtigingswet inkomensvorming en bescherming werkgelegenheid, 1974). This Act was a reaction to the perceived negative affects of the 1973 oil crisis. When the feared consequences of this crisis did not become a reality the Act was not renewed in 1975. The Enabling Act was responsible for several ministerial decrees, which enforced universal wage raises in the Netherlands (OSA 1995, 35-37). During this period government control over wage formation was again being questioned. This was especially the case with article 8 of the Wage Base

The Historical and Legal Context of the Dutch Collective Agreement 27 Determination Act, which gave the government the power to declare collective agreements non-binding. The SER had advised to drop this article because it restricted employers and employees in one of their principle tasks: to freely negotiate and come to an agreement. Another argument was that if wage measures were necessary due to economic conditions, they ought to be applied to all collective agreements. Article 8 was finally dropped in 1976 announcing a very brief period of silence regarding wage base determination, which lasted only until the second oil crisis of 1979-1980. Because of the number of bankruptcies and increasing unemployment rates articles 10 and 11 of the Wage Base Determination Act were altered in such a way that the Minister of Social Affairs had Enabling Act-like powers and was able to independently determine the terms and conditions of labour. This situation was to last until the end of 1982 when the Wassenaar Agreement (Akkoord van Wassenaar) was concluded. This was an agreement between the central employer and employee associations residing in the STAR, which materialized because of combined aversion to pay policy measures. The essence of the agreement was that unions would accept wage restraints in exchange for a shorter working week. This agreement was made in order to improve the profitability of trade and industry and reduce (youth) unemployment at the same time. In two ways, the Wassenaar Agreement was a milestone. Firstly, it ended a period of governmental control through pay policy measures, and secondly, it marked a new period of decentralization of bargaining terms and conditions of employment. The Wage Base Determination Act was also adapted after a while. In 1985 article 10 section 1 of the WLV, which gives the minister some means to control wages, was restricted to emergency situations only. International regulations (such as ILO- Treaty 87, which will be discussed in more detail later) have also played an important part in restricting government control over the wages. To some extent, the avv Act had been responsible for the process of centralization till 1970, as it effectively centralized collective agreements from a company level to a branch and industry level. Tros (2000a, 65) concludes that the Public Industrial Organization Act of 1950 (Wet op de publiekrechtelijke bedrijfsorganisatie, PBO) also played an important role in the centralization process. The PBO Act allowed companies that fulfilled similar or related functions in business to form a branch organization. These organizations were given powers by the authorities to issue sector wide directions regarding wages, training and fringe benefits. These directions were verified by the Ministry for Social Affairs and Employment up until 1992, and by the SER thereafter. 7 A continuing process of decentralization characterizes the nineties. The 1993 STAR agreement A new direction (Een nieuwe koers) is an example of this process. The agreement encouraged the government, the employers and the trade unions to promote diversity and tailor made terms and conditions of labour. The agreement also enhanced the possibility of part time work and the reduction of working hours, 7 This step, which consists of effectively allocating decision powers to a lower level, can be considered as a form of decentralization.

28 Developing Institutions because the unions gave up their veto against the flexible deployment of labour. Finally, it also encouraged union leaders to increase the amount of consultation and democracy within their organizations (OSA 1995, 38). Having been left alone since their introduction in 1927 and 1937 respectively, the Collective Agreements Act and the avv Act became the subject of social and political debate. The 1992 SER advice regarding the Collective Agreement Act and the avv Act, carefully questioned the legitimacy of both laws. From an economic perspective, the SER concluded, collective agreements effectively limit competition. This effect was amplified through the extension of certain collective agreements to a branch or sector. The downside to this was that a lack of competition could lead to an unnecessary rise of wages in the sector, which in the end could hurt the consumer. High prices also tend to negatively affect international competitiveness. Entrepreneurs could also have difficulties in having to pay the possibly relatively high entrance wages, which could indirectly hurt economic growth and levels of employment. According to the SER, collective agreements also have their upsides, which are, in the end, more important than the possible downsides. Collective agreements are ascribed a tendency of generating a stable social climate, which implies that instead of having to focus on competition over the conditions of employment, or having to deal with free-riding employers and unsatisfied striking workers, one can concentrate on matters dealing with the production process. Besides economic pros and cons, the social implications of both laws were also taken into consideration. The SER concluded that both laws have enabled a more dynamic interaction between collective agreements and legislation. Branches could have never have achieved the amount of self-regulation they were capable of developing now. The SER also concluded that the avv has broadened the process of collective consultation. The process of declaring regulations non-binding, in its opinion, can be removed from the avv Act (OSA 1995, 39-42). The 1992 SER advice was important enough for the government to leave both laws intact. It urged the government to withdraw even more and promote selfregulation. In 1993 the government also evaluated the collective agreement and avv and concluded appreciatively that both can stimulate adaptation, i.e. if the adaptive features of collective agreements are sufficiently used. In 1994 a STAR agreement was made regarding the policy of bargaining parties with respect to the avv. The agreement recommended bargaining parties to consider more explicitly what collective agreement-regulations are desired and legitimate for extension. It also recommended a policy regarding potential dispensation for entrepreneurs who might experience difficulties when entering a branch in which an avv is in effect. 8 A final recommendation was to develop a policy on the lowest wage scale. Stocktaking in the first year after the STAR agreement, 1995, shows that in quite a number of collective agreements provisions were made regarding the lowest wage scales, but that these 8 This is something that may also play a role in the IT industry in case collective agreements are extended.

The Historical and Legal Context of the Dutch Collective Agreement 29 were almost never used because the lowest wage scales lay far above the minimum wage (OSA 1995, 42-46). In July 2004, the largest Dutch employers association, the Confederation of Netherlands Industry and Employers (Vereniging van Nederlandse Ondernemingen- Nederlands Christelijk Werkgeversverbond, VNO-NCW), published a memorandum entitled 'The Netherlands must be more active. Work(ing) in the knowledge economy in 2010' (Nederland moet actiever. Werk(en) in de kenniseconomie 2010). At about the same time, the Christian democrat/liberal coalition government published a memorandum entitled 'Opting for growth' (Kiezen voor Groei). Both these documents contain proposals which could have an impact on the Dutch system of industrial relations, and particularly the extension of collective agreements, which should, according to these memoranda, be made more cautiously in the future. In June 2004, Minister De Geus expressed the government's view that collective agreements cannot be extended at the expense of the public interest. In order to tackle the economic crisis, the government had again attempted to moderate wage raises in an attempt to increase productivity. The government would call a halt to extending collective agreements if the current wage moderation ended, or if new collective agreements included provisions about supplementing sickness benefits during an employee's second year of absence. According to the VNO-NCW memorandum, VNO-NCW thought it necessary to review critically which provisions of collective agreements are deemed eligible for extension. Companies should also have sufficient opportunities for exclusion from such extensions, and periodic assessments must be carried out of the whole extension system. However, they agreed that the current mechanism for the binding extension of industry wide collective agreements contributes to maintaining stable industrial relations. 9 2.1.7 Conclusion Looking back at the history of the collective agreement in the Netherlands, we can see it was originally a matter of balancing power between the workers and their employers. The government soon discovered collective agreements could also be used as a tool for controlling economic life and used the collective agreement as an instrument to improve international competitiveness and increase productivity. Increasing international competition through wage moderation however introduce three problems according to Kleinknecht and Naastepad (2002, 624). Firstly, export prices are not solely determined through wages, but rather by cost per unit produced, which in turn depend on wages and labour productivity. Secondly, the authors stress 9 Van het Kaar, R., Employers and government propose industrial relations reform, European industrial relations observatory on-line, http://www.eiro.eurofound.ie/2004/08/feature/nl0408104f.html (accessed December 9, 2004).

30 Developing Institutions that innovation and the quality of product are at least equally important as the level of wages. Finally, wage moderation negatively affects competitiveness as it reduces the focus on innovation. This latter point is why according to Huizinga and Broer (2004) some economists argue that wage moderation reduces labour productivity by allowing marginal firms with low profits and little innovation to remain in existence. Huizinga and Broer (2004, 33) continue that it is probably best not to use wage policy at all as a tool to influence productivity. As a tool against unemployment, however, their analysis indicates that wage moderation can be a very effective and efficient tool. Though the government stresses the importance of innovation and increased productivity in their memorandum Opting for growth (Kiezen voor groei), their tool for reaching these goals, wage moderation, is a poor one in the long run. Obviously the unions complained at the threat by the government of not extending collective agreements anymore, their principal argument being it would most likely hurt those employees that need the extended collective agreement most. In a letter to the Department of Social Affairs, the unions collectively stated that the government was using the potential ban on the extension of collective agreements as an instrument, which is in violation of international treaties emphasizing the freedom of collective negotiations and social security, as well as fundamental principals of justice (FNV, CNV, and MHP 2004, 2). 2.2 The Legal Context of the Dutch Collective Agreement This section will deal with the legal context of institutions that regulate the terms and conditions of employment. Though each and every individual in the Netherlands has the so-called freedom of contract, the Netherlands has a vast array of institutions that directly or indirectly regulate the terms and conditions of employment. These institutions reside in a hierarchical context of rules and regulations. At the top of this hierarchical structure are the international rules, regulations, and treaties which define parameters for what can be regulated and legislated nationally. Some of the international regulations and treaties that are relevant to this discussion will be dealt with in this chapter. The hierarchy does not end here however and continues down to the employment contract, which is the individual contract between the employer and the employee. The employment contract can be based on a number of institutions besides being directly drawn up between the employer and the employee. An employment contract can be drawn up based either on a collective agreement or on a works council agreement. The collective agreement may in its turn stem from an avv which has a higher place in the hierarchy. 2.2.1 International Regulations and Treaties A nation s law system, at least not in Europe, is usually not without international influences. International treaties or resolutions of international organizations can impose guidelines and directives on a nation state with respect to legislation and

The Historical and Legal Context of the Dutch Collective Agreement 31 regulation. This influence can be both direct and indirect. Direct influence stems from international law, which has priority over national law. Indirect influences usually stem from international organizations. The most important international organizations dealing with labour law are the International Labour Organization (ILO), the Council of Europe, and the European Union. The ILO was founded in 1919, not long after the end of World War I. Its goal was to reduce the negative effects of international competition that hindered the improvement of industrial relations. The assumption was that this could only be realized if all countries were governed by the same guidelines. The ILO has played an important role in establishing and protecting the freedom of the employees organizations. ILO Treaty 87, for instance, concerning the freedom of collective bargaining influenced the Dutch governments pay policy. The Treaty imposes limits on the governments restrictions of collective bargaining. Such wage raise restrictions as the Dutch government has made between 1974 and 1976 were only permitted if they were exceptional, only go as far as is necessary, are valid for a reasonable period only, and are accompanied by measures protecting employees living standards. When in 1976 the union FNV filed a complaint against the Dutch government with the ILO regarding the Dutch pay policy, 10 the ILO concluded that the policy of the Dutch government fell within the margins of the treaty and rejected the complaint of the FNV. During the period 1980-1982, when the Dutch government again used pay policy measures, the ILO Committee looked more critically at the measures and even started a direct contact mission to investigate matters in the Netherlands. The Council of Europe was established in 1949 in order to form a stronger union between its members to protect the ideals and principals of their common heritage and to protect and develop economic and social prosperity. The Council of Europe has written the European Social Charter (ESC), which dates from 1961 and like the ILO-treaties contains minimum regulations. The ESC was ratified in the Netherlands in 1980. As a member of the European Community, the Netherlands has also been subject to a number of important treaties. EU-law also affects Dutch national law. In particular, article 81 of the EECtreaty is of interest here as it affects the Dutch Collective Agreement Act and avv Act indirectly through Competition Law (Mededingingswet). This article states that the agreements between firms, decisions of employer coalitions and mutually attuned dealings of firms are forbidden in case these agreements negatively affect competition (through prevention, falsification or limitation). Currently an article (article 16 Competition Law) exists that effectively excludes the collective agreement and extended collective agreements from article 6 Competition Law. Article 16, however, will be deleted in 2003 in order to comply with international regulations and this might affect the legitimacy of collective agreements and, moreover, extended collective agreements (Franssen et al. 2000, 324, 327-328). 10 The Dutch unions will be discussed in greater detail in section 2.3.4.

32 Developing Institutions 2.2.2 Government Regulations and Legislation Government regulations and legislation are the key determinants of the institutional framework. As was also pointed out previously when discussing international legislation and treaties, the government is restricted in its legislative interference and must refrain itself from unnecessary or abusive legislative intervention, but is free to decide how to implement rules and regulations within the internationally set boundaries. The Civil Code, especially book 7, is the foundation of labour law in the Netherlands. Next to the Civil Code a number of other Acts exist that influence industrial relations. Not all of the Acts influencing industrial relations will be mentioned here, only those which are interesting to note with regard to the subject of this study. One of these Acts is the Working Hours Act (Arbeidstijdenwet 1995, ATW), whose primary goal is to guard the safety, health, and well being of employees by determining legislation regarding working hours and rest periods. This Act tries to reach this goal by means of standard regulations regarding the maximum number of working hours and minimum number of rest hours, which hold for all sectors. A second goal of the Working Hours Act is to promote the possibility of combining work and care. Article 4:1 paragraph 1 of the Working Hours Act obliges the employer within reasonable margins to take employees personal situation into account. Finally, the Working Hours Act tries to make the regulations regarding working hours more flexible. In comparison to previous versions of the Working Hours Act, this 1995 Act now focuses at minimum levels of control and protection regarding working hours, instead of laying down a predetermined set of working hours. Another Act influencing industrial relations is the 1968 minimum wage and minimum holiday allowance Act (Wet minimumloon en minimumvakantiebijslag, WMM). The goal of this Act is to ensure that all employees have the right to a minimum income that can be considered acceptable in view of the general level of prosperity. The primary function of the (Dutch) Occupational Health and Safety Act (Arbeidsomstandighedenwet 1998, Arbowet) is to prevent negative effects from labour and to protect the health and safety of workers. The Act holds for both employers and employees. Employers should provide a working conditions policy and employees should be cautious and precise and do all that is possible for their own safety and health. The recent labour and care Act of 2001 (Wet Arbeid en Zorg, WAZ) accommodates employees leave arrangements in order to ease the combination of labour and care. 2.2.3 The Employment Contract The legal basis for the terms and conditions of employment resides in the employment contract. An employment contract is a contract between two persons, which on a national level is regulated by: (1) the law, (2) case-law, and (3) custom, and

The Historical and Legal Context of the Dutch Collective Agreement 33 if manifested on a collective level, by collective agreements. Article 7:610 of the Civil Code defines an employment contract as a contract by which an employee binds himself to an employer to do work for a certain amount of time and wage. The contract implies a number of things. Firstly, it entails that the employer is obligated to pay a wage. Secondly, it means that the employee is obligated to work personally. Third and lastly, it implies a hierarchical structure, where the relation of the employee to the employer is a subservient one. This last implication is important to distinguish the contract of employment from other contracts where labour is offered and the relation between the principal and agent is not a subservient one. If one of these conditions is not fulfilled, the contract is not a contract of employment. In order for both parties to draw up a legally valid employment contract, they have to meet certain conditions, which stem from general contract law (overeenkomstenrecht). These conditions include that both parties have to agree to the agreement and both parties must have legal capacity. This also implies that if a contract is changed, it must be done in agreement of both parties. There are circumstances under which the employer is allowed to alter a contract unilaterally. These circumstances are limited to those instances in which the interests of the employer are more compromised than the employee s interests. Another condition to which employment contracts have to adhere is that they agree to general rules of decency, public order, or an enforcing Act. If a contract does not comply with these conditions, it is void (Article 3:39 of the Civil Code and further). Both parties entering into an employment contract have legal obligations. In general, parties within an agreement have to adhere to general rules of reasonability and fairness. In case of an employment contract these general rules have been specifically mentioned in article 7:611 of the Civil Code, which, for instance, states that employer and employee should behave as a good employer and employee towards another. Employers also have to adhere to rules and regulations regarding discrimination. An employer is not allowed to discriminate on grounds of one s gender (articles 7:646 and 7:647 of the Civil Code) and number of working hours (article 7:648 of the Civil Code). Other grounds on which an employer cannot discriminate can be found in the equal treatment Act (Algemene wet gelijke behandeling, AWGB). These other grounds include: homo- or hetero-sexuality, belief, political affiliation, marital status, nationality, and, since 2004, age. As has already been stated, an employment contract implies that the employer pays the employee a certain wage in return for his or her services. Wage though can comprise a number of things such as money, usage of living space, education, a company car, bonds, et cetera. Under certain conditions, an employer has to pay wages for work not delivered. The most common of these conditions is time off. Employees are granted at least four times the number of days they work per week for holidays or leave per year. There exist a number of special cases under which the employer has to pay wages for work not done and which cannot be classified as holiday or leave. In some instances, as for instance in some collective agreements, conditions are mentioned under which an employee cannot work one can think of weather conditions in the case of tree surgeons.

34 Developing Institutions The level of wages is usually left to the employer and employee, but has to conform with regulations regarding the legal minimum wage. This is the minimum acceptable wage in the light of the general level of wealth in the Netherlands, against which an employee can offer their services. In principle employment contracts can be very basic and are enclosed by a vast array of rules and regulations laid down in various laws that cover most contingencies regarding the relation between employer and employee. These general rules and regulations are constantly being measured against the general interest and beliefs of society. In practice, employment contracts are usually a lot more detailed and go further than simply mentioning the number of hours per week an employee has to work and the wage to which he or she is entitled. The collective agreement itself will be discussed below. The principal relation between the employment contract and the collective agreement is that the terms and conditions of employment have been set by the employers negotiations with the unions and not by direct negotiations with the employee himself. Moreover, the collective agreement on which the employee s terms and conditions of employment are based may stem from an extended collective agreement. This is the case when a collective agreement is extended through an avv. The extension of collective agreements will be discussed after discussing collective agreements. 2.2.4 Collective Agreements Article 1.1 of the 1927 Collective Agreement Act defines a collective agreement as an agreement concluded by one or more employers or associations of employers possessing full legal capacity on the one hand and one or more associations of employees possessing full legal capacity on the other, primarily or exclusively regulating terms and conditions of employment that must be observed in contracts of employment. In principle anyone who is able to represent the employee or employer is allowed to negotiate over the terms of labour in a collective agreement if the negotiators allow for this. It is not possible to force negotiators to admit someone to the negotiations. The traditional parties in the Netherlands are the unions and the employer s organizations. The tasks of these unions include, amongst other things, the conclusion of collective agreements and the representation and defending of the interests of the employees as both a collective and as individuals. The legal basis of unions is rather minimal and very open and resides in the right of association. Practically anyone can start a union, there are almost no special permissions required by a government body, nor does a supervisory body exist. The only thing required is that they have a full legal capacity to represent the employees. The largest union federations in the Netherlands are the Confederation of Dutch Unions (Federatie Nederlandse Vakbeweging, FNV), the Christian National Labour union (Christelijk nationaal Vakverbond, CNV), and the Union MHP, medium and higher personnel (Unie MHP, Middengroepen en Hoger Personeel). These federations themselves never are a negotiating party, but instead coordinate

The Historical and Legal Context of the Dutch Collective Agreement 35 the work of member unions and operate on a national level above regions, branches, and sectors. The focus of union activity lies a level below these federations and is done by the member unions. Employers organizations are organized in a similar fashion to the unions. They have also merged into a few central organizations and operate at different levels. Their primary function is to advise individual organizations on both socioeconomic and legal matters and to coordinate collective agreement negotiations. The largest employers organizations are the Confederation of Netherlands Industry and Employers (Vereniging van Nederlandse Ondernemingen-Nederlands Christelijk Werkgeversverbond, VNO-NCW), the Royal Association MKB-Nederland (MKB), which organizes medium and small-scale businesses, and the Dutch Confederation of Agriculture and Horticulture (LTO Nederland). These three employers organizations work together in the Council of Central Employers organizations (RCO). Another major employers organization working closely together with VNO- NCW is the General employers association VNO-NCW (AWVN). Within companies another workers coalition can exist which is called the works council. Article 2 of the Works Council Act 1998 (Wet op de ondernemingsraden, WOR) stipulates that every company employing more than fifty people is obliged to set up a works council. 11 If an employer owns more than one company, each employing fifty or more people, which share a similar nature, structure, and management, the employer is compelled to also set up a common works council (gemeenschappelijke ondernemingsraad, GOR) if this improves the existing situation (article 3 WOR). If the employer employs more than ten people, but less than fifty, and a voluntary works council has not been assigned, nor was a works council compulsory through a collective agreement, the employer is obliged to initiate a representation of the staff if a majority of the employees so wishes (articles 35c WOR). A staff association also has several rights, but they do not go as far as those of the works council. Members of the works council are democratically elected and from this selection a chairman is chosen. The number of members can vary (3-25), depending on the size of the company. The works council has several rights, but it can only apply these rights to work in the collective interest of the company. This means that a works council can represent employees, but must also take the company s interests into account. The works council has a right to invite experts outside the company to participate in meetings, or give advice. This, however, does not imply that members can give information to just anyone. The employer can force them to keep matters secret from the public. The employer is obliged to provide the works council with sufficient funds in order to be able to operate. This includes time and space for meetings and education. The employer has to have the consent of the works council for alterations to those parts of the individual agreement that have to do with the organization of work 11 The concept of a company is more broadly defined in the WOR. A company is an independent unit operating in an organized context within society, where work is done in the context of an employment agreement or a public appointment. Article 1 of the WOR contains the definitions of relevant concepts.

36 Developing Institutions and wages, and also to most parts of the employment agreement that are not related to the primary conditions (pay and working hours). If matters are worked out in detail in a collective agreement however, the employer does not have to consult the works council (article 27 paragraph 3 WOR). The principle that anyone capable of representing the employee or employer is allowed to negotiate over the terms of labour in a collective agreement could imply that the works council can also be allowed to enter the negotiations as a party. This is a matter of great debate however since works councils are not an association, while it is mandatory for collective agreement negotiation parties to be associations. 12 Traditionally, solely the union and employer associations have negotiated over the terms of labour. It is an interesting question to determine which party, the unions or the works council, is best suited to represent the employees. The collective agreement, by nature, limits the role of the works council (article 27 paragraph 3 WOR). It is an entirely different question whether the works council is capable of negotiating the same terms as the unions can. This is as described above a matter of power. The unions possess a number of pressure tools, e.g. the possibility of strike action. One may also assume the unions to have more expertise and experience. On the other hand one may assume the works council has superior expertise regarding internal organizational matters. A union is only then competent of concluding a collective agreement if the statutes of the organization mention this competence expressly (article 2 WCAO). A collective agreement must also be in writing if it is to be regarded as such (article 3 WCAO) and must be public to all employees falling under its scope (article 4 WCAO). The text of the collective agreement must be submitted to the Minister for Social Affairs and Employment (SZW). In order for the collective agreement to become effective, the Minister (in practice the Labour Inspectorate to which the Minster has delegated this task) must notify the parties involved that he has received the text. The collective agreement becomes effective (except when it is being backdated) on the day this notification is dated (article 4 Wage Formation Act). Collective agreements are usually concluded for a period of one to two years, but may legally be concluded for a maximum of five years (article 18 WCAO). Two types of collective agreements can be distinguished in the Netherlands: the company collective agreement and the sector or branch collective agreement, the major difference between them being their scope. This is obviously not the only difference as matters need to be regulated differently at a sector or branch level than they are at a company level. The advantage of company level collective agreements is 12 In the Dutch IT industry cases are known (e.g. in the case of the first Origin collective agreement) where the works council has been a party to the collective agreement negotiations. This is the reason why works councils are added to the negation parties and negotiations section of collective agreements. See also Bijlsma & Van het Kaar (1999, 170-171), Van Liempt & Van Uffelen (2000), Van Liempt (2000), and Kamerstukken I 2000-2001, no. 9. The works council itself will be discussed in greater detail when the works council agreement is treated.

The Historical and Legal Context of the Dutch Collective Agreement 37 that they are potentially more suited to an individual company than are larger scale collective agreements. In order to continue the discussion of the collective agreement, figure 2.1 is presented, which shows the various possible clauses in relation to the parties involved in the negotiation process of collective agreements. I will first start by discussing the horizontal clauses, which are the regulations regarding the terms and conditions of employment and can also be called normative clauses of a collective agreement. There are two other types of collective regulations, the diagonal clauses of a collective agreement and the obligatory regulations of a collective agreement. Table 2.1 Clauses and commitments of parties concluding collective agreements Employers organization Obligatory clauses Union Membership Diagonal clauses Union membership Employers Horizontal clauses Employees Source: Fase (1980, 521). The horizontal clauses of a collective agreement have a normative effect and therein lies the most important legal effect of the collective agreement. The normative effect implies that the collective agreements carry over to the contract of employment. Two conditions need to be met before this carry-over effect can take place. Firstly, both employer and employee must be members of one of the parties that have concluded the collective agreement (articles 9 and 12 WCAO). Employees that are not bound in terms of articles 9 and 12 are covered through article 14 of the WCAO. Article 14 was introduced to do away with employees without such a membership being cheaper for the employer. Article 14 forces the employer to also offer the terms and conditions of the collective agreement to employees without a membership. The employee without a membership on the other hand cannot derive rights from the collective agreement, he can only accept (which in practice usually is done silently and automatically) the offer or refuse it in which case a different contract needs to be drawn up. Employees who are members though, do not have this option. A second condition in order for the carry-over effect to take place is the condition that the contents of the collective agreement must be applicable to the

38 Developing Institutions activities of the employer and employee. This means that the higher staff or some part-timers are not always covered by a collective agreement (article 9 WCAO). If these two conditions are met, the collective agreement has a compulsory substitutive effect (article 12 WCAO). The collective agreement also has a compulsory complimentary effect, which means that if terms and conditions found in the collective agreement are not found in the individual contract, they are automatically applied to the individual contract (article 13 WCAO). Schutte (1995, 6-7, 53) remarks that collective agreements cannot completely regulate the terms of labour for a number of reasons. Firstly, the content of collective agreements is inherently limited and a large part of this is due to the process of establishing a collective agreement. Negotiating parties are free to choose each other and the content of collective agreements is determined through a negotiation process. The power of the negotiation partners, which to a large extent is determined by socio-economic conditions, is thus an important factor for the outcome of these negotiations. Another fact of the negotiation process is that outcomes are usually compromises - less than ideal outcomes for either party. Secondly, the reach of collective agreements is limited. At most collective agreements are made to cover a certain type of industry. The collective agreement is by nature limited in its socioeconomic sphere, not everyone is covered by a collective agreement. Thirdly, the duration of collective agreements is limited. If the period of a collective agreement ends (or the collective agreement is cancelled), the terms and conditions of the collective agreement remain in effect until a new individual or collective agreement is concluded that changes the terms and conditions of the previous one. This, however only holds for the horizontal clauses of a collective agreement and not for the obligatory and diagonal clauses discussed further on in this section. Schutte also mentions a fourth limitation of the content of collective agreements, i.e. the legislative claim of the collective agreement. Non-organized employees are submitted to the collective agreement because of article 14 of the WCAO. This implies that employees have to accept a set of labour terms, which they do not necessarily support. Although this compromises the content and support of collective agreements to some extent, it is not a very strong limitation and does not quite match the other three limitations. Various reasons can be found for non-organized employees, e.g. it may be due to free rider behaviour or other reasons (Schutte 1995, 48-50). Next to the horizontal clauses, which affect the individual obligations between employer and employee, are those collective clauses that regulate collective relations. Two types can be distinguished: obligatory and diagonal clauses. Obligatory clauses regulate the mutual legal relationship between the negotiating parties and determine their rights and duties regarding the collective agreement and the negotiation process. Examples of obligatory clauses can be a peace effort clause where all parties agree to keep the peace during the term of the collective agreement. Diagonal clauses determine the rights and duties between the members of the employers or employees organizations with the other party or with a third party (e.g. a fund or a works council). Usually these clauses are duties for the members of

The Historical and Legal Context of the Dutch Collective Agreement 39 employers organizations for the benefit of employees organizations that are a collective agreement-party. Sometimes diagonal and obligatory clauses can overlap, as for instance can be the case with a peace obligation clause in a sector or branch type collective agreement (individual members of employers organizations may also be so bound). In a company collective agreement such a clause would be only obligatory if the employer accepts this obligation. Besides the horizontal clauses, which have the special characteristic of being collective agreements that affect individual contracts, the collective agreement as a contract has several other characteristics. The collective agreement is one of the few contract types that allows employment contracts to divert from specific (three-quarter mandatory) legal provisions. 13 Article 7:629 paragraph 1 Civil Code, for instance, states that an employee has, in principal, the right to continued payment of wages in the event of sickness. A three-quarter mandatory legal provision states that this holds true unless defined differently in the collective agreement. Another important characteristic of the collective agreement is that the Minister for Social Affairs and Employment can use article 2 paragraph 1 of the WAVV to extend regulations of a collective agreement to a whole sector or branch, provided that these regulations already hold for a majority of the branch or sector. In the next section the extension of regulations of a collective agreement is looked at more closely. 2.2.5 Extension of a Collective Agreement Through the WAVV The original goal of the 1937 WAVV was to protect and stimulate collective bargaining on a sector level. Extension of terms and conditions of employment can reduce the level of competition over terms and conditions of employment. The primary effect of extension is that industrial relations are regulated on a sector level. Extension of collective agreements can also contribute to the collective goods of a sector such as training and education. In game theoretic terms extension of collective agreements could bring Pareto optimal outcomes closer. 14 This might be so because the number of parties that can deviate from the central agreement (by underbidding for example) is reduced to zero, which improves the chances of making arrangements that do not have to take potential free riders or underbidders into consideration. Of course making agreements on such a scale makes an agreement rigorous, something that may not be desired by one or more parties. In order to solve this potential 13 Deviation from this type of legal provisions is also possible through an agreement on behalf of or by a qualified governing body. 14 Pareto optimal outcomes are outcomes by which no one can be made better off without making someone else worse off. Game theory studies how decisions of one party influence the welfare of others and vice versa.

40 Developing Institutions problem they are often formulated as framework agreements that function as a guideline. Only the horizontal and diagonal clauses of a collective agreement can be extended and not the obligatory clauses for these clauses regulate the mutual legal relationship between the negotiating parties (articles 2 and 3 WAVV). In order for a collective agreement to be extended, certain conditions must be fulfilled. Firstly, a request for extension must be made by one or more collective agreement-parties. Interested parties can bring forward objections against the request (article 4 WAVV). These objections are first shown to the collective agreement-parties who then have to comment on the objections. The STAR then reviews the comments and objections. In the case that the scope of a proposed extension overlaps with the scope of an existing extended collective agreement, it is automatically rejected. This urges the parties to gear the scope of their collective agreements to one another. Secondly, only those regulations that are already part of a branch or sector collective agreement can be extended. Company agreement regulations are never extended (article 2 WAVV). Thirdly, the collective agreement regulations under consideration for extension must already hold for a majority of the people operating in a branch or sector. 15 If these conditions are met the Minister can decide to extend the collective agreement, but he is not obliged to do so. On January the 1st 1999 a new framework of review for the extension of collective agreements (Toetsingskader) came into force. This framework has a twofold goal. First and foremost, it intends to give insight into the procedures and criteria of the framework by which requests for the extension of regulations of a collective agreement are reviewed. This includes the different mutual responsibilities of the collective agreement-parties and the Minister of Social Affairs and Employment. Further, the framework of review also tries to clarify both the criteria by which objections of third parties against the extension of collective agreements are judged and the objection procedure (Van Drongelen & Van Rijs 2002, 125). An extended collective agreement cannot be backdated as is the case with collective agreements and its maximum duration is two years (or shorter in case the collective agreement expires before that period). It is possible for companies to ask for a dispensation to an extended collective agreement. This is mainly the case when they object to the extension of the collective agreement, because they have a company agreement or fall under the range of a sub sector agreement. 2.2.6 Works Council Agreements A works council agreement can be defined as an agreement between an entrepreneur and a works council. Only recently has the works council agreements found a legal 15 This majority includes article 14 employees.

The Historical and Legal Context of the Dutch Collective Agreement 41 basis in article 32 paragraph 2 of the WOR. 16 Works councils are a generally accepted phenomenon in the Netherlands. The legislator supports this view, as evidenced by the additional tasks allocated to works councils by many new laws over the years - a process which is by no means complete. Works council agreements fit into the strategy of the Dutch government to decentralize industrial relations and the Works Council agreement is regarded as a useful instrument in providing tailor made agreements. Works council agreements can contain agreements regarding terms and conditions of employment, but more often contain procedural agreements regarding facilities and work. There is no set duration for a works council agreement and the agreements must be written down. Agreements made do not have the same normative effects as is the case with horizontal clauses in the collective agreement. Also when a collective agreement exists and the works council agreement regulates matters which have also been regulated within the collective agreement, the collective agreement has priority. The works council agreement does not have a carry-over effect, which means that once a works council agreement is concluded, it will not automatically hold for all employees. An agreement will have to be concluded with each individual employee in order to have the works council agreement applied to him or her. Important in this context is article 7:613 of the Civil Code, which stipulates the unilateral alteration stipulation. This article was introduced to protect the rights of the employee with respect to unilateral alterations by the employer, which the employer now can only make if he can prove that his interests are so high that they supersede those of the employee, who shall have to give way within boundaries of reasonableness and fairness (Van der Heijden & Van der Hulst 1995). 2.2.7 Conclusion Figure 2.2 has been created in order to put the Dutch legal structure into perspective. International rules determine, to some extent, how national rules can be implemented. The national regulations and legislation contain three specific Acts, which regulate how the works council agreement, collective agreement, and individual employment contract, come about. The shaded area shows which actors are involved in what type of negotiation. Every employee in the Netherlands has an individual employment contract, which may or may not be based on a works council agreement or collective agreement. 16 Works council agreements were already commonly used before 1998. See Van der Heijden, P.F. and A.C.B.W. Doup (1991).

42 Table 2.2 Developing Institutions The legal structure of the various terms and conditions of employment available in the Netherlands International Regulations and Treaties National Regulations and Legislation Works Council Act WCAO* WAVV** Negotiations Works Council(s) Union(s) Employee employer(s association) Works Council Agreement (Industry wide) Collective Agreement Employment Contract * Collective Agreement Act (Wet op de collectieve arbeidsovereenkomst, WCAO) ** Collective Agreement Extension Act (Wet algemeen verbindend/onverbindend verklaren van CAO s, WAVV)

3 History and Definition of the Dutch IT Industry This chapter has two goals: to sketch the history of the Dutch IT industry and to shed some light on the definition of the Dutch IT industry. Drafting a history of the Dutch IT industry is not an easy task as very little has been written about the history of the Dutch IT industry. The only useful source available that deals directly with the Netherlands is Van Oost et al. (1998). This publication, however, only deals with the industry s history until 1960. So in order to present a more up to date history other sources in the form of expert interviews and annual reports have been used. In a way this chapter can be considered a first step towards an overview of the history of the Dutch IT industry. Defining the IT industry is quite complicated as one of the major characteristics of the IT industry is its ability to integrate and converge with other sectors. In order to get a better understanding of this process of integration and convergence this chapter will start by shedding some light on the history and development of the (Dutch) IT industry. The historical account will be divided into three parts. The first part discusses the initial and increasing need for information technology. The second part focuses on the Dutch IT industry s most recent successful years (1996-2000) and the following period of depression (2001-2003). The final part of this historical account identifies three types of lineages in relation to industrial relations in the Dutch IT industry. Throughout its history the IT industry has branched out significantly and shedding light on these various branches is an important step forward in understanding the industry. In order to be able to create a workable definition of the IT industry, the various branches are linked to the 1993 Standard Industrial Classification system. Although this system is not ideal, it allows for a statistical description of the Dutch IT industry that is meaningful to this study. The statistical description that can be found at the end of this chapter does not include statistics about employee characteristics. These are dealt with in chapter five, where they are compared to the results of the Internet survey amongst Dutch IT professionals.

44 Developing Institutions 3.1 The development of the Dutch IT industry The IT industry is and has been expanding rapidly over the last fifty years. Today round about 80 to 90 percent of our business activities is influenced by information technology, which says a lot about the impact of Information Technology on our society. In this light it is interesting to turn to the roots of Information Technology and see how it developed into an industry that has revolutionized our way of life and the ways in which people communicate just as the printing press has. 3.1.1 The Roots of Information Technology At the root of the IT industry, or rather information technology, lie three different activities. According to Van Oost et al. (1998, 11-15) these activities comprise administrative data processing, technical and scientific computing, and regulation of production processes. Though these three activities have totally different applications, they share a common need that leads them to use the same information technology. The single most important reason for this demand was the societal and economic changes that took place at the beginning of the 20 th century. These changes led to the rise of large scale industries, increasingly complex technical systems and a strong growth of the service sector. In order to facilitate this rise in scale and complexity, calculators were introduced and, later on as demands rose, larger machines. This process of growth and demand for facilitating technology eventually led in 1946 to what some consider the basis for the modern computer, the ENIAC. The ENIAC (Electronic Numerical Integrator and Computer) was designed and built during the 2 nd World War in the United States and was the first and biggest fully automated calculator of its time, able to perform calculations at unprecedented speeds. During the 1950s machines like the ENIAC gained commercial interest and slowly started to be embedded into the three realms of activities mentioned above. The practical usability of these early computers progressed as they became easier to program due to the usage of memory and other innovations. Previously programming tasks into the computer was a time consuming and difficult process that included the usage of punch cards. In 1964 about 200 computers were installed and operational in the Netherlands of which about half were used for administrative tasks (Van Oost et al. 1998, 137). A characteristic of the development of computers, one which still holds today, is that computers became increasingly cheaper, smaller and were able to perform more tasks in a shorter time span. In those days however, computers were still too expensive for medium and smaller sized companies, which made purchasing a computer only viable for a few large organizations. Besides becoming faster, computers also became more versatile. IBM (International Business Machines), the market leader at the time, promised a

History and Definition of the Dutch IT Industry 45 revolutionary computer that would be capable of performing multiple tasks at the same time and separated system programming from application programming. 1 From this point on, the growth and usage of computers almost doubled every two years. In 1956 four computers were available in the Netherlands, 15 years later, in 1971, this number had increased to 1,544. Compared to the United States of America the Netherlands remained a relatively small player with regard to the number of available computers per million persons employed. In the Netherlands 12 computers per million persons employed were available in 1960. In 1961 this number almost doubled to 20. In the United States of America the number of available computers per million persons employed was significantly higher; 75 in 1960 and 116 in 1961 (Van Oost et al. 1998, 139-140). The cost of buying or renting a computer could be up to several millions of euros, which were astronomical figures at that time. Though computers promised a lot of things, they still needed proper housing and still required people to maintain, operate and program them. As has already been said, financially, computers were not a viable option for smaller businesses. The high costs of computers gave rise to service centres. These service centres were often exploited by computer suppliers and offered administrative and computational functions for smaller businesses that could not afford a computer of their own. In 1960 five of these centres existed in the Netherlands and their number increased to 46 in 1967. These service centres proved to be one of the first expansions of the IT industry (Van Oost et al. 1998, 140). In the 1960s, the IT industry was dominated by a handful of large companies which supplied computers. Companies such as IBM, Burroughs, Control Data, Honeywell, Sperry-Rand (previously known as Remington Rand and today known as Unisys), 2 offered a comprehensive solution for their customers. They researched and developed the technology, produced the hardware and the software, and provided all the services required with respect to their systems. These companies had establishments in various countries that sold their computers on local markets. IBM played an important role in establishing this full service model of selling information technology. The general notion of computers before the 1970s was that you could give it some input and it would provide output. A computer was thus perceived as a complete package and people at that time assumed that no one was ever going to make any money by solely developing software or providing services. This belief was mostly due to the success of IBM s business model. IBM was from the moment it started in the IT industry the dominant player. The reason for IBM s early 1 The difference between system programming and application programming is that the first type is software that is required to operate the system, whilst the latter is software targeted to run on a system. In this sense system software facilitates application software. This text, for instance, is written with application software that runs on operating system software. The introduction of system software allowed computers to run various programs rather than just one. 2 In a press release, which was sent out during 2001 as part of marketing campaign, Unisys stated that after 50 years they (Unisys) officially apologized for Creating Unintended Consequences of the Computer Age. One of their most important apologies was for Making it impossible for anyone to do more than five minutes' worth of work without being interrupted by an emailed joke, Top Ten lists, or chain letter. (http://www.unisysworld.com).

46 Developing Institutions dominance can be attributed to it already possessing the required expertise and competencies when it entered the market of selling computers (Norberg, 2002). IBM started out as the Tabulating Machine Company in 1896 and already had earned the trust of companies a long time before IBM started selling their first computers. IBM had always provided full service contracts to their clients and if clients bought a computer from IBM they would get a full service bundle, which included hardware, software and services. This business model dominated the IT industry until at least 1969 when antitrust legislation forced IBM to unbundle its software, services and hardware. Though other firms had attempted to sell software prior to IBM s unbundling process, Mahony (2002) suggests that IBM s case has been of singular symbolic importance for the future development of the sector into three separate areas: that of hardware, software, and services. 3.1.2 The Development of the IT Industry Between 1970 and 1990 The 1970s and early 1980s can be characterized as a period in which regular companies using information technology, rather than selling it, centralized their IT infrastructure. 3 Centralization was a natural tendency within companies because the implementation of IT was built around a mainframe that served the entire enterprise. In most cases this led to IT departments becoming separate departments within organizations. Those involved in IT within firms were usually out of touch with the rest of the firm who generally saw them as difficult people to communicate with and extraneous to the organization. Things, however, were about to change as a fourth generation of computers was introduced and information technology spread more freely throughout the company. In the history of computing, generally four generations of computers are discerned. The first generation of computers was produced between 1945 and 1956 and was typified by operating instructions that were made-to-order for the specific task for which the computer was to be used. The use of vacuum tubes for processing and magnetic drums for data storage was also a distinctive feature of this generation of computers. The second generation of computers operated between 1956 and 1963 and was characterized by the use of transistors and improved magnetic storage devices. This allowed for much smaller computers than the use of vacuum tubes ever allowed for. The third generation of computers was produced between 1964 and 1971 and distinguished itself by the use of integrated circuits. The integrated circuit combined electronic components onto a small silicon disc, which was made from quartz. Scientists later managed to fit ever more components on a single chip. As a result, computers became ever smaller as more components were squeezed onto the chip. An important effect of the introduction of integrated circuits was the so-called mainframe crisis. The mainframe crisis was a direct result of the introduction of the 3 CIO magazine (15 September 2001), Next stop: centralization, http://www.cio.com/archive/091501/centralization.html (accessed February 9, 2005).

History and Definition of the Dutch IT Industry 47 minicomputer most notably the IBM system/360. The first minicomputers were small, special-purpose machines that could work within the existing (IBM dominated) mainframe environment. By the early 1970s minicomputers had become inexpensive, full-function, general-purpose computing systems that enabled multiple users to interact with them. Minicomputer manufacturers such as Data General, Hewlett- Packard, Prime and Wang saw their sales grow by a factor ten during the 1970s, while IBM continued to focus on the mainframe market. In spite of its market dominance and close ties with its customers, IBM was late to recognize that smaller, interactive machines that could empower departments and satisfy the computing needs of workgroups and smaller businesses would be very much in demand. Because of its mainframe orientation IBM was unable to dominate the minicomputer market. It was slow to adapt to the shift in demand, and the industry did not wait.. 4 The minicomputers were a product of the third generation of computers. The fourth generation of computers is the generation of computers that is still produced today. These fourth generation computers were able to integrate millions of transistors onto a small area the size of a coin. Besides size, fourth generation computers distinguished themselves from earlier generations by also being a lot cheaper, more reliable, and more generally applicable, that is, computers could be used for more than one function. Computers therefore naturally spread throughout the enterprise as office computers. In combination with the introduction of distributed computing this has led to a decentralization process of the IT-departments within enterprises. 5 Besides decentralizing IT departments in enterprises, the introduction of office computers and minicomputers also allowed the branch of computer suppliers to expand. Most of the newer computer suppliers came from already established electronics firms such as Siemens, Philips, Hewlett Packard, NEC and CII Honeywell-Bull, 6 and the already mentioned Wang and Burroughs, but also newer companies appeared such as MAI Systems Corporation and Nixdorf. Also the way in which computers were sold and implemented changed. Computers were sold more and more by computer dealers rather than the manufacturers of computers and so a new branch was added to the rapidly expanding IT industry. Because of the introduction of system software, more diverse software could be developed and this allowed the software and services segment of the IT industry to develop and expand. Operating systems made computers much more flexible and diverse in usage. A lot of the larger well known software and service providers we know today in the Netherlands such as CMG, Cap Gemini, Ernst & Young, 4 Quote from http://faculty-gsb.stanford.edu/mendelson/computer_history/mini.htm (accessed February 7, 2005). 5 Searchwebservices.com defines distributed computing as: any computing that involves multiple computers remote from each other that each have a role in a computation problem or information processing. In business enterprises, distributed computing generally has meant putting various steps in business processes at the most efficient places in a network of computers.. 6 This was a global joint venture of the American company Honeywell Inc. and the French Compagnie des Machines Bull. Bull and Honeywell established a joined venture in 1975, but the joint venture was disbanded in 1991 when Honeywell Inc. left the computer market.

48 Developing Institutions Getronics, Baan, SAP, Oracle, originated from this period. At first these software and service providers created custom made applications, but after a while they started producing more general applications that could be bought by acquiring a license. Wide scale use of software licensing led to the expansion of the IT industry with independent software vendors (ISV). ISVs make and sell software products that run on one or more computer hardware or operating system platforms. The companies that make these platforms, like IBM, Hewlett-Packard, Apple, usually encourage and lend support to these ISVs because, in general, the more applications that are available on a platform, the more value the platform offers to customers. Some ISVs focus on a particular operating system; other ISVs specialize in a particular application area, such as engineering, and develop software primarily for workstation platforms. Contrary to the hardware manufacturers who were not doing so well at the start of the 1980s, software and service providers were doing relatively well. Service providers had known a few years of extreme growth where their annual turnover grew by 15 percent. The actual figures were closer to a growth of 25 percent per year, but this figure included the growth of labour costs and the additional hardware that was sold with the total service package. Not all companies in this segment were achieving these growth percentages. These figures were mostly achieved by larger American, capital-intensive service providers that could develop specialized services such as financial planning and forecasting. Also information was in high demand at the time and increased investments were made in databases and information distribution. This also helped increase the annual turnover of the service branch as a whole (Automatiseringsjaarboek 1982, 8-9). Body-shopping, the business of hiring out specialized and trained information technology personnel on an hourly or project basis, was one of the main income sources for software houses. Specialization in this segment also appeared to be the key to success as demand rose for specialized software that concentrated on expert knowledge. The focus until that time had been general automation which required more generally trained automation personnel. The demand for specialized software and personnel specialized in another field besides automation was increasing. Companies demanded that the focus should be on expert knowledge in the field of application rather than the field of automation (Automatiseringsjaarboek 1982, 9). On the hardware front expansions of branches could also be noticed. Examples of this are original equipment manufacturers (OEM) and value added resellers (VAR). An OEM is a company that builds components, which are used in systems sold by a VAR. The practice of a VAR selling products with components from OEMs is common in the IT industry. An OEM will typically build to order based on designs of the VAR. IBM is an example of a supplier to the OEM market and IBM is also an OEM itself since it uses other company's parts in some of its products. Generally, an OEM company differentiates itself from the company it buys parts from by adding features or using different selling concepts. VAR companies combine existing products into a unique product, usually in the form of a specific application for the

History and Definition of the Dutch IT Industry 49 product (for example, a special computer application), and resell it. 7 The difference between the two is that OEMs sell a general product, whilst VARs adapt these general products for a specific task or target group, hence the added value. Combined, the software and hardware branches allowed computers to be implemented more flexibly. The computer was regarded more and more as a universal product that could be given specific functions through the use of software by the user or supplier. Another important effect of this specialization of functions was that computers no longer needed to be operated solely by technically oriented people, but could also be used by a larger less specialized group of people. It was too expensive to replace the hundreds of thousands of type writers that were still in use, but the user base of computers was expanding (Automatiseringsjaarboek 1981, a.8- a.10). The micro computer, commonly referred to as personal computer or PC today, was becoming increasingly popular in the mid-1980s and several million had already been sold on a global scale at the time. The microcomputer had a twofold user base. On the one hand microcomputers were used by private users for word processing, database maintenance and spreadsheets, but also for recreational purposes. On the other hand microcomputers were becoming affordable for some small businesses as well. This became possible because of advances in system software, which allowed multiple users to access various applications on one computer. The initiative to implement microcomputers was not always taken by the automation departments within firms. Various business units at the time bought microcomputers without the knowledge of the automation departments because these departments were too slow and usually wanted to use minicomputers, which were relatively expensive and required more maintenance. These business units therefore bought microcomputers to become more independent of these centralized automation departments. At the time it was expected that microcomputers, because of their speed and capacity would replace minicomputers altogether. In 1983 20,000 microcomputers were sold for business use in the Netherlands. Earlier forecasts for 1984 estimated sales of 30,000 microcomputers, but the actual figure was 45,000. 8 There was a lot of competition in this segment. Besides the fact that there were well over 300 different systems available, new systems were being added continuously. This had the effect that, though sales were high, profits were relatively meagre because of low profit margins (Automatiseringsjaarboek 1984, xiii- xiv; Automatiseringsjaarboek 1985, xi). The use of networks became popular during the 1970s and 1980s. One type of network was the smaller scale local area network. This type of network could be used within firms to connect the various computers that operated in it and needed to communicate with each other. Wide area networks were another type introduced in this period. These networks could be used over longer distances and were not 7 The definitions of independent software vendors, original equipment manufacturer, and value added reseller stem from http://whatis.techtarget.com (accessed February 8, 2005) and http://en.wikipedia.org/wiki/original_equipment_manufacturer (accessed February 8, 2005). 8 Though the estimated microcomputer sales figure for 1985 was set on 75,000, the actual figure would be 60,000, which was considerably lower than the previously estimated figure (Automatiseringsjaarboek 1986, x).

50 Developing Institutions restricted to a single building. A classic example of a wide area network is Arpanet, the military network that served as the basis for the most popular wide area network known today, the Internet. The mid-1980s were characterized by a shortage of qualified employees and this could negatively affect the realization of various IT related projects, because there were simply not enough people with sufficient knowledge and expertise. Large IT projects had become so complex in terms of both application and technique that only a few people were qualified to design and implement such systems. This was not a problem that could easily be solved by training people. After training, they would still lack the required experience for the tasks they had to do (Automatiseringsjaarboek 1986, x; Automatiseringsjaarboek 1987, xix) At the end of the 1980s the first signs of convergence could be found between consultancy and software and system houses. 9 These consultants were almost all associated with large global accountancy firms and competed directly with the software houses who themselves set up consultancy branches as an effect of these competitive pressures. This partial convergence of consultancy and implementation had the effect that producers of hardware and software were forced to function more and more as suppliers only. The consultant who made the final decisions decided what software was to be used, thereby decreasing the direct influence software suppliers had on potential users (Automatiseringsjaarboek 1989, xxiii). 3.1.3 Specialized IT Training and Education During the 1960s - 1980s The shortage of skilled employees has been a major concern for the sector throughout its entire existence. In order to create skilled personnel, specialized training and education facilities were required. In 1969 a study was done in the Netherlands to determine whether the Netherlands was capable of producing enough IT professionals. 10 The study s major conclusion was that this was not the case and that action towards specialized formal educational facilities for IT professionals should be instituted as soon as possible. The shortage was responsible for the lack of development of computer applications in the Netherlands. The study produced some figures that were based on the number of computers available in the Netherlands. At the time four to five computer specialists were required per computer. By 1968 1,000 computers were operational in the Netherlands. This implied that roughly between four and five thousand people were required to operate these machines. Approximations for the year 1975 varied between 1,700 and 6,000 computers, which meant that the demand for computer specialist would increase significantly. Estimates at the time indicated that roughly 21,000 to 60,000 people were required in 1975. In 9 The word in use at the time in the Netherlands for consultancy was advice. 10 This study was titled Opleiding informatica, (computer science education) and performed by the committee Opleiding van deskundigen voor Automatische Informatieverwerking (Van Dael, 2001, 80-82).

History and Definition of the Dutch IT Industry 51 order to be able to meet these figures, 3,500 to 10,000 people should be trained annually. According to Van Dael (2001, 79-93) two types of formal education came about in the Netherlands during the 1970s and 1980s. The first type of formal specialized education and training aimed at the IT industry was that of the nonacademic higher vocational education. As had been the case during the origin of the IT industry, different fields required similar tools for different objectives. Administrators needed computers in order to automate their administrative tasks. Engineers and mathematicians on the other hand, required computers for complex calculations. This division had not changed during the course of history and analytically it would not be likely to change in the future since the tasks demanded different fields of expertise. Therefore two formal higher vocational studies were introduced. In 1971 two schools for higher computer science (Hogere Informatica Opleiding, HIO) were instituted that satisfied the demand for mathematical computer scientists in the Netherlands. In 1973 a school for administrative computing specialists (Hoger Economisch Administratief Onderwijs Bedrijfsinformatica, HEAO-BI) followed. Though the latter type of higher vocational education for administrative computing specialists had not been planned, proponents of this type of education, such as Philips, IBM, and Digital, persuaded the Minister of Education to do so anyway. The second type of formal education that was established in the Netherlands was the academic version of computer science education. The education of computer science on academic level in the Netherlands came about much later (1981) than the higher vocational educations. The reason for this was that mathematical academic studies had already begun educating mathematical computer specialists. In 1974 a structural plan (Structuurplan 1974) for academic computer science was finalized. In the plan both the mathematical computing specialists and the administrative computing specialists were regarded as members of the computer science profession and consequently should be educated in this field. The big question, however, was whether computer science should be taught by the mathematical faculties. As this would most likely not be the case, it explains why some of these faculties fiercely opposed the plan. To them, people who applied mathematical computer science were not automatically computing specialists. The mathematical computing specialist saw data processing as the core of informatics. To prove the reliability of a data processing system, knowledge of mathematics at the academic level was required. The administrative computing specialist was specialized in the social complexity of designing and implementing information systems that were used to manage economic, administrative and organizational problems. Due to the opposition of the mathematics faculties, a new plan was made in which a mathematical definition of computer science was used. The reason for this mathematical definition was that it guaranteed that computer science could be considered as a genuine academic study, something which apparently was not the case with the administrative definition of computer science, which was thought of as being too vague. The consequence of this pure mathematical definition was that on an

52 Developing Institutions academic level only the mathematical computing specialist was demarcated (Van Dael 2001, 93-110). The formal types of education that materialized in this period did not satisfy the total demand for education. It was therefore not surprising to see the field of informal types of computer science education and private training institutions expand rapidly during this period. The foundation of informal training and education during the 1980s was quite large and even included hardware manufacturers. These hardware manufacturers saw their sales rates dropping and needed to find additional sources of income. Hardware manufacturers therefore re-invented the notion of full service. The difference with the 1960s, when full service was a common thing and was included in the price of the whole system, was that in the 1980s every additional service would cost more money. At this point it should be noted that informal education mostly took place inside the company. A good example of this practice is Roccade, an IT service company which will be discussed in greater detail in chapter five. 3.1.4 The IT Industry During the 1990s The 1970s and 1980s were periods in which the IT industry expanded rapidly and acquired the basic structure as we know it today. This period is often referred to as the golden age of the computing industry. The period of rapid expansion and the virtual absence of competition would come to an end when the 1990s started. Far too many players had entered the market and consolidation was required to make the sector healthy again. The crisis that hit the sector at the beginning of the 1990s was not its first one. The first crisis in the IT industry had occurred at the introduction of minicomputers in the late 1960s and early 1970s and has already been discussed as the mainframe crisis. A second crisis, similar in nature, had taken place in the early 1980s when personal computers (or microcomputers as they were also called) were introduced and minicomputers were at a loss. This was so because almost every fair sized firm that could afford a mainframe or minicomputer already had one. The large non-automated base of smaller and medium sized firms still could not afford these systems and had to wait until computers became less expensive. Both these crises were overcome by the introduction of new technologies. This third crisis, at the start of the 1990s, was of a different nature and could not easily be overcome by introducing new technologies. Due to competitive pressures a great number of players disappeared or merged into larger companies. This process abruptly stopped the smooth entry of new firms into the IT market. Before, it had been relatively easy for firms to enter the market having relatively very little capital. Within a matter of years these firms could expand significantly and take a share of the market. The process of concentration and expansion, however, had resulted in larger players in the sector, players who had more capital and were stronger than smaller competitors. This concentration process took place in both the hardware and software oriented branches (Automatiseringsjaarboek 1989, xxiii-xxiv). In a matter of just a few years the face of the IT industry had changed because of this process of concentration. In 1989 well over a hundred takeovers took place most of which were initiated by large service

History and Definition of the Dutch IT Industry 53 providers such as Raet, HCS, BSO, Volmac, and Getronics. Through takeovers these firms were able to provide a wider range of activities and could maintain the level of returns they had been achieving in the previous years (Automatiseringsjaarboek 1990, xv). American and Japanese firms used the crisis, which predominantly hit Europe, to expand. Lots of firms were for sale at relatively low prices. The feeling at the time was that the Japanese were more rigorous than Americans firms who occasionally opted for collaboration rather than taking over a firm. Another difference between Japanese and American firms was that the latter were more focused on the software branches, whilst Japanese firms focused more on the hardware branches (Automatiseringsjaarboek 1992, x). Takeovers were the dominant growth strategy for IT companies in the 1990s. In 1995 for instance almost 100 billion dollars worth of takeovers took place in the global IT industry. 11 In the Netherlands, Getronics and Roccade acquired Raet. Takeovers had, and still have, a lot of advantages for companies. They allowed them to grow and enter new markets. Growth was also essential in order to remain independent. Being big was important in those days, an interviewee stated, it was either buy someone else, or be bought yourself. Outsourcing, an arrangement by which one company provides (IT) services for another company that used to be provided in-house, was also gaining popularity during the 1990s. Outsourcing was regarded as a necessary instrument to increase competitiveness. At the end of the 1990s outsourcing really took off. This was partly due to a number of factors such as the high level of economic growth which fed through into the IT industry. Most of the companies at that time did not have the capacity to do everything themselves and saw outsourcing as a means of solving their immediate problems. Later on at the beginning of the 2000s outsourcing would remain popular, but for different reasons. Companies used outsourcing as a means of lowering costs and increasing efficiency (Slangen 2004, 14). During the period of recession in the early 1990s it was not solely the hardware oriented branches of the sector that were hit by the economic downfall, but the software and services oriented lines of business as well, although not as badly as the system houses that dealt in PCs and minicomputers. This depression, which was characterized as the end of the golden era of computing, would last until the mid 1990s when the sector started picking up because of renewed investments in information technology. During these years package software had become the number one selling product within the IT market, increasing the market share of software suppliers in the sector. This is the period in which players such as Microsoft became the dominant forces they are in today s IT industry. Package software seemed to be the only solution for small-scale businesses to automate. Though these businesses could now afford computers, they still had to buy software that they could 11 Computable, March 29, 1996, Bijna 100 miljard dollar aan overnames in IT-sector.

54 Developing Institutions use. Bespoke software, 12 however, was too expensive, so they had to revert to package software. This also meant that instead of software being customized to fit the business, these businesses had to be tailored to the software. Things would eventually change in favour of these smaller businesses as bespoke package software would become more readily available (Automatiseringsjaarboek 1991, xii). The 1990s should not however be described as a decade of crises. The latter part of the 1990s would prove to be the sector s most successful period in its entire history so far. Part of this success can be ascribed to the renewed interest in IT when the Internet started taking off in the middle of the 1990s. At its start in 1969, the Internet was called Arpanet. It had received its name from the agency from which it stemmed, the Advance Research Projects Agency. The Arpa was established in 1957 as a response to the USSR Sputnik launches and was meant to ensure a US lead in (military) science and technology. In 1969 the first four computers were linked together but results were rather poor. Soon after Arpanet became more stable and different protocols by which information could be sent were developed. 13 The different protocols most commonly used today originate from various decades. Email, for instance, stems from the 1970s whilst the Hypertext Transfer Protocol (HTTP), 14 the protocol that is used for common web pages, originated in the 1990s. In the first twenty five years of its history the Internet has grown at a moderate pace, but the Internet really started taking off after the hypertext protocol was introduced. Figure 3.1 demonstrates this Internet explosion quite well through the increase in the number of hosts on the Internet. 15 Originally intended for communicating scientific papers between researchers, the easy means of describing and universally presenting information caught the attraction of the general public and, inevitably, commerce. Already in 1994, just three years after it was released by CERN, the first signs of commercial use of the Internet could be found as shopping malls opened up shops on the Internet and people were able to order pizza s at the Hut. 12 Whatis.com defines bespoke as a term used in the United Kingdom and elsewhere for an individually- or custom-made product or service. Traditionally applied to custom-tailored clothing, the term has been extended to information technology, especially for software consulting services. Typically, software consulting companies offer packaged (already invented and generally applicable) software and bespoke software for client needs that can't be satisfied by packaged software. In the U.S., bespoke software is often called custom or custom-designed software. 13 Most notably, the Transmission Control Protocol (TCP) and the Internet Protocol (IP). These protocols are used to connect computers on the Internet. 14 Http is a protocol that facilitates the transfer of hypertext-based files between local and remote systems. 15 A host is a computer system that has a registered IP-address. In order for information to be sent over the internet a source and destination IP-address must be available. An IP-address is comparable to a phone number.

History and Definition of the Dutch IT Industry 55 Table 3.1 Growth of the Internet between 1969 and 2003 1.000.000.000 100.000.000 10.000.000 Internet hosts Internet hosts (new survey) Internet websites 1.000.000 100.000 10.000 1.000 100 10 1 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Source: http://www.zakon.org/robert/internet/timeline/ (accessed February 8, 2005). An Internet host is a computer system with a registered IP address (an A record). An Internet website is a number of web servers (one host may have multiple sites by using different domains or port numbers) At its start the Internet was merely a text-box with minimal navigation and layout possibilities. As graphical browsers were introduced and the Hypertext Mark-up Language (HTML) developed, the Web became more attractive to a wider audience. At these initial stages the Web was an easily accessible place and soon small companies arose that provided web-services. The technology behind the Internet also progressed and people were able to access the Internet at higher speeds. Developments were also being made on the server front, the place were information is sent from, turning the Web into a more dynamic place and allowing it to become more than just a collection of graphics and text. Today the Web is produced by a wide variety of people and organizations and WebPages range from simple informative presentations to more sophisticated Web based applications. 3.1.5 The Rise of the Dutch IT Industry in the Late 1990s The rise of the Internet has played an important role for the IT industry. During the 1990s the personal computer had already entrenched itself in both household and business environments and without a new boost the market would turn into a replacement market. The Internet ensured that computer sales remained high. Also other parts of the hardware sector profited from the Internet hype. Suppliers of equipment to facilitate the flow of information on the Internet, such as Cisco who

56 Developing Institutions supplied routers and hubs, flourished. The huge increase in demand for bandwidth triggered lots of investments from which the sector profited. The Internet was not the only reason though why the sector was flourishing. Approaching the year 2000 people wondered what effect the year 2000 would have on systems that had been developed several decades before. The problem they discovered was that all of the software that was written, and which was now considered legacy software, used two digits instead of four digits to express a year. At the time using two digits was convenient enough and no one thought their software would still be in use 20 years later in a new millennium. This meant that with the arrival of the year 2000, all systems would assume it was the year 1900. Obviously, this could have a major impact on systems that e.g. calculated pensions. At the time it was hard to estimate exactly what would happen, but the most serious threat was the possibility of systems crashing. This potential threat triggered an enormous boost in investments in software and hardware. One may wonder why so much legacy software was still in use. The general view of IT is that it is a fast paced world were technological developments are an everyday business. Though this may be so, systems did run that long because, as it turned out, it was cheaper to modify existing hardware and software than replacing it altogether. 16 It was also less risky to adjust something that had already proven itself for such a long time, whilst a replacement could be buggy and cause serious problems for company s critical systems. The fact was that it worked. Systems did run without problems far longer than they were originally expected to do. It made no sense to invest a lot of money into something that was already doing its job. This is why, when the Y2K bug or millennium bug was generally acknowledged, so many systems needed to be updated at the same time and at such short notice. The Internet and the millennium bug were not the only reasons the sector was booming. In Europe the Euro was about to be introduced and this produced a similar problem to the millennium bug that had to be resolved in a similar fashion. All software that was specifically programmed for accommodating local currency had to be rewritten or adapted to be able to work with the Euro. These factors, combined with a general economic upheaval, increased the demand for qualified personnel. Especially in the software and services branches of the sector the demand for personnel was staggering. However, it proved to be very difficult to find qualified personnel to work in the IT industry. To solve this problem in the Netherlands a special taskforce was set up by the Ministry of Economic Affairs that had to study this problem. The taskforce estimated that in 2003 the demand for qualified personnel would reach 280,000, whilst the figure in 1999 was estimated at around 195,000. 17 This meant that in the three to four years following, the number of qualified IT specialists would have to increase by over 40 percent in 16 Most of the legacy software consisted of generally poorly documented COBOL software. 17 Taskforce ICT, Deltawerkers voor de digitale delta: Rapportage TaskForce 'Werken aan ICT', s- Gravenhage, Ministerie van Economische Zaken, 1999, 14-18.

History and Definition of the Dutch IT Industry 57 order to meet the estimated demand. This proved to be an almost impossible figure as it was very hard to find qualified personnel at such short notice. The problem seemed to be related to a number of factors. The first is the poor branch formation and the general opaqueness of the IT industry. The combination of these factors alienated people from the IT industry. Another factor that made the IT industry less attractive to people was the wave of layoffs that had followed the crash at the beginning of the 1990s. These large scale job losses were still in people s minds Another factor was a failing policy regarding the supply of training and education. Not enough students had been entering computer and information science studies in the years prior to the IT explosion in the late 1990s. Also the education level requirements for qualified IT personnel in the late 1990s where higher than those ten years before. While most of the vacancies in the late 1980s where filled with students of vocational and higher vocational training levels, the vacancies of the late 1990s required higher vocational and university level training. The market decided what type of specialization was in demand. 18 On the whole the IT industry failed to attract enough people from sectors outside the industry. It was not uncommon that a person working inside the IT industry in company X quit his job and started working at company Y in the IT industry. This typical situation is often described as the chain problem, where one hole is filled with another hole. 19 3.1.6 The Post-Millennium Crisis of the Dutch IT Industry The problem of the tight labour market, as it turned out, solved itself, because the demand for qualified personnel collapsed due to another crisis in the IT industry. Figure 3.2 shows the drop in vacancies in the Dutch IT services industry since the second quarter of 2001. 18 In the case of Baan for instance, Baan had developed Enterprise Resource Planning (ERP) software software that streamlined information within enterprises and specialists that could work with this software where in high demand for a period of time. After a few years though, when Baan was having problems and companies were abandoning their Baan support, these ERP specialists where confronted with an altogether different market situation and had to re-specialize in order to find work. This should have been a first indication of the capriciousness of the ICT sector s labour market when things started turning worse in 2001. 19 Arbeidsvoorziening (1998, iii-iv, 24).

58 Developing Institutions Table 3.2 Vacancies in the Dutch IT services (SIC 93 class 72) 12.000 10.500 Vacancies Created vacancies Filled vacancies 9.000 7.500 6.000 4.500 3.000 1.500 0 nov-96 feb-97 mei-97 aug-97 nov-97 feb-98 mei-98 aug-98 nov-98 feb-99 mei-99 aug-99 nov-99 feb-00 mei-00 aug-00 nov-00 feb-01 mei-01 aug-01 nov-01 feb-02 mei-02 aug-02 nov-02 feb-03 mei-03 Source: Statline (http://www.cbs.nl). A number of factors contributed to this latest crisis in the IT industry, which started in the period shortly after the millennium bug was worked out. As has already been discussed, the millennium bug triggered an enormous amount of investments and required a great number of people to patch the legacy code. Most of the people who were hired had no previous experience whatsoever with information technology and, prior to their training, had no programming skills. They were tested, recruited and trained within a six month period and were then put into action debugging the legacy code. After the millennium bug period the sector had to remould itself to the new market conditions and some companies were too slow to adapt to the new situation and hence had problems maintaining positive financial results. The millennium bug had required enormous investments and companies were weary of maintaining the same levels of investment. Instead of investing in software, investments were made in the Internet and telecommunication services. This made the dip for the sector as a whole less noticeable, but for most IT companies such as Origin (today called Atos- Origin), Cap Gemini, and Getronics, it turned out to be a difficult period. As these companies adapted to the new demands their profits returned. There are a number of factors that have played a role in the decline of the IT industry since 2001. These factors include the bursting of the Internet bubble combined with the economic downturn at the beginning of the third millennium. The deflation, or rather bursting of the bubble that was called the Internet hype or Dotcom hype, proved to be an important blow for the trust and belief people had in information technology. Dotcoms, basically companies that promised to sell something through the Internet, were companies that profited from the belief that a

History and Definition of the Dutch IT Industry 59 new economy was outperforming the traditional economy. Dotcoms managed to attract large sums of venture capital from investors that wanted stocks in return for their money. These investors were gambling on the moment these Dotcoms would go public and make their initial public offering (IPO). Venture capitalists were gambling on a rise of these stocks and a high return on investments. A large number of these companies did not survive because their business plan was not solid. One of the best examples of such a company was pets.com, which spent millions of dollars in advertising. 20 At the time people were blinded by the hype and hoped for easy money. At first people started buying shares of such companies without even checking the turnover and profit figures of such companies. When eventually people did start checking these figures, it turned out that these companies were making no money whatsoever and everyone wanted to sell the shares they had as fast as they could. During the spring of 2001 most of these companies went bankrupt, dragging down the whole IT industry with them as belief in IT was at an all-time low. Though the software and services segment of the IT industry was not flourishing at the time, the position of the hardware segment was even more dramatic. There are a number of reasons for this. The first has to do with the fact that in order for hardware to be produced, a lot more investments needed to be made in fixed capital such as factories, machines, and transportation, which meant fixed costs. Software and service providers can, under ideal circumstances, hire or fire personnel, but this is not an option for a hardware manufacturer and thus an economic decline has more impact in the hardware segment of the sector, because hardware manufacturers are less flexible in terms of increasing or decreasing their means of production. In part the hardware segment s misery is related to the general lack of investment in the software segment and vice versa. The burst of the Internet bubble and the general decrease in investment in IT have decreased the demand for hardware such as personal computers, mobile phones, and routers. This is why companies such as Cisco, Nortel, Lucent and Motorola, who had been doing very well in 2001, performed poorly. The decreased demand had its effect on their suppliers and the suppliers of these suppliers as well and chip suppliers such as AMD and Intel saw their profits dropping below the zero mark as did the figures of ASML, Applied Materials, and Agilent. Because of the poor hardware sales, a reverse effect could be noted in the demand for system software, which had also decreased. Landman and Veldkamp (2002, 10-17) have made similar claims and note a relation between hardware, software, and services segments in the IT industry. They have extracted empirical evidence by comparing IT expenditures in the hardware, software, and services niches to the Dutch Gross National Product. From these figures they concluded that when investments in hardware decreased, decreasing 20 Their most famous advertisement was that during half time of the Super bowl, the most expensive broadcasting time of the year. See also http://www.geek.com/news/geeknews/2000nov/gee20001107002827.htm (accessed February 8, 2005). A good Dutch example is World Online, whose stocks were initially offered at Euro 45, but would soon drop to Euro 15 (Automatisering Gids, February 14, 2001, ABN Amro wast handen in onschuld ).

60 Developing Institutions investments would follow in the software segments, and finally, in the services segments. These figures are indications that the various segments of the IT industry are interrelated. The last explanation of the recession in the IT industry finds its origin in the general economic decline since the middle of 2001. Though the IT industry s malaise started earlier than the general economic downfall, it has put a greater strain on an already struggling sector. Forrester Research has developed a theory, based on empirical evidence, which describes the relation between IT investments and economic cycles (described in terms of gross national product). Historical data indicate that investments in technology only take place if the gross national product (in the United States) has increased by 4.5 percent in the previous year. This percentage is called the tech economy tipping point. If growth of the national economy falls below or goes beyond this point, investments in IT will drop or rise as well (Financieel Dagblad, 18 November, 2002, 19). According to Landman and Veldkamp (2002, 10-12), the link between the economy and the IT industry has grown since the 1990s, as the 1970s and 1980s were typically growth markets, while the 1990s, when already 50 percent of all potential end users had a computer, 21 can be typified as a replacement market. Replacement markets are by nature more related to general economic trends than growth markets are, because these latter markets have a biased demand in respect to markets that are influenced by general economic trends. 22 Landman and Veldkamp (2002) conclude that there is a neglectable connection between the economy and the IT industry. 3.1.7 The Three Lineages of Software and Service Companies in the Dutch IT Industry So far this chapter has discussed the history of the sector, how it materialized, and how it has developed to the sector that it is today. What has not been discussed is the way in which the terms and conditions have developed in this sector. Looking at the Dutch IT industry three types of lineages can be noted with respect to history and culture of industrial relations. The first type of lineage is the oldest one and dates back to the earliest use of computers in mathematical and administrative centres. Over time these mathematical centres have been privatized and sold to other firms. What distinguishes this lineage from other lineages is that IT companies that originated from this period have a different culture that makes them more inclined to use collective agreements. The most notable incarnations of this lineage today in the Netherlands are PinkRoccade and Atos Origin. 23 The second type of lineage stems from the two periods of growth during the 1970s and 1980s, when companies started specializing into different branches of 21 The term end user is used to distinguish the person for whom a hardware or software product is designed from the developers, installers, and those who service the product (www.whatis.com). 22 Meaning they have both a structural replacement demand and an increasingly marginal growth demand. 23 The history of most of the companies mentioned here will be discussed in greater detail in chapter five.

History and Definition of the Dutch IT Industry 61 information technology. In combination with an increasing demand for computers and different types of software, a number of foreign IT companies established subsidiaries in the Netherlands. It is quite likely that during this period, when the IT industry was rapidly expanding, these foreign companies introduced home country effects: they imported their own corporate culture. Examples of these foreign firms are the French Cap Gemini, the British CMG, and the United States based EDS. The Unites States was already dominating the world IT market though US firms received fierce competition from Japanese businesses that were rapidly catching up, predominantly in the hardware manufacturing industries. These types of companies, of which IBM is a good example, had no interest whatsoever in the unions and even actively maintained an anti-union policy. 24 The last type of lineage stems from the massive numbers of small and medium sized businesses that sprouted in those periods when IT was booming business. There have been a few moments in the brief historical span of the IT industry in which it grew very rapidly and large profits were made. Obviously this attracted a lot of entrepreneurs. These types of firms usually restricted their terms and conditions of employment to a minimalist contract that only discussed the height of wages and contributed to a large extent to the bad name the sector has in terms of industrial relations as they usually were very inexperienced at managing a business. Having grown and expanded in a period of economic upheaval they were not well enough equipped for dealing with periods of decline and consolidation. Bankruptcies and loss of jobs was the consequence. Examples of entrepreneurial firms that have grown considerably and remained relatively successful in the Netherlands are Centric, Ordina and Baan. 25 Chapter four will continue to explore the industrial relations, position, history, and interactions between the unions, employers associations, and works councils in the Dutch IT industry. 3.2 A Definition of the IT Industry 3.2.1 The Many Branches of the IT Industry In this section of the chapter an account will be given of what defines and comprises the IT industry. In the previous historical section of this chapter it was demonstrated that the sector has grown in terms of numbers of companies and workers. Further, the number of activities IT has become involved in has increased significantly throughout the years. At first it was solely the hardware companies that dominated 24 IBM for instance circulated a Corporate Instruction Letter amongst its European establishments. The letter s message was that in so far as this was legally possible, any contact with the unions should be avoided as much as possible (Smit et al. 1995, 89). Even today most ICT companies are weary of union intervention (personal communication by various union representatives). 25 Though Ordina was originally a French company, it was sold to a Dutch entrepreneur in 1973 and has been a Dutch firm ever since.

62 Developing Institutions the sector as they provided hardware, software, and services. Later on, as system software was introduced that allowed others to write software, a process of differentiation began. First software houses and service companies were introduced. Next in line were the system houses, package software suppliers, PC dealers, and consultancy. During the 90s, due to a wider scale use of technological innovations and development of new business models, outsourcers, network suppliers, Internet providers, and application service providers were added. This combined process of growth and branching has today led to a quite complex web of companies involved in IT. In order to systemize this web of IT companies, they will first be categorized into a number of core segments of today s IT industry. These core segments, or categories, comprise the following branches: hardware, software, services, and branches that consist of firms providing hardware, software, and services. Hardware Branches The hardware branch in the IT industry invests in R&D in order to develop, produce, market, and sell computers, cables, connectors, power supply units, and peripheral devices such as the keyboard, mouse, audio speakers, and printers. Famous computer manufacturers still around today are IBM, Compaq, Sun, Apple, and Dell. Until the 1970s, when IBM had to unbundle software and services from its hardware, it had been normal for companies that produced computer systems to also provide software and services. Since the 1980s, computers progressively became a matter of components. As more hardware manufacturers entered the market and specialized their R&D in fields like graphics processing, companies like IBM have become more dependant on hardware suppliers. Even more R&D dependent are the suppliers of computer manufacturers. These are the companies that develop processors (AMD and Intel), hard disks (Western Digital, Seagate, and Maxtor), and network components (3COM and Cisco). The companies that produce equipment to manufacture IC s (ASM Lithography) can also be included here. Another important niche within this branch is that of peripherals suppliers, companies that develop printers and scanners (e.g. HP, Canon, Epson, and Lexmark), monitors (e.g. Iiyama, Philips, and Sony), scanners, etc. It is usually a matter of definition whether graphics and sound cards are categorized as peripherals or can be seen as an integral part of a computer rather than a side dish. Graphics cards in particular have become very complex in the last decade. During the 1990s most of the traditionally hardware oriented firms showed a tendency to move towards the services probably because of the poor results in the hardware niches at that time. Software Branches Software is available as a service (custom made software), as a product (package software), and as a shrink wrapped mass product. In the case of package software, developers design software at their own risk. The result of this R&D is a software package that is sold at a moderate (license) price to a variety of users. Roughly three

History and Definition of the Dutch IT Industry 63 types of package software can be discerned: system software, middleware, and application software. System software comprises software that is used to operate a system, e.g. Microsoft s Windows, IBM s OS/400, or Sun s Solaris, development tools, software that is used to design other software, and system management software, a set of tools used to manage a computer system. Middleware is a general term for any programming that serves to "glue together" or mediate between two separate and often already existing programs. A common application of middleware is to allow programs written for access to a particular database to access other databases. The term middleware is sometimes used to describe programming that mediates between application and system software or between two different kinds of application software. Typically, middleware programs provide messaging services so different applications can communicate. The systematic tying together of disparate applications, often through the use of middleware, is known as enterprise application integration (EAI). Application software is software that is designed to perform a specific function directly for a user or, in some cases, for another application program. Examples of applications include word processors, database programs, Web browsers, development tools, image editing programs, and communication programs. Applications use the services of the computer's operating system and other supporting applications. There s also a special type of application software, which is sometimes called multi-module application software and can also be regarded as an outgrowth of middleware. This category of application software is usually not intended for one user, but for companies. Customer Relationship Management (CRM) software is software that helps an enterprise manage customer relationships. Another well-known example within the industry is that of Enterprise Resource Planning (ERP) software, which helps businesses manage parts of their business, including product planning, parts purchasing, maintaining inventories, interacting with suppliers, providing customer service, and tracking orders. ERP can also include application modules for the finance and human resources aspects of a business. Services Branches The term service has a relatively ambiguous definition in the IT industry. This is primarily so for two reasons. Services in the IT industry often comprise a combination of hardware and software and thus transcend the common hardware/software distinction. Secondly the term services is at the same time ambiguously applied because it is debatable whether the provided service can be defined as a product or service. Take for example custom made software, which the industry regards as a service because the product (a piece of software, which is the result of a process or activity) is provided as a service. The number of services branches within the IT industry is quite large. Services in the IT industry are typically provided using (one or more of) the following four agreements: Firstly, services can be provided on an hourly basis (1). The client pays for each man hour required to complete the service. Secondly, services can be

64 Developing Institutions provided on a project basis (2). Depending on the contents of the agreement, the client will know how much a certain service will cost in advance. Thirdly, services are often provided on a posting basis (3). In this case the service provided can to some extent be regarded as an employment agency. Finally, services can be provided on a contract basis (4). This is usually the case when a client purchases hardware and wants it to be maintained for a fixed annual price. Table 3.1 demonstrates the types of the services provided by the IT services industry. The third column of the table lists the type of agreements that are most commonly used (listed from most common practice to least common practice). Table 3.1 also mentions the Standard Industrial Classification (SIC 93) class the type of service is categorized under. In section 3.3.2 the SIC system with respect to the IT industry is examined more closely. Table 3.3 Services provided in the Dutch IT industry Service Description Agreement SIC 93 Hardware maintenance Maintenance and repair of hardware. 4, 1, 2 72.50 Software house Development of application software based on specifications provided by the client. Services include design, programming, and implementation of the software. 1, 3, 2, 4 72.20 Service bureau Data processing. 4 72.30 72.40 System management Consultancy Training centres Outsourcing System integration Network installation Internet service providing (ISP) Application service providing (ASP) Installing and management of both software and hardware for a client. Advising clients concerning processes, purchases, implementation, and optimizing its IT systems. Private training centres providing training and education on all sorts of IT related aspects. An arrangement in which one company provides services for another company that could also be or usually have been provided in-house. Organizing the entire design of an automated information system. 1, 3, 2, 4 72.10.1 1, 2, 3, 4 72.10.2 4 72.10.1 4, 2 2, 1 72.10.2 Installation of physical networks within buildings. 2, 1 51.84 Organization providing Internet access to customers. 4, 1 64.20 Organization providing both hardware and software to clients over the Internet. Clients need not buy the hardware and software themselves, but use it over the Internet. 4 72.60 Internet hosting Organization which predominantly hosts Internet content. 4 64.20 Web design Design and actualization of content of web pages. 2, 1, 4 72.20

History and Definition of the Dutch IT Industry 65 Combined Branches It is a common practice in the IT industry for companies to offer a mix of the various services mentioned above. Many of the firms thus operate in multiple branches. Combinations of branches within a firm are often regarded as a branch of their own. Table 3.2 shows the combined branches that can be found frequently in the Dutch IT industry. Table 3.4 Combined services branches in the Dutch IT industry Branch Description SIC 93 System houses Internet suppliers IT services This branch provides, usually its own, software packages, the hardware that runs the software, as well as all the services and support required to keep the system running. Like to call themselves one stop shops. 72.10.1 This branch provides Internet services, hosting, and web design. 64.20 72.20 This branch combines the branches software houses, consultancy, outsourcing, and system integration. 72.10.1 72.10.2 72.20 3.2.2 Defining the IT Industry When talking about the IT industry in general in this study, those companies are included that fall under either the hardware, software, services, or combined branches as mentioned above. Companies that solely sell or rent software and hardware without having anything to do with the implementation or production of this software or hardware are excluded from the definition of the IT industry as is used in this study. The reason for this is that this study is mostly interested in the types of agreements in use at a specific segment of the IT industry, namely the IT software and services companies such as Getronics, Atos Origin, Capgemini, et cetera, and not companies such as the Free Record Shop, who sell CD s, DVD s, and computer games. Statistical agencies such as the Central Bureau of Statistics (CBS) in the Netherlands use a different definition. They classify economic activities of enterprises according to the Standard Industrial Classification (SIC 93). Table 3.3 demonstrates what classes apply to the IT industry. There are two major classes that cover the IT industry: IT manufacturing (30) and IT services (72).

66 Table 3.5 Developing Institutions The Dutch IT industry in terms of the Dutch Standard Industrial Classification (SIC 93) SIC 93 30 30.02 72 72.10.1 Description Manufacture of Office Machinery and Computers Manufacture of computers and other information processing equipment - Manufacture of automatic data processing machines including micro-computers: - digital machines - analogue machines - hybrid machines - Manufacture of peripheral units: - printers, terminals, et cetera. - magnetic or optical readers - machines for transcribing data onto data media in coded form Computer and Related Activities System houses - System analysis and development, software services, system maintenance, renting computer time, training and education, and support. - Posting of computer experts 72.10.2 Hardware consultancy - Consultancy on type and configuration of hardware and associated software application. Analyzing the users needs and problems and presenting the best solution. - Consultancy with regard to computer security NB. If consultancy within the company is combined with other activities such as hardware and software supply, software development, system analysis, these activities should only make up 20% of the total business of the company. Otherwise a company will not be categorized in this subclass. 72.20 Software consultancy and supply - Development, production, supply and documentation of ready-made (non-customized) software - Analysis, design and programming of systems ready to use: - analysis of the user s needs and problems, consultancy on the best solution - development, production, supply and documentation of made-to-order software based on orders from specific users - writing of programs following directives of the user - web page design 72.30 Data processing - Processing of data employing either the customer s or a proprietary program: - complete processing of data - data entry services - scanning of documents - Development, production, supply and documentation of ready-made (non-customized) software - management and operation on a continuing basis of data processing facilities belonging to others - web hosting - database related activities: provision of data in a certain order or sequence, by on-line data retrieval or accessibility (computerized management) to everybody or to limited users, sorted on demand

History and Definition of the Dutch IT Industry 67 72.40 Database activities - on-line database publishing - on-line directory and mailing list publishing - other on-line publishing - web search portals 72.50 Maintenance and repair of office, accounting and computing machinery 72.60 Other computer related activities One subclass 30.01 has been added, which is not really a part of the IT industry according to the strict definition that is used here. The reason for including class 30.01 is that most available statistics regarding IT manufacturing in the IT industry are categorized as class 30 and do not discriminate between the subclasses 30.01 and 30.02. Table 3.4 shows the details for this subclass. Table 3.6 Added SIC 93 subclass SIC 93 Description 30.01 Manufacture of office machinery - manufacture of manual or electric typewriters - manufacture of word-processing machines - manufacture of hectograph or stencil duplicating machines, addressing machines and sheet fed office type offset printing machines - manufacture of calculating machines, cash registers, postage franking machines, special terminals for issuing of tickets and reservations, etc. - manufacture of diverse office machinery or equipment: - Processing of data employing either the customer s or a proprietary program: - machines that sort, wrap or count coins; automatic banknote dispensers; machines that stuff envelopes, sort mail; pencil sharpening machines; perforating or stapling machines, etc. Two classes have been omitted for practical reason (see table 3.5). These classes are 51.84 and 64.20. Class 51.84 seems only to be available in the Dutch implementation of the International Standard Industrial Classification (ISIC) and is too loosely defined to be of any use. Class 64.20 is also too loosely defined to be of any use.

68 Table 3.7 Omitted SIC 93 classes Developing Institutions SIC 93 Description 51.84 Wholesale of computers, computer peripheral equipment and software This class also includes (according to SIC code 45.31 Installation of electrical wiring and fittings): - Installation of data- and computer networks 64.20 Telecommunications This class includes transmission of sound, images, data or other information via cables, broadcasting, relay or satellite: - telephone, telegraph and telex communication - maintenance of the telecommunications network - transmission (transport) of radio and television programs - Internet access provision 3.2.3 The Dutch IT Industry in Figures In this section of the chapter a few key characteristics of the IT industry are presented that will conclude this chapter. If possible the Dutch statistics are set against an international background in order to place the Dutch IT industry in a broader perspective. First of all, statistics regarding the number of IT companies are given and the type of employment agreements they most commonly use. Then relevant statistics regarding IT employees are given. This section concludes by presenting the Dutch IT industry s economic position. Table 3.6 shows that the number of IT companies in the Dutch IT industry has grown significantly in the past ten years; approximately 63.7 percent from 1993 to 1998 and again by 70.5 percent from 1998 to 2002. IT manufacturing has not grown as much as the IT services during this period. This has lowered the share IT manufacturing has in the Dutch IT industry to 2.1 percent from 3.1 percent in 1993. The IT industry in the Netherlands is dominated by IT software and service companies.

History and Definition of the Dutch IT Industry 69 Table 3.8 Number of IT companies in the Dutch IT manufacturing (30) and services (72) industries SIC 93 1993 Share 1998 Share 2002 Share Growth 1993-1998 Growth 1998-2002 30.01 10 0.2% 20 0.2% 15 0.1% 100.0% -25.0% 30.02 190 2.9% 325 3.1% 370 2.1% 71.1% 13.8% 30 200 3.1% 345 3.3% 385 2.1% 72.5% 11.6% 72.10.1 395 6.1% 865 8.2% 1,960 10.9% 119.0% 126.6% 72.10.2 1,810 28.0% 3,980 37.8% 5,985 33.4% 119.9% 50.4% 72.20 3,605 55.8% 4,475 42.5% 8,645 48.2% 24.1% 93.2% 72.30 305 4.7% 310 2.9% 245 1.4% 1.6% -21.0% 72.40 45 0.7% 170 1.6% 210 1.2% 277.8% 23.5% 72.50 70 1.1% 300 2.9% 430 2.4% 328.6% 43.3% 72.60 35 0.5% 75 0.7% 80 0.4% 114.3% 6.7% 72 6,265 96.9% 10,175 96.7% 17,555 97.9% 62.4% 72.5% 30 + 72 6,465 100.0% 10,520 100.0% 17,940 100.0% 62.7% 70.5% Source: CBS Statline (http://www.cbs.nl). The total number of companies in the Netherlands in 2002 was 689,625 and 598,150 in 1993. IT services and IT manufacturing today make up 2.60 percent of the total number of companies in the Netherlands. In 1993 this percentage was 1.08, which demonstrates the growing significance of the IT industry in the Netherlands. SIC 93 class 30.01, as we can see, is relatively insignificant and is unlikely to have much influence on the pureness of the statistics in relation to the narrow definition if the IT industry we are using here. On January 1, 2003, the number of IT companies in the Netherlands was 18,375. This seems like a lot for the Netherlands, but if we look closer at the demographics of IT companies (table 3.7), we can see that their number largely depends on companies without employees.

70 Table 3.9 Developing Institutions Demographics of the Dutch IT industry in terms of company size SIC 93 Year Total 0 1-5 5-10 10-20 20-50 50-100 100-200 200-500 30 1993 200 105 35 10 10 20 10 5 0 5 1998 345 210 75 25 10 10 10 5 0 0 2003 350 220 65 20 15 10 5 5 0 0 72 1993 6,265 4,210 1,320 265 175 185 55 25 20 10 1998 10,180 6,265 2,560 600 325 265 85 35 25 15 2003 18,025 11,865 4,035 810 565 460 165 60 35 25 Source: CBS Statline (www.cbs.nl). Interesting to note is the number of companies employing more than 500 people in the IT manufacturing industry in 1993: five to be exact. In 1998 this figure was reduced to zero, which can mean that these companies have gone bankrupt or in some other way disappeared. It may be that they have been reclassified as 72. 26 One may think here of large companies expanding their core business with providing services. Takeovers may also be a good explanation for their disappearance in IT manufacturing class. Many firms previously known for their involvement in hardware have shifted their scope and focus. IBM is a good example of such a firm. Table 3.8 shows the trade balance results for the IT services industry from 1996 until 1999. Particularly of note in this table is the increase in both the cost of personnel per full time unit (FTU) and the operating profit per full time unit. These figures corroborate the idea of a wage-price spiral in the IT industry since the mid 1990s where both the prices of IT products and wages of IT professionals rose significantly. Both profits and cost of personnel per FTU have risen during the four years. 500-26 Unfortunately the data provided by the CBS Statline is not detailed enough to show what companies were involved.

History and Definition of the Dutch IT Industry 71 Table 3.10 Trading results IT services (SIC 93 class 72) 1996-1999 1996 1997 1998 1999 Number of firms 10,200 11,850 12,260 13,410 Number of people employed 76,800 88,600 106,000 117,700 Number of Full Time Units (FTU) 72,900 82,200 96,500 107,300 Part time ratio* 0.95 0.93 0.91 0.91 Cost of personnel (millions of Euros) 2,699 3,435 4,294 5,172 Cost of personnel per FTU (Euros) 37,023 41,788 44,497 48,201 Net turnover (millions of Euros) 5,931 7,659 9,556 11,165 Net turnover per FTU (Euros) 81,358 93,175 99,026 104,054 Operating profit (millions of Euros) 640 840 901 1,310 Operating profit per FTU (Euros) 8,779 10,219 9,337 12,209 Source: CBS Statline (http://www.cbs.nl). No data available for IT manufacturing. * Number of FTU divided by number of people employed. If the part time ratio is 1, this implies everyone is working full time. The part time ratio increases as the figure moves closer to 0. The number of people employed in the IT services industry has risen significantly. To a minor extent this raise in numbers can be attributed to the increasing share of part time workers as the part time factor indicates in table 3.8. Figure 3.3 gives a clearer picture of how the number of companies is related to the number of jobs in the Dutch IT industry.

72 Developing Institutions Table 3.11 Growth in the number of IT companies compared to the growth in the number of jobs in the Dutch IT industry between 1993 and 2002 (Index 1998 = 100) 180 160 140 Netherlands IT manufacturing (30) IT services (72) Jobs of employees in the IT services (72) 120 100 80 60 40 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Source: CBS Statline (http://www.cbs.nl). As figure 3.3 demonstrates, the number of companies in the IT services industry has increased more than the number of jobs of employees in the same industry. This is predominantly so because of the large number of IT companies with no employees. The growth in the number of vacancies as shown in figure 3.2 shows the recent period of stagnation in the Dutch IT industry more clearly than does figure 3.3. Figure 3.4 shows the relative growth of jobs of employees in the Dutch IT services industry. The graph demonstrates that the relative number of females and part time jobs has increased in comparison to the total number of jobs. In 2002 there were 111,000 jobs in the Dutch IT services industry. 23 percent of these jobs were occupied by women and 19 percent of these jobs were part time. The average income was 42,500 Euros. Males earned 48,700 Euros on average and females 22,000 Euros. Full time jobs paid an average of 47,700 Euros and part time jobs 20,300 Euros. From these figures it may be safe to assume that women occupy most of the part time jobs.

History and Definition of the Dutch IT Industry 73 Table 3.12 Growth of the number of employees in the Dutch IT industry between 1993 and 2002 (SIC 93 class 72) (Index 1998 = 100) 180 160 140 Total number of jobs of employees Males Females Full time Part time 120 100 80 60 40 20 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Source: CBS Statline (www.cbs.nl). Statistical information regarding the characteristics of IT employees, union density, et cetera, can be found in chapter six, where they are discussed in comparison to the results of an Internet survey among Dutch IT professionals. 3.3 Conclusion Trying to compress a diverse history such as that of the (Dutch) IT industry into the above sections does not do credit to such a history. The object of this chapter, however, was to sketch the development of the (Dutch) IT industry and give insight into its diversity and complexity. This was necessary because the industry s history in the Netherlands may say something about why some companies have started using a collective agreement and why most have not. In this respect the differences between the three lineages of Dutch IT companies are interesting in the light of the limited choice set hypothesis, defined in chapter one. Whilst all companies, at least their direct descendants, that stem from the first lineage (mathematical and administrative centres) use a collective agreement today, this relation between lineage and usage of a collective agreement is not so clearly present in the other two lineages (foreign IT companies and entrepreneurial start-ups). A closer look at the individual cases is required in order to conclude to what extent these lineages are related to what type of agreement an IT company is using. Other interesting facts about the history of the Dutch IT industry are typical cycles of fast growth and steep decline the sector seems to be going through. Every

74 Developing Institutions seven to ten years new technologies give the IT industry a boost that allows these companies to promise new levels of productivity. However as time progresses, and general economic cycles urge companies to spend their money more carefully, IT is often the first area in which savings are made. This is noticeable in all three main branches of the IT industry. Firstly there are the savings on hardware and software (why invest in something new, when the old is still functioning well enough?), and secondly, the savings on services and consultancy (why do we need to hire so many of these expensive people?). Add to this the widespread presence of computers today within offices, firms, and homes. As such this branch of the market has become more of a replacement market than a growth market. The services companies are still able to grow, either by buying each other, or by insourcing the IT processes of other companies. 27 As such IT is becoming more and more intertwined with the rest of the economy. This is especially true in the case of outsourcing processes where companies not only outsource hardware and software, but also people, who thus come to work for a new employer, an IT company. This is something that needs to be looked further into, and is best done when discussing the individual case studies amongst IT companies in chapter five. 27 Insourcing and outsourcing are one and the same. It simply depends on the point of view. An IT company insources the IT infrastructure of a company that seeks to outsource this infrastructure.

4 Industrial Relations in the Dutch IT Industry This chapter is the first of three empirical chapters. Each chapter will cover a different level of industrial relations in the Dutch IT industry. This chapter focuses on the actors operating at sector level. Chapter five then descends to the company level and discusses ten cases of companies operating in the Dutch IT services industry. The sixth and final empirical chapter addresses industrial relations at the employee level and is based on an Internet survey carried out in 2003. In the Dutch IT industry a number of actors can be identified at sector level in the field of industrial relations. Some of these actors, such as unions and employers organizations, have already been discussed in chapter two. In this chapter they are subject to a more detailed historical approach of their activities and interactions with respect to the Dutch IT industry. The focus of the historical analysis will be on the period between 1990 and 2004. The reason for focusing on this period is that little to no information is available prior to this period. Finally this chapter discusses the largest collective agreement in the Dutch IT industry, which during 2004 covered approximately 35,000 people, the ICK collective agreement. This is an industry wide collective agreement for the IT hardware and office equipment industry. This chapter is based on interviews and e-mail exchanges with key figures of relevant organizations. Other sources were archive research and a review of journal articles during the period 1996-2004. 4.1 Employers Associations and Branch Organizations in the Dutch IT Industry Employers organizations seek to promote the interests of employers through collectivization. As discussed in chapter two, many of the current major employers organizations and workers organizations formed in the same period in the first part of the twentieth century. Employers organizations were established as a reaction to employee organizations beginning to take shape in those periods. When employers organizations first came about in the Netherlands it was predominantly because of reducing uncertainties with regard to the workers coalitions and protecting the

76 Developing Institutions freedom of employers in an ever increasing domain of legislation and regulation. Their activities have since increased and today include: Consultation, bargaining and concluding agreements with unions (collective agreements, social plans, collective recommendations for companies). Lobbying at local, national, and European level (influencing rules and legislation, and policy). Influencing public opinion through the mass media. Representation in advisory and civil / public administrations, and funds at national and sector levels (e.g. training and education, pensions). Advice, information supply and other services to its members regarding political developments, legislation, collective agreements, market developments, management, and organization of insurances. Opinion formation by developing contacts between employers, initiating collective research, product development, etc., within the industries. Research and policy development. Coordination with other organizations (Tros 2003, B1200-15). In the Dutch IT industry two employers organizations are active. They are the General Employers Association VNO-NCW (Algemene Werkgeversvereniging VNO-NCW, AWVN), 1 which operates on a national level in multiple industries and branches, and the Employers Association ICT (Werkgeversvereniging ICT), which operates predominantly in the IT (hardware) industry. Both employers organizations are associations and provide their services to members only. 4.1.1 AWVN AWVN originates from the General Employers Association (Algemene Werkgeversvereniging, AWV), which was founded to counterbalance the increasing power of the unions. The AWV differed from other employers associations in that it organized employers from different industries. The reason behind this was that although this diversified interests in some respect because of the different nature of various industries, it increased the number of organized employers beyond the scope of one industry, and thus their relative bargaining power. In later years the AWV became less of a defence line against unions, and more of the modern type of employers organization known today. The relationship between AWVN and the IT industry began in 1990 when the National Computer Centre (Rijks Computer Centrum, RCC), was privatised, becoming known as Roccade a few years later. At this point RCC became the AWV s first member from the IT industry. It was not until 1996 that a second IT company, Getronics, established relations with the AWV. Other current AWVN members in the IT industry include IBM, Atos Origin, PinkRoccade, Bull, and T-systems. 1 The AWVN works closely together with VNO-NCW and specializes in industrial relations and terms of employment. VNO-NCW protects the general interests of employers.

Industrial Relations in the Dutch IT Industry 77 Relations between AWVN and the IT industry were elaborated when the AWV merged in 1996 with the Labour Consultancy Agency (Adviesbureau Arbeidszaken, AA) of the NCW (that just months before had merged with VNO). One of the members of the AA was VIFKA, an IT employers organization that will be discussed later. From that point on AWVN established relations on a sector level by assisting VIFKA in their negotiations over the ICK collective agreement, the sector collective agreement for the information, communication, and office technology branch. The ICK collective agreement will be discussed in section 4.3. Beside a few larger companies, AWVN covers a number of the smaller and medium sized IT businesses which they help indirectly by providing services to the Werkgeversvereniging ICT. AWVN s actions as an employers organization are restricted to advisory and mediating functions with regard to industrial relations. They are not a lobby organization and have no interest in representing the IT industry as a whole. AWVN provides social services, not commercial services. AWVN respects the wishes of each individual member, such as the way members shape the terms and conditions of employment or the manner in which they conclude contracts with their employees. AWVN has no preferences with regard to collective agreements, works council agreements, or individual agreements. The sole interest of AWVN is that the social policy and industrial relations within companies are in good order and the belief that the interests of their members are well served. AWVN is concerned about the position of the unions in the IT industry. In other industries the unions have more members. Also throughout the years AWVN has built a strong relationship with the unions. In the IT industry, however, employees are more frequently represented through the works council. This is a source of concern for AWVN as it negatively affects the professionalism within firms with regard to industrial relations. The strength and expertise of a works council is often dependent on the manner in which a firm supports its works council or other factors such as changes in the composition of a works council. An example of the latter is the Atos Origin case that will be discussed in chapter five. Atos Origin had, when it started using a collective agreement, a very strong works council. After a few years though, a number of expert works council members had left and this compromised the works council s expertise for a period. In the Netherlands, MKB-Nederland promotes the interests of small and medium enterprises. The interests of MKB-Nederland in the Dutch IT industry are very minimal and are limited to informing their members of the availability of the ICK collective agreement. They have no specific opinions or ideas which are directly targeted at the IT industry. 2 2 Interview with a spokesperson of MKB Nederland.

78 Developing Institutions 4.1.2 Branch and Employers Associations Until 2001 The history of branch organizations in the Dutch IT industry is a colourful one and presents a clear demonstration of the many different interests that are prevalent in the industry. In fact it was not until 2001 that the Dutch IT industry had finally established one organization that covered the largest part of the IT industry called Nederland-ICT (Netherlands ICT). This association was founded in order to become the umbrella organization of branch and employers associations in the Netherlands. At the time, the branch and employers associations did not have a professional reputation with regard to the representation of the Dutch IT industry. 3 Figure 4.1 sketches the history of the most important branch and employers associations in the Dutch IT industry. Each of the shaded areas, three in total, represents an umbrella organization that was dominant throughout a specific period. The first employers organization operating in the Dutch IT industry was called the Association of Importers and Manufacturers of Office Machines (Vereniging van Importeurs en Fabrikanten van Kantoormachines, VIFKA), which was originally founded in 1922 and consisted of a number of branch organizations. The reason VIFKA became the employers association for the Dutch IT Industry was because throughout the years a number of branch associations had joined VIFKA which were closely associated to the IT industry. 3 Computable, Fenit en Vereniging ICT Nederland beginnen eindelijk koepelorganisatie., June 29, 2001. In the same article a joke was made regarding the solidarity in the IT-industry. They wrote: Put two IT professionals in a room and three parties will arise..

Industrial Relations in the Dutch IT Industry 79 Figure 4.1 Three periods of branch and employers associations in the Dutch IT industry VIFKA (1922) VIFKA OPS (1995) V-ICTN (1997) VIFKA OPS (1995) Nederland ICT (2001) VIFKANTEC (2001) VIFKA Telecom (1992) VIFKA Telecom (1992) ICT Telecom (2001) VIFKA Informatica (1990) Vereniging Computer Service en Software Bureaus (1971) / COSSO (1990) FENIT (1996) FENIT (1996) Werkgevers-vereniging ICT (2001) ICT Milieu (2001) VONIT (1998) I.T. UNITED (2002) I.T. UNITED (1997) Though the first umbrella organization, VIFKA, dates from 1922, it was not until the 1980s that VIFKA became active in the manufacturing niche of the Dutch IT industry. At the beginning of the 1990s a process of renewal began within VIFKA, which specialized VIFKA s activities into three branches: the hardware branch, the telecommunications branch, and the office products and services branch. VIFKA Informatics (VIFKA Informatica) was the first of these specialized branch organizations to come about. It was founded in 1990 and had 175 members (predominantly larger hardware companies) responsible for an annual turnover of 1.59 billion Euros (Smit, Schilstra, and Pauwe 1995, 88). The telecommunications branch was next. In 1992 VIFKA Telecommunications (VIFKA Telecommunicatie) was founded. Its members were active in the field of (wireless/mobile) telephony, (wireless/mobile) data communication, paging, satellite communications, etcetera. In 1994 VIFKA Telecommunicatie merged with the Association for Mobile Communication (Vereniging voor Mobiele Communicatie, VERMOCON), and doubled its member pool to 55. The last branch to merge into a central organization

80 Developing Institutions was that of the office products and services. In 1995 the branch associations VIFKA Document (VIFKA Document) and VIFKA Office and Firm Interiors (VIFKA Kantoor en Bedrijfsinrichting, VIFKA KBI), founded the new branch association VIFKA Office Products, Systems & Services (VIFKA OPS). When it was founded it had approximately 160 members, which supplied hardware, systems and services for postal centres, copy supplies, faxes, printers, office and business interiors, computer supplies, office supplies, et cetera. Besides providing services to the three branches mentioned above, there were also a number of activities taking place within specialized committees which transcended branch activities. These committees, which existed on a central level within VIFKA, included social affairs (conditions of employment, social security, industrial relations, labour law, training, education, and the labour market), legislative affairs, small and middle sized companies, communication, and international affairs. VIFKA was the first umbrella branch organization to negotiate on behalf of the employers regarding the ICK collective agreement. The software and services branch, which in the Netherlands became significant during the late 1960s, founded their own branch association in 1971. The Association for Computer Service and Software Bureaus (Vereniging Computer Service en Software Bureaus) started out relatively small, but had close to 50 members during the first half of the 1990s. In 1990 this branch association dubbed itself COSSO. COSSO remained strictly a branch organization, and unlike VIFKA, did not have ambitions as an employers association. COSSO restricted itself to protecting the commercial interests of its members and not the social interests (Smit, Schilstra, and Pauwe 1995, 88). COSSO and VIFKA Informatica started working together in 1993 and this led to a full merger in 1996. The branch organization then became known as the Federation of Dutch IT (Federatie Nederlandse IT, FENIT). Through this merger FENIT immediately became the largest branch organization for the whole IT industry. FENIT, however, would have nothing to do with social affairs and restricted itself to looking after the commercial interests of its members and strengthening the good reputation of the IT industry. 4 The reason for VIFKA Informatica to leave VIFKA had been the structure of VIFKA. VIFKA s structure was that of an integral organization where all members were part of the umbrella organization with secretarial offices for each branch. VIFKA Informatica wanted to preserve the identity of a branch organization, not that of an umbrella organization (Lievense 1998). VIFKA had been operating as an employers organization, whilst the associated organizations had been operating as branch organizations. During 1997 FENIT, VIFKA OPS, and VIFKA Telecommunicatie had been negotiating a new umbrella organization that included both FENIT and VIFKA. The organization was meant to become the personification of unity in the Dutch IT industry. As negotiations progressed however, FENIT increasingly became aware that this was not what its members - predominantly a segment of the older 4 As stated by FENIT representative Machiel the Rooij in 1998 (Lievense 1998).

Industrial Relations in the Dutch IT Industry 81 COSSO members, which comprised the larger software and services firms - wanted. FENIT s members desired a federative structure, whilst the two collaborating VIFKA branches wanted a more integral structure as VIFKA had been. Finally, in 1997, VIFKA OPS and VIFKA Telecommunicatie, without FENIT, founded a new employers organization named Dutch ICT Association (Vereniging ICT Nederland, V-ICTN). Both parties, VIFKA and FENIT, made accusations against each other. The latter claimed there were going to be differences in interest between the various branches. As an example the Millennium problem was given, which was supposed to polarize the software and telecommunications branches. The latter was predominantly a user of IT, whilst the former was predominantly a producer of IT. A conflict of interest was bound to stem from this. The secretary of VIFKA Telecommunicatie at the time saw matters differently and regarded the whole problem as stemming from cultural differences between VIFKA and FENIT (Lievense 1998). When VIFKA Informatica left VIFKA as a federative member, it remained a regular member of VIFKA, and when V-ICTN came about, they became a member. VIFKA Informatica s members, predominantly hardware IT companies, were dependent on the services of V-ICTN as an employers association. Leaving V- ICTN would mean they would loose access to the ICK collective agreement. FENIT did not provide services in this respect and was not going to provide them in the future. So VIFKA Informatica had to remain member of V-ICTN at the same time. V-ICTN s commercial and lobby activities, however, were aimed at the Telecommunications and Office and firm interior branches. Extending the coverage of these activities to include the IT hardware branch would imply direct competition between FENIT and V-ICTN (Lievense 1998). The two associations, V-ICTN and FENIT, would not become direct competitors though and eventually overcame their differences. During 1999 an agreement was made in which the two parties stated that they would discuss common problems with each other on a regular basis. Both parties felt that the IT industry should improve the way it had been looking after its interests. Both organizations also promised to be more open towards other branch organizations such as the Association of Cable Companies (VECAI), Association of Suppliers of Professional Telecommunication Equipment (Netelcom), and the Branch Association of Dutch Internet Providers (NLIP). 5 Despite FENIT providing special services for smaller IT companies, 6 typically companies employing three people at the most, two new branch organization for smaller IT firms were founded. I.T. United came about in 1997 and the Association of Independent IT Specialists (Vereniging van Onafhankelijke ITspecialisten, VONIT), in 1998. Both associations were founded because of unclear criteria regarding the independency of self-employed IT specialists. A lot of self 5 Computable, ICT-Verenigingen weer op een lijn, March 9, 1999. 6 These special services included reduced contribution fees (25 percent of the regular fees) and offering a special starters zone, where smaller companies were given a mentor from a larger company.

82 Developing Institutions employed IT specialists were regarded as employees by the government and had to pay social security charges. They protested against these charges claiming they were not employees, but rather self employed. VONIT also tried to increase the number of successful acquisitions for smaller firms. By bringing smaller IT firms together, they would stand a better chance of acquiring larger customers that were used to doing business with larger IT companies. In 2002 VONIT and I.T. United merged together and named themselves I.T. United. 7 The principal goal of this new collaboration was to enforce their position as self employed IT specialists. 4.1.3 Nederland-ICT In 1999, Annemarie Jorritsma, then minister of Economic Affairs, reproached the IT industry for the lack of being able to speak with one voice. Henk Broeders, the chairman of FENIT and managing director of Cap Gemini at that time, shared Jorritsma s criticism regarding the fragmentation of the Dutch IT industry, which should present itself as one entity to the outside world. Broeders admitted though that at the same time IT was so intertwined with the outside world, that presenting itself as a whole would be very difficult. Broeders wondered what it really was what the IT industry does. The range of activities in the IT industry is so wide that it is hard to put under one umbrella. Despite the difficulties and complexities involved, Broeders claimed that a new umbrella organization would come about in the future, and that V-ICTN and FENIT would be the founding organizations. 8 Broeders claims became, to some extent, a reality in the form of Nederland- ICT, a federation established in 2001 and comprising three branch associations, FENIT, ICT Telecommunicatie, and Vifkantec, and two interest groups, Werkgeversvereniging ICT and ICT Environmental (ICT Milieu). 9 The three branch associations stem directly from their predecessors: FENIT and V-ICTN. ICT Telecommunicatie was previously known as VIFKA Telecommunicatie and Vifkantec was previously known as VIFKA OPS. The two interest groups also stem from previous activities of V-ICTN, which used to operate as an employers association. Within Nederland-ICT these employers association functions have been moulded into a foundation and an employers association. The foundation ICT Milieu is set up for all IT companies that have to meet regulations regarding the recycling of packaging and equipment. The employers association Werkgeversvereniging ICT is responsible for the ICK collective agreement. 7 The original plan was that also the Platform Zelfstandige Professionals in de ICT binnen het Nederlands Genootschap voor Informatici (Platform for self-employed IT professionals within the Dutch Fellowship of IT Professionals), or NGI-ZIT, but in the end this association remained within the NGI (Automatisering gids, Zelfstandigenclubs gaan samen verder IT-United, VONIT en NGI-ZIT fuseren na de zomer, June 10, 2002). 8 Computable, FENIT en Vereniging ICT Nederland kondigen krachtige lobby in Den Haag aan, November 5, 1999. 9 So in the end, the FENIT got what it wanted all along: a federative structure and not an umbrella organization. The umbrella structure of the V-ICTN had been the greatest objection of FENIT against joining the V-ICTN.

Industrial Relations in the Dutch IT Industry 83 During 2001, Nederland-ICT experienced difficulties defining its role in the Dutch society as a whole and in the IT industry in particular. Voices were raised that the IT industry itself was losing sight of its role as so many problems seemed to arise for which information technology should provide a solution. Branch organizations are not equipped to deal with the potentially increasing demand for their services. This is an effect of the growing reach of IT, which increasingly encompasses most aspects of society. Nederland-ICT wondered whether questions such as solving the problem of long waiting lists in the health care, should be put on the plate of a branch organization. Nederland-ICT thought and still thinks these questions should be dealt with by a specialized council for the use and implementation of information technology, which they dubbed the ICT council (ICT-Raad). Despite considerable lobbying throughout the years by both Nederland-ICT and the EPN, 10 an ICT-Raad still did not exist in 2005. A specialized IT-Council for Organizations Fighting Disasters (ICT-Raad Rampenbestrijding) has been planned but politicians are afraid an IT-Council will undermine their authority and therefore want first to make sure this is not the case. 11 4.1.4 Werkgeversvereniging ICT The Werkgeversvereniging ICT serves the interests of companies that fall under the scope of the ICK collective agreement, members of both ICT Telecom and Vifkantec, and companies that voluntarily use the ICK collective agreement. This encompasses a total of 31,000 employees working in 200 companies. Besides the ICK collective agreement, the Werkgeversvereniging ICT serves the interests of its members in a number of other ways. They operate a helpdesk for answering questions with regard to collective agreements, social legislation, employee participation, reorganization, layoffs, et cetera. They also provide a number of framework agreements covering different areas such as day care, workplace ergonomics, company cars, et cetera. Most of these framework agreements provide benefits aimed at smaller businesses. A number of committees and study groups are active within the Werkgeversvereniging ICT: the policy group social affairs, which is the think tank for the board; the project group branch pension scheme, which is a work and study group for the development of the branch pension scheme; 10 EPN is the abbreviation of Electronic-Highway Platform Netherlands, a non-profit foundation, established in 1994 in order to gain a "firm and balanced foothold for the electronic highway in Dutch society", not only for the benefit of the business community but also for that of the general public. Within EPN businesses, governmental organizations, local governments of the larger cities, politicians, trade unions, employers' and consumers organizations, and universities work together to reach this goal (EPN website (1997 and 2004)). 11 Computable, Pleidooi voor ICT-Raad Nederland ICT worstelt met conflicterende eisen van de samenleving, November 30, 2001. Automatisering gids, Breed pleidooi voor nationale ICT-Raad, week 23, 2002. Automatisering gids, Afgezwakte vorm voor ICT-raad Rampenbestrijding, week 4, 2004.

84 Developing Institutions the standing committee, a committee for solving problems regarding the ICK collective agreement; a negotiations delegation, a group of people that conduct the negotiations over the ICK collective agreement and which monitors the progress of the projects that result from the agreements made in the collective agreement; an editing committee, which translates agreements made during the negotiations over the collective agreements into new collective agreement texts; the collective interest fund, which deals with special projects that stem from agreements made during the negotiations. It manages the financial account; the terms and conditions of employment committee, which encompasses a group of employers whose wishes regarding the negotiations of the collective agreement are heard and communicated; the project group Budgetplan, consisting of a group of employers that investigate the ways in which individuals can be given a personal development plan, creating a budget connected to an individual s life phases; the project group competence management, a group of employers gathering information regarding competence management and seeking to apply this knowledge in the ICK collective agreement; and finally, the project group development of complementary function profiles, a group which develops function profiles and connected wage scales. A final aspect of the Werkgeversvereniging ICT activities is to check whether its scope of action overlaps that of other branches. This is a constant process of determining whether association with other industry collective agreements is required. The IT industry s scope of business is extensive and as such its borders are under the constant scrutiny of other industries such as the graphics industry. The Werkgeversvereniging ICT is in permanent negotiations with other branches over the delineation of the ICK collective agreement, and to keep it as clear as possible. Other bordering branch organizations include the Federation of Technology Branches (Federatie van Technologiebranches), FHI, the Dutch Association of Reprographic Companies (Vereniging voor Reprografische Ondernemingen Nederland, Repro). The Werkgeversvereniging ICT describes the culture in the IT industry as individualistic. Not all members of Nederland-ICT, especially not the members stemming from the software, services, and telecommunications branches, want a sector level collective agreement such as the ICK collective agreement. Members of Nederland-ICT are, however, free to use the ICK collective agreement should they wish to.

Industrial Relations in the Dutch IT Industry 85 Although the services provided by these branch organizations seem extensive, 12 a great number of IT companies still see no benefit in joining a branch organization. Becoming a member of a branch organization feels like a waste of good money to these firms. Branch organizations seem unable to make themselves attractive enough for these firms simply because they are unable to convince these firms that their investments will be returned through the activities of these branch associations (Lievense 1998). 4.2 Unions and Works Councils in the Dutch IT Industry 13 Union power stems mainly from the number of people unions are able to organize. Through collectivism, unions seek to counterbalance the power of the employers. The power of employees lies in the fact that they are able to refuse to work by means of a strike. Although calling out a strike is an individual right and can be called by any worker, only the unions have a special fund that serves as a financial backup for such actions. Works councils, for instance, do not have such a fund. The activities of unions are, not surprisingly, quite comparable to those of employers associations: Consultation, negotiation and contract formulation with employers and employers associations (collective agreements, social plans, influencing the bargaining process through strikes, occupation of a company, etcetera.). Consultation, negotiation, and making agreements with the Government. Influencing public opinion through the mass media. Representation in advisory and civil / public administrations, and funds, at national and sector levels (e.g. training and education, and pensions). Advice, information supply, and other services to its members, executives, and works councils. Offering possibilities for participation and opinion formation (by means of a platform). Research and policy development. Coordination with other unions (Tros 2003, B1200-8). Negotiating the terms and conditions of employment has always been the major activity of unions and through the conclusion of collective agreements, unions are able to influence industrial relations. Originally, union influence depended on union 12 FENIT claims that its 200 branch members are responsible for 80 percent of the total gross annual turnover of the total Dutch IT industry (Fenit website 2004). Nederland ICT has more members, because they also include members from ICT telecommunications and Vifcantec. Recalling table 3.6, 17,940 companies operated in the IT services and manufacturing industry in 2002. This means that in terms of numbers, the representation of branch associations is very low. 13 This section is partly based on interviews with various (former) union leaders from the FNV, CNV, and Unie, working in the Dutch IT industry.

86 Developing Institutions density. Union membership, however, has slowly declined over the years. During the 1970s, membership amounted to 40 percent of the active working population in the Netherlands. Today this number has almost been halved to 23 percent (see Table 6.4). Also the average age of union members is rising. The current average age of union members is 46, 14 while the average age of the active labour population is 38. Regardless of these figures, the number of people that are covered through collective agreements, has increased to 85 percent (see Table 6.6). This can to a large extent be explained by the relative success of tripartite consultation, as was described in chapter two. Today, three unions operate in the IT industry : FNV Bondgenoten, which is part of the Federation of Dutch Trade Unions (Federatie Nederlandse Vakbeweging, FNV), and has a general/socialdemocratic character. FNV has 14 affiliated trade unions and a total membership of 1,226,000; CNV Dienstenbond, which is part of the Christian Trade Union Federation (Christelijk Nationaal Vakverbond, CNV), with 11 affiliated trade unions and a total membership of 355,000; and De Unie, which has no specialized background, offering their services to middle and higher educated employees. Unie has a special industry group for the IT industry. Within their IT industry group, De Unie discriminates between three different branches in the IT industry, which are computer services bureaus, the office machine business, and telecommunications. With regard to De Unie an additional remark should be made: De Unie is part of the Federation of Managerial and Professional Staff (Vakcentrale Middelbaar en Hoger Personeel, Vakcentrale MHP), to which are affiliated three individual trade unions and two groupings of unions, with a total membership of 196,000. Cooperation is one of the characteristics of the Dutch trade union movement, which has always known a high level of internal coordination. However, in recent years, cooperation between Dutch trade union movements came under the pressure of an increasing number of conflicts over both access to collective bargaining and unity in concluding collective agreements. Trade unions are increasingly seeking to enhance their own distinctive profiles, increasing rivalry between them, while the established unions are facing greater competition from newer independent organizations. 15 14 Table 6.4 is based on CBS figures. In order to be able to compare these figures to the average age of union members working in the Dutch IT industry, these figures have been corrected for age. Only those between the age of 15 and 64 have been counted. 15 Frank Tros, Competition mounts between trade unions, EIRO, http://www.eiro.eurofound.eu.int/2003/02/feature/nl0302103f.html (accessed on December 8, 2004). Tros continues to explain that there have been numerous case reports of individual employers and smaller sector-specific associations having signed collective agreements with 'independent' trade

Industrial Relations in the Dutch IT Industry 87 This lack of cooperation between unions is not apparent between the unions activities in the IT industry. They seem to be aware that breaking up the alliance will only weaken their collective bargaining position. Throughout the years, unions have used different strategies to get a grip on IT industry. Sometimes their vision collided, but on the whole they have managed to stick together, despite small rivalries over member numbers. 4.2.1 The Big Sleep: Union Activities in the Dutch IT Industry Until the 1990s Industrial relations in most Dutch industries have traditionally been characterized by consultation. This has already been discussed in detail in chapter two. The IT industry has always been an outsider in this respect. While collective agreements in the IT industry have been around since the industry s early days, 16 they are still not a prominent feature in the Dutch IT industry today. A lot of new companies entered the growing IT industry during the 1970s and 1980s. These companies quickly outnumbered the few IT companies employing a collective agreement in those days. 17 Unions were not able to develop an adequate strategy in response to the rapid growth of IT companies. Most of the new IT firms were either small, fast-growing start-ups, or foreign companies that were used to a different industrial relations setting, and simply went about their business as they did in their countries of origin. Also the relatively high wages that were paid in the IT industry have probably helped keep the unions outside. The Heidemij, a Dutch state owned IT company, paid its IT specialists higher wages than non-it employees, in order to be able to keep them. The reason for this was that Heidemij s IT professionals could easily leave the company and join another, or start one of their own, which in fact happened quite frequently (Van Dael 2001, 126-127). Therefore it seems unlikely that the contemporary unpopularity of unions in the IT industry has anything to do with a recent trend. It is more of an apparent feature of the industry s cultural and historical background. IT professionals generally are higher educated, are more assertive, individualistic, and have faith in the future. They are thought to be capable of negotiating their salary and fringe benefits, and they trust their employers with regard to pensions and other collective goods. According to a union leader this is mostly due to opportunism: union members first need to feel the effects of a reorganization or mass layoff before they are willing to appreciate the unions for what they, the unions, can do for them. The shakeout in the Dutch IT industry, which started in 2001, indeed seems to have had a unions, on which basis they have been able to avoid sector-wide collective agreements. Obviously this severely undermines the unions activities. 16 Two of the first IT companies in the Netherlands, RMA and Raet, had been using collective agreements since they started their operations in the industry. More on these companies in chapter five; in particular the Getronics and PinkRoccade cases. 17 See chapter three.

88 Developing Institutions positive effect on union membership in the IT industry; at least until 2003. 18 Since 2003 union membership has only been rising relatively and not in absolute numbers due to the amount of layoffs in the industry. 19 The initial effect of the early 2000s crash was quite similar to the one that occurred in the IT industry at the beginning of the 1990s. Both times, union membership in the IT industry increased right after the IT industry collapsed. 20 4.2.2 A Wake Up Call for the Unions: The Crash of the Dutch IT Industry at the Beginning of the 1990s Until the beginning of the 1990s, unions had had little influence in the Dutch IT industry. An interviewee stated that in 1980, there had once been talk of a collective agreement at IBM. At the time the conditions of employment were quite terrible at IBM in the Netherlands, and employees demanded a change in the form of a collective agreement. IBM opposed these suggestions and went even as far as writing threatening letters to specific people. Though the collective agreement never really became a topic of discussion, the works council at IBM was taken far more seriously since that time. The reason the unions started to get more influence in the Dutch IT industry during the 1990s was because of a shakeout in the first half of the 1990s. The reason for this shakeout was quite similar to events leading up to the recent crisis in the IT industry, which have already been mentioned in chapter three: a fall in demand for IT products and services. Everyone up to that point had been confident that information technology could solve most problems. In reality this had not been the case and when the crisis started in the 1990s, companies looked for means to save costs and decreased their investments in IT products for a while. This had devastating effects for the sector. During 1991-1994, many thousands of IT professionals were fired because of bankruptcies and reorganizations. Typical cases of companies that have disappeared were Infotheek and HCS. Major companies such as IBM, Raet, Digital, ICL, and Triple P had to lay off many hundreds of employees also. The unions played, just as they had in the period 2001-2003, an important role in the 18 Unions are usually not keen on giving exact data regarding member numbers in the Dutch IT industry and give general indications at the most. This is partly due to competition between the unions and partly because their records are sometimes not specific enough, which makes it hard to tell whether a union member is a member working in the IT industry. 19 This is not extraordinary, because traditionally union membership declines during an economic crisis. The reason for this is that union membership fees usually become too high to bear for unemployed people. 20 A union leader of the CNV claimed that between 1991 and 1995 their IT based membership doubled. Since 1995 things continued normally in the sense of a moderated growth in union membership in the IT industry (Automatisering gids, IT-sector moet op termijn gewone bedrijfstak worden, week 22, 1998). Also in recent interviews with union leaders, unions claim membership has been increasing significantly since 2001. Table 6.4 shows that unions may not be exaggerating. Union membership in the IT services industry, the least union favourable branch within the IT industry, is now approximately 12 percent.

Industrial Relations in the Dutch IT Industry 89 development of redundancy packages, which were predominantly aimed at outplacement trajectories. The unions suggested an association called Trans-IT, which was supposed to institutionalize outplacement trajectories and provide training and education, but this association was never founded. 21 The number of parallels between the early 1990s and the first years of the 2000s is striking if one looks at the rise in union membership and the causes and effects of the crises. Specialized IT professionals in the first half on the 1990s had incomes that were close to 45,000 Euro. When these specialists became obsolete, because of technological developments, they were asked to cut back their wages. 22 As has been described in chapter three, the same thing has been happening in this recent period. Two examples are AINO where a 20 percent pay cut was demanded and CGEY, 23 where on average a ten percent wage reduction was requested. 24 In both cases the unions advised employees not to agree with the salary reduction because it would not guarantee them their jobs and, in the event they got fired, they would end up with lower unemployment benefits. 25 The wake up call of the 1990s crash got the attention of the unions which were now seriously competing for IT professionals. Competition between the unions is a normal thing. All unions are dependent on union members and although the unions may focus on different groups of people, they are usually fishing in the same pond. This competition takes on different shapes and even includes strategic alliances in order to outmanoeuvre other unions. During 1993 CNV noticed that FNV and the Union for Higher Personnel (Vereniging voor Hoger Personeel, VHP), had made a private understanding that the two of them would conquer the IT market and would share the profits. Obviously market and profit refer to potential union members. The period of the early 1990s was after all a period in which IT professionals were looking for means to decrease their job insecurity. FNV and VHP had agreed that the VHP would get all higher educated IT professionals and FNV all lower educated IT professionals. This may have seemed like a bad deal for FNV, because of the number of higher educated IT professionals, but there were also a great number of lower and medium educated IT professionals working in the IT industry. In addition FNV thought that their chances with higher educated IT professionals were slimmer than those of the VHP, which made a special case of providing services for higher educated employees. The goal of the collaboration was to cut out the two competitors of that time, CNV and MHP. Both had members predominantly residing in the 21 Computable, Werkgever veronachtzaamt oudere IT ers, September 5, 1997. 22 A union leader mentioned figures that even went as high as 50 percent. 23 A pay cut was demanded because a collective lay off would be too expensive. If the employees did not accept a pay cit of 20 percent, Aino would go bankrupt (Automatisering gids, Aino-personeel levert salaris in, July 29, 2002). See chapter five, section 5.3.1.1, for a more elaborate discussion of what happened at Aino. 24 Unlike the Aino case, only a selection of employees was asked to cut wages. Predominantly those employees that were specialized in certain area that was not profitable anymore. 25 Which is based on 70 percent of the last received pay check. Unemployment benefits are levelled down in case 70 percent of someone last received daily wage is still higher than the maximum daily wage.

90 Developing Institutions middle segment. This alliance, however, was not an easy one to maintain. As soon as CNV found out about it and started to complain, it was allowed to join the forces. They did not regard CNV as a threat because CNV predominantly attracted members based on Christian ideology. In the end the alliance did not really matter as there were enough IT professionals joining the unions at the time. 4.2.3 The Heydays of the Unions in the Dutch IT Industry: 1996-1999 The period 1996-1999 was a remarkable moment in the history of the unions in the Dutch IT industry. Up until then there had only been a few collective agreements. In the hardware branch a sector level collective agreement existed since the late 1980s, but it was hopelessly outdated and did not accommodate the demands of the IT manufacturing industry well enough. A number of company level collective agreements had existed up till then as well: Bull, a French IT manufacturing company, had been using a collective agreement for more than forty years in the Netherlands; the State Computer Centre (Rijks Computer Centrum, RCC), which originated from the RMA and was a State owned computer centre, had been using the collective agreement for public servants since 1950; Raet, a software developer operating in the agricultural industry, had been using a collective agreement since it gained independence from Heidemij; 26 and finally, Agridata, another IT services company in the agricultural industry, was using a collective agreement for its 170 employees. In 1996 and the following years a number of new collective agreements came about. The first of the IT services companies that started using a collective agreement was Debis IT Services, which concluded a collective agreement in 1996. Debis IT Services used to be Fokker s IT department, but became independent after Fokker went bankrupt in 1996. Fokker had already been using a collective agreement, so it was not out of the ordinary for Debis to keep using a collective agreement. Today Debis IT services has been taken over by the German based T-Systems. Debis IT Services, however, was not a very big company. In the same year, little short of a revolution occurred. Getronics, a company that previously had been known as a fierce opponent of union involvement, 27 started negotiations with the unions over a company collective agreement. In 1997, the Getronics company collective agreement was finalized and implemented. The unions were on a winning streak, because in that same year Roccade also concluded a collective agreement after a two year delay. 28 In 1998, finally, two more collective agreements were concluded with Origin and EDS. Origin stemmed from a merger 26 See chapter five for a more elaborate discussion of Raet. 27 In 1993, many of the employees within Getronics were fed up with the bad social policy of their employer and called upon the unions. This has led to a fierce conflict because Getronics CEO, Ton Risseeuw, refused to speak with the unions. See the Getronics case study in chapter five for more information. 28 They had been following the collective agreement for public servants.

Industrial Relations in the Dutch IT Industry 91 between BSO/Origin and Philips C&P in 1996 and was using a variety of employment regulations, which it sought to harmonize. EDS was a daughter company of General Motors and at the time was in the process of acquiring the IT department of the Dutch railways, which already was covered by a collective agreement. 29 In retrospect, an interviewee asked himself whether this sudden increase in the number of collective agreements had not been a simple case of la dernière mode. The IT industry has always been a trend-sensitive industry and even dogmas with regard to personnel management and unions can be subject to this. It is undeniable that having a collective agreement has been fashionable for some time, especially between 1997 and 1999 when the IT industry was booming business again. Stiff competition for employees between employers led to the believe that having a collective agreement could have positive effects on attracting potential employees. Employers were keeping a close eye on the conditions of employment other employers were offering. During 1999 for instance, Roccade, Origin, EDS, Bull, Datelnet and Getronics were studying possibilities of introducing a choice model into their respective agreements, a means for employees to buy or sell days off. 30 4.2.4 The Development of Union Strategies Since the 1990s Obviously collective agreements were taken more seriously than a fashion trend. Using a collective agreement would mean having to deal with unions, something about which employers in the IT industry were still not keen. The reason that several employers have made the transition to a collective agreement was in part due to the unions overcoming a number of major obstacles that included employers, employees and works councils, and themselves. FNV was one of the first unions to take the IT industry seriously in the first half of the 1990s. This was still very late though considering the age of the Dutch IT industry, which in the Netherlands, took off on a larger scale during the 1970s. Already in 1986 a plan de campagne was drawn by a FNV union leader, which informed other union leaders about the IT industry s special position. Until 1986, FNV had regarded the IT industry as part of the wholesale business. Somewhere around 1993 this changed, and two FNV unions, Manufacturing Industry Union FNV (Industriebond FNV) and FNV Services Union (FNV Dienstenbond), set out to expand their knowledge of the IT industry. 31 In 1993 a research bureau was asked by FNV to make a explorative and descriptive study of the sector, its employers, and 29 In chapter five a more detailed analysis will be given regarding the origins of these collective agreements. 30 Computable, Invoering flexibele CAO blijkt lastig, April 2, 1999. 31 Predominantly driven by the need to do so. Many IT companies were doing very poorly and the unions were often consulted.

92 Developing Institutions its employees. 32 FNV was not the only union to take the IT industry seriously though. A few years before, in 1987, CNV had also commissioned a study, 33 but it was very much outdated in 1993. CNV also commissioned their own study in 1993. 34 The unions did not know of their respective studies until they were published. The definition of the sector that was used in FNV study was very wide and included the implementation and production of hardware, software, and services. A detailed description was given of the activities of the various branches within the industry, which informed union leaders what the IT industry was all about. The study provided detailed information regarding the state of the terms and conditions of employment in the various branches. Most of the branches were doing poorly at the time, but the hardware branches were having especial difficulties. The study reported that employees in the hardware branches were more open to unions. Even works councils were inclined to call in the help of the unions. The problems they were having included work pressure, unclear job descriptions, career, and long travelling times. In the software branches the problems were slightly different: part time work, pension schemes, pressure of work, and unclear function descriptions. The study made the following suggestions: Unions should familiarize themselves with the developments in the branch; Unions should adapt to the culture of the IT industry; Unions should treat IT professionals differently from employees in traditional industries; Unions should focus on themes such as: pressure of work, child care, parental leave, and early retirement (STZ advies & onderzoek 1994). Through these studies, unions had a clear understanding of what was expected of them if they were to be taken seriously in the IT industry. It was now a matter of shedding the old skin and starting to wear a suit. 35 Apparently the unions succeeded in changing their approach, because in the years to come they were quite successful in terms of concluding collective agreements. Changing had not been very easy for the unions though, and there were also other internal problems. Some people within the unions felt that the union leaders within the IT industry were making too many compromises. Performance related pay was one of the hottest debated topics within the unions and the biggest concession they had to make. To unions, performance related pay brings back bad memories of piece-wages. Another source of conflict within the unions was the fact that unions leaders operating in other industries were 32 Bouwman, T. Van De Camp, A. & T. Tom (1994), Net werk. Over werken in de automatisering, Amsterdam: STZ advies & onderzoek i.s.m. FNV Dienstenbond en Industriebond FNV. 33 Kant (1987), Informatici en vakbeweging, Nijenrode, in opdracht van CNV. 34 The CNV had used the same subsidy by the Ministry of Social Affairs and Sciences as the FNV had received, and also used it for a staudy. The CNV, however, used their own research centre, rather than an independent one. 35 In the survey, which was also a part of the FNV study, respondents answered that wearing a suit would have positive effects on the unions acceptance in the IT industry. It was part of adapting to the culture within the industry.

Industrial Relations in the Dutch IT Industry 93 complaining that employers in those industries were jealous of the flexible approach of the unions in the IT industry. These employers wondered why it was not possible for them to introduce performance related pay. The conflicts between union leaders (sometimes even representing the same union) and their rapid successions was very confusing for employers having to deal with the unions. In may 1996, two FNV unions, Industriebond FNV and FNV Dienstenbond, started a new collaboration project which they called Industry Group Information Technology and Electro Technology (Bedrijfsgroep IT/ET). This project was a test case for both unions in the light of the institutional merger that was about to take place in 1998 and form FNV Bondgenoten. The goal of this collaboration was to strengthen the position of FNV in industries that were important to the unions. The strategy was to increase the number of union members that they had thus far been unable to reach through traditional methods. During this period FNV wanted to examine the potential of both the Internet and a new manner of providing so called two way -information. 36 In order to set their agenda for the coming years and to think of strategies, a new study was commissioned. 37 Despite the relative success of unions at this time, a great number of white spots - companies that refused to talk to the unions in IT industry - still existed. Membership numbers in the IT branches still lagged behind national figures. This study looked similar in design to that of 1994 and reported on many of the same subjects. The suggestions made by this study were somewhat different. This was in part due to the fact that one of the study s objectives was to determine whether it was a logical step for FNV to merge IT and ET together, because the two branches were very different. The study reported that the merger was a logical step. Both unions, however, should retain a sector approach and not a regional approach, at least in the coming years. The value of the merger lay in the fact that ET often crossed the boundary to IT and it would pay off if union leaders kept informing each other of the movements in the two industries, between which, according to the study, obvious synergies existed. The study indicated a few weaknesses in the unions and made some suggestions. Increased involvement of union activists in the IT industry was required. This would, however, lead to a greater demands on union leaders who already had to deal with the job pressure of their regular jobs. Coaching of individuals would also become more important. For this a union leader s skills should shift from negotiating to consulting. An interesting remark was made regarding the nature of the problems of IT professionals, which may not always appeal to union leaders. Union leaders will have to ask themselves whether they want to have to deal with the working time problems of 31 year old, unmarried men earning a triple average national income. Again union leaders were advised to increase their knowledge of developments in the IT industry. Dealing with someone who knew what he was talking about, was a 36 By this the seminars were meant that the FNV organized during 1999 and 2000. These will be discussed below. 37 Lambregts, J.W.P., Niehoff, P.E., Poortman, J. & Suijkens, M. (1998), Rapportage onderzoek: bedrijfsgroep IT-ET, Utrecht: Basis & Beleid.

94 Developing Institutions strong issue for IT professionals. In light of these conclusions a number of suggestions were made: define yourself, tell employees what unions have to offer; exploit unrest and discontent in companies (e.g. during reorganizations); actively deploy union activists in recruitment activities. Union activist were an important source of information and unions should include union activists in their strategies. Union activists could become the ambassadors of the unions within their respective companies; negotiate a sector level collective agreement; use works councils as an entry point; increase the visibility of union (leaders), even in smaller companies; recruitment campaigns; pro actively pay attention to developments in the IT industry and try to adapt policy to these developments. Unions should be more flexible with regard to these developments. Finally an advice was given regarding the level of membership fees, which was supposedly too high for IT professionals. 38 The IT-ET project ended in 2000. In the spirit of the 1998 study, a number of activities had been developed during the project. Seminars were held with a wide variety of topics, such as the participation of women in the IT industry, 39 job specifications, and teleworking. A study based on a survey for IT professionals was commissioned. 40 This study was not an internal study as the 1994 and 1998 studies had been, but was targeted at a greater audience to highlight the unions position amongst IT professionals. 41 A special website was built, targeted at the IT industry, and, during 1999, a number of brochures and other materials were developed to inform IT professionals. 42 The effects of this campaign were quite noticeable. The website was visited very frequently and had attracted a considerable number of new members. 43 The occasions on which FNV had been in the media also positively affected membership numbers in the IT industry. This meant that FNV s strategies had been working. Other unions at this time were operating in the shadows of FNV in terms of union membership and number of available unions leaders. Table 4.1 shows figures with regard to unions size and union density. These figures demonstrate that, in terms of size, FNV has been dominating union presence in the IT industry for a long time. The reason FNV had more union leaders available had to do with the number of members, but also with the IT-ET project. Due to this project two additional union 38 Certainly an interesting advice considering the fact that the triple average national income remark was made on the same page. 39 This seminar was organized in light of a research project the FNV had commissioned. See Van Den Brekel, C., Van Klaveren, M., and Tijdens, K. (1999), Absence of women in the ICT-sector, Amsterdam: AIAS, UvA. 40 AIAS (2000), Werken in de ICT-sector: onder welke voorwaarden?, Amsterdam: Computable/FNVpers. 41 The number of respondents was 1689, but included a lot of union members because not solely subscribers to the Computable magazine were selected, but also union members known to work in the IT industry were approached, which may have biased the outcomes of the study. 42 http://web.archive.org/web/*/www.ict.fnv.nl (accessed March 4, 2004). 43 Figures may not be disclosed.

Industrial Relations in the Dutch IT Industry 95 leaders had been added to the already existing pool of union leaders, which was two to three times larger than that of the other unions. This does not mean that other unions were completely invisible during this period, but in terms of sheer size, capacity, and financial backing, they were simply outnumbered. Table 4.1 demonstrates that, in comparison to the other unions, FNV has managed to win the most members during this period, predominantly at the cost of Unie and other unions. 44 Though, relatively spoken, the tips may have scaled between the unions during this period, FNV lobby undeniably also had positive effects on the other unions recognition. The acceptance of performance related pay and choice options in collective agreements could not be attributed solely to FNV. Performance related pay and choice options were only accepted because employers would have never accepted an agreement without them. The unions simply did not stand a chance in this matter. The only thing they could bring in against these terms was the way they were implemented in the collective agreements. By accepting pay for performance, unions had managed to overcome their own reservations; by accepting choice models, they had managed to give a place to the individual in collective agreements; and finally, by concluding a number of key collective agreements in the IT industry, they had managed to gain the trust of at least a few of the employers. Table 4.1 Relative size of unions and union density in the Dutch IT industry 1 1987 2 1998 3 2003 4 FNV 41% 44% 67% CNV 11% 10% 17% Unie 5 27% 33% 15% Other unions 21% 13% 2% Union density 8 to 9 % 6 N.A. 7 12 % 1 Figures are based on three totally different sources, and give at most an indication of actual relative sizes. The ways in which the 1987 and 1998 data were gathered were based on interviews and a survey amongst union activists. Unions, however, seem to highly value these figures, which stem from union activists estimates. 2 CNV study by Kant (1987). 3 FNV study by Basis&Beleid (1998). 4 Berenschot (2003). Figures for the IT services industry. These Berenschot figures, which are based on an Internet survey amongst approximately 1300 IT professionals, will be discussed in greater detail in chapter six. 5 The 1987 figures are that of the MHP, the federation of which Unie is a part. The 1998 figures are the combined figures of the VHP and Unie. In 2000 the VHP and Unie have merged to become one union. 6 This was a CNV estimate. 7 No figures were given in Basis&Beleid (1998). 44 The Unie, however, states that between 1998 and 2003 the number of Unie members in the Dutch IT industry has practically doubled. This would mean that the FNV and CNV have had even higher growth rates. Obviously it may also be due to an inaccurate assessment by the Unie or validity problems with the Internet survey on which the figures for 2003 in table 4.1 are based. In chapter six the validity of the Berenschot Internet survey will be discussed in greater detail.

96 Developing Institutions 4.2.5 2001-2004, A New Crisis in the Dutch IT Industry: Hello Works Council, Goodbye Unions? Since late 2000 it became apparent that, again, the Dutch IT industry had to deal with a decreased demand for their products. In the following years the number of vacancies dropped staggeringly and IT companies reorganized as if there were no tomorrow. The crisis has been having two major effects on the unions: on the one hand unions were asked to help draw up social plans, and on the other hand unions were taken less seriously. Due to a decreased demand for IT products and services, employers increased their negotiating power at the cost of both employees and unions, 45 which changed the nature of negotiations over collective agreements. Employers had less money to spend and unions were not about to give in so easily, so both parties were forced into showing their teeth. This has led to increased tensions between employers and unions in the IT industry. The contacts between employers and employees have been established in a period in which periodic wage increases were affordable. During this period the unions dared to make performance related pay agreements as the industry was doing so well that most people would be getting periodic raises anyway. Now that unions were starting to act in a more traditional way once again, employers felt compromised and were wondering why they had to listen to the unions demands. 46 Their members were predominantly the ones they had been meaning to get rid of anyway. The relatively increasing number of predominantly older members had been sort of an adverse selection process. Younger people were more flexible in cutting their losses and figured they d be seeing more prosperous times again. Older IT professionals, however, were not as flexible and urged the unions to take a more direct and urgent approach. AWVN states it was, and is, not pleased with the events at the time. They prefer negotiating with the unions, because they are a strong negotiator and both organizations are familiar with each other s agenda. Negotiating with a works council is less attractive to AWVN as real professionalism cannot easily be found within companies. The relation between works councils and unions has always been an ambiguous one. When the WOR was introduced in 1950, 47 the Dutch Government s goal was to improve the functioning of companies by increasing collaboration 45 Most recent figures (2003) indicate that in the industrial category financial and business services (under which the IT services industry is categorized) the number of jobs that was lost was 39,000 (CBS, Aantal banen bedrijfsleven daalt verder, http://www.cbs.nl/nl/publicaties/persberichten/2004/pb04n032.pdf (accessed March 4, 2004). 46 See for instance NRC, Vakbonden nog onmisbaar, August 6, 2002. In an online survey of the Automatisering gids, visitors could respond to the suggestions of VNO-NCW of using works councils as negotiation partners in the future. Fifty-one percent responded they wanted to negotiate over their terms and conditions of employment individually, 17 percent agreed with the proposal, and 32 percent indicated it wanted the unions to do the negotiations for them (Automatisering gids, Wie kan het beste onderhandelen over arbeidsvoorwaarden?, August 16, 2002). 47 See chapter two for a more elaborate discussion of the Wet Ondernemingsraden (Works Councils Act).

Industrial Relations in the Dutch IT Industry 97 between employers and employees. This would have had serious effects on unions, were it not that works councils were not allowed to negotiate over subjects that were the major domain of unions such as wage and working time. Regulations which are made by negotiation parties of a collective agreement can and may never be superseded by agreements made by the works council. There are a number of arguments for this: The union is more experienced and has more knowledge with regard to negotiating. Many works councils know from experience that negotiating well is not easy. A frequent counter argument heard in the IT industry is that most IT professionals are more highly educated than industrial workers and are therefore more likely to be able to negotiate for themselves (or for colleagues for that matter); Unions often know more about the general situation in the industry or branch than a single works council. Though this may be true in other industries, the usual response from the IT industry is that this is not the case. Also a number of works councils in the Dutch IT industry have set up a discussion platform for themselves, the SOI, which will be discussed below; The unions are independent from the industry, branch, or company with which they are negotiating. Union leaders do not have to deal with the fact that the managing director is not only a party at the negotiation, but also their employer. This holds valid even in the IT industry where IT professionals are the main source of income and not product they are working on; 48 Works councils take an ambiguous position. On the one hand they are supposed to promote the interests of employees. On the other hand, though, they are supposed to promote the business of the employer. These two functions may contradict one another at times; As has already been mentioned at the beginning of this section, unions have a means of power not available to the works council, namely the availability of a strike fund. To date however there have been no official strikes in the IT industry; 49 A works councils composition is not constant and its functioning is very much dependant on individuals. From this, one can conclude that a works councils strength may be subject to change, both for better or for worse. Unions also have to deal with changes though. Unions operate through union leaders, which means relations are often established on a personal basis; Companies employing fewer than 50 employees are not required to have a works council. Moreover, there is evidence that a lot of IT companies that 48 This is the biggest difference between service providers and software developers. The former type of IT professional earns money for the employer by charging hours to a client. The more hours he charges, the more money is earned. The latter type of IT professional earns no money for the employer by working extra hours. The product that is made and sold is making the money. 49 Though the CNV, Unie, and FNV have held actions at Getronics and the CNV at CMG.

98 Developing Institutions should have a works council, based on the number of employees in their organisation, do not have one (Van Het Kaar & Bijlsma 1999, 154-155). 50 Since 1998 works councils are able to make more extensive agreements with their employers in the form of works council agreements. 51 These works council agreements do not have the same status as the collective agreement, and employees are only covered by the works council agreement when this is stipulated in the individual agreement. In a 1994 SER-advise simplification of the coverage by a works council agreement was proposed. 52 If regulations from a works council agreement would be extendable to include all employees in a company employing this type of agreement, employers would not have to negotiate individually anymore. Unions obviously opposed to these proposals as it would seriously undermine their current position. To date these proposals have not been accepted. Tros (2000a, 69-71) concludes that this may have had something to do with the Government s hesitation with regard to making changes to the system of industrial relations. They highly value the effects of collective agreements with regard to industrial peace. Unions in the IT industry admit that works councils are often quite capable and professional and consider works councils to be a valuable partner, which they will help in any way they can. This seems paradoxical, as works councils are more often than not direct competitors of the unions. Moreover, works councils are used as an excuse by employers not to have to negotiate with the unions. Their claim is often that unions do not represent the workers of the firm, while the works council does. There are also some works councils that simply refuse to talk to the unions. 53 Unions, just as the employers, try to use the works councils for their own goals when given the chance. Unions know that getting union members to join the works council of a company will increase their chances of ending up meeting the employers in these companies. Unions have two strategies to interest the works councils. They try to get union members to join the works council, 54 or they try to help the works council in the hope that this will establish a good relationship. FNV, for instance, has developed a special information package for works councils in the IT industry and has also given special courses for works councils. 55 During 1993-1995 two works councils platforms were established. One for the hardware industry called Sector level Consultation for Works Councils in the IT industry (Sectoraal Overleg Ondernemingsraden in de IT-branche, SOI), and one for the software industry called Software Works Council Platform (Software 50 In the Dutch IT industry in 2000, of the 480 companies that are supposed to have a works council, only 150 actually have one (Siefert, F., (2000), ICT er wil meer dan leasebak, in OR Informatie, p. 30). 51 This has also been discussed in chapter two. 52 SER (1994), Het arbeidsreglement; het instemmingsrecht van de OR, Den Haag: Sociaal Economische Raad. 53 See for instance the CMG case in chapter five. 54 Which in some companies does not work because union members are not voted into the works council. 55 Much to the dismay of Formaat, a periodical for works councils in the Netherlands, who saw the union s actions as a breach into their field.

Industrial Relations in the Dutch IT Industry 99 Ondernemingsraden Platform, SOP). The hardware platform has been around since the beginning of the 1980s and was originally set up to exchange information regarding popular arrangements such as the lease car. During that period the platform depended predominantly on the works councils of two companies (IBM and Bull). After a brief demise, the original SOI started up again in 1993, this time with a larger pool of participating companies. 56 In the past an attempt had been made to include the software platform into the hardware platform, but this was not a success and during 1995 a special software platform was created. 57 It was agreed was to send each others meeting reports and have people sit in both platforms meetings. The unions were also invited to join the works council platforms. This had advantages for both parties. The unions 58 could bring in expertise 59 and could introduce themselves to the works councils of the major IT companies of that time that they had not been able get in contact with. In 2004 the platforms merged into Platform for IT Works Councils (IT Ondernemingsraden Platform, ITOP), because the subjects of both platforms converged and companies that had traditionally belonged to the IT manufacturing industry had been gradually including services into their arsenal. This is why more and more companies where joining both platforms. In 2004 32 companies met each other every six weeks. 60 FNV and De Unie were still regular visitors of the platform at the time. CNV had not been attending the meetings for a few years. The reason for this was that they had been developing a new strategy to approach the IT industry. CNV had set up a, in the words of a union leader, unique structure that other unions did not have. This structure was based on two pyramids. The first pyramid was shaped by three layers of CNV union members and activists. The bottom of the pyramid consisted of union members in various companies that were able to anonymously report to an intermediate layer of union activists that ideally were members of the works council of the respective company. These union leaders then reported to a board of super union leaders for the IT industry. This board then reported to a second pyramid of IT union leaders of CNV. These union leaders finally reported to the main IT union leader of CNV. This was not a one way trajectory. From within the second pyramid all sorts of activities were organized for the first pyramid. The pyramid model is based on the philosophy that only union members should get support from a union. According to CNV this model suits the new generation of union members. Though CNV admits it has also been getting older members in the last few years, it claims its membership in the IT industry is younger 56 Unisys, Bull, Wang, IBM, Digital, SiemensNixdorf Informatiesystemen, Hewlett-Packard, ICL, Getronics, and NCR (Smit et al. 1995, 90). 57 Raet, BSO/Origin, Multihouse, Cap Volmac, Exact, RCC, Logica, Kramers & Ruijs, Datex, and Getronics (Smit et al. 1995, 90). 58 FNV, CNV, and VHP. 59 Which they did, as well as sometimes hiring experts to give presentations at the meeting. 60 CGE&Y, EDS, Exact, Getronics, Atos Origin, Pink Roccade, Compuware, Oracle, Logica/CMG, Ordina, IBM, NTNT Nederland, Sylis ICT, Altium, HP/Compaq, Fujitsu, NCR, Bull, Unisys, NEC, and Sioux.

100 Developing Institutions than the average union member in the Dutch IT industry. 61 Based on their social philosophy, FNV wants to negotiate for each and everyone in the sector and make no difference between members and non-members. According to CNV this is not a good strategy. One can wonder how much CNV s strategy is dependant on their shared Christian ideology. How many people not sharing this ideology feel comfortable with a union focusing on their ideology? They may have moved to become the second largest union in the IT industry but this seems to have been more the effect of FNV doing well, and De Unie doing not so well. FNV has been dealing with a number of problems since 2001, predominantly the poor financial situation. In order to save costs, the old structure, which was based on a sector level approach, has been replaced by a regional approach on a multi-sector level. On the one hand this increased the total number of union leaders available. On the other hand, however, these union leaders now have to deal with more industries, reducing the focus on one typical sector. An advantage of the regional approach is that union leaders have significantly reduced their travelling times, but at what expense? The sacrifices FNV has made go directly against the advices of the studies FNV has been commissioning, and FNV agrees that the changes have had negative effects for its presence in the IT industry. The period 2002-2004 has been dominated by reorganizations and social plans, something IT companies rather avoided negotiating with their respective works councils, and little time was left for the FNV to work on their name in the IT industry. Despite FNV s current situation in terms of membership numbers, it has not been doing too badly. Table 4.1 clearly shows the progress FNV has made in terms of attracting union members compared to other unions. It is strange that De Unie has lost their second place, because on paper they appear to be the most attractive union to IT professionals. De Unie specializes in middle and higher educated employees and concentrates on individual mediation. De Unie states it has been doing well the last few years and have not experienced the same problems CNV and FNV have been having of attracting predominantly older employees. According to De Unie, it has also been attracting younger higher educated members. De Unie is confident it can reclaim its second position in the IT industry and base their confidence on the fact that it has a better financial position than the other two unions in the IT industry, plus it is able to provide individual services. Also the way De Unie is organized, a De Unie representative mentions, allows them to be more effective per union leader than other unions. De Unie admits though that there is room for improvement in terms of reputation and familiarity. Like CNV, De Unie is not inclined to set up a collective strategy with CNV and FNV. Both CNV and De Unie fear such a collective strategy would benefit FNV the most. Moreover, De Unie wants to set a different tone than CNV and FNV. A tone, 61 Table 6.5 in chapter six seems to support this claim, but the original dataset is too small to make such a claim for a particular union. The number of respondents is too low to allow for very detailed categorizations such as union members by union and by age.

Industrial Relations in the Dutch IT Industry 101 which, according to De Unie, suits the demands of professionals in the IT industry better than that of the other unions. 4.2.6 Summary This section concludes the first part of this chapter where a historic account has been given of both the employers associations and unions operating in the Dutch IT industry. Nothing has yet been said about how these parties are related to the various types of agreements that exist in the Netherlands. Figure 4.2 shows that the unions are only directly involved in the case of company collective agreements and the industry wide collective agreement, the ICK collective agreement. Furthermore unions can advise works councils, or work together with them, but they are not involved in concluding other types of agreements besides collective agreements. Employers associations can serve as advisory bodies, but are generally not involved in industrial relations. Exceptions are AWVN, which advises individual companies, and the Werkgeversvereniging ICT, which negotiates the ICK collective agreement on behalf of their members. AWVN advises the Werkgeversvereniging ICT on the ICK collective agreement. Employees are only directly involved in one type of agreement, the individual agreement. These individual agreements may however refer to other types of agreements on which the employee s influence is absent or indirect through the works council or union membership. 62 Finally the employer s position in the Dutch IT industry is quite strong as the only agreement he is not able to negotiate over directly is the ICK collective agreement, which is not extended and does not cover the entire IT industry. In most cases the employer can call in the help of AWVN or Werkgeversvereniging ICT if he is a member of those organizations. 62 See also chapter two.

102 Developing Institutions Figure 4.2 Actors and the type of agreements in the field of industrial relations in the Dutch IT industry Employers AWVN Werkgeversvereniging ICT Individual agreement Personnel Guide Works council agreement Company collective agreement ICK collective agreement Works council Employees Unions 4.3 The ICK Collective Agreement Up till now the discussion of the ICK collective agreement has been postponed. The reason for this has primarily been that it is the only case where employers associations and unions have been, and still are, working together in the Dutch IT industry. 63 The ICK collective agreement is also a good case study of how the relationship between the unions has developed between 1993 and 2004. 4.3.1 Origins of the ICK Collective Agreement ICK stands for Information, Communication, and Office technology, and originates from the office machines and equipment branch. The ICK collective agreement originated in this branch and was known as VIFKA collective agreement, named after the 63 There have been a few other instances in which the unions and employers organizations have collaborated, such as the Centrum voor Arbeidsmarktvraagstukken ICT (Centre for ICT Labour Issues), or CA-ICT, which was set up in 1997 by a number of employers (including Roccade and Oracle), trainings institutes, employers associations, and unions. The centre manages the funds that come from the government for industry related education (Computable, Platform schoolt 200 werkzoekenden, October 3, 1997).

Industrial Relations in the Dutch IT Industry 103 employers organization VIFKA. In 1993 FNV, CNV, 64 and VHP, together with the Sector level Consultation for Works Councils in the IT industry (SOI), started thinking about extending VIFKA collective agreement, which at that time covered roughly 20,000 employees in the IT manufacturing industry, to also include the software and services industry. At the time the unions were only marginally successful at attracting members and only two company level collective agreements existed in the IT services industry. 65 VIFKA at the time was asked by its members to create a new collective agreement, which would protect them from the sphere of action of collective agreements of other industrial branches. VIFKA turned to the unions, who lacked a significant foothold in the sector as a whole and saw VIFKA collective agreement as an opening. In a discussion paper the plans for the future of VIFKA collective agreement were laid out. 66 The paper was based on informal negotiations between the WSOI and three unions (FNV, CNV, and VHP). The Study group Sector level Consultation for Works Councils in the IT industry (Werkgroep Sectoraal overleg Ondernemingsraden in de IT-industrie, WSOI), was a special study group of the SOI that sought to improve the current VIFKA collective agreement. Employers that participated in the WSOI and were working with VIFKA collective agreement agreed that it did not suit the demands of the IT industry very well. Therefore the interests of the SOI coincided with that of the unions who thought they could now (re)structure the industrial relations in the IT industry with a new industry wide collective agreement. The type of collective agreement proposed by the unions to the WSOI was called a multi level collective agreement (etage-cao). This type of collective agreement should meet two important requirements: 1. it should meet the demand for individuality; and, 2. it should offer the same advantages of scale as a regular collective agreement. The multi level collective agreement would house two types of agreements: 1. central agreements that are made with the employers organization; and, 2. decentralized agreements that are made on a company level. The multi level collective agreement would also be capable of offering room for individual choices to employees in the form of à la carte choices and custom-made agreements. 67 The multi level collective agreement thus represented a framework collective agreement on an industry level. According to the unions, this multi level collective agreement, should meet a number of their criteria on three levels. On an industry level the collective agreement should, first and foremost, have the status of a minimum agreement for all small companies in the IT industry. This can only be guaranteed by means of extending the 64 CNV and MHP were initially not invited to join, because of their presence in the Dutch IT industry at the time. 65 The previously mentioned Raet and Roccade Groep collective agreements. 66 FNV,VHP, and CNV (1993), Discussiestuk over de totstandkoming CAO IT-branche (Discussion paper on the realization of a collective agreements for the IT industry), August 9, 1993. 67 À la carte collective agreements imply that some elements of the collective agreement are optional. Employees can select, within boundaries, a number of elements that suit their personal preferences.

104 Developing Institutions collective agreement to the industry level (avv). Secondly, the collective agreement should be a framework agreement. Agreements that were to apply to the whole industry are agreements that have to do with old age (pension funds and early retirement), illness (prevention of, and social legislation), and minimal collective wage raises. Finally, the SOI should have an advisory role on an industrial level. On the company level, bargaining was to take place on subjects that have not been regulated centrally. These negotiations should have the status of collective agreement negotiations. Obviously the unions were aware of the fact that there would be a lot of companies that would be covered by an extended collective agreement. Because not all firms could bear the same costs, the unions proposed to negotiate over the minimum specifications rather than the desired specifications. The unions suggested an intermediary level between the company and the individual employee. The unions hereby meant to empower the works council of larger companies, who would then be able to negotiate over the optional elements of the à la carte collective agreement. The unions suggested a strategy to get the plans implemented. The unions proposed to follow one of the following three strategies: Extension of the sphere of action of the already existing VIFKA collective agreement, both in a qualitative and quantitative manner; Promoting company level collective agreements on the short term and harmonization of these in the long term; Creating a brand new industry wide collective agreement. The third option was favoured most by the unions. Predominantly, because they would not have to deal with the loose ends of the outdated VIFKA collective agreement, which was more suited to the old office machine branch than it was for the IT industry. They also favoured option three over having to deal with multiple company level collective agreements. These would cost a lot of time and still meant they would not reach their ulterior goal: one collective agreement for the entire IT industry. A modern flexible collective agreement suitable to the individually oriented IT industry. 4.3.2 The First ICK Collective Agreement: Conflicts, Compromises, and Disappointments A gap of almost four years spans the initial proposal of 1993 and the first ICK collective agreement, which was concluded in 1997. In the meantime the old VIFKA collective agreement would continue to be in effect. The agreement the parties were working to should cover the roughly 20,000 people working in the IT manufacturing industry. This agreement was then to be expanded to the approximately 80,000 people that were working in the entire IT industry. Employers and unions agreed that they would give each other one year in which the ICK collective agreement would be extended to the entire sector as was proposed. Both employers and unions knew that this would be difficult, because

Industrial Relations in the Dutch IT Industry 105 almost none of the software and services companies wanted to be covered by a collective agreement. As was written above when discussing the employers organizations, VIFKA did not have a lot of members and the members they had stemmed from different branches in the industry. This put VIFKA in an awkward position as it had to satisfy conflicting interests. A collective agreement that would not hold for the entire sector, but only for those who wanted a new collective agreement, was the best solution for VIFKA. What the unions wanted, an extension of ICK collective agreement, was also not possible because VIFKA simply did not have enough members. 68 In April 1995, an agreement in principle was made. The unions had high hopes of a successful conclusion. The following negotiations, however, turned out to be much more difficult than they had anticipated. The employers were not at all charmed by the idea that the new collective agreement would empower the works council, because it would mean they would have to negotiate specific topics with their works councils. As such, they refused to accept the flexible components of the agreement. 69 The employers also refused to make central agreements concerning overtime and shorter working weeks. As the differences between the unions and the employers seemed insoluble, VIFKA rejected the 1995 agreement in principle in October 1996. The unions were shocked and rejoined that they had been very willing and cooperative until then. 70 The reason, they said, that VIFKA rejected the principal agreement was that they were afraid it would hamper their merger with FENIT. 71 Four major IT companies (Cap Gemini, CMG, Origin, and Roccade) indicated they were not at all interested in a industry wide collective agreement. FENIT at the time indicated that the IT industry did not need a collective agreement, but rather new concepts such as flexible regulations with individual components, flexible employment of workers, employability, ongoing training, performance related wages, and an improved image of the industry. 72 The unions were also very surprised to learn that such a modern industry, which stresses the individuality of their employees, could have such a traditional outlook, and vote against introducing the proposed new concepts. 73 In January 1997, unions and VIFKA restarted negotiations. Though VIFKA was careful to use the term negotiations, the unions claimed they could no longer be ignored by VIFKA. As has been mentioned earlier in this chapter, the unions, after 68 Requirements for extending a collective agreements include majority rules. In order for stipulations of a collective agreement to be extended, they should, according to the Minister of Social Affairs and Employment, already hold for an important majority of the employees of a specific industry for which the extension is intended. A majority of 60 percent or more is always regarded as important. 69 Now named layered, rather than multi level, collective agreement. 70 They had agreed to longer working days if these were compensated as well as working on Saturdays. Also the proposed shackles with regard to wages, were loosened. 71 Independent of the unions accusations, the proposed merger was not going to happen anyway. This is however a matter of hindsight. 72 Internal Getronics study. 73 Computable, Vifka laat cao voor IT-branche afketsen Werkgevers zien niets in gelaagde overeenkomst, October 11, 1996.

106 Developing Institutions their first ICK fiasco, had changed their strategy and took a more company oriented approach. They did so together with the SOI. This led to a major success for the unions at Getronics, which until that time had been relatively hostile towards unions. Things were also looking very good with regard to EDS. 74 Because of these successes, the unions were confident of their future position in the IT industry. So far in this discussion of the ICK collective agreement, it has been assumed that unions are a homogenous group sharing similar ideals and ideas. As it turns out, the ICK collective agreement, at least the first one, demonstrated that this was definitely not the case. The first ICK collective agreement was only signed by the two smallest unions, CNV and De Unie. FNV and the VHP did not sign the agreement because of two objections. The first objection dates back to 1995, when it became apparent to the unions that the coverage of the ICK collective agreement would be far smaller than the unions had anticipated. VIFKA represented a much smaller proportion of the IT industry than they made the unions belief. The second objection for FNV and VHP stemmed from the manner in which they perceived the results of the negotiations over the ICK collective agreement. Unions are represented in an industry through one or more union leaders. The number of union leaders depends on the size of the union and the size of the industry. Because of their relative independence, union leaders have quite some power. Though unions may decide to a union leader back, union leaders determine to a large extent the strategies and dealings that go on in the industry they are representing. This was no different in the IT industry. In the case of a collective agreement, which involves multiple unions, union leaders will try to coordinate their actions, but they are by no means required to do so. As it turned out, this was also the case during the negotiations over the ICK collective agreements. FNV and the VHP, the two largest unions of the four unions that took part in the negotiations, generally regarded the results of negotiations more negative than did CNV. CNV also felt that FNV and VHP were listening too much to the demands of some of the SOI members that were comparing the results of the ICK collective agreement to the terms and conditions of employment that they were used to at their own employer, IBM, which offered better terms and conditions of employment than most other companies in the IT manufacturing industry. The reason for this was that both FNV and the VHP had introduced new union leaders and according to CNV this meant a loss of experience and insight. It also meant a greater dependence on union members residing in the SOI on behalf of the unions. CNV at one point decided that the process had already taken enough time and they had to think of their members, most of them working in smaller companies where the ICK collective agreement was most desired. CNV therefore started informal negotiations with VIFKA and soon, after CNV consulted its members, it became apparent that the negotiations could be formalized. Within a few months De 74 A more detailed account of the actual proceedings will be given in the next chapter. As it turned out, both Getronics and EDS had their own reasons for wanting a collective agreement. For instance, in the case of Getronics, the unions had to make relatively rigorous compromises to get the actual agreements drawn.

Industrial Relations in the Dutch IT Industry 107 Unie joined sides with CNV and the first ICK collective agreement was concluded. De Unie regarded the first ICK collective agreement as an infant that would grow in its coming incarnations. Though the ICK collective agreement was primarily targeted at hardware companies, 75 the hope remained that software and service branches in the IT industry would follow. In section 4.4.4 the original proposals that were made by the unions, will be compared to the first and 2003 versions of the ICK collective agreement. This will demonstrate just how much the ICK collective agreement has developed throughout its six incarnations until September 2004. 76 4.3.3 Extending the ICK Collective Agreement: A Mission Impossible? 77 As has been written above, already in 1993 plans existed to try to create a collective agreement for the entire Dutch IT industry by means of extending the revised version of VIFKA collective agreement. By 1995, the unions realized that VIFKA did not represent the entire IT industry and that extending the ICK collective agreement to the entire IT industry would legally not be possible because of the majority demand. 78 During 1997 an attempt was made to extend the ICK collective agreement solely to the IT manufacturing industry, but this turned out to be impossible because of the overlap with other industries. 79 This was, however, not to be the end of the story of trying to get the ICK collective agreement extended to the IT manufacturing industry. A second serious attempt was made after the second ICK collective agreement of April 1999 had come about. This second ICK collective agreement was now backed by all four unions. Both FNV and the VHP still thought the ICK collective agreement was a meagre one, but not being part of it would mean having no influence over it at all. One of their major sources concerning the current collective agreement was that it was too noncommittal. Companies still were able to use their own pay- and wage scale system, which gave them the right to not follow specific agreements contained in the collective agreement. 80 A stronger focus should be put on minimum arrangements, and new arrangements should be introduced dealing with part time work and day care. To some, this endeavour of making the regulations of the ICK collective agreement more rigorous was considered a risk as it would make the collective agreement less attractive to the IT software and services 75 In their definition, companies that sell, assemble, distribute, repair and/or maintain (parts of) hardware. 76 The first ICK collective agreement was valid from April 1997 until March 1999, the second from April 1999 until March 2000, the third from April 2000 until March 2001, the fourth from April 2001 until March 2002, the fifth from April 2002 until May 2003, and the sixth from June 2003 until September 2004. 77 See chapter two for an account on the proceedings with regard to extending a collective agreement. 78 See note 68 above. 79 The manufacturing industries insofar as related to the assembly of metal components and micro electronics and retail insofar as it is related to the selling of office supplies. 80 This would change in the 2000-2001 ICK collective agreement. There it was agreed that only companies who had a works council, were allowed to use their own pay- and wage scale system.

108 Developing Institutions industry. 81 Though this may have been true, it did not mean that such a move made the ICK collective agreement less attractive to other types of firms. During 1999 a number of telecommunications providers indicated they wanted to start using the ICK collective agreement. 82 The unions were positive that this second ICK collective agreement would be extended for two reasons. Firstly, they thought they had covered the problem of the ICK collective agreement overlapping with other industries, and secondly, the ICK collective agreement was now signed by all four unions and not just two of the smaller ones. Getting the ICK collective agreement extended would mean that all companies, including the ones which were not members of V-ICTN, 83 would have to accept the ICK collective agreement. V-ICTN was also optimistic that extending the ICK collective agreement would be successful this time. Not everyone at V-ICTN was equally pleased with the plans for extending the ICK collective agreement though. 84 There were, however, more advantages than disadvantages regarding collective agreements, according to V-ICTN representative Hans Keukelaar. He stated that a collective agreement can have organizing effects for the industry. It can reduce the growing complexity of the IT industry and it can raise the general quality of the terms and conditions of employment in the IT industry. Moreover, it will give the IT industry a certain prestige that can have a positive effect on the labour market. 85 The attempt of both V-ICTN and the unions to get the ICK agreement extended was not met with enthusiasm by FENIT, the employers association for the IT software and services industry. At the annual congress of the Federation of Technology Branches (FHI), 86 the chairman of FENIT at that time, Henk Broeders, spoke out against extension of the ICK collective agreement. Broeders stated that: This industry is illustrative for the post collective agreement age without extensions. He continued to argue that an extended collective agreement was too restrictive: The fast technological developments and the changing role of the employee demand a different approach towards terms and conditions of employment characteristic for traditional industries. At the same congress, the chairman of FHI made similar claims and stated that the freedom of choice and flexibility of employees cannot be laid down in collective agreements anymore. 87 In a response 81 The author of a news article in the Automatisering gids ( Onderhandelingen ICK-CAO van start, April 16, 1999). 82 Only larger telecom provided such as the KPN (Royal Dutch Telecom Provider), remained using their own collective agreements (Automatisering gids, CAO wellicht geldig voor hele ICK-sector, July 2, 1999). 83 During 1997 the VIFKA was replaced by a new umbrella organization for the IT industry, the V-ICTN (see earlier on in this chapter). 84 Most notably, the chairman of the V-ICTN at that time, Pieter Liefkens. 85 Automatisering gids, Snel AVV-aanvraag hardware-cao, December 17, 1999. 86 The FHI, the Federation of technology Branches, was mentioned in the first part of this chapter. The FHI represents firms in Industrial Electronics, Industrial Automation, Lab technology, and Medical Technology. 87 Computable, IT heeft geen behoefte aan algemene CAO: FENIT-voorzitter Henk Broeders is wars van onnodige keurslijven, December 10, 1999.

Industrial Relations in the Dutch IT Industry 109 the unions countered these statements by arguing that collective agreements have increased the capacity to accommodate for bespoke work, that they shape solidarity amongst employees, and that they are good for the image of the IT industry. 88 The unions, however, promised to narrow down the coverage to not include the companies covered by FHI. Despite the effort of both the unions and V-ICTN, an extension of the ICK collective agreement would not come about this time either. It turned out to be a useless endeavour as the ICK collective agreement they were attempting to get extended would only last for another six weeks. Though the unions and V-ICTN had been optimistic with regard to confining the coverage of the ICK collective agreement should it be extended, the negotiations, especially the ones with the parties to the various metal collective agreements, proved to take too much time to be able to successfully apply the agreement for extension. 89 A third attempt to get the ICK collective agreement extended was made at its third incarnation. This third version had a totally renewed salary system. Fifty-three of the most prominent functions are included, which are divided over twelve function families. For each function a minimum wage was set. Because the previous versions of the ICK collective agreement still allowed companies to use their own salary systems, roughly 100 of the 300 companies that were using the ICK collective agreement at the time would have to raise their wages, because according to the new salary system, they were not paying sufficiently high wages. A child care arrangement was also introduced. The unions, however, still felt this regulation was too meagre and only accepted it because it was a first step in the right direction. 90 The discussions over the coverage of the ICK collective agreement proved again to be very difficult and slow. A representative of an employers organization of a competing industry, the Association of Electro Technical Contractors and Retailers (UNETO), states that the reason for this is the incompatibility between the metal and technology industry and the IT industry. A common framework simply does not exist and parties use different definitions. If one party speaks of installation, configuration, or assembly, he or she often means something different than the other party. This is also due to the fact that the different industries are in constant flux. The goal of the parties is to mark out the different scopes of action in such a way that the whole process of demarcation does not have to be repeated each and every year. 91 As it turned out, however, the negotiation process would, yet again, take too much time. The unions put the blame on the two competing branches and stated that even a snail is faster compared to these two branches 92 88 Computable (online edition), December 8, 1999, IT heeft geen behoefte aan algemene CAO. 89 Automatisering gids, AVV hardware-cao voorlopig van de baan, February 25, 2000. 90 Employers reserved 0.1 percent of the total wage sum for day care for their employees, whereas according to the unions, a 0.2 to 0.25 was usually required. No special collective day care fund was created, it was a company level arrangement. Automatisering gids, Salarisstelsel vernieuwd in principeakkoord ICK-CAO, June 9, 2000. 91 Automatisering gids, Hardware-CAO blijft steken in discussie werkingssfeer, September 15, 2000. 92 Automatisering gids, FNV: ICK CAO niet algemeen verbindend, February 7, 2001.

110 Developing Institutions Negotiations over the fourth ICK collective agreement were atypically smooth, and an agreement in principle was made as early as March 2001. Things went so smoothly that, for the first time, the ICK collective agreement was sent to the Ministry for Social Affairs and Employment for extension. At the time the ICK collective agreement covered roughly 30,000 people and if it would be extended, it would cover about 50,000 people. 93 The Ministry, however, turned down the request for extension, because in their opinion it was simply not appropriate. Many sections of the ICK collective agreement were delegated to a company level body or the works council. The company level arrangements that stemmed from these delegated paragraphs could not be extended to other companies, and thus the ICK collective agreement, as a whole, could not be extended. 94 After the turndown by the Ministry, the unions, predominantly FNV, were ready to throw in the towel. The negotiations over the fifth ICK collective agreement had again turned for the worse. The IT industry was not doing very well, and employees representatives were careful not to bite the hand that fed them, which led to a meagre wage paragraph. On other fronts, such as the planned pension fund, also no progress was made and parties again had to negotiate over the scope of action of the ICK collective agreement. In order to accommodate the unions, the employers association, now called the Werkgeversvereniging Nederland-ICT, cooperated in a special committee that would investigate the possibility of whether or not the ICK collective agreement with the flexible paragraphs could legally be extended. The unions indicated that they were also interested if the ICK collective agreement could be extended when the particular paragraphs were adapted to accommodate such an extension. FNV had the impression the employers were not doing their utmost with regard to getting the ICK collective agreement extended, and wondered whether the employers were still willing to fulfil their initial promise made so many years ago. Until now, only the point of view of companies in competing industries has been discussed. Companies falling under the scope of the ICK collective agreement were in danger of falling under the scope of another collective agreement, the metal electro collective agreement to be precise, which had been extended in 2002. For these companies this could mean that they would have to pay extensive amounts of money to the various social funds that existed in these competing industries. So in order to protect the interests of those firms VIFKA, in 2000, had promised the unions that it would help them getting the ICK collective agreement extended, which was after all in the interest of both parties. 95 As it turned out, the representatives of the various industries had managed to solve this problem without the need for an extended collective agreement. This meant their principal argument for getting a collective agreement was gone. FNV, again, threatened to refrain from signing the 93 Automatisering gids, Verbindendverklaring ICK-CAO aangevraagd, June 6, 2001. 94 Automatisering gids, Loonbeleid grootste struikelblok ICK-CAO, April 2, 2002. 95 In an interview with Werkgeversvereniging Nederland ICT, it was stated that this was the principal reason for other IT manufacturing companies, who were against such an extension, to agree with a path towards extending the ICK collective agreement.

Industrial Relations in the Dutch IT Industry 111 new ICK collective agreement if the research by the special ICK extension committee concluded that the Werkgeversvereniging Nederland-ICT had not worked hard enough in trying to get the ICK collective agreement extended. The Werkgeversvereniging Nederland-ICT on the other hand was not impressed by this statement of FNV, and said it had no problem signing the new ICK collective agreement without FNV. 96 Though the ICK collective agreement still was not extended during its fifth incarnation, FNV did sign the agreement, as it would also sign the sixth version. During the discussions over the sixth ICK collective agreement, wages again proved to be a hot potato. The Werkgeversvereniging ICT wanted to get rid of the structural wage raises. They instead wanted to introduce a result dependent wage raise structure. In their opinion this was necessary because the IT manufacturing industry who had been hit hardest by the depression in the Dutch IT industry. The result dependent wage raise structure should have a maximum wage raise budget of 1.5 percent. Another point the employers association made was that the result dependent wage raises should be determined at company level. This frustrated the unions, who had demanded a structural wage raise of 2.5 percent. In the end the employers association agreed with a 2 percent structural wage raise, on the condition that they would install a special committee to investigate result dependent wage raises. Werkgeversvereniging ICT informed both their members and the unions that result dependent wage raises are a common practice in the IT software and services industry. 97 4.3.4 Collective Agreements Connected to the ICK Collective Agreement Together with the first ICK collective agreement of 1995, one other collective agreement was drawn up, involving the same parties. This was the so-called Collective Interest Fund (Fonds Collectieve belangen). Funds like this are common in industry wide collective agreements in the Netherlands and usually include special resources for training and education, and retirement schemes. The collective interest fund for the ICK industry was set up to reach a number of goals: 1. Providing information with regard to the collective agreement; 2. Paying for the tasks of the permanent committee; 98 3. Initiating research of the labour market and employment; 4. Initiating education projects, or projects related to employability; 5. Working towards improving the image of the IT manufacturing industry; 6. Working towards improving the working conditions. 96 Automatisering gids, FNV stelt toekomst van ICK-CAO ter discussie, Week 19, 2002. 97 Automatisering gids, eindelijk CAO voor ICK-sector, week 39, 2003. 98 The permanent committee is an arbitration board made up of four representatives of the unions and four representatives of the employers association. Their task was to make decisions with regard to companies not following the lines of the collective agreement or wished to deviate from them.

112 Developing Institutions Although the collective interest fund was initiated in 1997, it was not until 2001 that the employers actually started contributing to this fund. 99 This was interesting as the fund was introduced as something that should have given the ICK collective agreement a social face at its introduction. 100 The negotiations over the ICK collective agreement that took place in 2001 went smoothly for the first time. This was a sign that things were going well and parties, especially the Werkgeversvereniging ICT were showing their social face. Over the years various project were financed through the collective interest fund, of which some have already been mentioned, such as the project on result dependent wage raises, and the project relating to the extension of the ICK collective agreement. Until 2002 the ICK collective agreement had no arrangements with regard to pensions. This was something left to the individual companies. Within the industry a variety of pension regulations existed. Social partners eventually introduced an industry-wide pension plan because the Dutch government planned on terminating the early retirement scheme (VUT) in 2004. The VUT scheme turned out to be very expensive and rather than financing their own early retirement, people were financing those who had already retired early. As more people would use the possibility to retire early, the system would go bankrupt, so it had to be changed. In order to accommodate the employees in the ICK industry, the unions and employers had plans for a newly devised early retirement regulation, but this was something that could not be changed in the short run, so they devised a transitional regulation named the Flexi-VUT, which still meant that employees were paying VUT premiums, but were at the same time already contributing to a private saving fund that they could access at a later date. The Flexi-VUT had the status of collective agreement and was compulsory for all the companies that were using the ICK collective agreement. 101 The original plan was to introduce the ICK pension scheme on July 1, 2003, but it was decided to postpone it until January 1, 2004, when the new ICK collective agreement was supposed to be extended. 102 As it turned out, the ICK pension scheme again had to be postponed for a minimum of six months. The reason for this was that the Ministry of Social Affairs had to make the industry pension scheme compulsory for the whole ICK industry, which is comparable to the extension of a collective agreement. This proved not to be possible because both the employers and unions had not done their paperwork properly and the Ministry needed some questions answered with regard to the representativeness and coverage of the pension scheme collective agreement. In order to be able to introduce the pension scheme at all, the 99 Interview with an union representative in 2004. The contribution was originally set at an annual contribution of 0.05 percent of the total wage sum of employees earning less than 42,655 Euro (ICK collective agreement April 1997- May 1999), but was raised to an annual contribution of 0.06 percent of the total wage sum of employees earning less than 51,455 Euro in 2001 (Automatisering gids, Nieuwe ICK-CAO snel rond, March 27, 2001). 100 Computable, Fonds moet nieuwe CAO gezicht geven, April 11, 1997. 101 Automatisering gids, De ICK-sector ontwart langzaam zijn pensioenknopen, week 10, 2002. 102 Automatisering gids, Invoering van bedrijfstak pensioenfonds uitgesteld, week 20, 2003.

Industrial Relations in the Dutch IT Industry 113 Werkgeversvereniging ICT planned on introducing the scheme voluntarily on January 1, 2004. 103 The unions were against a voluntary introduction because they thought it would work against getting the ICK pension scheme made compulsory. They agreed in the end because delaying the introduction would mean that both employers, of whom most had already terminated existing contracts, and employees would have no pension scheme. The unions estimated that roughly 95 percent of the companies would accept the ICK pension scheme voluntarily, and that 5 percent would refuse to accept it. These were mostly companies who already had little to no social regulations. By not accepting the pension scheme, they would have an unfair competitive advantage to companies who did accept the pension scheme. 104 On June 4, 2003, the official request for making the ICK pension scheme compulsory, was filed. From that point on, interested parties could file objections against the ICK pension scheme. The Association of Insurers (Verbond van Verzekeraars), a league for insurance companies, 105 was one of them and objected that the official request to make the pension scheme compulsory, was too vague and that it was not clear which companies would be covered by it. There were more organizations having similar problems with the request, and again the ICK employers association and the unions were forced to negotiate over its coverage. 106 Half a year later the ICK pension scheme had still not been made compulsory. 107 4.3.5 A Closer Look at the ICK Collective Agreement 108 CNV signed the first ICK collective agreement because it felt an obligation towards its ranks. Despite the fact that it was a relatively meagre collective agreement, which was still missing a number of important areas, such as a payment system, CNV felt the initial agreement could and would be expanded in the future. In this section, when possible, the initial 1993 proposal for a sector collective agreement will discussed and compared to the first ICK agreement of April 1997 - May 1999, 109 and the sixth ICK collective agreement of June 2003 - September 2004. The boxes will show what both the Werkgeversvereniging ICT and unions have stated in respective interviews: nothing really much had changed in the ICK collective agreement over the years. 103 Automatisering gids, Bedrijfstakpensioen opnieuw uitgesteld, week 38, 2003. 104 Automatisering gids, Bedrijfstakpensioenfonds toch begin volgend jaar van start, week 46, 2003. 105 Of which some of the members probably used to provide company level pension schemes for some of the ICK companies prior to the introduction of the ICK pension scheme, and now saw a source of revenue disappear. 106 Automatisering gids, Pensioenfonds onder vuur, week 4, 2004. 107 Automatisering gids, Pensioenfonds nog niet verplicht gesteld, week 25, 2004. 108 See chapter two for an account on the proceedings with regard to extending a collective agreement. 109 FNV,VHP, and CNV (1993), Discussiestuk over de totstandkoming CAO IT-branche (Discussion paper on the realization of a collective agreements for the IT industry), August 9, 1993.

114 Developing Institutions The ICK collective agreement was originally envisaged as a multi level collective agreement, which meant it was a framework agreement for companies, and a bespoke agreement for employees. As has been mentioned a few times throughout the above section, the ICK collective agreement distinguishes between compulsory and optional regulations. The non compulsory regulations mean that these regulations still need to be determined by the employer and the works council or the unions. These optional regulations can have three different statuses. Table 4.2 lists the types of regulations found in the ICK collective agreement. These regulation types were present in the ICK collective agreement from the beginning and have never been changed. Table 4.2 Types of regulations in the ICK collective agreement Indicator Status Additional status A1 Compulsory These regulations must be implemented. A2 Partly optional Should be regarded as a framework regulations. Companies are free to change these regulations, as long as they remain within the confines of the framework. These regulations can have a compulsory status in case consultation has not led to a regulation or no other negotiation party is present. A3 Optional No framework is provided. Companies can elaborate this regulation together with its negotiation party. A4 Optional These regulations are considered as recommendations and have no status. For A2 and A3 regulations, examples are provided that can be used during enterprise consultation. In the 2003-2004 agreement there is no mention of a timeframe within which companies should have introduced company level regulations. In the 1997-1999 agreement, the timeframe was six months. Perhaps such a timeframe was required at the initiation of the ICK collective agreement, and is now superfluous. On the employee level, the ICK collective agreement offers a number of source elements that can be exchanged for target elements. See table 4.3 for a comparison of this à la carte element of the ICK collective agreement. Over the years a number of target elements have changed, and a new source element has been introduced.

Industrial Relations in the Dutch IT Industry 115 Table 4.3 The à la carte terms and conditions of employment 1993 proposal The initial proposal thought individualization of workers preferences to be a priority. 1997 ICK collective agreement A choice model for terms and conditions of employment is optional for companies employing a works council. The source elements of the choice model in this collective agreement are: reduction of working hours, days off, overtime and overtime bonus, and inconvenience bonus. Target elements included: reduction of working hours, deposit in the company level personal early retirement plan, deposit in savings account, deposit in an annuity, educational, leave, calamity leave, care leave, or combinations. 2003 ICK collective agreement The choice model has been expanded in terms of source and target elements, which are now diverse. Source elements include: reduction of working hours, days off, overtime and overtime bonus, inconvenience bonus, and hourly wage. Target elements now include: gross hourly wage, personal early retirement fund, savings wage, paid leave, day care (companies could voluntary use a day care system), personal computer for home usage, and union contribution should this be fiscally possible. The generic wage scale and job rating systems were not introduced until 2000 and have not been changed since (see table 4.4). The unions showed some leniency towards pay for performance arrangements in their proposal. They would, however, later on refuse to accept this proposal as a generic part of the agreement.

116 Table 4.4 Developing Institutions The pay system and job rating system compared 1993 proposal Companies should be able to maintain their current pay system, but agreements should be made with regard to some aspects of these pay systems (the number of function levels and scales, wage levels, and individual deviation margin (15 to 20 percent)). The core of the proposed pay system is that the hourly wages should be fixed. Though companies were to be able to retain their current job rating systems, agreements should be made on the specifications of these systems. Subsequently bandwidths should be agreed upon in which the evaluation of comparable functions should be categorized. 1997-1999 ICK collective agreement Both the pay system and job rating systems that are used, are inherited from its predecessor, VIFKA collective agreement. These systems are not compulsory and companies can use their own systems. Parties have agreed to negotiate over a more generic pay system and job rating system during 1997. 2003-2004 ICK collective agreement Companies are still able to use their own pay system and job rating systems, provided they have a works council and that are not inferior to the generic systems of the ICK collective agreement. The ICK collective agreement provides examples for companies on how to install a performance based pay system. The wage scale system and job rating system have not developed over the years, but as has already been mentioned, the employers have expressed a desire for a generic result based pay system, which is opposed by the unions. Changes with regard to working hours have also been minimal (see table 4.5).

Industrial Relations in the Dutch IT Industry 117 Table 4.5 Working hours 1993 proposal With the help of an inconveniency matrix, the seven days of the week are to be divided in time units. Per unit an indication is given of the bonus for each hour worked during that time unit. Also a definition should be given of overtime, per day, week, period. Based on this matrix the monthly salary can be calculated. 1997 ICK collective agreement - The maximum structural number of working hours quarterly is 520, based on 13 weeks of 40 hours. The maximum incidental number of working hours is 598 per quarter. Companies are free to determine their own working hours. - Part time work is possible, given that it is not contrary to the interests of the company. - Employees earning between 27,227 and 42,655 euro annually, and employees having a software and services function within the ICK industry, are able to buy spare hours to a maximum of 80 annually. - Reduction of working hours - Overtime is defined as person has worked more than 520 hours per quarter. Overtime can be exchanged for a variety of elements, such as the regular wage, savings account, a variety of leaves, spare time, etc. An inconveniency matrix is present, as are bonus percentages for each time unit in the matrix. 2003 ICK collective agreement - Compensation should regard travelling time made for work longer than one hour, as regular working hours. Over the years only a compensation for travelling time has been added. During 2003 the reduction of working hours (ADV), which until then had been a matter of course, became a source of heated debate. The ICK collective agreement contains one employability measure, which is the two days of paid training leave, each employee is entitled to. It is interesting to note that a technology driven industry such as the ICK industry, has such a scanty employability regulation. 110 4.4 Conclusion This chapter has given a historical account of the industrial relations system and its most important actors in the Dutch IT industry. AWVN, the IT industry transcending expert on industrial relations, does in the IT industry what it is doing in all of the other industries, which is advising employers. AWVN indicates it sometimes feels the absence of unions. To them unions are a natural negotiation 110 Van Velzen (2004, 153) notes that contrary to many other Dutch industries, the ICK collective agreement does not provide for a training and development fund for workers in the industry.

118 Developing Institutions party that can increase the efficiency of negotiations because the parties know each other s agenda. The umbrella organization Nederland-ICT is still mainly a lobby organization that tries to protect the interests of an entire industry. So far it has not been doing so with a great deal of fervour. Internally the forces seem too divided. Even employers sometimes do not know of which employers association they are a member of. It is confusing to them that being a member of Fenit, automatically makes them a member of Nederland-ICT as well. In order to solve these problems Nederland-ICT decided to change itself from an umbrella organization to a single employers association named ICT-Office. The merger took take place on January 1, 2005. Fenit, Vifcantec, and ICT Telecom since then have become one single organization. ICT Milieu and the Werkgeversvereniging ICT have not merged as that would have meant that companies that became a member of ICT-Office would also have fallen under the grasp of the ICK collective agreement, something which they might not have preferred. The reason for this merger was to strengthen the ICT-lobby and to become an opinion leader, to improve the image of the IT industry, and to upgrade the member services. 111 The parties indicated they have resolved their internal struggles and were ready to face the future. It is questionable whether they now will meet all of the preconditions to make the new employers association successful. The employers association indicates it is now able to speak with a single voice, but have they done something about the restrictions that their members impose on the organization? Their members are obviously interested in specific services, but can and will the companies work together on important matters such as a stable development of the industry? There hardly seems to be any indication that this has changed. In the first half of 2005 and the demand for experienced IT professionals is again increasing. 112 These same professionals were only recently laid off by IT companies. The crisis in the Dutch IT industry of the early 2000s has also had a negative impact on students thinking about a career in the IT industry. One can wonder if the employers association can be blamed for this as it is the companies themselves that refuse to work together to get a grasp on the seven to ten year cycles they seem to be going through since the origin of the industry. Interviews with employers have indicated they do not expect such a crisis to happen again, because they think the customer has become more aware of information technology. However, without new students entering the industry, IT firms are destined to face another labour shortage in the coming decade. Unions could help stabilize the IT industry during periods when the demand for IT professionals rises and they are able to demand higher wages. In that respect the unions may prove to be better allies for the employers than the employers association, which seems powerless to do anything about the cycles because they are 111 Automatisering gids, Nederland-ICT heet straks ICT-Office, November 5, 2004. 112 Automatisering gids, Goede automatiseerder is alweer schaars goed geworden, October 1, 2004. Computable, Te weinig ouder personeel in ict-sector, November 26, 2004.

Industrial Relations in the Dutch IT Industry 119 supervised in a way that prohibits them to enforce any regulation that may hamper an employer in any way. The unions, however, seem to be treated with an even greater disregard, which on occasion borders on hostility. Unions have repeatedly indicated during interviews that employers have no respect for them as negotiation parties. The employers main arguments against the unions are that they do not represent their employees because the union membership figures are too low. Employers indicate they regard their works councils as the negotiation party, and works councils happily agree. This often has the effect on unions that they have to prove themselves to two parties; the employer and the works council. Though the above problems for unions predominantly reside in the IT software and services industry, they also hold true in the IT manufacturing industry where the unions and the Werkgeversvereniging ICT negotiate over the ICK collective agreement. The history of these negotiations is very repetitive and shows signs of a union wrestling with its conscience whether the negotiations over the ICK collective agreement are worth the effort. Instead of improving these terms and conditions of employment over the years, it has mostly been a struggle to protect what they had. Where the unions are focussed on extending and improving the agreement, the employers association s focus is on maximizing the flexibility and minimizing the costs for its members. Oddly enough works councils of companies using the ICK collective agreement have indicated they are pleased with the agreement. Their main argument is that the ICK collective agreement offers them a handle to make arrangements with their employer. The ICK collective agreement empowers works councils that would otherwise have had a hard time making arrangements. These sentiments were expressed at a meeting between works council members and Unie on June 20, 2003. It is hard to determine whether these feelings hold true for all of the companies working with the ICK collective agreement. Some of the companies, of which some are even as large as Cisco, do not even have a works council at all. The drama of getting the ICK collective agreement extended seems never ending. The most recent attempt that was made during this study was August 13, 2004, and this attempt has failed again for the umpteenth time. Though an accountant has positively verified that the ICK collective agreement meets the minimum 60 percent demand for representation of the IT manufacturing industry, it was still rejected because of border conflicts with other industries. It is a miracle that the unions are still able to muster enough energy to enter in a process that seems doomed to fail. The initial incentive for the employers association to get the ICK collective agreement extended has disappeared because they have made agreements with the competing industries. The only reason they are officially backing up the process of extending the collective agreement is that they promised this to the unions many years ago. A representative of the employers association is positive the extension will come about, but thinks it will last only two years. None of the employers has any interest in extending the ICK collective agreement anymore. This means the unions are on their own. When reflecting on what has been said in this chapter so far, one gets the idea of living in a pre-collective Agreement Act society. It is as if industrial relations

120 Developing Institutions in the Dutch IT industry are similar to those in the early twentieth century. It will be interesting though to take a closer look on how unions interact with companies on a company level in chapter five. So far, this is the only area where the unions seem to have been successful.

5 Case Studies This chapter can be considered the crux of this study. Ten case studies will describe the reasons and causes underlying the introduction of company level collective agreements in the Dutch IT industry. Chapter four showed that the role of IT unions at the industry level is limited and that unions had to adjust their strategies to a company level approach. This chapter will continue the discussion of the cases mentioned in chapter four. Many of the case studies that will be discussed in this chapter have already been briefly mentioned in earlier chapters. In chapter three, when discussing the three lineages of the modern Dutch IT industry, a hint was dropped that these lineages may to some extent explain the current existence of collective agreements within some companies in the Dutch IT industry. Obviously, the same may hold true for the absence of collective agreements in other companies. Before going to the cases though, the process of selecting the ten cases will be discussed. 5.1 Introduction and Methodological Account The cases selected for this study are predominantly operating in the software and services industry. This is the segment of the Dutch IT industry, which is thought to be the least likely to introduce collective agreements, due to the higher educated workers it employs. The Dutch IT manufacturing industry is covered by the ICK collective agreement, which was discussed in the previous chapter. 1 The cases have been selected based on a number of criteria. The primary indicator was the presence of a collective agreement in a company. In the Dutch IT software and services industry, six company level collective agreements existed at the time of selection: T-Systems, Atos Origin, PinkRoccade, Getronics, Torex-Hiscom, and EDS. 2 Of these six companies, five companies were selected based on interviews 1 This is not entirely true, as there exist nine company level collective agreements in the Dutch IT manufacturing industry: Bull, Imtech, Alcatel, Cinram, Detron, Lucent, Océ, Xerox, and NRG Benelux. Also the ICK collective agreement is not extended, and the employers association Werkgeversvereniging ICT only covers about 250 firms and 31,000 employees, which is roughly 75 percent of the total IT manufacturing industry according to the Werkgeversvereniging ICT. 2 These companies were still using collective agreements in August 2004. In 2005 Getronics and PinkRoccade merged, reducing the total number of collective agreements to five. The merger process was still ongoing whilst writing this chapter so it has not been incorporated in this study.

122 Developing Institutions with various union leaders. The reason for not selecting T-Systems, which employs 200 people, was that it was the smallest of all six companies employing a collective agreement. In order to be able to say something about IT companies not employing a collective agreement five other cases were selected. The selection of these five noncollective agreement cases was based on the principle that they should be comparable to the five collective agreement cases in terms of: size, country of origin, and type of product or service. The following five companies were selected based on these criteria: CMG, Cap Gemini, Ordina, SSA Global Baan (Baan at the time of selection), and Centric. Table 5.1 shows some of the companies similarities and differences. Most companies, notably the larger ones, offer a range of products and services. Those products and services mentioned in the table are a selection of specializations. Together the ten companies make up roughly 40 percent of the total number of jobs in the Dutch IT software and services industry.

Case Studies 123 Table 5.1 Overview of the cases Collective agreement EDS Country of origin Size Specializes in Growth strategy : United States of America : 1,900 : services, outsourcing : takeovers Works council agreement Ordina Country of origin Size Specializes in Growth strategy : : : : France, Dutch since 1985 3,550 services, outsourcing, consultancy takeovers Atos Origin Capgemini Country of origin Size Specializes in Growth strategy : the Netherlands, French since 2001 : 9,000 : ERP, outsourcing, manufacturing : takeovers Country of origin Size Specializes in Growth strategy : France : approximately 5,000 : consultancy, outsourcing : takeovers Torex-Hiscom (isoft since 2004) Country of origin Size Specializes in Growth strategy : the Netherlands, British since 2002 : 300 : Hospital and GP IT systems : autonomous Baan (SSA Global since 2003) Country of origin Size Specializes in Growth strategy : : : : the Netherlands approximately 400 ERP autonomous Getronics (has taken over PinkRoccade in 2005) Country of origin Size Specializes in Growth strategy : the Netherlands : 5,100 : consultancy, network infrastructure : takeovers LogicaCMG Country of origin Size Specializes in Growth strategy : : : : United Kingdom approximately 5,500 consultancy, network infrastructure autonomous PinkRoccade (taken over by Getronics in 2005) Centric Country of origin Size Specializes in : the Netherlands : 7,400 : IT solutions for local governments Country of origin Size Specializes in : : : the Netherlands 2,850 IT solutions for local governments Growth strategy : takeovers Growth strategy : takeovers Figures stem from the respective websites or most recent annual reports available on July 1, 2004. The case studies are based on a number of semi-structured interviews with each firm s works council and HRM department. Also magazine and newspaper articles

124 Developing Institutions were used, 3 as well as annual reports, and various documents provided by both the HRM department and works council of the firms. After each case was worked out, it was sent to all the parties involved, including the unions, to verify or expand the factual representation of the case. Though all of the works councils of the ten cases have been interviewed, the HRM departments of some of the companies decided not to cooperate with this study. The HRM departments of the companies that did not cooperate with this study were EDS, Torex-Hiscom, both companies using a collective agreement, and Centric, which does not use a collective agreement. At Centric, the works council representatives preferred to remain anonymous. Later on another interview was held with a different representative of the Centric works council. Combined with the fact that some companies were more generous in providing information than others, some cases are more elaborate and detailed than others. Though a considerable effort was made to obtain a minimum level of information, not all cases are as rich in terms of detailed information. Especially for companies not employing a collective agreement, far less information was available in terms of company bargaining processes. 4 The most important information in respect to the presence or absence of a collective agreement, however, was retrievable, but potentially biased because only one party, in all cases the works council, was the only source of information within the company. Finally, while many of the companies also operate outside the Netherlands, it should be noted that the focus in the case studies is on the Netherlands. This chapter will start with a discussion of the ten cases and conclude with a comparison of the individual regulations contained within the terms and conditions of employment of the ten cases. Each case will start with a description of some of the key characteristics of the company and is followed by a description of the company history and labour relations. Finally, if enough information was available, the position of the works council within these firms is discussed. If there was not enough information, the position of the works council was discussed within the main text of the case. 5.2 Case Studies of Companies Using a Company Level Collective Agreement 5.2.1 Getronics Getronics is an IT supplier of services and solutions. Getronics employs 22,000 people worldwide and 5,100 in the Netherlands. The annual turnover in 2003 was 2.6 3 Both Computable and Automatisering Gids, weekly magazines for the Dutch IT industry, have proven to be an invaluable source of information. 4 In some cases works councils preferred not to make specific information public, despite talking about it during interviews. Interestingly enough, they also mentioned to having had problems with unions who brought some matters out into the public.

Case Studies 125 billion Euro worldwide. Getronics core competency is the implementation of combinations of hardware and software solutions. Getronics develops, implements, integrates, and controls IT infrastructures and business solutions. Their clients are both globally and locally operating firms. The services Getronics provides are spread over two lines of business. Business Solutions encompasses consulting, software solutions, and branch specific solutions. Infrastructure Solutions includes the design, implementation, and running of IT infrastructures. Outsourcing, the internalizing of IT operations of other firms, also falls under Infrastructure Solutions. 5 Company History 6 Getronics sets the date of its origin to 1887 with the corporation Groeneveld, Van der Pol & Co's Elektrotechnische Fabriek NV. Groeneveld acted as an installer of control and technical equipment for the utilities construction and ship-building industries. In 1950 Groeneveld formed the corporation Technisch Verkoop Kantoor Groenpol. Groenpol was a separate sales division that distributed technical equipment ranging from micro switches to steam turbines. In 1968 Groenpol NV merged with the Geveke NV to form Geveke and Groenpol NV, which was acquired in 1970 by Steenkolen Handelsvereniging NV (SHV). SHV divided Geveke and Groenpol NV's operations into a sales division (Groenpol Industrial Sales) and an installation division. In 1972 the name Groenpol Industrial Sales was changed to Geveke Electronics. 1975 was a special year for the future of Getronics. In that year Ton Risseeuw, the person that would eventually lead Getronics until 1999, started working for SHV. During 1976, when the Dutch economy was in decline, SHV decided to focus on its core competencies. Risseeuw was asked to close down the doors of Geveke Electronics. Geveke Electronics was making serious losses at the time, but Risseeuw discovered that Geveke Electronics still had a number of economically viable activities, such as the installation and maintenance of hardware. Also at this time communication between computers was becoming more important, as was the installation of local networks. In 1983, after managing the company for a few years, Risseeuw, with the help of a number of investment companies, took over Geveke Electronics by means of a leveraged management buy out, 7 reducing SHV's shareholding to a minority position. All subsidiary companies were brought under one parent company, Geveke Electronics International NV (GEI). 8 The takeover turned out to be a clever move, because in its first year GEI made a profit of 1.2 million Euro with 300 employees. 9 5 The company websites: www.getronics.nl and www.getronics.com. 6 This section is largely based on the company history, which can be found on www.getronics.com/global/en-gb/getronics/profile/history/history.htm (accessed August 17, 2004). 7 Leveraged buy out (LBO) means buying a company using borrowed money to pay most of the purchase price. The debt is secured against the assets of the company being acquired. The interest will be paid out of the company s future cash flow (Source: www.economist.com). 8 Computable, January 23, 2004, Ton Risseeuw vindt bedrijfstak nog niet volwassen. 9 Annual report Geveke Electronics 1983.

126 Developing Institutions In 1988 GEI changed its name to Getronics NV. Under the leadership of Risseeuw, Getronics had adopted an aggressive policy of expansion by means of internal growth, acquisitions, and strategic investments. Throughout the years a number of large takeovers were made. In 1988 Datex was bought, which employed 800 people and had an annual turnover of 53 million Euro, almost as much as Getronics itself. In 1996, together with Roccade, whom Getronics planned on taking over in the near future, 10 Getronics took over one of the eldest IT companies in the Netherlands: Raet. 11 Both parties agreed on buying 50 percent of the shares of Raet. Risseeuw s final takeover before he left Getronics in July 1999, was that of Wang Global systems. Through this takeover, Getronics was able to expand its international activities significantly. Wang, similarly to Getronics, was a provider of solutions and services for network and desktop environments. 12 Despite Risseeuw s efforts to turn Getronics into a depression proof IT company by focusing on continuity, rather than booming business, Getronics stocks have been taking a fall since the second half of 2001. The reasons for this fall were, amongst others, the two subordinated bond loans that were made in order to pay for the various takeovers such as Wang Global, which book value had halved in just a year time. Over the years, in order to make Getronics more solvent, parts of Getronics that did not belong to the core competencies, such as HR Solutions, were sold, and people dismissed. At the end of 1999, Getronics employed well over 34,000 people worldwide, of which 22,000 were still employed in 2003. 13 In May 2005 Getronics took over PinkRoccade and became the largest IT company in the Netherlands. On November 1, 2004, Getronics expressed its interest in taking over PinkRoccade as they thought it would improve Getronics position. On November 22, 2004, Ordina, another Dutch IT company, expressed its interest in PinkRoccade as well and made a better offer for PinkRoccade s shares than Getronics had. Despite a better offer though, PinkRoccade preferred to be taken over by Getronics. The takeover meant that Risseeuw s wish for taking over Roccade in 1997 was finally realized. 14 Collective Bargaining at Getronics Getronics has known three different phases with regard to its policy concerning conditions of employment. The first phase was called the building phase and lasted from 1983 until 1992. This phase was characterized by a non-harmonizing approach to HRM activities and policies concerning conditions of employment. It was a 10 This would never happen. The official reason for this failure at the time was the Millennium problem. Unofficially, however, according to Risseeuw, it turned out that Roccade was not doing too well. Risseeuw later on mentioned that in order not to harm Roccade, Getronics took the blame for the failed takeover. 11 See the conclusion of this chapter for a more detailed account of Raet and what happened after they regained independence from Getronics in June 2003. 12 Computable, August 28, 1998, Ton Risseeuw: Visionair met harde hand. 13 Getronics annual reports 1999 and 2003. 14 The takeover of PinkRoccade by Getronics will not discussed here in detail as the takeover was still too fresh to analyze in the context of this study.

Case Studies 127 common strategy to maintain and respect the existing conditions of employment in the case of a takeover. This first phase had led to a rather chaotic situation where Getronics was made up of 26 companies. Each of these companies had its own HRM department and is own regulations. Therefore a system integration phase was introduced, which lasted from 1993 until approximately 2000. This second phase was aimed at merging the existing systems of conditions of employment to five basic systems aimed at different business units (telecom solutions, consultancy, software, system integration, and installation). The five basic systems meant that Getronics now had five different works council agreements. The process of harmonization, however, would not end with the introduction of the five systems. In 1994 Getronics made a lot of its employees unhappy by refusing wage raises despite positive results of the company. The unions, who were alarmed by a number of union members at Getronics, reacted by suggesting to apply the VIFKA agreement. 15 Getronics took notice of these signals and started thinking about a single system for the entire group. 16 During 1995 an end was made to the autonomous development of the five basic systems. Instead they were fitted into a new group policy framework, named Entrepreneur and Social Policy Getronics Group (Ondernemer en Sociaal Beleid Getronics Groep). The second phase culminated in the introduction of the Framework Agreement (Raamwerk ArbeidsOvereenkomst, RAO) in 1997. This collective agreement was named Framework Agreement on purpose, to avoid the collective connotation, as Getronics wanted to emphasize the individual components. Beside the collective agreement, two other collective agreements were in use at Getronics at the time. The employees of Wang still fell under the scope of the ICK collective agreement, and Raet still had its company level collective agreement. Even when Getronics started using a collective agreement in 1997, these two companies held on to their respective collective agreements. 17 The third phase, which was a business integration phase, started in 2000 with the introduction of a new framework collective agreement (Raamwerk CAO). The new framework collective agreement was, even more than its predecessor, aimed at integrating existing employment regulations. With the introduction of the new collective agreement, also previous Raet and Wang employees were now falling under its scope. The philosophy of this phase was that Getronics was to become one company with one regulation for its terms and conditions of employment. Ton Risseeuw had been renowned for his dislike of unions, 18 and when Getronics started using a company level collective agreement 1997, this was considered to be 15 The predecessor of the ICK collective agreement. See chapter four. 16 More on these thoughts below, when a number of options, Getronics considered, will be discussed. 17 A member of the works council stated that this was predominantly so, because these regulations were better than the collective agreement of Getronics. 18 He is even known to have called a union leader a tattooed ape once, because of the way he dressed (Automatisering Gids, January 22, 1999, Risseeuw zwaait dit jaar af bij Getronics ).

128 Developing Institutions the first true success for the unions in the Dutch IT industry. 19 Two events led to the introduction of a collective agreement at Getronics. The first event was the potential threat of an extended and renewed VIFKA collective agreement. At the time Getronics, or at least a considerable part of it, fell under the scope of this collective agreement. When the works council of Getronics infrastructure mentioned that its conditions of employment were poorer than that of the VIFKA collective agreement, Getronics immediately cancelled its membership of the VIFKA in order to prevent to have to upgrade the conditions of employment for Getronics Infrastructure. The extension of the VIFKA collective agreement, which would eventually be called ICK collective agreement, could lead to a relative loss of control over the terms and conditions of employment and Getronics sought for ways to overcome this potential problem. Getronics came up with three options. The first option was to associate Getronics with the new ICK collective agreement, which was still in development at the time. Getronics knew about the plans and liked the idea of a collective agreement for the entire IT industry. Especially since the ICK collective agreement was advertised as a framework agreement, which allowed for individual elaboration at the company and business unit levels. Getronics had, however, problems with the ICK collective agreement. First of all, it was not clear whether this collective agreement would come about at all. Also FENIT, the employers association, was not inclined to have the ICK collective agreement extended to the software and services branch. Getronics therefore thought of a second option, which was to create a company level collective agreement. On August 25 1995, Getronics had its first meeting with CNV and FNV to discuss the possibilities of a company level collective agreement. After only a few meetings with the unions, which by now also included De Unie and VHP, Getronics decided it was indeed fruitful to continue on the path towards a company level collective agreement. Getronics, however, also had a third option in mind, a collective agreement shared with likeminded companies. In secret Getronics started negotiations with companies such as Volmac, KPN, and IBM over such a collective agreement. The advantages of this third option over the other two options were that Getronics would not have to organize the compensations and benefits alone, which would be costly in terms of manpower, as it could call in on the expertise of the unions and share the required manpower with other companies. This was the most preferred option, as it could combine the advantages of a sector level collective agreement (economies of scale) and a company level collective agreement (terms and conditions of employment that fit the particular company). There were, however, simply not enough companies willing to play along with Getronics, so the option disappeared. The potential threat of the VIFKA collective agreement was not the only reason for Getronics to start thinking of a company level collective agreement. Risseeuw had already made it public that Getronics was interested in taking over both Raet and Roccade. Both companies were using a collective agreement, and especially 19 See chapter four for a more elaborate discussion of the successful 1996-1999 period for unions in the Dutch IT industry.

Case Studies 129 taking over Roccade, a state owned IT company, which originated from RCC, could become problematic without a collective agreement. Getronics never admitted that Roccade was the principal reason for starting the negotiations for a collective agreement with the unions. Instead Getronics claimed a collective agreement could help facilitate their outsourcing processes. Collective agreements can assist outsourcing transitions by inspiring confidence in employees of the company that is outsourcing its activities. 20 A fourth option, no collective agreement at all, was only a theoretical option. Getronics sought to expand its activities and did not see a way to achieve this without a collective agreement. The negotiations for the first Getronics collective agreement went relatively smoothly. During the negotiation process, Getronics called in the help of the AWV. 21 In an evaluation report of the negotiations between Getronics and the unions, an interesting image was created of both parties. 22 The study mentions the positive atmosphere that existed between the parties, as well as how much the unions valued the contacts with Getronics. Getronics was aware that for the unions concluding a collective agreement was a high priority. Getronics, however, was surprised that the unions took Getronics initial proposal, with regard to the preconditions and starting points of the collective agreement, as a diktat, which took away any possible room for negotiations. Getronics stated these were merely a framework for negotiations. Getronics was ready to talk about anything except a few non-negotiables: performance related pay, a la carte terms and conditions of employment, flexible pensions, 40 hour working week, and different terms and conditions of employment for different subsidiaries negotiated with the works council. The unions agreed with the main lines of the proposal, such as the framework character of the collective agreement, the maintaining of functional differences, the reference to other condition of employment systems, and the decentralized elaboration of agreements. Some points, however, remained a topic of discussion, such as the influence of unions on the contents and application of pay systems, collective wage raises, working time, and the coverage of the collective agreement. The unions were willing to accept a performance related pay system, but the format of such a system should be carefully chosen - particularly the quality of the evaluation system. In the same AWV report the interests of both the unions and Getronics were compared, which gave a clear indication of the position of both parties. Table 5.2 shows the view of the AWV with regard to the interests of both parties. 20 According to Getronics, at the time of the first interview in July 2003, they had been recently evaluating the position of the collective agreement and concluded that the outsourcing argument was still valid. 21 Today called the AWVN. See chapter four for a more detailed description of the general employers association (AWV). 22 A non-public evaluation report by the AWV (1995) of the negotiations with the unions over a collective agreement for Getronics.

130 Table 5.2 Developing Institutions Interests of the negotiating parties with regard to a company level collective agreement Unions Getronics - A collective agreement for Getronics increases - A good relationship with the unions is the credibility of the unions towards their important for Getronics. members, which can have a positive effect on - A company level agreement will be positively union membership. regarded by employees. This will also benefit - A collective agreement for Getronics (to unions industrial peace. the company in the software and services - A company level agreement will have a role sector) will radiate within the IT industry and model function for a possible sector level may induce other companies to start using a collective agreement. Switching to a sector collective agreement as well. level agreement in the future remains a - Simultaneous negotiation for both Getronics possibility. and the ICK collective agreement is of strategic - Negotiating with professional (and importance. Unions want to have results before independent) unions will bring clarity to the April 1, 1996. functioning of the works council. - Formalizing contacts with Getronics will open - A company level collective agreement other doors in the industry. guarantees a more optimal negotiation results - The image of unions will be improved. for Getronics than a sector wide collective agreement. Would as it would water down interests Source: 1995 internal AWV report of the negotiations with the unions over a collective agreement for Getronics The AWV stated that there were three options open to Getronics: 1. total rejection of the collective agreement, which would hurt both the position of Getronics and its relation with the unions. Getronics had already progressed too far in the negotiations with the unions. 2. Temporary rejection of the collective agreement could be an option if the reasons for postponing were backed up with good arguments. The unions might take postponement as a play for time, which could hamper their cooperation in other areas, such as the takeover process of Raet. 3. Start up formal negotiations with the unions, which Getronics eventually did and which resulted in a collective agreement that was valid for two years, from January 1997 until December 1998. The first collective agreement at Getronics held for the entire company except for: (1) the division Electric Engineering, which fell under the scope of the extended collective agreement of the light engineering industry; (2) the former Raet division, which was still using its own collective agreement; and (3) employees earning more than 42,000 Euro annually, which was between 8 to 9 percent of the employees. In a similar fashion to the ICK collective agreement, the unions thought of the Getronics collective agreement as a collective agreement that needed to grow. This was also the principal reason for the two year duration of this first collective agreement. The unions wanted to have enough time to renegotiate and fine tune the job evaluation system, the performance related pay system, wage system, and pension system. The first collective agreement had both a structural wage raise component, as

Case Studies 131 well as a result dependent component, which was an addition to the structural wage raise. 23 During 1997 negotiations over performance related pay took place and a modified system was introduced at the beginning of 1998. Getronics was not the only company where the unions accepted a pay for performance system, earlier they had already accepted such a system for the Roccade collective agreement. 24 Moreover, Getronics had already been using a performance related pay system for a long time. The system now included an at least regulation, which meant that if a person performed well enough he would get a wage raise of at least a predetermined amount. In October 1998, the period of relative peace and stability that so far had defined the relationship between Getronics and the unions was interrupted. In a newspaper article, the unions stated that they felt the negotiation process for the next Getronics collective agreement was proceeding too slowly. Two years after the initial agreement, Getronics still had not elaborated a job evaluation system, wage system, and pension system. The only thing that both parties had agreed upon so far was a performance related pay system. With regard to the pension system, Getronics responded that it had already been working on a flexible pension system for seven years. The unions stepped right into the middle of this process and felt they were only playing a marginal role. The unions were also not satisfied with the role of the works council, which in their opinion was too marginal. Again Getronics countered the critique, stating that some of the divisions had been managed poorly and this has negatively affected the works council s right to assent in some divisions. 25 The fact that these matters were made public by De Unie was a source of great concern to Getronics, and almost led to the refusal to negotiate with De Unie in the next negotiation rounds. 26 In 1999 pessimism amongst the unions was growing. The negotiation parties had not been able to successfully conclude a new collective agreement at Getronics so far. With regard to the new collective agreement the unions disagreed on a number of points with Getronics: 27 1. Getronics refused to agree upon a wage raise budget as Getronics felt this would give their competitors too much information. 2. Getronics agreed upon a number of minimal wage raises, but these were lower than what the unions had asked for. 23 Computable, January 24, 1997, CAO-akkoord Getronics doorbraak IT-branche. Risseeuw, who had only joined the negotiations when cost related matters were discussed, was positive of the expertise of the unions, and was pleased with the results of the negotiations (Automatisering Gids, March 14, 1997, Risseeuw ziet bijrol Fenit bij CAO-overleg ). 24 Computable, January 8, 1998, Overeenstemming CAO Getronics. 25 Computable, October 9, 1998, Traagheid irriteert bonden. 26 Computable, November 6, 1998, Getronics uitgepraat met vakbond de Unie. 27 Computable, January 22, 1999, Groeiend pessimisme over CAO Getronics.

132 Developing Institutions 3. Getronics did not want to go as far as the unions in terms of a choice model regarding conditions of employment. Getronics refused a proposal to enable employees to buy or sell days off. 4. Getronics only wished to make limited arrangements for part time work. 5. Getronics did not want to introduce any child care regulations if this would mean additional costs. 6. Getronics did not wish to make special arrangements with regard to calamity leave and union facilities. 7. Getronics wanted to use the loose regime with regard to the Working Time Act, which allowed them to use employees more often, and at less attractive moments, such as Sundays. 8. Getronics also wanted to make special arrangements with regard to the recently introduced Flexibility and Security Act. This Act allowed collective agreements to deviate from the standard arrangements of the Flexibility and Security Act. 28 Getronics wanted to make special arrangements that would negatively effect employees with temporary contracts and the term of notice of contracts. Due to an article in a Dutch magazine Computable, 29 Getronics refused to negotiate further with De Unie and CNV. Informally, however, the negotiations continued with FNV, 30 and a month later, in February 1999, a new collective agreement was closed. FNV continued to consult the other unions, because not doing so would hurt their relationship in the long run. In the end both the unions and Getronics were moderately pleased with the results. An average wage increase was agreed upon and agreements were made to introduce a cafeteria system in 2000. Also agreements were made with regard to employability and the possibility for employees to work part time. The unions were pleased to learn that Risseeuw would leave Getronics later that year. They hoped this would lead to smoother negotiations in the future. According to the unions, it was a well known fact that Risseeuw had been directing the negotiations for Getronics in the background. 31 The unions relief because of Risseeuw s departure is interesting as Risseeuw, at least after 1995, on various occasions, had publicly stated that conditions of employment in general and collective agreements in particular, were an important development for the maturation of the Dutch IT industry. 32 The negotiations for the January 2000 collective agreement went relatively smooth considering the fact that the new collective agreement would introduce a 28 In Dutch law these special arrangements are known as driekwart dwingend recht. 29 This was titled growing pessimism over the Getronics collective agreement (Computable, January 22, 1999, Groeiend pessimisme over CAO Getronics ). 30 FNV was not responsible for the article. 31 Computable, January 29, 1999, Getronics breekt CAO gesprekken af, Computable, February 26, 1999, Bonden gereserveerd over CAO Getronics, and Automatisering Gids, February 18, 1999, Nieuwe CAO bij Getronics. 32 Automatisering Gids, February 26, 1999, Getronics en bonden sluiten een CAO akkoord.

Case Studies 133 brand new choice model, which enabled employees to buy or sell days off, day care, extra pension, and a savings wage - topics that could polarize both parties. The problems this time, however, did not stem from the unions, but rather the union members working at Getronics that were also a dominant party in the works council of Getronics. Employees that used to have their own collective agreements within Getronics, the former Raet and Wang Global employees were afraid that the new collective agreement, which from 2000 on would also apply to them, would offer them poorer conditions of employment. 80 percent of the recently purchased Wang Global employees voted against the new collective agreement. Though the conditions of employment in the new collective agreement were not poorer, as such the employees were afraid this would only be a matter of time, despite the fact that Getronics had offered them guarantees. 33 In a public letter to the Wang Global employees, Getronics criticized the works council for passing judgment on the new collective agreement. Getronics stated that it was only a small segment of workers that were responsible for slowing down the integration process, and that the rest simply wanted to work. The unions were offended by Getronics response to the situation, and condemned Getronics for not appreciating the critical stance of its works council. The unions also admitted they could have done a better job at informing the former Wang and Raet employees on what was going to happen and what the actual content of the new collective agreement was. 34 The 2001 collective agreement finalized the framework collective agreement that was introduced in 2000. next to the framework collective agreement, Getronics introduced three works council agreements in 2000. Part one held for those employees earning more than 46,000 Euro annually and was virtually similar to the framework collective agreement. Part two contained agreements which had not been made with the unions but, amongst others, with the central works council (COR). The regulations contained in this second part held for all employees. Finally a third works council agreement existed, which contained agreements that could vary between the various divisions of Getronics, such as regulations with regard to overtime. Until 2002 all Getronics collective agreements contained some form of collective wage raises. At the end of 2001 however, Getronics announced a salary freeze because of the poor market conditions. This was a source of concern for the unions, who immediately stopped negotiations, because Getronics was only willing to offer a wage increase of 1.5 percent for employees who performed averagely. The unions demanded a minimum wage raise of 3.35 percent. Nearing the second half of 2002, still no solution was found to the impasse, and the unions were preparing the first ever union actions in the history of the Dutch IT industry. The actions took place at the gates of Getronics and included a petition to the management. 33 Computable, January 28, 2000, CAO Getronics opnieuw omgeven door problemen and Computable March 31, 2000, Personeel Wang wantrouwt Getronics. 34 The Computable of April 14, 2000, printed the public letter of Getronics and asked the unions to reply to this letter (Getronics: Kleine groep medewerkers vertraagt integratie, FNV: mail van Voorst is een gemiste kans, De Unie houdt even gas in, and CNV: kritiek OR serieus nemen.

134 Developing Institutions Approximately 150 of roughly 7,000 employees, which fell under the scope of the collective agreement, showed up for the demonstrations, and consisted solely of union members. The unions were angry with Getronics because it had given management a wage raise of 42 percent the previous year. 35 The unions returned to the negotiation table when most of the reorganizations were over and the HRM manager, which had led negotiations with the unions for Getronics since the departure of Risseeuw, had departed. 36 At the end of 2002, the parties decided to extend the duration of the 2001 collective agreement until July 2003. The 2001 collective agreement s validity had already been extended for a year, and if no agreement was made with the unions, Getronics would have been forced to conclude individual agreements with each and every employee. 37 In an interview, the works council stated it was asked by the management whether it was willing to cooperate to establish a works council agreement, which would enable Getronics to introduce more individual terms and conditions of employment. The central works council decided not cooperate, because the works council noticed the difficulties the unions were already having during the negotiations, and they thought they could not negotiate better than the unions. Because the central works council did not want to cooperate, Getronics was forced to reopen negotiations with the unions. An additional argument for Getronics to restart the negotiations with the unions were the ongoing reorganizations. Between 2002 and 2004, the number of employees in the Netherlands again grew smaller from 6,600 to 5,100, due to reorganizations and the sale of divisions. In July 2003, Getronics drew up a social plan for 600 people. This was the first time the unions worked together with Getronics on a social plan and contained a regulation that gave union members a bonus. The special bonus for union members had not appeared earlier in the Dutch IT industry before. According to a CNV union leader, the reason union members were treated separately was because they had asked for it: We have been pulling the cart for so long, it is time to do something extra for us. 38 According to all parties, the relations between the unions and Getronics had never been better. During November 2003, the unions and Getronics successfully concluded a collective agreement, which was valid from July 2003 until December 2004. The new collective agreement had both a collective and a pay for performance based wage raise component. 39 Position of the Works Council The position of the works council at Getronics has been less than optimal since 1983 and can be explained by the presence of Risseeuw at the top of Getronics. Risseeuw considered the works council as a tool of management, which in the words of the works 35 Computable, June 26, 2002, Medewerkers Getronics ongelukkig met sfeer and Automatisering Gids, June, 24, 2002, Werknemers Getronics voeren actie. 36 The reorganizations had reduced the company size from 8,000 to 6,600 employees. 37 Automatisering Gids, December 12, 2002, Principeakkoord over verlenging Getronics-CAO. 38 Automatisering Gids, week 36, 2003, Vakbonden zoeken wegen om leden te belonen. 39 Automatisering Gids, November 10, 2003, Collectieve loonsverhoging bij Getronics.

Case Studies 135 council implied that it was not considered to be a self thinking unit. Risseeuw had problems with the fact that the works council was able to take an independent position. He was accustomed to moulding things according to his insights. This was not so easy in the case of a works council as it was with his managers, as the works council was protected by law. This was the ground for a great number of conflicts between the works council and Risseeuw, which the former party most of the time lost due to individual fear of works council members of unemployment. 1999 turned out to be an important year for the Getronics works council, because of the departure of Risseeuw. At that point a new HRM manager was appointed for the Netherlands, who participated in most of the negotiations with the works council. At the time, the works council of Getronics was a bit jealous of the works councils of both Raet and Wang Global, who had good relations with their respective HRM managements. The 2000 framework collective agreement introduced a new system that was intended to make things easier for employees if they had any questions regarding their conditions of employment. The system offered an overview of negotiation procedures and the level at which they took place at the level of the central works council, or at the level of the works council of a division. However, instead of simplifying matters, in reality this cross system caused a series of problems, because management interpreted rights of consent differently to those asking the questions. The relationship between the works council and the unions has progressively grown since 1999. The fact that seven out of nine members of the works council were also a member of an union played a role in this aspect. The relationship with management had also progressed over the years, but had its occasional setbacks due to incidents regarding the right of consent, which the management, for some reason, often seems to forget about. The 2004 reorganizations have again improved relations between management and works council. The company and the works council in 2005 started negotiating over a new works council structure that solely consisted of a central works council. The outcome of has yet to be determined at the time of writing. 5.2.2 EDS According to its website, Electronic Data Systems (EDS) is one of the largest IT service providers in the world with 132,000 employees worldwide in 60 countries, and an annual turnover in 2003 of US$ 21 billion. In the Netherlands in 2003, EDS employed 1,500 people in four subsidiaries. EDS s trademark is that of co-sourcing, which according to EDS is a form of outsourcing that spreads the risks involved with outsourcing between the insourcer and outsourcer. Its stock of clients in 2003 included amongst others ABN AMRO, Delta Lloyd, D-Reizen, DSM, Nerefco, Nederlandse Spoorwegen, NUON, ProRail, Rijkswaterstaat, Vodafone, and Xerox. The services EDS provides are divided over four lines of business. Operations Solutions focuses at outsourcing of traditional IT services, which can go as far as outsourcing of complete company processes and functions. Solutions Consulting is the consulting branch of EDS, and addresses CEO-level concerns through strategy, organization,

136 Developing Institutions operations, IT and executive search services. A.T. Kearney offers strategic advises targeted at e-commerce. PLM Solutions offers product lifecycle management, which encompasses product development, manufacturing planning, product data management, and collaborative commerce offerings supported by a broad set of product-driven services. 40 Company History 41 On the company website, 42 EDS portrays itself as the first ever IT services company in the world. EDS was set up in 1962 by Ross Perot with a venture capital of just US$ 1,000. EDS earned its first money by renting unused mainframe time for information processing. Later on EDS would buy its own computer. During the 1960s EDS developed a custom made IT system to process insurance claims and payments for the Medicare program in Texas. Other versions of this piece of software were soon used in different companies throughout the United States. In 1968 EDS went to the stock market. At the time 303 people worked at EDS and a profit of US$ 7.7 million was made. Within a year the number of employees quadrupled because of a large contract with Blue Shield California. This success continued and in 1978, 15 years after EDS was established, 6,000 people were working for EDS and the turnover was US$ 164 million. 1978 was also the year EDS had its first commission in the Netherlands. Lacking an office in the Netherlands, operations for this project were controlled from the US. The first project was an integration project for the Sociaal Fonds Bouwnijverheid. As part of EDS Benelux, EDS opened its first office in the Netherlands in 1980 at the former RSV Data computer centre. In 1984 EDS was taken over by General Motors. In practice this meant that the former IT department of GM became part of EDS, which grew from 13,000 to 30,000 employees worldwide. In 1990 A.T. Kearney, which at that time was the second largest value management group in the world, was acquired by EDS. During the 1990s a new company philosophy, co-sourcing, was introduced, which is a form of outsourcing based on the idea that industrial processes should be facilitated at the company providing the outsourcing services. In practice this simply meant taking over the employees of the outsourcing company. Outsourcing enabled EDS to grow rapidly worldwide, as well as in the Netherlands. In 1994, through outsourcing, EDS took over ICIM, the IT department of Rijkswaterstaat (comparable to the department of public works), and CVI, the computer centre of the Dutch Railway services (NS), followed in 1996 by the IT department of the Koninklijke Schelde Groep. In 1996 it was the turn of Agridata, and in 1997 of Nedlloyd Computer Services. During these years the company size of EDS in the Netherlands expanded from 300 to 1,800 employees. This proved to be big enough for EDS Netherlands to become an 40 http://www.eds.nl and http://www.eds.com (accessed March 4, 2004). 41 This historical account of EDS is predominantly based on the historic information provided on the company website: http://www.eds.nl (accessed March 4, 2004). 42 See note 40.

Case Studies 137 independent subsidiary. Until that time EDS had fallen under the scope of EDS Benelux. The period 1997-1999 was a period of consolidation, which meant a temporary stop to acquisitions. The first significant IT outsourcing order EDS Netherlands got since 1999 was that of ABN AMRO in 2002. At the time of writing, Achmea s IT department, which had been outsourced to EDS Netherlands in the first half of 2004, had been the latest major takeover. Company Bargaining at EDS Netherlands EDS s company history classifies it as a typical takeover company. Before EDS started using a collective agreement in the Netherlands, the company was using a hotchpotch of company regulations and individual agreements. In all cases, takeovers and outsourcing contracts meant the introduction of additional regulations in order to facilitate these transitions. Throughout its history, EDS has taken over companies that both were and were not using collective agreements. Two of the most important takeovers in this respect were that of the CVI, the computer centre of the Dutch Railway services (NS), in 1995, and Agridata, a developer of software and IT services for the agricultural industry, in 1996. Employees of both companies were accustomed to using a collective agreement. In order to accommodate these transition, EDS designed special transitory collective agreements that were extended to the rest of the company until EDS started using its own collective agreement. This collective agreement was introduced for three reasons. The first was the takeover bid for Roccade, which would fail in the end. EDS was one of many companies in the market for this state owned IT software and service provider. 43 Using the same argument as Getronics, namely that a company using a collective agreement would be a preferred takeover candidate compared to companies who did not, EDS turned to the unions. Roccade was however not the most important reason. As has already been explained, EDS had taken over a great many companies that had been using a collective agreement. This is not surprising as 84 percent of all Dutch companies uses either a company level collective agreement, or falls under the scope of an industry wide collective agreement. 44 As most of EDS s takeovers consisted of IT departments of public bodies, most of the companies EDS was taking over were using a collective agreement. In order to deal with the negative Wild West image commonly associated with the IT industry, EDS decided it wanted to start using a collective agreement of its own. The takeover of Nedlloyd Computer Services in 1997, which was also using a collective agreement, and initially refused to being taken over by EDS, may have been the proverbial last drop that made the cup run over. A third reason for using a company wide collective agreement was the fact that EDS wanted to harmonize the hotchpotch of employment regulations that were being used throughout the company at the time. In order to restructure EDS, a 43 See both the Getronics and PinkRoccade cases. 44 Author s calculations based on SZW 2004 (5,321,500 people fell under an industry wide collective agreement and 716,000 under a company level agreement) and CBS data for the active labour force between April 2004 and June 2004 (7,077,000 people).

138 Developing Institutions consolidation of the multitude of divisions was required. A company wide collective agreement was thought of as an efficient means to accomplish this. 45 The consolidation process at EDS, of which the collective agreement was part, was not always very well received. EDS was roughly shifting 160 of its 1,800 employees around, and this was a source of concern for both the unions and the works council, but which was out of their hands. During the initial negotiations over the collective agreement, the unions therefore wanted to strengthen their own position, as well as that of the works council. 46 All in all, the negotiations between the unions and EDS took about a year. The first EDS collective agreement had a relatively long duration: it was valid from April 1998 until December 2000. As had also been the case in both the ICK collective agreement and in the first Getronics collective agreement, many topics still had to be worked out in the collective agreement, such as the bonus for working in shifts, harmonization of health care insurances, mobility policy, and a choice system that allowed employees to buy, sell, or trade days off. Despite EDS and the unions had also not been able to agree on the pay system, they agreed that such a system would be implemented on the first of January 2000 at the latest. Until then employees that were performing well or more than well, were guaranteed a minimum wage raise of 2.5 percent. Though the unions were not overly pleased with having to accept a performance related pay system, EDS refused to go without such a system. EDS had already been using such a system for a long time. The unions were only willing to accept such a system under the condition that it would be combined with a proper assessment system. 47 Next to the collective agreement, EDS also started using a works council agreement named Wegwijs. EDS and the unions, together with their members at EDS, negotiated over which subject would fall under the collective agreement, mostly the primary conditions of employment, and which would fall under Wegwijs, mostly fringe benefits. During 1999 EDS started a new round of reorganizations, which initially meant am internal shift of about 100 people, but eventually led to a lay off of 67 people. EDS had promised to make the reorganizations as pleasant as possible and spoke of retraining those people that needed to. The unions stated that it could do nothing about the lay offs because it was not convinced it had enough support from the EDS employees to be able to protest against the lay offs. EDS stated that the unions had two options, either to agree with the lay offs, or to watch the lay offs from the sideline without any means of influencing the process. EDS stated it had attempted to prevent lay offs, but it turned out to be impossible to re-educate of replace some people. These were the ones that needed to be laid off. 48 Two other 45 Document EDS used to inform its employees about the plans for a company level collective agreement, February 1998, Op weg naar een moderne CAO voor EDS. 46 Computable, September 19, 1997, CAO voor EDS in de maak. 47 Automatisering Gids, week 25, 1998, Eén EDS en het personeel is straks één groot volk. 48 Computable, March 23, 1999, EDS Nederland zet mes in organisatie and Computable, April 23, 1999, EDS Nederland zet 67 mensen op straat. Also all employees older than 58 were offered a settlement to leave (Computable, January 27, 2000, CNV: IT waardeert ouderen onvoldoende ).

Case Studies 139 reorganizations, aimed at the simplification of the management structure, would take place that same year. In 1999 EDS Netherlands was cut loose from EDS Benelux, which meant that the Dutch top management at EDS Benelux was laid off. In November another 40 people, predominantly managers and people in overhead positions, were laid off. 49 In January 2001, EDS and the unions signed a two-year collective agreement. This was again longer than was common in most other IT companies using a collective agreement. Whilst managers were excluded from the previous collective agreement, in the new collective agreement only the board was excluded. The wage system, which had already been worked out whilst the previous collective agreement was in effect, was based on salary scales that were adjusted annually to market conditions. This is fairly common practice in the United States, but not in the Dutch IT industry. Every year individuals meet with their team leaders or managers to discuss their individual bonus for that year. Assessment is based on criteria such as professional development. EDS also uses other types of assessments, which are based on customer satisfaction and project profitability, amongst other factors. During the period of the previous collective agreement, EDS management had tried to implement a new rule, which stated that 73 percent of each and every employee s time should be billable. This rule, however, was blocked by the companies works council based on the fact that such a rule would hurt employees that were between projects. Whilst the 2001-2002 collective agreement contained several improvements for employees, such as the day care and employability arrangements, the continuing focus on harmonization caused problems. EDS wished to harmonize both the conveyance regulations and pension schemes. With regard to the conveyance regulations, the unions and EDS agreed that mobile workers should get a lease car, whilst the other employees should get a personal budget, which they could spend on a bike, car, or public transport. EDS employees, however, declined the offer, because they felt it was too meagre, and it was decided to make a new offer next year. EDS also introduced a new pension scheme, which was based on 70 percent of the average wage during an individual s stay at EDS. This used to be 70 percent of the last earned wage. EDS employees older than 50 were offered a choice to keep the old regulation, or accept the new one. Because of the various bonuses people could receive annually, it could be possible that the average wage regulation was an improvement for some. 50 Nothing special went on with regard to the 2003-2004 and 2004-2005 EDS collective agreements, except for the fact that the minimum wage raises for those performing in the range well to good, were lowered to 1.5 to 2 percent, as opposed to 3 to 4 percent in previous years. EDS, like other IT companies, had to deal with losses during this period and started reorganizing again. Until June 2004 EDS laid off 200 people, as well as most of the temporary workers that had come with the takeover of the IT department of ABN AMRO in 2003. ABN AMRO decided it wanted to 49 Computable, November, 26, 1999, EDS Nederland schrapt 40 banen. 50 Computable, January 19, 2001, Heibel rond CAO s.

140 Developing Institutions outsource its IT department and had agreed with EDS that EDS would cut back the costs of the department. EDS thought that laying off a number of temporary employees was the easiest way to meet the agreement with ABN AMRO. At the time of introduction, the plans for a collective agreement were not well received by the international management of EDS. The takeover of the IT department of ABN AMRO, though, had proven that the EDS collective agreement had been helpful in convincing ABN AMRO and its employees that EDS was a reputable company. According to EDS s works council, the HRM department was still very pleased with the collective agreement in 2003. Position of the Works Council The introduction of the 1998 collective agreement was well received amongst the employees of EDS. The management, however, was less enthusiastic, as it had been deprived of a number of responsibilities, which were now dealt with by the collective agreement. Whilst the employees enjoy the relative safety of the minimum regulations the EDS collective agreement offers, managers often regard these minimum regulations as rigid. They feel it obstructs their scope of action. Managers also frequently interpret regulations differently from workers, which has led to problems with the works council. That is why in the 2003-2004 collective agreement a number of regulations have been written down more clearly. EDS has a single works council for the entire company. EDS used to have a more elaborate works council structure, but this structure had often led to coordination problems and so it was decided to harmonize the structure of works councils.. Within EDS two forms of consultation exist for the works council. The principal form of consultation is the works council consultations, which take place every six weeks. Every six weeks the works council also meets with HRM management, in order to discuss problems regarding the regulations that are written in Wegwijs, the companies works council agreement that exists next to the collective agreement. Two members of the works council are also allowed to attend the collective agreement negotiations between the unions and EDS. The relation with HRM management has become more formal in recent years. This is mostly due to the present conditions in the Dutch IT industry, which have also affected EDS and have led to lay offs and company restructuring. EDS Netherlands reports to and receives orders from the US based EDS international. National policies are more and more determined by international policies and this has led to a number of conflicts with the works council. According to the works council, the collective agreement has an empowering effect on both the works council and the Dutch HRM division because of the WCAO, to which Dutch management has to abide despite guidelines set out by the international HRM department.

5.2.3 Atos Origin Case Studies 141 On the company website, 51 Atos Origin presents itself as an international business and technology integrator, providing integrated design, and IT solutions. Atos Origin offers full business and technology integration services to its clients, through three lines of business: Consulting, Systems Integration, and Managed Operations, which includes outsourcing operations. In 2003 Atos Origin generated an annual turnover of 3 billion Euros, and employed in excess of 26,000 people in 50 countries. In the Netherlands Atos Origin employed 8,424 people in 2003. Atos Origin s clients include ABN AMRO, Akzo Nobel, Alstom, BNP Paribas, BP, Ericsson, EDF, Euronext, Fiat, France Telecom, ICI, ING, KPN, Lucent, Philips, Renault, Royal Bank of Scotland, Saudi Aramco, Schlumberger, Shell, Telecom Italia, UBS, UK Department for Work and Pensions, Unilever, Vivendi Universal and Vodafone. Company History Atos Origin was the result of a merger between the French Atos and the Dutch Origin in August 2000. Both companies were relatively young. Atos was the result of a merger in 1997 of two French-based IT services companies (Axime and Sligos), both of which had come forth from earlier mergers. Origin was a subsidiary of Philips Electronics, which had come about in 1996 as the result of a merger of BSO/Origin and Philips Communications & Processing Services. Of the two companies, Origin was the largest at the time of the merger with 16,000 employees in 30 countries, where Atos employed 11,000 people in France. Following the merger, Atos Origin landed three major outsourcing contracts with KPN and took over the running of KPN Datacenter in the Netherlands in July 2001. This has now become Atos Origin Telco Services BV, a part of the Managed Operations service line of Atos Origin. In August 2002, Atos Origin acquired KPMG Consulting in the UK and the Netherlands, trading as Atos KPMG Consulting and providing the Group with a major presence in the Consulting segment of the IT services market. Company Bargaining at Atos Origin Netherlands 52 Prior to the first collective agreement at Atos Origin, which was then still called Origin, the works council of the former BSO/Origin played a significant role. After BSO/Origin and Philips Communications & Processing Services merged in 1996 to become Origin, Origin had two works councils. The two works councils differed significantly. The works council of the former Philips C&P was used to a strict division of tasks between the unions and the works council. Philips used to speak with 51 http://www.atosorigin.nl (accessed March 4, 2004). 52 This case study is partly based on an earlier case study for the ILO in 1999, which was published in 2002 (Van Liempt, 2002b). A more up to date version of this case study appeared in Van Liempt (2000). New interviews have been taken with both HRM and the works council of Atos Origin to bring this case up to date.

142 Developing Institutions the unions about the settlement of the terms and conditions of employment, and with the works council about the different aspects of social policy. The former BSO/Origin staff, however, were used to individually negotiate over their terms and conditions of employment. This had the effect that the works council of the former BSO/Origin was a much more active and mature negotiator than the former Philips C&P works council, which had developed much less experience and responsibility. The former BSO/Origin works council was a stranger to the unions and preferred not having to deal with them. The presence of multiple conditions of employment regimes was a thorn in the side of the HRM department of Origin, and they started thinking of an agreement that would hold for all employees. At the time there was absolutely no intention of using a collective agreement as a tool to harmonize the various conditions of employment. The board of directors of Origin, but also the majority of the management, shared the opinion that collective agreements were something of the past and did not belong in the IT industry. Before the negotiations started, parties first sought unison within their own ranks. Origin installed workgroups that each were to prepare a part of the future harmonized agreement. The mission of these internal workgroups was to design a new package of conditions of employment, which would be able to deal with the dynamics of the market. The works councils in the meantime started negotiating over a mutual basis of cooperation between each other. This was necessary as both works councils stemmed from different cultures, which made the task ahead a substantial one. The consolidation process was going to have to harmonize: two collective agreements of the former Philips C&P, which had an A- and a B- collective agreement for two classes of employees (higher and lower skilled); a works council agreement of the former BSO/Origin; and thirty plus different arrangements, which stemmed from multiple, smaller takeovers. Both works councils wondered whether such a harmonization venture would be organizationally possible. They also wondered whether they had enough expertise to pull this venture off. The same held for the management, which shared the desire to take the negotiations to a professional level by bringing in expertise, which both the works councils lacked. Despite the fact that the former Philips works council was the one familiar with unions, it turned out to be the former BSO/Origin works council which proposed to invite the unions to the negotiations. The former Philips C&P works council simply wanted to use the former Philips collective agreement. The former BSO/Origin works council, which was made up of 50 percent original BSO ers and 50 percent of people stemming from takeovers, had already been discussing whether or not to use a collective agreement prior to the merger. This is why the BSO/Origin works council was more familiar with the problem of harmonization than the Philips C&P works council was. The BSO/Origin works council had already figured out that in order to successfully harmonize the various employment regulations, it would be necessary to involve the unions.

Case Studies 143 The management of Origin, at first, was opposed to bringing in the unions, but after a few introductory rounds in which the unions had the chance to demonstrate their expertise and feeling for the IT industry, they came to an agreement. A single representative for all the unions, which consisted of FNV, CNV, VHP, and De Unie, was allowed to join the negotiations. The reason was that Origin wanted to negotiate as efficiently as possible, and forced all parties to reduce their numbers at the negotiation table. FNV was the largest union in terms of members, so it was FNV who brought forth the representative for the unions. Also it was agreed that this representative would inform all employees at Origin, and not solely the union members of the three unions. All negotiation parties, which now also included the unions, were allowed to come up with a number of conditions before the negotiations would start. Origin s conditions included that it did not want to discuss annual collective wage raises and reduction of working hours. Origin had the intention of introducing a performance related pay system without the traditional, collective component. Also Origin s main source of income was billable hours of its employees, and, in a tight labour market, it did not want to start experimenting with a reduction of working hours. Origin also wanted the total package of conditions of employment to fit the goals and culture of the company. The works councils only had one condition, which was that it would be accepted as a full member to the negotiations. The unions had three conditions. The first was that they would accept other parties conditions only if these conditions were accepted by the majority of the employees. The unions did not want to present itself as a party that would think for the people it represented, but rather, with the group it represented, namely, all of the Origin employees. In order to make this possible they demanded that they could openly and regularly consult Origin s employees. The second condition the unions made was that the signatures of both works councils should be present on the final agreement. An important argument for the unions was the high percentage of nonunion members. Through this condition, it was guaranteed that non-union members would also have a say, and that the works councils would be equally responsible for the results of the negotiations. The third and final condition of the unions was that the final agreement would have the structure of a framework agreement. This would allow for the possibility of bespoke work of the conditions of employment on multiple levels such as that of the business unit, location, or individual. The unions wanted to prevent a rigorous harmonization, which in their view often led to an average of sums that would hold for all, but with which no one was content. By offering individual choices, the largest differences between the various employment regulations themselves could be harmonized more easily and with less frustration to the individual. The modules, 53 which the workgroups had come up with before the introduction of the unions, were now being used to structure the process. After an 53 These modules were: working hours, pay, bonuses, pension, employability, services, choice system, and general provisions.

144 Developing Institutions exploratory phase in which the three parties (Origin, the works councils, and the unions) threw light on their individual backgrounds and the conditions they had set, Origin launched an initial proposal. This proposal was then discussed by the unions and the works councils during public meetings with the Origin staff. After everyone was informed about Origin s proposal, three months of negotiations started, which were concluded with an agreement in principle, which again was put before the Origin staff. These public meetings always received a lot of attention. In total 1,500 employees had visited these meetings or had responded to the proceedings through the mailboxes of the works councils and the unions. At some point the new Origin agreement was the principal topic of discussion in all business units and locations. The works councils were working almost full time on responding to emails and drawing up comparisons between the various proposals. As was to be expected of a harmonization venture of this magnitude, many individual and group agreements came to light, all of which had to be dealt with and incorporated into the final package, or for which a transitory arrangement would have to be made. This hotchpotch of agreements and the potential threat of individual procedures made Origin realize that harmonizing towards a works council agreement would cause serious problems because of the carry-over effect of the earlier agreements. Both Origin s legal department and the unions had informed Origin management that the easiest way out would be a collective agreement, which unilaterally applied to all employees. The use of a collective agreement would prevent Origin from having to explain changes to each and every individual, every time changes were made. Origin management, however, had strong emotional arguments against a collective agreement, and it was only at the very last moment that they decided for a collective agreement for reasons of practicality. During the process of negotiating not everything went as smoothly. The relationship between the two works councils for instance had been ambivalent at times. On the one hand they had to cooperate and work together, but on the other hand it proved to be difficult at times for the representatives of the Philips C&P works council in the negotiations to convince their fellow works council members of the positive negotiation results, which they did not regard as positive themselves. At some point the Philips C&P works council even threatened to stop the negotiations. Intervention by the unions prevented this from happening. At times the works councils and the unions had also been having problems amongst each other. These had been mostly due to the works councils negotiating with management without including the unions. In the end loyalty to one another prevailed, strengthened by a common goal. The unions had solved their problems with management at a very early stage. They were brought into the negotiations in order to professionalize these, not to empower the works council in relation to management. Obviously the unions had their own ideas, but once they proved they could think beyond the boundaries of the blue collar worker, they were well received. The unions had to make a number of sacrifices though, amongst others the introduction of a performance related pay system without a collective component. At Getronics they had at least been able to introduce a collective wage raise component.

Case Studies 145 In the first Origin collective agreement, which was valid from May 1998 until December 1999, the function levels of employees were divided into ten skill levels, which were based on a description of tasks, level of autonomy, responsibility, and skills and competencies. Wage classification was dependent on the relative position of an employee on the salary scale. For each and every employee the relative position was determined based on a model function. Origin annually adjusts the minimum and maximum margins of the salary scales, based on macro-economic developments such as the price index, the relative position of Origins salary with respect to other IT companies, and the economic position of the company. The effective wage raises were an individual matter and were determined on a personal reference point of the employees and the evaluation of both performance and the individual development of skills and competencies. Everyone who performed well received a minimum wage raise. Promotions were possible if an individual had proven to meet the main requirements of the next salary scale. Demotions were considered if an employee did not meet the requirements of his wage scale for an extended period of time. The unions, at the time of the negotiations, had been afraid that an opaque wage system would arise. Origin had the intention of individualizing the pay system and rewarding people for performance and development. According to the unions, the pay system should hold on to a collective part, and should be connected to a verifiable and clear assessment system. In addition assessment should not be purely based on results, but also on competencies and the potential for development. In order to lessen these worries, an agreement was made that everyone at Origin should receive an Individual Development Plan (IOP), which would be the responsibility of both the manager and the employee itself. This IOP could then form the basis for the short term development (one year) and the career possibilities in the longer run (three years). In the end, the collective part never made it, but the IOP was successfully introduced. Other novelties that were introduced because of the unions in the first Origin collective agreement were the possibility to work part time, pension build-up during unpaid leave, and a rudimentary day care regulation. A number of differences existed between the negotiations over the first and second Origin collective agreement. At the negotiations over the second collective agreement a different works council was present. The former two works councils had merged immediately after the successful conclusion of the first collective agreement. Several indicate that this new works council was much less active and proficient than the former two combined had been. The works council was not the only party that had changed during the negotiations for the second collective agreement. With the exception of FNV, most of the union leaders had changed and the HRM department also had a new manager. So although the parties may have been the same, only four people had actually been present during the negotiations for the first collective agreement. This, and the fact that Origin felt it had been too compliant during the negotiations over the first collective agreement, changed the atmosphere during the negotiations over the second collective agreement. Origin wanted to set things straight again. The

146 Developing Institutions negotiations became much more formal and strict. There were positive aspects as well though, the negotiations were much less hectic, and management was much more involved than the first time. The number of changes between the first and second collective agreement were quite modest. The most important change was made to the performance related pay system. During the first collective agreement, the unions had stipulated that those with a first time mediocre performance would get a minimum wage raise. This was abandoned during the second collective agreement, because it made evaluations more thorny. From now on people were assessed twice a year. If one would get a mediocre rating during the fist half of the year, this person could change things and perform better in the latter half of the year. If the person still performed below which was required of him, he would not get anything extra. During 2001 Origin was taken over by the French company Atos. This takeover has put a mark on the negotiations over future collective agreements significantly. The reason for this is that the negotiations over the wages have been separated from the negotiations over the terms and conditions of employment. HRM of Atos Origin want to negotiate over the wages at the latest possible moment, which is September. The negotiations over the fringe benefits generally take place during the first quarter of the year. This is so because the financial situation of the company is an important aspect for determining the space for wages. If Atos Origin were to determine the space for wages during the middle of the year, it would make commitments to the unions, which it did not desire. HRM acknowledges this has negative effects for all parties (Atos Origin, the unions, and the works council) on the negotiations over exchanges between primary and secondary conditions of employment, but the downside is not strong enough to counter the positive effect for Atos Origin, which is more certainty regarding its financial position. This problem has been pestering the negotiations ever since and has led to a great number of postponements of concluding the collective agreement. The management of Atos Origin has considered two alternative solutions to this problem, the first being to move the secondary conditions of employment to a works council agreement that the unions would have no say over. The second alternative was to get rid of the collective agreement in its entirety. Atos Origin also decided against this second alternative, because the collective agreement had a number of upsides: 1. Not having a collective agreement could mean Atos Origin had to negotiate with each and every employee individually. In order to pull this off, Atos Origin would have had to think of a way to get this alternative accepted by all the employees. This was considered too risky. 2. The collective agreement had also proven itself during a great number of successful takeovers, amongst them the takeover of the KPN computing

Case Studies 147 centre during 2002, 54 during which 2,350 employees moved to Atos Origin. Without a collective agreement this would have been more difficult. 3. A collective agreement also allows for deviation of the law at times. 55 In the collective agreement regulations are present that allow Atos Origin to give a person an indefinite number of fixed term contracts, which would not have been possible without the collective agreement. The take over of KPN has led to two new unions entering the arena, Union for telecom Employees (Bond voor Telecompersoneel, BTP) and Professionals for Professionals Union (Vereniging van Professionals voor Professionals, VVP), which were competing to get accepted to the negotiating table. Atos Origin HRM was afraid that having six parties (the two new unions, the three old unions, and the works council) at the negotiations could be unworkable. Atos Origin, however, was not about to refuse the new unions at the negotiation table beforehand. Because the traditional unions who had been negotiating at Atos Origin had proven to be so difficult to make compromises with, Atos Origin thought it better to see what the position of these new unions was. VVP for instance had indicated it wanted to participate in the negotiations. VVP stated it has a keener eye for the interests of a company than other unions. 56 In the end it was BTP that was invited to the negotiation table. Origin thought that one of two unions was a fair compromise. Position of the Works Council The position of the works council within Atos Origin has been and still is unique as has been described in the previous section. Already before Origin started using a collective agreement, the works council of BSO/Origin had more power than the average works council had based on article 27 of the WOR. 57 Over the years, the works council has lost some of its strength, compared to the strength of the works councils that were present during the negotiations over the first collective agreement. This is partly due to the number of takeovers since that period. As each new company became part of Atos Origin, a new member was added to the works council. This had the effect that the works council had to protect multiple interests, which has had a diluting effect. Another reason was that except for the wage regulations, virtually no room was left open for improvements to the collective agreement. The works council still plays an important role in the background. It is much more familiar with detail regulations such as the day care regulation, mobility, and part time employment, than the unions. The reason for this was and is the low union presence at Atos Origin (somewhere between 10 and 15 percent), which is spread 54 According to a union representative this takeover improved the position of the unions as there were relatively a lot of union members among the new KPN employees. 55 See chapter two on driekwart dwingend recht. 56 Automatisering Gids, week 9, 2004, Schoonheidswedstrijd Atos Origin zet kwaad bloed bij vakbonden. 57 See chapter two.

148 Developing Institutions over five unions. Unless a union has well informed members seated in the works council, it has a hard time obtaining detailed information about the company. 5.2.4 PinkRoccade 58 PinkRoccade employed 7,436 people at the end of 2003, its net revenues at that time amounted to 739 million Euros. Though PinkRoccade has operations in Canada, Belgium, and the United Kingdom, it is predominantly a nationally operating firm. Twenty-five percent of its shares are owned by the State, and fifty percent of its clients stem from the public sector and local governments. Roughly half of the employees of PinkRoccade stem from outsourcing activities. PinkRoccade offers two types of core services: Infrastructure Management and Application Services. Infrastructure Management, which stems from the former Pink Elephant, provides services related to the installation and management of IT infrastructures and databases. In Application Services, which stems from the former Roccade, software is developed, maintained, and renewed. Next to these two core activities, Pink Roccade has a number of subsidiaries which are categorized as Competence Centres, which offer expertise on a number of areas such as outsourcing, information technology, and quality care. Company History 59 PinkRoccade, in part, stems from one of the oldest IT companies in the Netherlands, National Mechanical Administration Centre (Rijkscentrale Mechanische Administratie, RMA) for the processing of large financial records and budget systems. This was way back in the punch card days. During the 1950s and 1960s, RMA grew from 30 to 120 employees. On 4 July 1969 the Council of Ministers decided to rename RMA to National Computer Centre (Rijks Computer Centrum, RCC). Where RMA had been responsible for the processing of punch cards, RCC became responsible for providing services in the field of automated information processing. In 1970 RCC employed 293 people and from then on steadily grew in worker numbers to about 600 in 1982. Because RCC was a State owned company, it felt the effects of the Government policy of reducing the size of publicly owned companies. RCC was not allowed to hire any more personnel. This forced them to hire roughly 200 people from companies such as Volmac and BSO in the following years. Because RCC fully supported training and education, it was named the training institute for IT professionals in the Netherlands in the 1980s. In 1985 66 percent of its employees took part in one or multiple trainings. 60 The Government s policy of reducing the number of publicly owned companies eventually led to RCC gaining autonomy on 1 November 1990. RCC was now incorporated as a public company, wholly owned by the State. RCC also got a new task. At the time RCC was renowned for its single line 58 http://www.pinkroccade.nl (accessed March 4, 2004). 59 PinkRoccade has published a book on the history of PinkRoccade to celebrate its 50 th anniversary (Blauwhof & Van der Krol, 2001). This section is predominantly based on this publication. 60 See also chapter three.

Case Studies 149 of service. The autonomous RCC was given the task of diversifying the company s services. In order to develop the market, RCC, Roccade from 1990 on, took over companies, such as Maatschappij voor Informaticadiensten, Instituut voor Enquête Verwerking, Bouwfonds Informatica, and Pink Elephant. Pink Elephant s history was a completely different one, and traces back to a bunch of students working for 4.5 Euro an hour for an employment agency, making backups and processing printer output in the after hours for Dow Chemical. At some point the students were offered to do the same work 24 hours a day, and they started their own firm, hiring a person to do the work during the day, whilst they were studying, and continuing doing the job in the after hours. At the start of the 1980s, Pink Elephant started attracting new clients such as the Dutch Reassurance Group in 1981, and Shell in 1983. The business was still the operation and maintenance of computer centres. In 1983 15 people worked at Pink Elephant, of which five were full timers. From then on the company grew and professionalized. Until 1985 the company grew predominantly through part timers, which made up 86 percent of the 160 employees Pink Elephant had at that time. From that point on though, because of the increasing complexity of the task, Pink Elephant started growing by employing more and more full timers. In 1994 the absolute number of part timers had grown slightly, whilst the total number of employees had increased to 800, which lowered the relative number of part timers to 20 percent. During the 1990s Pink Elephant started realizing it was not big and diverse enough to land some contracts. This tempered growth. In order to keep growing three alternative strategies were put forward. The first was to keep on growing autonomously, the second was a combination of autonomous growth and growth through takeovers and alliances, the third strategy was to merge with a larger company. In 1992 RCC took over 25 percent of Pink Elephant. Six months later this was expanded to 49 percent, and in 1993 Pink Elephant was taken over in its entirety. In 1997 the company decided it should be quoted on the stock exchange and wanted to become autonomous from the State. A number of takeover candidates were available including CSC, KPN, EDS, IBM, and Getronics. Getronics, which was preferred by the management of Roccade, was the most serious takeover candidate for a long time, but the takeover failed which brought Roccade into an awkward position for a while. This ended when in July 1999 the company, now branded as PinkRoccade, was quoted on the Amsterdam Stock Exchange. The stock quotation gave PinkRoccade the opportunity to make large acquisitions, such as TAS (1,500 people) and ASZ (1,000 people), and to win large outsourcing deals. With the takeover of Computeraid Services Ltd in 2001 (1,300 people), PinkRoccade acquired its first company in the UK, which was considered to be the first step in a development towards a second European home market. Pink Roccade s UK activities include IT management, computer centre services, maintenance, and consultancy. Other significant takeovers since 1999 included the IT departments of Campina and PCM. On March 7, 2005, PinkRoccade was taken over by Getronics. Already in 2003 the State, which still was the biggest shareholder in PinkRoccade at the time,

150 Developing Institutions indicated it wanted to sell the rest of its shares in PinkRoccade. PinkRoccade had not been doing too well since 2001 and was suffering from international competition. The State thought it best if PinkRoccade would merge with another IT company. 61 5.2.4.1 5.2.4.2 Collective bargaining at PinkRoccade 62 Roccade, being state owned almost throughout its entire history, had a long tradition with respect to collective agreements and dealing with unions. The opposite held for Pink Elephant, which had a much more pronounced individualistic culture. It is interesting to read Blauwhof and Van der Krol (2001: 92, 109-110) in this respect, where they cite one of Pink Elephant s founders, Martin van Kesteren, who says in response to a proposed collective day care regulation by a works council member: The works council, with which I had to negotiate, told me But getting children is important, is it not? That may be true, but it means you have to pay for someone else s children as well. You the sole wage earner! This means you will earn less money, because we all have to pay for day care. That s the issue. Money for me is not the issue. It was about principles. About loyalty, which you had to have. The relevance of this quote is that it identifies the problems involved in combining collective action with individual rationality. Paying for collective goods such as child care was not as normal to Pink Elephant employees as it was to Roccade employees. PinkRoccade s culture, however, is not defined by Roccade and Pink Elephant alone. Over the years some significant takeovers had taken place. If a distinction between collective and individualistic cultures is made, 60 percent of all employees could be categorized as having an individualistic background. Despite this diversity of backgrounds, it was unthinkable in 1997 that Roccade, as it was still called then, would go without a collective agreement. History and culture alone were not the only reason for sticking to a collective agreement. Over the years the collective agreement had proven to have had positive effects during takeovers. The collective agreement sent out a signal that the terms and conditions of employment were not those of a Wild West company. These takeovers were a necessity to grow and remain independent. In 1997 a framework collective agreement called BRAVO (Basis Roccade ArbeidsOvereenkomst, Basic Roccade employment agreement), which held for the entire company, was combined with a number of collective agreements and works council agreements called DAVO s (Deel ArbeidsOvereenkomst, Partial Employment agreement), which held for the subsidiaries. The reason for this structure was to accommodate all parties involved. Some parties simply felt more comfortable negotiating without the intervention of a union and some did not. The BRAVO functions as a framework regulation, the DAVOs are there to fill in the details within the margins set by the BRAVO. This structure has made it relatively easy to integrate 61 More on this takeover can be found in the Getronics case. A detailed analysis, however, will not be available as events took place after finalizing these cases and were still too fresh for parties to make comments about in relation to this study. 62 This section is partly based on an article by FNV union leader Veenhuizen (2001).

Case Studies 151 companies that had been taken over. The à la carte system has proven to be another helpful tool in this process. The first framework collective agreement at PinkRoccade did not come about easily. Many union members did not like the fact that collective wage raises were replaced by a performance related pay system. Employees who now fell under the scope of a collective agreement for the first time had mostly been keeping quiet, and the union members from former state owned subsidiaries thought they were paying for the changes made to the agreements they were used to. The unions on the other hand had done everything in order to empower the works councils during the first negotiations. During the negotiations a temporary union of works council members was founded (VHP Roccade) to make it legally possible to officially take part in the negotiations. This temporary union had been formed because the unions did not accept the works council during the negotiations. Only by becoming a union of their own, could they join the negotiations- something which the HRM department tried to prevent. Later on this temporary union merged with the VHP. 63 At the introduction of the BRAVO/DAVO structure in 1997, a great many subsidiaries existed, and their number was expanding because of takeovers. Each takeover meant the introduction of a new subsidiary and at the heyday of takeovers, 19 subsidiaries in total existed. This also implied the presence of 19 DAVOs. Although there were already plans in 2000 for decreasing the number of subsidiaries, partly because of the poor development in the IT industry, PinkRoccade waited until 2002 before it fired almost the entire group management and reorganized the subsidiaries into six sectors plus a number of competence centres. At the start of 2004, these six sectors were reduced to two core activities: Infrastructure Management and Application Services. A process of harmonization had also taken place in the terms and conditions of employment. The new subsidiaries that were formed around market segments, after the first harmonization round, had to cooperate and coordinate their actions more extensively. This improved the internal mobility of labour and pushed the necessity to harmonize the conditions of employment between these subsidiaries. This process of harmonization took place in two ways. The first was that the different subsidiaries took over best practices from each other. For instance, the introduction of a day care regulation, which until then had been a strict no-no in the individualistic culture, was introduced. The second way in which the harmonization process took place called for an adaptation on the part of the unions, which from now on were absent from all DAVO negotiations, 64 and instead focused on the BRAVO level which was increasing in size and became less of a framework agreement and more of a regular collective agreement. To some extent the unions themselves had suggested the movement towards the harmonization of conditions of employment. Having to 63 It is interesting to note that the unions had not done something similar in the case of Origin, where the works council was (and is) also a party of the negotiations. 64 This meant that some of the DAVOs lost the status of collective agreement, and instead became works council agreements.

152 Developing Institutions negotiate over a number of DAVOs in addition to the BRAVO was very time consuming for both works councils and management of the subsidiaries. A relatively unique introduction was made in the second BRAVO, namely the possibility for employees to work part time, which the unions had successfully negotiated. From now on (the BRAVO was valid from November 1999 until March 2001), each and every request to work part time was granted. The limit was 80 percent of a full time working week though, which consisted of 40 hours. 65 The conclusion of the third BRAVO collective agreement proved to be the most difficult so far. As had been the case in the whole IT industry, 2001 had proven to be a very difficult year and the years ahead were not looking too bright. Though PinkRoccade had promised a four percent wage raise for well performing employees, they decided not to back their promise due to poor results and bad market conditions. Also they planned on firing 400 so called couch sitters, people who were doing nothing because there was no project to second them to. PinkRoccade was the first company in the Netherlands to do this on this scale. They had to because most of their money was earned through secondment. PinkRoccade had offered these people individual arrangements whilst the unions wanted a social plan. In October 2002 the parties finally came to an agreement. For PinkRoccade this meant it had been able to maintain a zero percent wage raise for nine months. The three percent to which the unions agreed in October, meant an effective wage raise in 2002 of 0.75 percent. Social plans with the unions were made because of the mass lay offs. If a subsidiary fired more than 20 people, the unions would be asked to join the negotiations over the conditions under which the firings would take place. The fourth collective agreement, which was valid from June 2003 until December 2003, was an interim collective agreement, and included an average wage raise by 1.7 percent. Immediately after concluding the negotiations over the fourth collective agreement both unions and PinkRoccade management started negotiating over the next collective agreement, which would be the beginning of the end of the BRAVO/DAVO structure that had existed since 1997. Not all topics were moved from the DAVOs to the BRAVO immediately. At the time of writing experts were determining the impact of a full transition of all topics, after which the results would be discussed with both the works council and the unions. The unions had already been suggesting such a move for some time. The move had advantages for all parties: employees could switch more smoothly from one division to another; unions would be empowered because all conditions of employment were now negotiated at the central level; and human resource management would be easier to implement centrally. There would still be two wage systems, one for subsidiaries which work with billable hours, and another for subsidiaries which do not. It is the intention to merge these two systems in 2005. 65 Automatisering Gids, December 17, 1999, PinkRoccade voert vierdaagse in.

Case Studies 153 Position of the Works Council Because of the BRAVO/DAVO structure, the works council, which was comprised of a central works council and a works council for each subsidiary, was used to having a central position in the negotiations with management. The current harmonization process, which has led to the abandoning of the BRAVO/DAVO structure, may have compromised the position of the works councils operating at the subsidiary levels. The works councils indicate that they still have to get accustomed to this new situation, which is not strange as they were used to having a great deal of autonomy. The relation of the various works councils with the unions is dependent on the number of union members in the various works councils. The relationship between the central works council and the unions is not optimal. This is so because the unions do not think a works council has anything to say about the BRAVO, which they consider their territory. A representative of the central works council of PinkRoccade thinks the union attitude may have to do with the empowerment of the central works council due to the harmonization process, something which the unions may see as a threat. 5.2.5 Torex-Hiscom (isoft since 2004) Torex-Hiscom is a relatively small company employing 300 people. Its core business consists of three information systems; one for hospitals, one for general practitioners, and one for hospital pharmacies. Its systems cover well over 2,100 GP practices and 70 health care institutions, amongst which are 7 academic hospitals, general hospitals, and mental health care institutions. Altogether 50 percent of all hospital beds in the Netherlands are supported through software by Torex-Hiscom. Torex-Hiscom also operates in Belgium, Aruba, and Curacao. Company history and collective bargaining at Torex-Hiscom Torex-Hiscom s history started in 1971 when a group of programmers under the leadership of prof. dr. Bakker began developing an operating system and applications for the DEC PDP-11 mainframe system, which was used by the Leiden Academic Hospital. The reason for writing their own operating system was that they wanted to make it very efficient. It should be an operating system that the developers would know by heart. At the time they did not know that this operating system would still be used 25 years after it was initially developed. Despite the announcement of a strategic alliance with Microsoft in April 2004 this operating system is still the core of Torex-Hiscom s hospital information system. This will probably change as the whole purpose of this alliance is to develop a new generation of information systems based on Microsoft technology. In 1976, after a government grant that would make automation processes possible for hospitals, the small group of developers established a foundation named Hospital Information System (Ziekenhuis Informatie Systeem, BAZIS). BAZIS consisted of a number of IT departments of various hospitals. These participants were categorized into A and B participants. The A participants were the participants who helped develop the hospital information system, whilst B participants were solely

154 Developing Institutions using the system. The supervision of the development still took place at the Leiden Academic Hospital. Until 1995 BAZIS had been using the collective agreement for hospitals (CAO Ziekenhuizen, CAOZ) and the ABP pension fund for employers and employees in service of the Dutch Government and the educational sector. BAZIS did not function as an employer, all A participants were employed by their respective hospitals. In 1995 BAZIS took over the hospital branch of Raet and renamed itself to HISCOM. HISCOM was not a foundation anymore, but a private limited company, and became a commercial company. With the switch to HISCOM, a lot of changes occurred such as the installation of a works council. BAZIS never had had a works council, only a staff association. Also because all of the staff of the Leiden Academic Hospital who had been involved in the development of the hospital information system were now employees of HISCOM, a new collective agreement for HISCOM was made. Initially the intention was to completely copy the existing CAOZ, but this proved to be impossible because of the commercial character of HISCOM, which had now become a profit organization. The wage system, for instance, was now result dependent. This first collective agreement was valid from March 1996 until December 1997 and was signed by the HISCOM management and two unions, FNV and CFO 66. At the end of 1996 227 people were employed at HISCOM who were mostly working full time. The part time factor was 0.93. 67 In 1998 HISCOM itself was taken over by the Vanenburg business unit of Baan. The reason for this take over was that HISCOM wished to expand its market. The Dutch market had proven to be too small to be able to afford extensive product development. Unfortunately two months after the take over Baan was suffering from poor results which had a negative effect on the plans for incorporating Baan technology and expanding the market. The conditions of employment had not changed for HISCOM employees. They kept on using the same collective agreement. An attempt had been made to introduce the Vanenburg function evaluation system but these plans never got beyond the stage of an initial exploration. Plans were also made to introduce a choice model, which allowed employees to trade wage for leave days, work part time, and to experiment with teleworking. In this period HISCOM stopped the co-development of software with the A participants and the software development solely took place at HISCOM/Baan. Because Baan Group did not regard HISCOM s activities as core business, HISCOM was sold to TOREX in 2002. TOREX was an English IT company, which operated in the same market as HISCOM and wanted to become an important European player. In the Netherlands, TOREX had already taken over SMS Nederland, which wrote software for general practitioners. Purchases had also been made in Germany, so TOREX was well on its way to becoming a European player. 66 CFO stands for Christian Federation of Government personnel. Today the CFO is called CNV Publieke Zaak. 67 HISCOM annual report 1996.

Case Studies 155 Through the takeover of HISCOM, TOREX became the largest supplier of software to hospitals and general practitioners in the Netherlands. Since the takeover by TOREX, the collective agreement had been a topic of discussion for the management. They thought a collective agreement was expensive because non-performers also profited from the collective wage raises. This was why management stated it wants to introduce a performance related pay system. In the 2003 and 2004 collective agreements, however, a performance related pay system was never introduced. TOREX itself was taken over in 2004 by the UK based isoft, an international health care IT supplier with establishments in the Netherlands, the United Kingdom, Ireland, Germany, Switzerland, Australia, and Asia. It was isoft who instigated Torex-Hiscom s strategic alliance with Microsoft in 2004. Position of the Works Council The works council of Torex-Hiscom has never felt like, nor been, a negotiation party for collective agreements. All negotiations have taken place between the unions (FNV an CNV) and the management. In 2003 Torex-Hiscom wanted to introduce a performance related wage system. This was a moment for the works council to get involved, which felt that a performance related pay system was out of place for Torex- Hiscom, which did not depend on individual, but rather on collective performances. Union membership at Torex-Hiscom is uncommonly high for an IT company and is close to 40 percent. Management also wanted to make employees sell two of the holiday leave days in order to increase productivity. The works council sees support for Torex-Hiscom s development branch as one of its most important tasks, as is guarantees the continuity of Torex-Hiscom. 5.3 Case Studies of Companies Not Using a Collective Agreement 5.3.1 LogicaCMG LogicaCMG is an international supplier of IT services and wireless telecommunication. IT services includes: management, IT consultancy, systems integration, and outsourcing. Clients stem from various markets such as Energy & Utilities, Financial Services, HRM & Payroll Solutions, Industry, Transport Distribution, Mobile Business, Offshore Services, Outsourcing, Public Sector, and Telecommunications. 68 In 2003 LogicaCMG employed 19,749 people in 34 countries, of whom approximately 5,500 people were in the Netherlands. The global annual turnover in 2003 was approximately 2,550 million Euros. 69 68 http://www.logicacmg.nl (accessed March 4, 2004). 69 Annual report LogicaCMG 2003.

156 Developing Institutions Company History and Collective Bargaining at LogicaCMG 70 Both Logica and CMG were founded in the United Kingdom in the 1960s. CMG was founded in 1964 by three friends, Collins, Mills, and Gorman, who developed a program to record working hours. Collins left the newborn company before they started hiring new people, and Mills and Gorman renamed the company to Computer Management Group. Most of these newly hired people were seconded to CMG s customers, setting the future base of operations of CMG. In 1968 Readers Digest, a Dutch publisher, was looking for a direct mailing system. This was to be the first project of CMG in the Netherlands. A local office was opened in a flat in Amsterdam. CMG proved to be very successful in the Netherlands and became larger in the Netherlands than in its home country, the United Kingdom. 71 Logica was founded five years later than CMG in 1969. Logica s first base outside the United Kingdom was also in the Netherlands. During 1973-1974 it wrote a piece of software, a national gas pipeline management system, for Nederlandse Gasunie. Logica had also grown in the Netherlands, but not as much as CMG. A member of the works council of CMG states that until 1995, when CMG became quoted on the stock market, employees and management were a tight group. On the one hand employees were paid very well, had excellent conditions of employment, and were in principle employed for life. On the other hand the employees offered their unconditional loyalty to CMG. In 1995 something of the informal atmosphere of CMG was lost. Stock quotes made the company more careful of what was brought out in the open and could hurt the company. CMG had big plans and in 1999 it took over the British company Admiral, which was almost as big as CMG. This was the first time that CMG grew significantly because of a takeover. Until then it had grown predominantly due to autonomous growth, which had always guaranteed the unique culture of CMG. The takeover of Admiral had no effect on the culture of CMG in the Netherlands, but it left CMG with a serious debt, which eventually led to the merger with Logica in 2003. 72 Throughout its history CMG and Logica have never used a collective agreement in the Netherlands. The company s HRM states it has no need for unions. Most of LogicaCMG s employees are highly trained and skilled people, perfectly capable of negotiating over their own conditions of employment. The cooperation with the works council also works very well. Another argument for not inviting the unions to the negotiation table was that unions would prolong the negotiation process. Changing a regulation in the company s works council agreement already takes half a year; if the unions were to intervene, this would probably take even longer. 73 The unions have never spoken with the HRM department of LogicaCMG, 70 The interviews with both HRM and the works council took place before CMG and Logica merged. The initial case selection was CMG. Therefore this case predominantly reflects the position of CMG. 71 Archived version of the company website: http://web.archive.org/web/*/www.cmg.nl (accessed March 4, 2004). 72 Logica in the United Kingdom called the merger a takeover. CMG in the Netherlands called the takeover a merger. The stocks were shared 60/40 in the advantage of Logica. 73 Unions think the reverse is true: unions will speed up the negotiation process they say.

Case Studies 157 but they have spoken with the board of directors and the works council. LogicaCMG s policy has always been that HRM deals with employees and the works council, not the unions. In 2001, CMG changed its policy with regard to the primary conditions of employment and froze the wages. In the past CMG usually offered wages 10 percent above the average market wages. CMG had an autonomous growth ambition and this had been a successful strategy. It was a growth strategy that only worked if the market allowed for it. As the market shrunk, so did the wages. 2001 was not the first year in which CMG had decided to freeze wages. It had also happened in the early 1990s during the previous crisis in the IT industry. 74 CMG also tried to be competitive in terms of secondary conditions of employment. For instance in the case of part time employment, CMG in 2000 announced it would start offering its employees the possibility of working part time, to a minimum of 32 hours per week. Until then this had been out of the question for the company. In 1998 a company manager had said in an interview that part time employment was non-negotiable at CMG. 75 The idea was that their clients should be able to receive services at all times. The reason this policy had changed was that clients would be more sympathetic because they were more accustomed to it and more understanding. If ten percent of the 5,500 CMG employees at that time would start working part time, that would mean a loss of 25,000 working days annually. CMG hoped it could be compensated by new employees who would opt for CMG because of this very measure. Employees were also offered the possibility of earning 20 extra days off if they would work one hour extra per day, and the possibility was introduced that allowed workers to buy or sell vacation days. 76 During 2002 CMG tried to dismiss 118 people, but did not mention this to the unions because in the view of CMG their role within the company was not significant enough to become a party. 77 In the Netherlands, companies which fire at least 20 people within the same region within three months are supposed to file a report to both the CWI (Centre for Work and Income) and the unions. 78 The CWI then gets a month to decide whether such a mass dismissal is valid, and whether for instance, government support can prevent or reduce the extent of this mass dismissal. If the unions are not informed, the CWI is not required to take the request for mass dismissal into consideration. 79 Within this month, CMG managed to find a new position for nine people. So in the final request for mass dismissal, 109 people were left. What made this final request interesting in the light of this study is that CMG 74 Computable, Oktober 18, 2001, CMG bevriest de lonen. See also chapters three and four with regard to the crisis in the Dutch IT industry of the early 1990ties. 75 Computable, March 6, 1998, Geen plaats voor deeltijdwerk. 76 It is interesting to note at this point that buying and selling of days of leave was already possible in the collective agreements (ICK collective agreement, Getronics, Origin, and Roccade) at that time. 77 The works council estimated the union density at CMG at 2 percent at that time. In an interview with a different works council member in July 2004, the estimated figure was 8-10 percent. 78 Articles 3 and 4 of the Reporting of Collective Dismissal Act (Wet Melding Collectief Ontslag 1976, WMCO). 79 Articles 5 and 6 WMCO.

158 Developing Institutions did not try to dismiss these people as regular employees, but rather as temp workers. The reason for CMG doing so was that the regulations for discharging temp workers are more lenient than they are for regular workers. The Dutch Civil Code articles 7:690 and 691, which define the temp worker s position, mention that because of the particular triangular relation of the temp worker, temp agency, and the employer, more lenient regulations exist with regard to dismissal. These leniencies include, amongst others, the seniority principle. Normally in the case of a dismissal, a company is required to use the last in, first out -principle. This would have meant that CMG would not have had been able to fire this particular group of people. The reason they wanted to fire them was because they had been sitting on a couch for some time. 80 IT companies like CMG, still had AINO freshly in mind. AINO was a Dutch IT company that just recently had gone bankrupt. They had attempted to cut costs by laying off roughly one hundred people. AINO s lawyers did so based on the seniority principle, which was the legally correct manner. What made the lay offs at AINO so dramatic was that most people that were being laid off had been the ones earning the money for AINO. They were working at the clients AINO still had. The bench sitters on the other hand were still employed. Obviously this upset the AINO customers. Seeing the people that AINO sent to them initially leave and being replaced by people that could not do the job as well made a lot of them decide to end their relationship with AINO. This effectively cut the last strings to which AINO was hanging. Things were not as dramatic for CMG as they were for AINO though. Still CMG preferred to lay off people that were not working as opposed to people who were. The CWI accepted CMG s proposal and allowed CMG to dismiss these people based on them being temp workers. 81 The unions, predominantly CNV, did not accept this and took the matter to court: a case which they lost. In March 2003 the Labour Foundation (see chapter two) made dismissal through Dismissal Resolution schedule B impossible. From that point on Dismissal Resolution schedule B was only applicable to the temp agency branch. A solidity clause ( hardheidsclausule ) was introduced though for companies such as CMG. This solidity clause allowed for some leniency for companies that seconded people to their customers with regard to the seniority principle. Obviously this court case did nothing to improve relations between the unions and CMG. At the end of 2002 Logica and CMG merged to become LogicaCMG. The reason for the merger was efficiency and economizing. At the time of the merger both companies expected to save 94 million Euro annually. The overlap between the two companies was not too big and limited to four companies amongst which Shell and Vodafone. The largest economizing source was staff, globally 1,400 people could 80 A couch sitter is someone who is in between projects. He is still employed, but not doing any work. It is very common for IT companies to have couch sitters as in normal economic circumstances, these couch sitters ensure an internal flexibility. During a crises however, couch sitting is taken more seriously as it can mean that no new project is available. 81 They were fired based on Dismissal Resolution schedule B (Ontslagbesluit B). For a more elaborate discussion of the legal issues and implications of this ruling, see Beltzer (2003).

Case Studies 159 be laid off. The unions were not consulted in this merger process. The unions stated they thought the merger to be a bad move because both companies were not doing well and they did not see a reason this would change when the companies merged. 82 In March 2003, LogicaCMG announced it would dismiss another 800 people worldwide, because of a better overview of the overlap. In the meantime, LogicaCMG had asked the CWI to lay off 415 people, again using Dismissal Resolution schedule B for those employees amongst the 415 that were seconded. The unions protested against basing the dismissal on resolution B. Both the works council and the board of directors offered to keep the unions informed of the progress. They even offered the unions a paid advisory role. The unions however, only wanted to talk when they were considered a negotiating party. They stated that it was hard to defend the interests of employees when they were paid, but not present during the negotiations. The presence of the unions at the negotiations was unacceptable for the works council, because of the low union representation at LogicaCMG, which was estimated by the works council at two percent. When a draft version of the social plan was finished, it was sent to the unions for comments. The unions made a number of suggestions, and these were, nine times out of ten, taken over by the works council. After handing over the final social plan to both the CWI and the unions, the unions, to the surprise of the works council, still refused to accept the final social plan. The unions felt the social plan was too directive and almost forced people to accept the conditions of the social plan. This was very frustrating for the works council who simply did not understand how this was possible. 83 Unions do not agree with the image LogicaCMG creates of itself, its employees, and its relation to the unions. The events since 2002 have changed the way LogicaCMG employees regard the unions and LogicaCMG. According to the unions this can be demonstrated by the increase in union membership at LogicaCMG, which the unions claim has increased fivefold since 2002. 84 Though two unions members in the works council had resigned their union membership, during the works council elections of May 2004 two new union members were elected with quite convincing figures. 85 A works council member mentions that the loyalty that CMG had from its employees, at least until 2001, disappeared to a large extent because of the wage freezes and their being considered temp workers. It was as if LogicaCMG finally had shown its true face. 82 Computable, November 8, 2002, Logica koopt CMG. 83 Computable, April 3, 2003, OR Logica verwerpt kritiek vakbond and Automatisering Gids, week 8, 2003, OR LogicaCMG en vakbonden in conflict. 84 The unions claim the union density is somewhere between 12 to 15 percent. If you compare these to the estimate of the works council, which was two percent, then perhaps the unions claim comes close enough to the real figure, which cannot be determined because unions refuse to publicize the number of members they have at LogicaCMG. 85 One received 118 of 248 votes at a subsidiary, the other 75 of 180 votes. The overall attendance rate at this works council election, was 26 percent.

160 Developing Institutions Position of the Works Council 86 As can be seen in the previous section, the works council at LogicaCMG takes a relatively strong position within the company. The works council is regarded as a valuable tool for the management, as through it, management has easy access to sentiments within the company. The opinion of the works council is highly valued and its advice is adopted. A works council member says the reason for the works council s strength is the people who join it. They have a very good idea of what is going on inside the company, are very open and accessible to other employees. This is also the reason why the works council has such a good notion of what is going on inside the company. People inside the works council are often re-elected. New people who enter the works council are coached and are generally well informed of what goes on in the works council even before they join it. Things do not always run smoothly though. Currently LogicaCMG is far more hierarchically organized than CMG used to be. Things have changed because CMG used to have a decentralized hierarchical structure. Each subsidiary formed an island and was capable of making independent decisions. These islands again were controlled by sector managers, who themselves had to report to the board of directors. This was also the principal reason why not HRM, but rather the board of directors had contact with the unions. Sometimes problems could occur between the sector management and the works council. This was so because the sector management and the works council never negotiated and its (the sector management s) actions were also never discussed. Sector managers often acted before they consulted others, and this has led to a few problems. The works council is also limited because LogicaCMG has a stock market quotation. It is often not possible to communicate certain information to its supporters, their fellow colleagues outside the works council, due to an embargo. This sometimes gives the supporters the idea that the works council is not doing anything. The unions agree that the cooperation between the works council and HRM at LogicaCMG is strong. They add though, that this is dangerous because it could exclude outside influences. Members of the works council are regarded as an extension of management. This defeats the purpose of a works council, which should try to combine both the interests of the employees and the employer. One works council member remarks that this is not true. Throughout the years a number of surveys amongst employees have indicated that employees are happy with the works council s performance. Employees are also pleased with the manner in which the works council communicates: openly, clearly, and regularly. Another works council member, however, has a different opinion. Though he agrees the position of the works council so far has been subjective, he thinks the works council really has tried 86 Because the works council member that was interviewed in 2003 and interviewed unions shared a completely different view of events and their relative positions, a new interview was held with another works council member in July 2004.

Case Studies 161 to do its utmost to protect the interests of the employees. Take for instance the results of the social plan, which the unions thought were poorly negotiated. Even a labour jurist agreed the social plan was not very bad. 87 Perhaps the works council is good, but not good enough to negotiate the maximum of what was theoretically possible. The unions state that because the employees have virtually no means of addressing their interests via the works council, they turn to the unions. The unions consider the election of two the union members and the increasing union density at LogicaCMG as evidence for their growing popularity amongst LogicaCMG employees. The unions mention further on that employees are on the one hand highly trained and assertive people, but on the other hand, their individual managers determine where employees are seconded, what they do, at what tariffs, and what training they will have. The managers also assess how well they have performed. The employees have no means of complaining. The works council does not deal with individual cases and managers are always protected from above. Though there is room for questions, there is no way to contribute. This is frustrating for employees. Both works council members disagree with the unions. Employees at LogicaCMG have a number of means to contact the works council, and their complaints are always heard and dealt with. Their frustrations, if any, stem from other sources. One of the works council members claims that the works council can only be blamed for two things. Firstly, they may have been too sympathetic to the position of the management and company, though management also received a lot of support outside the works council, and secondly, on a number of occasions they have informed the employees of negotiations results only after these negotiations had been concluded. 5.3.2 Ordina Ordina is a Dutch IT service provider of IT and management consultancy. In 2003 the annual turnover was 345 million Euro, and 3,550 people were employed. Ordina has four business units targeted at four markets: Finance, Utilities-Telecom & Media, Public, and Trade, Transport & Industry. 88 Ordina manifests itself as a full service provider. Its services are provided on an hourly basis, project basis, or based on outsourcing. The various competencies include: Management Consultancy, Architecture & Design, Software Development, Network & Infrastructure, Operation & Management, and Business Solutions Development. 89 87 See Automatisering Gids, week 15, 2003, Krachtmeting ondernemingsraden LogicaCMG en vakbonden in nieuwe fase. 88 Ordina annual report 2003. 89 http://www.ordina.nl (accessed March 4, 2004).

162 Developing Institutions Company History and Collective Bargaining at Ordina 90 Ordina was founded in 1967 in France. The name Ordina stems from the French word for computer ordinateur. Ordina opened several subsidiaries in France in the next years, and in 1970 it opened its first subsidiary in Belgium.The first Dutch office was opened in October 1973. In 1976 the Dutch office became a full subsidiary of Ordina Benelux. In 1980 Ordina was taken over by the French SG2. SG2 was part of the Société Générale, one of the biggest banks in France. SG2 offered Ordina the financial backing to expand. The bureaucratic structure of SG2 however, affected Ordina during a period in which it could have expanded rapidly, but was instead held back. In 1984 negotiations started for a management buy-out by the board of directors of Ordina Netherlands. The negotiations were successful and in 1985 Ordina became an independent Dutch IT company employing 200 people. In the following years Ordina grew predominantly through strategic takeovers. Ordina focused on its core competencies and in 1988 Ordina Software was sold, and Brown Consulting Holland was bought. In 1989 companies were taken over that enabled Ordina to expand her system house activities, which was one of the fastest growing markets at that time. Through a reverse takeover Ordina expanded its consultancy activities. 91 In 1995 Ordina employed 800 people. Though takeover has remained the main growth strategy, since 1995 Ordina has been actively recruiting new personnel through intensive campaigns in Dutch IT magazines and the Ordina Open Tennis Tournament. 92 What has characterized Ordina over the years is its focus on the Dutch market. Ordina has always used a works council agreement. Though in the past regulations existed in addition to the works council agreement, these have since been abandoned. At some point they were considered too archaic. The bylaws consisted of regulations with respect to the times one was supposed to be present, or when the office would open and close. The reason Ordina has never used a collective agreement is because of the sense of entrepreneurship, which exists throughout the company. This sense of entrepreneurship can be traced back to the 1985 management buy out. The sense of creating your own company and making it grow, is still very powerful within the company. Ordina wishes to minimize external influences. Introducing a collective agreement would imply introducing the unions. The so called critical role of the unions was not accepted, and only regarded as external meddling. Harmonization plays an important role at Ordina. Over the years, Ordina had taken over a number of companies, which were controlled by a holding. At some point, this became unmanageable due to the number of companies and a process of centralization was started. All of the companies were put into four sectors. The same held true for the conditions of employment. Working with a multitude of conditions of employment is not easy. Ordina harmonizes conditions of employment by 90 Because of the limited amount of information about the works council, this information has been integrated into the main text. 91 The history of Ordina is largely based on a 20 th anniversary publication by Ordina (1993). 92 Ordina annual report 1995.

Case Studies 163 translating exogenous conditions of employment to the Ordina conditions of employment. This process of translation means the alien conditions of employment are categorized into soft and hard components. The hard components are objective in nature, and can easily be quantified. The opposite holds true for the soft components, which are more subjective. During negotiations with the works council or personnel representation of the party that was taken over, these soft components are quantified. If this quantification would lead to a deterioration of the conditions of employment, the gross wage is compensated. If however a person does not accept the new Ordina conditions of employment, he/she can keep the original conditions. These, however, will not be adjusted annually meaning this person can never have a wage raise. Sometimes the conditions of employment of a taken over party are so high, they will not be harmonized. Usually in these cases the adjustments over the years are lower than that of the rest of the company. Though Ordina considers its works council to be an important means to keep the unions out, it thinks the works council still needs to develop. This is partly due to the way the company has been treating its works council, something which the company says it is improving. The works council itself states it is more than capable of negotiating over the conditions of employment within the company. Its confidence stems from the lack of complaints from the employees. The works council thinks the unions have no surplus value to offer. Moreover, in the last couple of years, the unions have tried to put Ordina in a bad light in the media. The first example is a social plan made in 2002 when Ordina had to fire 150 people. The unions were invited to act as an advisory party. Later on, however, FNV publicly criticized the very social plan to which they had been an advisory party. A similar experience occurred in 2003 when De Unie complained Ordina was trying to fire 225 people in a way which allowed Ordina to avoid both the CWI and the unions. 5.3.3 Capgemini Capgemini is one of the world's largest providers of consultancy, technology services, and outsourcing services. Capgemini employs 48,304 people world-wide (55,576, Transiciel included) of which 8,098 (8,540, Transiciel included) in the Benelux, and reported 2003 global revenues of 5,75 billion euro of which 13 percent stems from the Benelux. Company History 93 The roots of Capgemini lie in France. At the beginning of the 1960s, two companies started operating using the same abbreviation; CAP. The first company to do so was the French based Centre d Analyse et Programmation, which started operating in 1962. The second company to do so was the English Computer Analysts and Programmers, 93 Company history based on a webpage, which was available on an earlier version of the company website. This website can be found using the Internet archive: http://web.archive.org/web/*/www.cgey.nl (accessed March 4, 2004).

164 Developing Institutions which started in 1963. In 1967 the two companies formed CAP Europe, a company whose mission was to start new IT companies outside the United Kingdom and France. This led in 1968 to CAP Nederland, a subsidiary of CAP Europe. The 1960s holds a few other roots of the Capgemini known today. These companies are: the Dutch Volmac, the French Sogeti, the American Gemini, and the Dutch Pandata. Sogeti, which stood for Société pour la gestion de l enterprise et le traitement de l information, was founded in 1967. Sogeti expanded rapidly in the years after it was founded. As it was soon outgrowing France, it took over CAP France in 1973. Through CAP France Sogeti gained a 50 percent interest in CAP Europe. This was not the last international move for Sogeti, as in 1975 it took over the financially troubled Gemini, an American based IT company. From that moment on the merged company was named Cap Gemini Sogeti. In the Netherlands, Gemini had owned Pandata, which would remain using its own name until 1989. In the 1970s the cooperation between the English and French CAP became rather problematic. The growth of CAP Europe stagnated and the only really good operating subsidiary within CAP Europe was CAP Nederland. After a number of differences of opinion, the English and French stockholders of CAP Europe decided to discontinue CAP Europe and to divide the various subsidiaries of CAP Europe. A source of conflict was CAP Nederland, which was worth the most. In August 1975 all of the employees of CAP Nederland were asked whom they preferred. At first they preferred the English Cap over the French, but after the English had informed them of their plans (in the future all decisions would be made in England), the preference moved to the French who had a vision which gave Cap Gemini Nederland relative freedom of control. During the 1980s nothing much really happened. Both Pandata and Cap Gemini Nederland developed themselves autonomously and completely independent of each other to fully grown and prospering companies. Only in 1989 did both companies decide to merge all operations into a new company named Cap Gemini Pandata. In the meantime the consultancy branch of Cap Gemini Sogeti had also developed. A number of important acquisitions had been made, predominantly in the United States where during the 1980s five large consultancy companies were taken over. From 1991 on these companies continued under the name of Gemini Consulting. At the beginning of the 1990s the largest IT company in the Netherlands was Volmac. Volmac was founded in 1966 by three former IBM employees: Van Oostrom, Leerkamp, and Mol. The three wanted to start operating internationally, and were successful when the Cap Gemini Sogeti group took a 58 percent interest in Volmac in 1992. After the merger between Cap Gemini Pandata and Volmac, the company renamed itself to Cap Volmac. In 1996, management decided to drop the name Sogeti, and renamed all of its worldwide subsidiaries to Cap Gemini. In 2001, a large merger took place between Cap Gemini, Gemini Consulting, and Ernst & Young Consulting. Until 2004 the company was named Cap Gemini Ernst & Young. Since 2004 it is named Capgemini.

Case Studies 165 Collective Bargaining at Capgemini 94 Capgemini s income stems from three major sources: services, consultancy, and outsourcing. This means that Capgemini s income depends predominantly on billable hours made by its employees. Though this holds also true for other companies discussed in this study, each company has found its own way of dealing with this situation. Capgemini s way of dealing with the multitude of seconded personnel is to focus on the individual from a market perspective. At Capgemini the idea of lifetime employment is virtually meaningless. A person works at Capgemini as long as he represents a market value for Capgemini s customers. Individuality and flexibility are important for Capgemini and are internalized in both the conditions of employment and assessment of their employees. Though the primary and secondary conditions of employment have been fixed to some extent, the tertiary conditions of employment are offered à la carte. Certain events, such as special leave and extreme mobility compensation, 95 are not fixed. Both events are solved on a personal basis. Assessment of employees is based on three factors: performance, development, and market value. Employees are assessed by so called people managers, these are most of the time senior managers who guide junior employees and make arrangements with regard to training and educations. A people manager determines how well a person has performed, how professional a person has become, and what can be expected of him in the future. The market value, which is determined twice per year, in combination with the previously mentioned factors, determines whether a person will receive a wage raise. So it is possible a person performs in an excellent manner without getting a wage raise, because his market value has decreased. The market value is determined by the amount a customer is willing to pay for a certain service. Capgemini s works council states that a lot is demanded of Capgemini employees. Every employee is expected to continuously develop himself. This implies that the working week starts Sunday evenings and ends Friday nights. Everyone is expected to work 40 hours per week. Also no distinction in region is made, which means that someone living in Utrecht may end up working in Groningen. 96 On average a Capgemini employee spends 55 work related hours per week because of work, training and travelling. On the other hand, this is something for which they are well compensated. During the years of growth employees on average had wage raises in excess of 10 percent annually. People enjoy working for Capgemini because they can develop themselves. Capgemini s wage policy can be considered as reasonably stable. This has both advantages and disadvantages. Because they are rewarding people in both prosperous times and less prosperous times in the same manner, they do not have to fire people as much as other companies have to. On the other hand this has the effect 94 Because of the limited amount of information available, the information about the works council has been integrated in the main text. 95 Compensation when an employee has to travel abroad or very long distances. 96 A single trip of 130 miles.

166 Developing Institutions that they do not hire people as easily, and evaluations are made more carefully. Also it has lead to wage decreases. Though wage decreases are not common in the Dutch IT industry 97, in 2003 Capgemini asked approximately 275 specialists, whose market value had decreased, to take a step backwards and accept an average wage decrease of 10 percent. Management states this had predominantly caused consternation outside the company. The unions were particularly concerned with the fact that employees who had earned a great deal of money for Capgemini in the past, were now being sidetracked. 98 Within the company employees were very familiar with their market value, so the decrease did not come as a surprise to them. Interestingly enough, on the forums of a Dutch IT magazine, a number of Capgemini employees complained about the wage decrease. 99 During 1998-2000 the client used to come to an IT company with a specific question. For that question, a solution was found, and this solution came at a price. Since the market has changed, clients now look for solutions within a smaller budget. This creates problems for Capgemini as it now had to find a solution within the specific budget. This could only be solved through innovation and connecting wages to the market value. Since the 1970s, Capgemini has had a tradition of negotiating over the conditions of employment with their works council. In a letter to Automatisering Gids, 100 a Dutch weekly IT magazine, in 2002, both the chairman of the works council and the director of people and professions outline the position of the works council and the unions within Capgemini. Capgemini believes works councils are changing and becoming more professional. The Capgemini works council is a good example of this change. The times when conditions of employment were a separate topic on the agenda lie far behind at Capgemini. The works council is actively participating and involved in all discussions concerning the interests of the company. Capgemini states that: once management and the works council are on their way in such a model, there is little room left for the one-sided input of the unions (x percent, or else!). Unions should stop focusing on single topics and off-the-peg clothes for a grey mass, where the development is made-to-measure on a company level. 101 Despite the fact that harmonization plays an important role at Capgemini, and that Capgemini has grown considerably through both takeovers and outsourcing projects, Capgemini does not use a collective agreement. Capgemini says it never saw a reason to negotiate with the unions. They have three arguments for not negotiating with the unions. Firstly, negotiating with the union has no surplus value. The negotiations with the works council were never problematic. This is because the people who are working at Capgemini all are highly educated and are more than capable of acquiring and processing information without having to make enormous 97 The last known salary decrease had been AINO in 2002, which was on the verge of a bankruptcy and asked most of its employees to accept a lower wage to save the company. 98 Computable, January 9, 2003, Salarisverlaging CGEY valt slecht bij CNV. 99 http://forum.computable.nl/messages/32/1479.html?1080907213 (accessed March 4, 2004). 100 Automatisering Gids, week 48, 2002, De kracht van medezeggenschap. 101 Perhaps the reply was directed at the unions in general as the unions in the IT industry had already been promoting choice systems since 1993 (see chapter four).

Case Studies 167 time investments. Secondly, the representation of the unions is very meagre. More people present themselves as members for the works council than there are union members within Capgemini. Thirdly, negotiating with the unions may compromise the company s autonomy. Capgemini admits collective agreements can have benefits over a works council agreement. Particularly if a company has a lot of outsourcing activities, or takes over a lot of companies, a collective agreement can pacify the party being outsourced or taken over. The conditions of employment always represent the culture of a company. 102 Another reason for using a collective agreement, according to Capgemini, would be the fact that each and every employee must accept its conditions of employment. In practice though, there are hardly any employees not accepting the works council agreement. Despite not accepting the unions as a negotiation party, Capgemini HRM accepts the unions as professionals. To quote the HRM director of Cap Gemini: Such a professional relationship would be undesirable for a works council. Take for instance the situation of a mass dismissal. If you would negotiate over such an event with the works council, there would be a conflict of interests, and the works council could not possess the required objectivity. For some matters, the unions are simply a more easy negotiation party. 5.3.4 Centric Centric activities include consultancy, IT solutions, software engineering, e-business, systems integration, managed ICT Services, and training. Centric provides total IT solutions for customers in various branches such as government and local government, financial services, housing associations, and wholesale. Centric is the largest Dutch IT company not quoted on the stock exchange. In 2003 the net revenue was 235 million Euro, and 2,850 people were employed. Outside the Netherlands, Centric has offices in both Belgium and Germany. 103 Company History and Collective Bargaining at Centric 104 Centric is Gerard Sanderink and Gerard Sanderink is Centric. The person and the company are so intertwined, even more so than Ton Risseeuw at Getronics, that the two cannot be discussed separately. The company history is therefore also the history of Sanderink. Sanderink started in the IT industry in 1974 on a research fellowship to work on the transmission of data over a permanent connection. From 1975-1977 Sanderink was an IT consultant for Stork, and from 1977-1978 for BSO. A colleague 102 A good example of a difficult outsourcing project was that of Accu in 1998. Accu was the computing centre for the university of Utrecht. Most employees refused to be taken over by Capgemini, and when after 14 months of negotiating, Accu was finally taken over, 25 percent of its employees had left (Computable, March 11, 1999, Cap Gemini rondt overname Accu af ). 103 http://www.centric.nl (accessed March 4, 2004). 104 Because of the limited amount of information, the information about the works council has been integrated into the main text.

168 Developing Institutions and future partner of Sanderink (Hans Quellhorst) mentioned to Sanderink the plan to start his own business. Quellhorst was dissatisfied with the amount of work he was doing compared to the pay he was receiving. Because the two, Sanderink and Quellhorst, were complementary they thought it was a good idea to go into this venture together. In a radio interview Sanderink mentioned he would not have started his own business at that time were it not for Quellhorst. 105 The company they both started in 1978 was named ICT Automatisering. Their first job was to second themselves to Philips Medical Systems, where they worked on the development of medical appliances. Quellhorst and Sanderink eventually broke in 1992 because of a difference in view. In the same year Sanderink started the Sanderink Group, which grew through autonomous growth, but predominantly because of acquisitions. Since 2000 the company is named Centric. Sanderink is a self made man with a sober personality. One of Centric s works council members says about Sanderink: He appears to attribute the same qualities to his employees, whom he considers to be entrepreneurs that have to bear the same risks as he himself. Centric has no board of directors, nor does it have a supervisory board. All decisions are made by Sanderink who admits to consulting three to four managers within the company. 106 Centric s primary growth strategy is based on acquisitions. In June 2004, Sanderink took over his 44 th company since the beginning of Centric in 2000. Usually the companies that are taken over are not doing too well. Sanderink steps in and cuts all costs, weeds out the clients, steps through all of the personnel files and fires anyone who does not fit a certain profile. All dismissals are given on a personal basis by Sanderink himself. Each employee gets two interviews. The first is to inform the person being dismissed, the second to negotiate the terms of the dismissal. This is to avoid emotions during the negotiation process. Sanderink claims that none of the people he has laid off bear hard feelings. 107 Not all takeovers are as harsh though and sometimes companies are acquired simply to emphasize a certain niche within the company. Though these new companies are reorganized as well, the aim is not so much making the company profitable, as they usually already were, but more at fitting the company into the Centric structure. Centric does not have a collective agreement. Instead it uses three personnel guides, one for each of the three holdings of the Centric Group. Sometimes one of the acquired companies falls under the scope of the ICK collective agreement. As soon as such a company is taken over, the membership of the Werkgeversvereniging ICT is resigned, and after some time the company will start using the personnel guide of the division the company now falls under. Harmonization as such is not such a big issue at Centric as it is a continuing process. Sanderink though feels strongly about a company s culture. In order to fit an acquired company into the Centric Holding, Sanderink mixes up the employees. A number of employees are taken out of 105 EO, June 24, 2004 9:04 am, radio 1, De Ochtenden: het interview, http://www.eo.nl/portals/programs/episode.jsp?episode=4614952 (accessed July 12, 2004). 106 Ibid. 107 Ibid.

Case Studies 169 the acquired company and put somewhere else into the holding. This also happens the other way around. Sanderink feels it benefits the company. If after some time a person still does not fit within the company, he is fired. Control is of the utmost importance for the success of the company. Also, without rigorous measures, it is impossible to make an acquired company profitable. 108 Centric has no central works council. The reason for this is that Sanderink does not want to give a central works council the same negotiation rights as he currently does to the works councils. Though within Centric the possibility of a central works council has been discussed, Sanderink does not want such a body to be able to have a say in his acquisition decisions. Currently Sanderink informs the management of the various subsidiaries, which then communicate this information to their respective works councils. This way Sanderink does not have to negotiate directly with the works council. Because Centric consists of a number of firms, actions usually do not have a company wide effect. This has the effect that changes in one subsidiary do not affect sentiments at a company level. It also affects the way in which the two works council members that have been interviewed perceive the company. One member of the works council stated for instance that most of Centric s employees in his subsidiary desired a collective agreement, whilst the other member stated that union sentiments were usually short lived and not widespread throughout the company. An example of this was the introduction of an average wage pension scheme at one of the subsidiaries. Because of all the takeovers, Centric had a lot of different pension schemes. Some of these pension schemes were based on a so called final wage. This implied that the pension one would receive, would be based on 70 percent of the last wage earned. If someone had made a number of promotions, this could have a very positive effect on his final pension, which the employer would have to pay. These costs are called back service costs. Centric thought the final wage pension scheme too expensive and unfair to other subsidiaries that did not have the same pension scheme. Centric sent out a memo informing people they could opt for either a new pension scheme based on stocks, and allowing for future wage raise, or maintain the final wage system, without Centric paying for the back service costs. The latter option effectively meant not getting a wage raise, because the wage raise was used to cover the back service costs. Introducing this option meant that changes to conditions of employment were a salary problem, and not a pension problem, which meant that the works council had no voice in this matter. They had attempted to involve the unions in this matter, but this had no effect in the end. Unions are sometimes also more welcome at subsidiaries because of a lack of wage raises at that subsidiary. Centric has no collective wage raises. Instead, wage raises are based on the performance of a subsidiary. If a subsidiary does not make a profit, employees of that subsidiary will get no wage raise, which is sometimes hard to bear. In both instances the unions have never managed to get a company wide foothold at Centric because most of these occasions are restricted to a single subsidiary. 108 Ibid.

170 Developing Institutions 5.3.5 Baan (SSA Global Since 2003) Together with Torex-Hiscom, Baan is the only software development company amongst the cases presented here. Baan is currently in the hands of two American investors General Atlantic Partners and Cerberus Capital management, which made Baan a subsidiary of SSA Global. Baan employed 400 people in 2004. 1998 was Baan s best year. At the time it employed 5,101 people full time world wide, of which approximately 1,200 in the Netherlands. The annual turnover in 1998 was US$ 735 million. Company History and Collective Bargaining at Baan 109 Baan s origin lies in 1978 when Jan Baan started a financial and managerial support company. Baan first used a programmable calculator to refine his models for the financial management of a company. In 1979 Baan acquired a computer and hired a programmer to write software for the financial management services. This developed into a business where Baan was selling both (micro)computers and software as a total package. So, Baan was now both an advisor and a dealer. The philosophy was to offer the client a total package, without risks involved, because the client knows what the whole package will cost him. Later on the financial management activities were sold and Baan became a system house. The personnel consisted solely of programmers who worked in the evening hours. In 1980 the first full time programmer was hired. In 1981 Baan renamed itself Baan Automatisering and was operating at a national level. Baan s expertise focused more and more on software to assist operational management. In 1985 45 people were employed and in 1987 120. In 1981 Jan Baan s brother Paul also joined the company. During 1981 Jan Baan visited Silicon Valley for the first time, something he would continue doing twice a year, and became familiar with the operating system Unix, which allowed programs to run on different computers that were using this operating system. This meant software could be used on a larger scale and did not necessarily have to be written for one specific system. Because at the time no one in the Netherlands was interested in Unix, Baan experimented with the system. At the same time Baan worked at industrializing its software. Software at the time was predominantly bespoke work and written bottom up for each client. This was tremendously labour and cost intensive. What Baan usually did was write an application for client A and customize this application for client B, and so on. During 1985-1989, Baan expanded tenfold and now employed 450 people. In 1987 Baan was one of the leading Unix system houses in the Netherlands. 1987 is also the year in which Baan s product were exported to England and Germany. There was 109 Baan s history is predominantly based on Post (1996) and Baan s annual report of 1998. For a more detailed analysis of Baan s rise and fall see Mark Houben and Jeroen Wester, 2001, Baan: Opkomst en Ondergang van een Uniek Softwarebedrijf, Business Contact/Het Financieele Dagblad. Because of the limited amount of information, the information about the works council has been integrated into the main text.

Case Studies 171 also a multimillion dollar deal with Olivetti, who started using Baan s products as a strategic international product for the integrated production control market. In 1989 Baan started focussing on its core business again. The Dutch IT industry was slowly declining at the time, and Baan thought it was time to return to the industrial production of software. The number of employees was reduced to 215 in the Netherlands. The personnel department helped find new jobs for the redundant employees and those that gave up their job voluntarily were rewarded with a check for four month s pay. The reorganization came as a shock and was criticized by the Reformed Societal Union (Reformatorische Maatschappelijk Unie, RMU). The RMU was founded in 1983 and was not a traditional union, as it also offered its services to employers and one-man businesses. Because of its reformatory nature and the reformatory nature of Baan, it was the largest union at Baan. The RMU addressed Baan in the press, stating that Baan s reorganization was premature and that Baan should have informed the unions about the mass redundancies. Baan replied that it only expressed the intention of redundancies and was in fact searching for alternatives. In December 1989 most of the people that were being made redundant had found new jobs and mass redundancies were not required anymore. The reorganization process was finished in 1990. In 1990 Baan started to reclaim the market again. It split up its operations into Baan International, which would focus on the ongoing development of Baan software, and Baan Info Systems, which focused on the distribution of Baan software in the Netherlands. It was also in 1990 that a works council was introduced at Baan for the first time. Baan started expanding rapidly again both nationally and internationally. In 1994 the company employed close to 1,000 employees and further increased its personnel base fivefold in the period up to 1998. In 1994 and 1996 Baan managed to land large contracts at both Boeing and Ford. Despite 1998 being Baan s best year, it also was the start of a new and difficult period. Together with the departure of the family Baan from the company in 1998, the market Enterprise Resource Planning (ERP) market was in serious decline. In Baan s 1998 annual report, the company listed a number of factors that played a role in this decline. Clients were more and more focussing at solving the millennium related problem, which had the effect of clients postponing investments, and there was more overall competition, which affected the price clients were willing to pay. 110 In 1998 Baan dismissed 120 people in the Netherlands and a recruitment freeze was introduced. But by the middle of 1999 Baan started recruiting again, specifically for its research and development team. This respite was only temporary though. Baan s stock prices had been unrealistically high, and when American stock authorities demanded a different manner of reporting, the stocks started plummeting and Baan was in serious trouble. Four scenarios were introduced to save Baan, which included Baan being (partially) taken over, the hiving off of subsidiaries, an external financer, or bankruptcy combined with a restart. In June 2000 Baan was fully taken over by the British Invensys. The takeover meant a loss of approximately 200 jobs in the 110 Baan annual report 1998.

172 Developing Institutions Netherlands. The unions were pleased with the takeover as bankruptcy would undoubtedly have led to a more serious loss of jobs. The labour market conditions at that time were very positive for IT personnel, but for the non-it ers solutions had to be found. During September 2000 a new, this time more serious, reorganization, was announced. This time, instead of having to make mass dismissals, personnel was already finding its own jobs outside Baan. In June 2002, 1,000 people were left at Baan in the Netherlands, and 3,400 worldwide. Invensys, in the meantime, was not doing so well itself and also started reorganizing. In July 2003, Baan became a subsidiary of SSA Global Technologies and new reorganizations were announced. During 2000, because of the takeover by Invensys, Baan had started a centralization process. Until then every subsidiary within Baan had had its own conditions of employment, making internal transitions difficult. It also had the effect that two people having the same function at different subsidiaries could have a different income. Part of the centralization process was the introduction of a central works council. This was to overcome the problem of the various works councils trying to negotiate the best conditions for their subsidiary. Since the departure of the brothers Baan, the relationship between the HRM department and the works councils had intensified. In 2002 a company wide works council agreement was introduced named General Conditions of Employment Regulation (Regulering Algemene Arbeidsvoorwaarden, RAA). The central works council and HRM are the only negotiation parties to this regulation. The works council is often warned in advance of HRM decisions. Experience has taught that it is much more efficient to discuss problems earlier on. Because of the numerous reorganizations at Baan, the works council has gained a lot of expertise. Being involved early on in the decision making process means that everything does not boil down to the actual moment of decision. Though relations with the RMU is quite good, both the works council and HRM indicate they do not think union presence is required and have never seen the point of nor need for a collective agreement. In the recent past, when during reorganizations social plans needed to be drawn up, the unions informed the works council that if they (the unions) would do the negotiations, they would fully take them over. The works council felt it was quite capable of conducting these negotiations by itself. Moreover, the works council stated that the unions only represented 15 percent of the employees, and not 100 percent as was in the case of the works council. 5.4 Flexibility: Collective Agreements vs. Works Council Agreements In this section the various works council agreements and collective agreements of the selected cases are compared. The goal is not so much to make a thorough analysis of its contents, but more a matter of determining whether the collective agreements is as rigorous as it is often labelled. This will be done through a brief summary (see tables 5.3 thru 5.7) of the individual and flexible components present in the various agreements.

Case Studies 173 Table 5.3 The flexibility regulations contained within the EDS collective agreement and Ordina works council agreement EDS Ordina EDS uses both a collective agreement and a works council agreement named Wegwijs. Most of the regulations found in Wegwijs, were also present in the collective agreement. Where the collective agreement outlined most of the regulations, Wegwijs used to contain the details. In 2004 EDS, together with the unions, started harmonizing both agreements in favour of the collective agreement. Parties to the collective agreement: - Electronic data Systems (EDS) International B.V. - FNV Bondgenoten - Dienstenbond CNV - De Unie Scope of Action: - All employees with a Dutch contract of employment Flexibility in primary conditions of employment: - Performance related pay - Part time working Flexibility in secondary conditions of employment: - A la carte system sources: Salary, holiday hours, employment related hours, and EDS hours. - A la carte system destinations: Saving of employment related hours, saving EDS hours, purchase of additional hours of leave (max. 64 per year), save-as-you-earn regulation, sabbatical leave, and pension scheme. Ordina uses a general works council agreement for all its subsidiaries. Parties to the works council agreement: - Ordina (subsidiaries of the Ordina Groep) - Works council Scope of Action: - All employees with a contract of employment which states the general works council agreement applies to it. Flexibility in primary conditions of employment: - Performance related pay Flexibility in secondary conditions of employment: - A la carte system sources: Salary, extra allowance, end of year payment, bonus or percentage of profits, holiday hours, overtime bonus, anniversary payment, bonus for bringing in employees, positive balance lease car, and positive balance alternative mobility regulation. - A la carte system destinations: Salary, holidays, pre pension, negative balance lease car, tax free contribution to day care, and private PC regulation. Parties to the works council agreement: - Electronic data Systems (EDS) International - Works council Sources: - EDS: CAO EDS, valid from January 1, 2003 till January 1, 2004. - EDS: First version of the harmonized collective agreement and works council agreement Wegwijs, July 2004. - Ordina: Algemene arbeidsvoorwaarden Ordina, versie 8.0, 2003. - Ordina: Flexreglement Ordina, versie 5.0, 2003.

174 Developing Institutions Table 5.4 The flexibility regulations contained within the Atos Origin collective agreement and Capgemini works council agreement Atos Origin Atos Origin uses a collective agreement, which it calls a conditions of employment manual. Parties to the collective agreement: - Atos Origin Nederland B.V. - FNV Bondgenoten - Dienstenbond CNV - De Unie - The works council of Atos Origin Nederland B.V. Atos Origin allows the works council to be a party based on article 32, member 1 of the WOR. Scope of Action: - All employees with a contract of employment Capgemini Capgemini uses a set of 14 regulations, which together make up a general works council agreement. The core of the works council agreement is the so called PRM charter, which defines the culture and vision of Capgemini, but also includes primary and secondary conditions of employment. Parties to the works council agreement: - Cap Gemini Ernst & Young Nederland B.V. - Works council Scope of Action: - All employees with a contract of employment which states the general works council agreement applies to it. Flexibility in primary conditions of employment: - Performance related pay in combination with employability regulation - Part time working - Flexible working hours Flexibility in secondary conditions of employment: - A la carte system sources: Salary, holiday allowance, end of year payment, labour market bonus, holiday hours, overtime hours, anniversary payment, bonus for bringing in employees, and premium saving amount. - A la carte system destinations: Long term leave, early retirement, holidays, higher pension, staff shop tickets, cash, bicycle plan, and private PC regulation. Sources: - Atos Origin: Arbeidsvoorwaardengids 2001-2003, versie 1.1, June 2002. - Capgemini: PRM information map, July 2002. - Capgemini: PRM charter 2003, January 2003. Flexibility in primary conditions of employment: - Performance related pay related to an individual s market value - Flexible working hours Flexibility in secondary conditions of employment: - A la carte system sources: Salary, percentage of profits, leave days, holidays, and mobility budget.. - A la carte system destinations: Gross taxed benefits, such as pension, holiday bonus, unemployment benefit and disablement insurance benefit, death benefit, moving expenses, and anniversary payment. Net taxed benefits, such as mobility, teleworking, day care, training additional health costs, and moving expenses.

Case Studies 175 Table 5.5 The flexibility regulations contained within the Torex-Hiscom collective agreement and Baan works council agreement Torex-Hiscom (isoft since 2004) Baan (SSA Global since 2003) Torex-Hiscom uses a collective agreement. Parties to the collective agreement: - Hiscom B.V. - FNV Bondgenoten - CFO Scope of Action: - All employees with a contract of employment Flexibility in primary conditions of employment: - Part time working Flexibility in secondary conditions of employment: - A la carte system sources: Days of leave, holidays, and result dependent bonus. Not all sources can be used for all destinations. - A la carte system destinations: One time gross allowance, holidays, day care, sabbatical, bicycle plan, and private PC regulation. Sources: - Torex-Hiscom: CAO Hiscom B.V., August 2002. - Baan: RAA 2003, 2003. - Baan: Compass policies 2003, 2003 In 2002 Baan started a harmonization process of its employment regulations. The RAA 2003 is the first general works council agreement of Baan. Next to the RAA, Baan also has a policy compass, which is a general code of conduct, but also contains details regarding the various regulations mentioned in the RAA. Parties to the works council agreement: - Baan Development B.V. - Works council Scope of Action: - All employees with a contract of employment. Individual conditions supersede the RAA 2003. Flexibility in primary conditions of employment: - Performance related pay - Flexible working hours - Part time working Flexibility in secondary conditions of employment: - A la carte system sources: Salary, and holidays. - A la carte system destinations: Salary, sabbatical, holidays, and private PC regulation.

176 Table 5.6 Developing Institutions The flexibility regulations contained within the Getronics collective agreement and LogicaCMG works council agreement Getronics Getronics uses a collective agreement. Parties to the collective agreement: - Getronics Nederland B.V. - FNV Bondgenoten - Dienstenbond CNV - De Unie Scope of Action: - All employees with a contract of employment at one of the subsidiaries mentioned in article 1 and earning less than 51,000 Euro annually. Flexibility in primary conditions of employment: - Performance related pay - Part time working Flexibility in secondary conditions of employment: - A la carte system sources: Incentives, salary, bonus, over time, availability days, and days of leave. - A la carte system destinations: Days of leave, extra pension deposit, extra income, day care, and private PC regulation. LogicaCMG LogicaCMG uses a general works council agreement. Parties to the works council agreement: - LogicaCMG - Works council Sources: - Getronics: Raam CAO 2003-2004, January 2004. - LogicaCMG: CMG Personeelsgids deel I en II, 2003. - PinkRoccade: Basis Arbeidsvoorwaarden Overeenkomst PinkRoccade NV (BRAVO), 2003. - Centric: Personeelsgids Centric IT Solutions, versie 2003.1, 2003. Scope of Action: - All employees with a contract of employment. Flexibility in primary conditions of employment: - Performance related pay - Part time working Flexibility in secondary conditions of employment: - A la carte system sources: Holidays, salary, and days of leave. - A la carte system destinations: Holidays, salary, and days of leave.

Case Studies 177 Table 5.7 The flexibility regulations contained within the PinkRoccade collective agreement and Centric works council agreement PinkRoccade PinkRoccade uses both a collective agreement, and multiple works council agreements for each subsidiary. PinkRoccade is harmonizing the works council agreements into the collective agreement to save costs. This harmonization process will take a number of years, but the goal is to have one single collective agreement for the entire company. The collective agreement used to only demarcate the outlines, whilst the works council agreements would fill in the details per subsidiary. Parties to the collective agreement: - PinkRoccade N.V. - FNV Bondgenoten - Dienstenbond CNV - De Unie During the negotiations over the first company wide collective agreement at PinkRoccade, the works council also was a party because they had established a temporary union (VHP Roccade). Centric Centric uses a works council agreements for each of its three subsidiaries. Parties to the works council agreement: - Centric IT Solutions B.V. - Centric Business Solutions Rotterdam B.V. - Works council Scope of Action: - All employees with a contract of employment. Individual conditions supersede the works council agreement. Flexibility in primary conditions of employment: - Performance related pay - Part time working Flexibility in secondary conditions of employment: - Each subsidiary has its own specific regulations. No à la carte regulations available. Scope of Action: - All employees that have been categorized in a salary class. Flexibility in primary conditions of employment: - Performance related pay - Part time working Flexibility in secondary conditions of employment: - A la carte system sources: Shorter working hours, profit sharing, salary, holidays, and holiday money. - A la carte system destinations: Holidays, day care, parental leave, pension build up, sabbatical leave, and salary. Sources: - PinkRoccade: Basis Arbeidsvoorwaarden Overeenkomst PinkRoccade NV (BRAVO), 2003. - Centric: Personeelsgids Centric IT Solutions, versie 2003.1, 2003.

178 Developing Institutions Tables 5.3 thru 5.7 demonstrate that the differences in flexibility between collective agreements and works council agreements amongst the ten cases are only marginal. Most of the companies use a system of performance related pay, except Torex- Hiscom, and most of them use an à la carte system, except Centric. The lack of a performance related pay system at Torex-Hiscom can be explained by the nature of the work, which is basically programming. The topic has been discussed at Torex- Hiscom and ideas such as counting the number of bugs a programmer produces have been considered. Working part-time was once considered an impossibility, but as the results show, almost every company allows its employees to work part time. Part time working is mostly restricted to a minimum of 80 percent of a regular working week of 40 hours, and usually dependent on a manager s decision. It is interesting to note that a number of companies employing a collective agreement also use a works council agreement, such as EDS and PinkRoccade. It is also interesting to note that both companies are attempting to incorporate these works council agreements into their respective collective agreements. It seems that their original function, to accommodate individual preferences at various subsidiaries, is losing out to the costs these extra negotiations bring with them. Companies now have to negotiate with both the works council and the unions, which costs extra time and money. The works council agreements may also have had a transitory function, allowing the various parties to get accustomed to collective arrangements. This holds true for both EDS and PinkRoccade, which were made up of various companies with different cultures. Though on average the position of the works council is restricted to article 27 of the WOR, the works council has been a negotiation party at two collective agreements. It is interesting to see that the negotiation parties (Unie, CNV, FNV, MHP, and Origin) were very accommodating to the works council of Origin, and not (CNV, FNV, and PinkRoccade) at PinkRoccade, where the works council had to revert to a trick to get accepted at the negotiation table. In that sense, though the works council s power may decrease in the coming years, the collective agreements have not had a negative impact on the position of the works council in general, and in some cases even had an empowering effect. 5.5 Conclusion As has been stated at the beginning of this chapter, the above case studies provide the most basic and direct information as to why companies are using a collective agreement or not. There are two main explanations for the presence or absence of a collective agreement in the companies discussed above. The first explanation has to do with the lineage, or origin, of a company. The second explanation is geared towards the takeover strategy of a company. There is a third explanation that resonates with these two basic explanations and has to do with corporate culture. These three explanations will be briefly discussed below, but are expanded upon in

Case Studies 179 the concluding chapter of this study. There these explanations will also be linked to the hypotheses set out in the first chapter. The first explanation has to do with the lineage of a company. These lineages are an important feature of the Dutch IT industry. In chapter three, a number of lineages were identified: (1) companies that originated from the early data and mathematical centres, (2) foreign IT companies that settled in the Netherlands during the 1970s and 1980s, and (3) early entrepreneurial start ups that grew exponentially during 1970-1990. If the above ten cases are compared to these lineages it is interesting to note that all of the decedents of the first lineage are still using a collective agreement today. The first lineage is associated with the (public) mathematical and administrative centres that first started using computers in the Netherlands, and which later on were privatized. These privatized companies all started out using either a legacy collective agreement from the previous company they belonged to, or started using a new collective agreement. PinkRoccade and Torex-Hiscom today are still using a collective agreement, because most people working in these companies have been falling under the scope of a collective agreement through their entire employment. Companies belonging to the other two lineages, foreign companies and entrepreneurial firms, had different starting points. Foreign companies tried to do what was normal in their home countries and offered individual agreements or works council agreements. These companies refused to deal with unions and contributed to the non-union sentiments in the Dutch IT industry. Examples of such companies include LogicaCMG, Capgemini, and Ordina. EDS could have been added to this list as well, but eventually started using a collective agreement. This will be explained below. Entrepreneurial firms had similar sentiments as the foreign companies and apparently thought: Why bother with the unions? We created our own work, built our own firms and managed perfectly without the unions. Baan and Centric are good examples of such companies. The first explanation, lineage, cannot cover for all the cases. Atos Origin for instance is difficult to place, but also companies such as EDS and Getronics that use a collective agreement despite their non-union origins. Why have they started using a collective agreement? In order to explain this a second factor needs to be introduced, which is growth strategy. There exist three primary growth strategies for companies in the Dutch IT industry: (1) autonomous growth, (2) growth through takeovers, or (3) growth through insourcing projects. Autonomous growth means companies grow by attracting new people instead of taking over other companies to expand the number of employees. Primary examples here are Torex-Hiscom, LogicaCMG and Baan. This does not hold true for companies that grow predominantly through takeovers or through insourcing projects. Both these latter growth impulses imply that these companies have to try to fit the new employees stemming from these companies, or parts of companies in the case of insourcing projects, into the existing company. Both Getronics and EDS have started using a collective agreement because a collective agreement allowed them to grow or attract new business. Getronics wanted to take over Roccade, but could not do so without a collective agreement. EDS did most of its insourcing projects amongst public organizations who wished to outsource their

180 Developing Institutions IT departments. Having a collective agreement meant there was less resistance amongst employees of these insourced IT departments against EDS becoming their new employer, which meant that EDS could attract more business. How then are these two explanations related to one another? The answer lies in the differences in corporate culture between companies. Following Kreps (1990) and Denzau and North (1994), corporate culture was identified in chapter one as ideologies and mental models formed from common experiences. Corporate culture is a set of collective templates that can be used for action. Company culture is not something material, but rather something represented through the ideas of people working in a company and the values they share. The notion of corporate culture can easily combine the two explanations discussed above. It explains that companies stemming from a different background (and thus sharing different common experiences!) have different corporate cultures. First lineage companies cultures have developed in the presence of a collective agreement. Second and third lineage companies have developed without such a presence. This does not imply that all corporate cultures of companies sharing the same origin are the same, but they share a common factor in their respective histories: the presence or absence of a collective agreement, and thus dealing with unions as well. Corporate culture also accounts for the second basic explanation, that of growth strategy. Autonomous growth allows a company to select new employees. It is also likely that potential employees partly select the company they want to work for based on its culture. In the case of takeovers or insourcing projects this is entirely different. Large groups of employees that have totally different cultures can enter a firm. In the case of the merger between Pink Elephant (a third lineage company) and Roccade (a first lineage company) this led to various adaptive strategies of both employees and the employer. A complex system, which included a collective agreement and many works council agreements was introduced to accommodate both camps. Only after number of years, when the cultures had adapted to each other, PinkRoccade tried to revert to a system with a single collective agreement. 111 Corporate culture also explains why Getronics and EDS started using a collective agreement. EDS was largely made up of insourced employees that were used to working with a collective agreement. EDS slowly introduced the possibility of a collective agreement in the company to see how it fitted with the employees. In the case of Getronics, the employees were happy to finally have the unions negotiate for them as they thought it would strengthen their negotiation position. In this respect it is also interesting to look at Atos Origin, which partly stemmed from Philips Communications & Processing Services, a company which has been employing a collective agreement for a long time. Yet, Atos Origin, or rather its predecessor Origin, only started using a collective agreement after the works councils, of the former BSO/Origin and Philips Communications & Processing Services, lobbied for the unions, and the legal department admitted that harmonizing the various existing agreements would be easiest using a collective agreement. BSO/Origin had been the 111 This harmonization process had not finished at the time of writing this chapter.

Case Studies 181 dominant party in the takeover of Philips Communications & Processing Services, and as such its culture was imposed upon that of the employees of the former Philips Communications & Processing Services. So cultural dominance plays an important role in this respect. It will be very interesting to see how company culture is reflected by the employee through preferences and attitudes. Chapter six will focus on characteristics and preferences of IT professionals, and can potentially shed some more light on the topic of company culture.

182 Developing Institutions

6 A Survey Amongst IT Professionals Almost every actor involved in industrial relations in the Dutch IT industry has been discussed in detail so far except for the IT professionals themselves. A closer look at the IT professionals is required for a number of reasons. Firstly, it very difficult to test some of the hypotheses without more information about which type of regulation IT professionals prefer. Secondly, there are many ideas about IT professionals which have so far not been verified in this study, for instance, that they are generally highly skilled, highly trained workers and are therefore less inclined to join a union than less skilled, less well trained workers. This assumption rests on the premise that higher skilled workers are more capable of negotiating their terms and conditions of employment than lower skilled workers. This chapter will answer a number of questions about IT professionals. Are IT professionals the young, well earning males we often think they are? Are they non-unionized, and what type of regulation do they prefer? Do they prefer a collective agreement, or do they prefer a different type of agreement? The answers to these questions are based on an Internet survey, carried out during 2003 by Berenschot, a management consulting firm, in cooperation with De Automatisering gids, a Dutch weekly periodical for IT professionals. 1 The survey is called Automatisering gids Salary Survey 2003 and is based on an Internet survey amongst subscribers of the Automatisering gids. Every year a special set of questions is added to the basic survey which are combined with two major variables: gross income per month and job title. 2 In the 2003 survey this special set of questions, specifically meant for this study, which focused at union membership and type of labour contract. This chapter will start by checking the representativeness of the Internet survey that was used. This will be done by comparing some of the data from the Internet survey to other statistical sources available. After discussing the general characteristics of IT professionals, possible differences between non-union and union 1 Berenschot employs over 400 people in the Benelux. One of their products is the Berenschot Salarisenquête, a survey amongst employers and personnel managers, which gives insight into well over one hundred occupations. De Automatisering gids has 74,000 subscribers, predominantly IT professionals. 2 Unfortunately, the list of questions used in the survey was not the final list of questions, but a conceptual list of questions. Some of the conceptual questions used in the survey were too ambiguous to be used in this study and have been omitted from the final results.

184 Developing Institutions members will be analyzed. This chapter concludes by examining the types of agreements IT professionals in the Internet survey actually have, and seeing whether these are in alignment with their preferences. 6.1 Representativeness of the 2003 Berenschot-Automatisering Gids Salary Survey Before trying to say something about IT professionals based on an Internet survey it is important to discuss the representativeness of the Internet survey that lies at the heart of this chapter. In order to do this, firstly the advantages and disadvantages of an Internet survey are discussed, and to what extent these advantages and disadvantages apply to this survey. In addition, the key characteristics of the survey are compared to more representative sources. This is to determine whether the survey actually represents the respondents we want to have represented in this study, namely the IT professional working in the Dutch IT industry. 6.1.1 The Problems With Internet Surveys According to Heerwegh (2001) Internet surveys have a number of advantages and disadvantages. The most important advantages are the speed of response, and the low costs involved compared to for instance traditional survey methods. Disadvantages of Internet surveys are coverage errors (Are those who respond to the survey on the Internet, representative for the population?), sampling errors (has a proper selection of the entire population been made?), and non-response (people not responding to the survey for whatever reason). With regard to potential coverage errors being made; all of the subscribers to the Automatisering gids was informed about the survey in both the magazine and the magazine s website on the Internet. This means that a potential number of respondents of roughly 74,000 people was informed about the survey. Since 50,000 of the magazine s members are paying subscribers, it is very likely that subscribers are directly or indirectly related to the IT industry. Of all respondents, fifty percent indicated that they are working in the IT services industry. The other half has an IT function outside the IT services industry. Another fact that should be taken into consideration is that the Internet survey was open to anyone visiting the Automatisering gids website. It is entirely possible that non-subscribers, but people with an interest in IT, answered the survey as well. This could affect the representativeness of the Internet survey. It is harder to determine whether a proper sample of personnel in the IT services industry has been drawn. It is, however, possible to compare a few key figures with different statistical sources. One of these sources is statistical data from the Dutch labour force survey of 2002. The Dutch Standard Occupation Classification system (SOC) of 1992 distinguishes five different occupational groups for the IT

A Survey Amongst IT Professionals 185 industry. For three of the five groups statistical data was available online in the CBS Statline database (classification codes between brackets): intermediate IT white-collar occupations (514), higher IT white-collar occupations (714), and scientific IT whitecollar occupations (914). Data for intermediary electro technical IT occupations (466) and higher electro technical IT occupations (666) were not available online. The reason for these groups not being present in the online database is the fact that there are too few respondents in the survey to be published. CBS (2003: 186-187) published two tables about active IT professionals, which were defined as SOC 92 codes 514, 666, 714, and 914. From the data in these tables we can estimate that about 14,000 people were active as higher electro technical IT professionals (SOC 92 666) in 2002. 3 A third source with which the Automatisering gids survey data can be compared is the official Berenschot salary survey. Berenschot claims this survey is a much more representative source of information because it is not answered by employees, but by personnel managers and is thus less likely to contain errors. The problem with that survey, however, is that some of the definitions differ greatly from that of the Internet survey, so its relevance for comparison with our Internet survey data is relative, and restricted to a few general indicators. Internet surveys are more prone to people entering false data than regular paper surveys, because it costs little to nothing to fill in a web based form. Internet surveys also often include the possibility of winning a prize, or offer a different kind of incentive to fill in a survey. These incentives, in combination with the ease and cost free entry of an Internet survey may bring people to fill in garbage rather than authentic information. In order to compensate for such errors, and thus increasing the representativeness of the entries, Berenschot closely examined all entries and compared these to more general data available to them from their official salary survey. If people entered too high or too low wages in comparison to their, age, position, function then these people were omitted from the final list. This brings us to the matter of non-response. The total number of respondents left over after the cleaning process by Berenschot was 1,290. This means that of a total number of possible respondents (74,000), 1.7 percent answered the Internet survey. A response rate of less than two percent is low for even the standards of the Automatisering gids survey. The previous two times this survey was taken more than 1,500 people responded and in the 2004 survey more than 1,700. The reason for this relatively low number of respondents in 2003 is difficult to explain, and may very well be related to the fact that it is an Internet survey, which have a more optional character. The 1,290 survey respondents work in IT occupations both within and outside the IT industry. The first questions that respondents were asked were related to the type of industry their company, organization, or institute operated. Four main 3 Data from the publication from1995-2001 was compared to data extracted from Statline (1995-2002) which had similar information except SOC 92 group 666. The least squares method (LSQ) was used to extrapolate the 1995-2001 data to 2002.

186 Developing Institutions categories were available. Six percent of the respondents worked in trade and transportation, 10 percent in trade and industry, 12 percent for the government and in the public sector, and finally, 71 percent in commercial services. All of these four main industrial categories had subcategories. Obviously this study is more interested in those people actually working in the IT industry as it has been defined in chapter three. In the survey, 635 respondents answered that they were employed in the subcategory IT services industry and less than 10 percent were working in the IT manufacturing industry. The rest (643 respondents) worked in other industries, outside the IT industry as defined in this study. 4 Because of the relatively low number of respondents working in the IT manufacturing industry (one percent) they will not be analyzed separately from those respondents working in the IT services industry. The reason for separating the IT services professionals from the rest is that they fit the unit of analysis in this chapter, the IT professional working in the Dutch IT services industry, more closely than does the total survey. 6.1.2 General Characteristics of IT Professionals According to the Dutch labour force statistics in 2002 the total number of IT professionals working in the Netherlands (excluding the electro technical IT professions) was 256,000. The total number of employees working in the Dutch IT services industry in 2002 was 111,000. From these two figures we can, to some extent, set the proportion of IT professionals working in the IT services industry (111,000) to 43 percent. In the survey 50 percent of the respondents worked in the IT services industry. Obviously a number of reservations needs to be made in order to compare these figures. Firstly, not all of the 111,000 people working in the IT industry can be considered IT professionals. This is a drawback of having to use company based data. Secondly, as was already stated, the 256,000 figure does not comprise all IT professionals in the Netherlands: the electro technical IT professionals are omitted. Finally, the CBS statistics are from 2002, while both the official Berenschot survey and the Internet survey date from 2003. How much of a drawback this really is can only be guessed at. Figure 3.2 in chapter three on the growth and decline in vacancies in the Dutch IT services industry showed signs of stabilization for 2003. The figures, however, give some indication that in the 2003 survey, IT professionals working inside the IT services industry are slightly overrepresented compared to the national statistics. The Dutch labour force statistics also contain data about sex, age, educational level and part time work, which can be compared to the Internet survey data. Table 6.1 lists data from three sources: the Dutch labour force statistics (2002), the official Berenschot survey (2003), and the Internet survey (2003). From this table we can see that people who responded to the Internet survey, compared to the national statistics for IT professionals, are more often male, are relatively younger, have received a higher level of education, and are more often working more than 35 4 See appendix C, table 1.

A Survey Amongst IT Professionals 187 hours per week. Both the official Berenschot survey and the Internet survey show that males are higher educated, and who work more than 35 hours per week. Table 6.1 General characteristics of Internet survey respondents compared to other statistical sources (figures are percentages of the total respondents) CBS active population (n=256,000)* Berenschot survey (n=1,040)** Internet survey (n=1,290) Sex Males 87-94 Females 13-6 Age 15-24 7 3 6 25-34 37 38 48 35-44 35 37 31 45-54 17 15 13 55-64 4 2 2 Highest attained level of education Prevocational education 8 2 1 Vocational education 39 18 21 Higher vocational education including university education 53 80 78 Number of working hours per week Less than 35 hours 12 7 6 35 hours and more 88 93 94 * SOC 92 classes 514, 714, and 914. Figures 2002. ** Only IT professionals were selected, which make up 18.8 percent of their total survey. The Berenschot data is based on employee files. Sometimes these files do not discriminate between male and female employees and in the survey they are all classified as males. Therefore this data is not available. Data for age was not available for three percent of the respondents. Figures 2003. What then can be concluded regarding the representativeness of the Internet survey? Though the following analysis will be a quantitative one, its representativeness is qualitative at best, and any conclusions should be regarded as such. Figures should be regarded as indications that help shape the image of the IT professional. On the one hand, the number of people responding is simply too low to make any general claims or statements. On the other hand, the group of IT professionals that is selected and analyzed, those IT professionals that are working in the IT services industry, resemble the group of higher educated employees that was assumed to be working in the IT industry. Therefore the results of the following sections can give an indication of what the opinions of IT professionals are who were working in the IT services industry during the middle of 2003. The following analysis will therefore describe

188 Developing Institutions rather than analyze and will only make use of simple frequency tables and matrices, and not of more elaborate statistical functions available. 6.2 IT Professionals in the Internet Survey This section will focus on the basic characteristics of IT professionals working in the IT services industry. Occasionally this group will be compared to IT professionals working outside the IT industry. 6.2.1 IT Professionals Inside and Outside the IT Services Industry Table 6.2 shows the same categories as table 6.1 with the difference that now IT professionals inside and outside the IT services industry are compared. The most notable differences here are that IT professionals working in the IT services industry are on average younger and have on average a higher level of education. Also when looking at these aggregate figures, there is almost no difference in income in IT professionals work inside or outside the IT services industry. There is, however, a difference between the two groups as the average age of respondents outside the IT services industry is slightly higher, and thus could positively influence the average income for this group. Working Hours and Part Time Work With regard to the number of working hours per week, both IT professionals in- and outside the IT services industry have high part-time ratios. Table 3.8 in chapter three showed the part-time ratios for people working in the IT services industry between 1996-1999. Part time ratios indicate the relative difference between the number of jobs and the total number of working hours. A ratio of 1 indicates that there is no part time work. The ratios in table 3.8 showed the ratio was becoming lower, from.95 in 1996 to.91 in 1999, which indicated that the number of part-time workers in the IT industry had been increasing during this period. The figures from the survey, however, indicate something quite contrary and may be explained by the fact that the relatively lower part-time ratios from table 3.8 merely state that these figures are influenced by non-it professionals working in the IT services industry. These probably work part time more often than the IT professionals themselves. With regard to part time work a special opinion question has been asked in the questionnaire, namely whether IT professionals thought that corporate culture somehow limited their freedom to work part time. 56 percent of the IT professionals in the IT services industry answered they did not agree with this, 23 percent were neutral, and 21 percent agreed that corporate culture did indeed restrict their possibilities to work part time. Women tend to agree more often (38 percent) than men (20 percent). The part time factor in both cases is very low. The part time

A Survey Amongst IT Professionals 189 factor is the total labour volume divided by the total number of people at work. 5 For the entire Netherlands this figure is much higher. In 2002 there were 7,058,500 people at work, and the total labour volume was 5,595,400. 6 The national part time factor thus was.79, which is significantly lower than that of the two groups in the Internet survey. The CBS also provides the same data for SIC 93 class 72, and IT professionals working in the IT software and services industry nationally have a.94 part time factor. Company Size IT professionals working in the IT services industry tend to work in either larger (43 percent work in a company employing 1,000 and more people) or smaller companies (33 percent work in a company employing less than 50 people). IT professionals working in other industries are more evenly distributed amongst size categories except at companies employing 1,000 and more people where 45 percent of the respondents resides. 5 Total labour volume is the total number of hours worked in term of full time jobs. 6 CBS, Statline, Labour market statistics 2002 temporary revision.

190 Table 6.2 Developing Institutions General characteristics of Internet survey respondents working inside and outside the IT services industry IT services industry (n=635) Other industries (n=655) Sex Male 95% 93% Female 5% 7% Age 15-24 6% 7% 25-34 53% 43% 35-44 30% 32% 45-54 10% 16% 55-64 2% 3% Average age 34.44 35.79 Highest attained level of education Prevocational education 1% 1% Vocational education 19% 24% Higher vocational education including university education 80% 75% Number of working hours per week Fewer than 35 hours 6% 7% 35 hours and more 94% 93% Part time ratio*.98.95 Average gross annual income including bonuses 48,342 48,752 Company size Fewer than 50 employees 33% 13% 50 and more 67% 87% *Number of worked hours divided by total number of employees working full time (40 hours). 6.2.2 Job Level and Highest Attained Level of Education Table 6.3 shows how the variables highest attained level of education and job level relate to one another. 56 percent of the respondents within the IT services industry are working in a job level higher than staff member. Outside the IT services industry this is 58 percent. One would expect that a higher attained level of education leads to a higher function. Although a far from perfect fit, table 6.1 shows that this expected relation is stronger in the respondents outside the IT services industry and a higher attained level of education more often leads to a higher function. Relatively more IT professionals working in the IT services industry have a university degree. This can also be seen in table 6.2.

A Survey Amongst IT Professionals 191 When looking at the average ages for both job levels and level of education, two distinct features arise. The higher the level of education, the older respondents are on average. Also job level is somewhat related to age in the sense that staff on average are younger than those with higher job levels. Within the IT services industry though, this is not entirely the case. This may be so because people in the IT services industry are younger on average than in other industries. Table 6.3 Job level versus highest attained level of education and average age Directorate Management Coordinator, advisor, or chief Staff Column total Average age IT services industry (n=635) University education 3% 25% 41% 32% 28% 35.75 Higher vocational education 6% 19% 32% 44% 52% 34.57 Vocational education 3% 11% 23% 63% 19% 32.28 Prevocational education 0% 0% 33% 67% 0% 31.00 Total 4% 19% 32% 44% 100% Average age 40.43 41.13 35.38 30.34 34.44 Other industries (n=655) University education 10% 29% 38% 23% 23% 37.21 Higher vocational education 4% 19% 37% 41% 52% 36.16 Vocational education 1% 10% 23% 66% 23% 33.31 Prevocational education 13% 25% 25% 38% 1% 40.63 Total 5% 19% 34% 42% 100% Average age 45.60 41.70 37.35 30.82 35.79 Type of Function Respondents to the questionnaire were also asked what their main function was. Table 6.4 shows the results. The differences between the two groups are not very big except for people who work in IT projects or system/network management. The function IT general seems to have been used by those people whose functions are less well defined.

192 Developing Institutions Table 6.4 Distribution of main functions IT services industry (n=635) Other industries (n=655) Application management 1% 6% Application development 10% 7% Database management 2% 1% Database development 1% 1% Helpdesk 2% 3% Information analysis 6% 5% Internet application management 0% 0% Internet development 4% 2% IT general 30% 31% IT operational 6% 5% IT projects 25% 16% IT training 1% 1% System / network management 10% 18% System / network development 3% 4% 6.3 Union Membership Amongst IT Professionals Recalling the assumption of higher skilled employees and their propensity for individual negotiation, it is interesting to see the union membership figures for the IT industry. Union membership in the Dutch IT industry has long been estimated in the region of four to seven percent. 7 Employers and unions have differed greatly in their estimates concerning actual union membership figures for IT professionals. Where employers have been estimating these figures within the pre-mentioned range, unions have been claiming union membership figures of at least ten percent since 1991. 8 6.3.1 Union Membership in the IT Industry Compared to National Figures Table 6.5 shows the national union membership figures and compares these to the union membership figures of IT professionals in- and outside the IT industry in the Internet survey. From these figures it seems that unions have not been too optimistic 7 Cf. Smit et al., 1995 and Van Liempt & Van Uffelen, 2000. 8 A former union representative of the CNV (John Koudijs) is quoted in a 1998 edition of the Automatisering gids (Number 22, May 29, 1998) where he claims union membership figures have doubled to ten percent during 1991-1995, because of the crisis of the IT industry during the first half of the 1990s in the Netherlands. Also in recent interviews in light of this study, unions have claimed figures of ten percent.

A Survey Amongst IT Professionals 193 with regard to their IT union membership estimates in recent years. Table 6.5 shows that the union membership figures (12 and 16 percent) amongst IT professionals are still relatively low compared to national figures, but higher than one would estimate based on popular images about IT professionals. Perhaps the timing of the survey is relevant. Chapter three discussed the crisis the Dutch IT industry is in since 2001 and it is likely that sour times increase the inclination of IT professionals to seek job security through unions. If this is true, one can wonder who these people are that seek increased security through union membership. These people may be on average older, may be highly specialized and expensive to re-educate, or they may stem from the batch of people that have only been hired because of the extreme need for people during the software and services boom during 1998-1999. More on this when table 6.6 is discussed below. With regard to the gender of union members nothing very interesting can be said because so few women were present in the Internet survey that these figures may easily change with a different set of respondents. The largest Dutch union, the FNV is also the largest union within the Internet survey respondents. Table 6.5 Unions and union membership of active respondents aged 15-64 Nationally* IT services industry Other industries (n=645) (n=621) Total 23% 12% 16% Sex Males 27% 11% 17% Females 17% 18% 12% Union FNV 63% 67% 49% CNV 18% 17% 17% MHP 9% 15% 19% Others 10% 2% 16% Sources: CBS active population survey (2002 estimates) and CBS union members survey 2003 * Figures from the Union members survey (2003) were divided by figures from the Active population survey (2002 estimates). From the union members survey, only those were selected that were part of the active working population (age 15-64). 6.3.2 Union Membership, Age, Income, and Job Level When taking a closer look at the union members in the Dutch IT industry using table 6.6, one can see that there is quite a difference between those working in the IT services industry and those working outside it. Consider age for instance. In the IT services industry 40 percent of the union members are older than 55. Only 11 percent are younger than 35. It is the same for unions members outside the IT services industry where only 9 percent of the union members are younger than 35, but only

194 Developing Institutions 22 percent are older than 55. The difference in age between union members and nonunion members becomes even more apparent when looking at the average ages of both groups. Union members are on average four to six years older than non-union members. Nationally the average age of union members is 46, so union members within the IT industry, be it inside or outside the services branch, are relatively young in comparison: 38 and 41 respectively. It would make no sense to compare the differences in income without relating these to age. Table 6.6 shows that union members on average earn less money (up to 4,000 euros annually), than non-union members, if one takes age into consideration. One can wonder why this is the case, because if these figures are compared to job level and highest attained level of education, there is no strong evidence that union members generally have lower positions. Their function levels look fairly similar to those of non-union members both inside and outside the IT services industry. With regard to the highest attained level of education there are indications that, at least inside the IT services industry, union members have a general lower level of education. It is questionable whether it is possible to conclude from these characteristics that union members inside the IT services industry, for a large part, have become union members because they wish to increase their job security. The fact that union members on average are older than non-union members, however, may be an indication that unions have thus far not succeeded in appealing to younger age groups.

A Survey Amongst IT Professionals 195 Table 6.6 Union membership amongst IT professionals inside and outside the IT services industry in percentages of the total respondents IT services industry (n=621) Other industries (n=645) Union membership and age 15-24 3% 0% 25-34 8% 9% 35-44 17% 22% 45-54 18% 33% 55-64 40% 22% Average gross annual income including bonuses per age group Union members Non union Union members Non union 15-24 21,581 27,364-27,584 25-34 39,492 40,795 39,616 41,029 35-44 51,682 59,187 50,192 54,922 45-54 61,854 69,393 65,177 65,248 55-64 63,479 70,488 62,766 76,387 Average income 49,241 48,076 52,937 47,928 Average age 38.0 34.0 40.9 34.8 Job level Directorate 4% 5% 6% 4% Management 25% 17% 25% 18% Coordinator, advisor, or chief 29% 33% 37% 33% Staff 42% 45% 33% 45% Highest attained level of education University education 24% 29% 23% 23% Higher vocational education 44% 53% 58% 51% Vocational education 32% 17% 19% 24% Prevocational education 0% 1% 1% 1% 6.4 The Terms and Conditions of Employment of IT Professionals Approximately 25,400 people are covered by a collective agreement in the Dutch IT services industry. This figure is based on the sum of employees working in the six IT services companies in the Netherlands that are using a collective agreement. 9 9 Employment figures at the end of 2003: EDS 1,400, T-systems 280, PinkRoccade 8,800, Torex- Hiscom 330, Getronics 5,600, and Atos-Origin 9,000. Figures stem from the companies respective websites and annual reports.

196 Developing Institutions Compared to the CBS figures for people working in the IT services industry it is then possible to estimate the relative number of people falling under a company level collective agreement in the IT services industry, which is roughly 23 percent. 10 In table 6.7 the figures on the usage of the four available types of terms and conditions of employment are mentioned for the two groups of the Internet survey and the Netherlands as a whole. The national figures speak for themselves: 85 percent of the Dutch active population is covered by a collective agreement. Moreover, 73 percent (87 percent of the total population covered by a form of collective agreement) do so with an industry wide collective agreement. This implies that the situation in the Dutch IT services industry varies significantly from that of the rest of the Netherlands. It is difficult to estimate the total number people working with an industry wide collective agreement in the Dutch IT industry because it is difficult to ascertain whether all the companies that are covered by the ICK collective agreement can be categorized as IT companies in the sense of the definition that is used in this study. 11 It makes sense that IT professionals working outside the IT services industry are covered more frequently by a collective agreement as the other industries group of respondents is more inclined to reflect national characteristics. In this light the 60 percent may even be perceived as relatively low. Amongst the respondents, the relative number of IT professionals working in the services industry, 16 percent, does not come close to figure based on the estimate of 23 percent based on CBS figures and employment figures of companies using a company level collective agreement. This group is thus underrepresented in the Internet survey. 10 The number of people working in SIC 93 code 72 for 2002 was 111,000. (25,400 / 111,000) * 100 = 22.88 percent. 11 See chapter four for more information on the ICK collective agreement.

A Survey Amongst IT Professionals 197 Table 6.7 Distribution of types of agreement amongst employees in the Netherlands* Nationally** IT services industry (n=622) Other industries (n=647) Collective agreement 85%*** 24% 60% - company level collective agreement 10% 16% 32% - industry wide collective agreement 75% 8% 28% Works council agreement 30% 19% Individual contract (no other 15% agreement) 51% 27% * People can have both a works council agreement and a collective agreement. So the sum of the columns can be higher than 100 percent. ** Author s calculations based on SZW 2004 (5,321,500 people fell under an industry wide collective agreement and 716,000 under a company level agreement) and CBS data for the active labour force between April 2004 and June 2004 (7,077,000 people). *** 10 percent of the 85 percent of the employees that are covered by a collective agreement, are covered by a collective agreement for the public services. 6.4.1 Company Size and the Type of Agreement Table 6.8 tries to show whether some of the results of table 6.7 can be explained by company size. Company Level Collective Agreement Looking at company level collective agreements, three percent of the respondents indicate they are work in a company employing between 50 and 100 people. This seems highly unlikely as there are no IT companies of this size in this segment of the Dutch IT industry that use a company collective agreement. These respondents must have made a mistake filling out the form. They may have selected the wrong type of employment arrangement or selected the wrong company size. Industry Wide Collective Agreement People covered by an industry wide collective agreement seem to fall predominantly into two categories of companies, those employing less than 50 people and those employing 1,000 and more people. This latter category may seem unlikely, but large companies such as IBM and Apple are members of the Werkgeversvereniging ICT and therefore automatically use the ICK collective agreement. 12 Works Council Agreement Works council agreements are quite common in the IT services industry. Works council agreements can usually only be found though in companies employing more than 50 people, because those companies are, by law (see chapter two), compelled to have a works council. Nine percent of the IT professionals indicate they are working 12 See chapter four.

198 Developing Institutions in a company employing fewer than 50 people. This could be possible if there was one works council for multiple companies. It could, however, also mean that the respondents have made an error in their estimate regarding the size of their company, or that they have misinterpreted the status of their contract, or that these companies may simply be using a type of contract that is drawn up for all employees, but is not backed up by a works council. Individual Agreement With regard to individual agreements there is a strong relation with company size. The smaller the company is in terms of numbers of employees, the more likely it is they are covered solely by an individual agreement. Table 6.8 Company size and type of agreement in the IT services industry Industry wide CA (n=47) Company level CA (n=103) Works council agreement (n=185) Individual agreement (n=318) Fewer than 10 17% 0% 1% 19% 10 50 21% 0% 8% 19% 50 100 4% 3% 9% 34% 100 250 6% 4% 6% 8% 250 500 6% 5% 4% 5% 500 1,000 9% 7% 5% 11% 1,000 and more 36% 82% 66% 4% 6.4.2 Type of Agreement Versus Terms and Conditions of Employment An important question now is whether the different types of agreements differ greatly in terms of wage, fringe benefits, and perks. Wages and Bonuses When checking the average annual wages including bonuses of the four types of agreements in table 6.9, one can see that within the IT services industry the differences are small. Only those IT professionals covered by an industry wide collective agreement seem to earn less. The similarities between the wages may indicate that employers (and employees) in the IT services industry tend to watch each other carefully in terms of what the average wages are. This is quite possible because already in chapter three it was indicated that IT companies keep a close watch to each other in terms of what they are paying their employees. Employees themselves do the same and are usually well informed of what their colleagues make. The wage similarities may also indicate that differences between having unions, works councils, or individuals negotiating over the terms and conditions of employment, seems to be less significant in terms of wages and bonuses (at least for the IT services

A Survey Amongst IT Professionals 199 industry where one may assume companies to be direct competitors 13 ). It may be that company size is responsible for this discrepancy. Table 6.8 showed that 38 percent of those working with an industry wide collective agreement did so in a company employing less than 50 people. 38 percent of those working with an individual agreement, however, also work in similar sized companies and they are not negatively affected in this respect with regard to company size. It is interesting to note that individual bonuses are used more commonly in company level collective agreements than in individual agreements where one would expect a relatively higher individual component. Then again this may already have been a component of the regular wage, which is a reflection of individual performances. Payment of a 13 th month is found most frequently within companies using a company level collective agreement. Fringe Benefits and Perks With regard to fringe benefits and perks, one can notice that companies using a company level collective agreements and works council agreements are particularly inclined to offer child care, funds for training and education, and a save-as-you-earn plan. They also tend to offer a lease car and laptop more frequently. When looking at table 6.9 one can see that a large share of those covered by an individual agreement or industry wide collective agreement work in companies employing fewer than 50 people. The most likely explanation for this is that smaller companies simply lack the required funds to be able to provide some of these benefits and perks. The second part of table 6.9 only shows IT professionals that are working in companies employing 50 and more people. When correcting for company size, the average gross annual wage and bonuses immediately rise for those covered by an individual agreement. 13 The differences between companies in different branches may be more significant than the similarities in type of work or employee outside the IT services industry.

200 Table 6.9 Developing Institutions Type of agreement and company size of IT professionals in the IT services industry Industry wide CA Company level CA Works council agreement Individual agreement Total Less than 50 employees (n=18) (n=0) (n=12) (n=169) (n=199) Gross annual wage including extras 40,751-45,898 44,541 44,280 - Bonus 9% 0% 5% 14% 9% Average bonus in Euros 1,925-3,964 6,365 5,736 - Share in profit 0% 0% 2% 5% 3% Average share in profit in Euros - - 2,167 2,038 2,058-13 th month 7% 0% 1% 5% 3% Average 13 th month in Euros 1,983-1,850 2,543 2,390 - Lease car 9% 0% 6% 33% 19% - Laptop 17% 0% 5% 33% 19% - Mobile phone 19% 0% 6% 35% 21% Savings wage 21% 0% 6% 34% 21% Day care 0% 0% 3% 5% 3% Training and education 13% 0% 4% 19% 12% 50 and more employees (n=29) (n=103) (n=142) (n=149) (n=423) Gross annual wage including extras 48,082 48,634 49,198 52,066 49,994 - Bonus 24% 41% 38% 21% 29% Average bonus in Euros 8,856 6,001 6,374 9,733 7,706 - Share in profit 9% 13% 9% 3% 7% Average share in profit in Euros 1,788 1,896 1,984 3,502 2,336-13 th month 13% 50% 13% 8% 16% Average 13 th month in Euros 3,379 3,061 2,737 3,160 3,039 - Lease car 53% 77% 85% 41% 59% - Laptop 45% 78% 76% 36% 54% - Mobile phone 43% 77% 58% 30% 45% Savings wage 55% 95% 88% 42% 63% Day care 32% 78% 46% 18% 36% Training and education 40% 76% 60% 29% 45% Looking at table 6.10, which shows the job level, highest attained level of education, and annual turnover of the company the respondent is working for of IT professionals working with an individual agreement in the Dutch IT services industry, one can see that companies employing more than 50 people employ relatively more managers and fewer staff members than companies employing fewer than 50 people. This may be a logical effect of smaller companies needing fewer managers. Another bias is that of the highest attained level of education where IT professionals in larger companies generally are higher educated than in smaller

A Survey Amongst IT Professionals 201 companies, at least this holds true for those working with an individual agreement. Finally, table 6.10 demonstrates that companies employing fewer than 50 people indeed generally have a lower turnover per employee. Table 6.10 Characteristics of IT professionals working with individual agreements in the IT services industry by company size Less than 50 employees (n=169) 50 employees and more (n=149) Function level Direction 5% 7% Management 12% 23% Coordinator, advisor, or chief 27% 27% Staff 55% 43% Highest attained level of education University education 24% 30% Higher vocational education 47% 54% Vocational education 28% 16% Prevocational education 1% 0% Annual volume of trade of the company Less than 5 million Euros 82% 7% 5 10 million Euros 14% 15% 10 25 million Euros 3% 15% 25 50 million Euros 0% 9% 50 100 million Euros 0% 11% 100 250 million Euros 0% 13% More than 250 million Euros 0% 29% Terms and Conditions of Employment of IT Professionals Outside the IT Services Industry The differences between wages and bonuses amongst IT professionals working outside the IT industry seem bigger. There those covered by either a company level collective agreement or an individual agreement seem to earn significantly more than those covered by either an industry wide agreement or a works council agreement. With regard to fringe benefits and extra perks, the differences are much smaller compared to IT professionals working in the IT services industry where also predominantly larger companies were able to offer these perks and benefits.

202 Developing Institutions Table 6.11 Type of agreement of IT professionals versus primary and secondary terms and conditions of employment Industry wide CA Company level CA Works council agreement Individual agreement Total IT services industry (n=47) (n=103) (n=154) (n=318) (n=622) Gross annual wage including extras 45,358 48,634 48,941 48,067 48,166 - Bonus 33% 41% 42% 35% 38% Average bonus in Euros 7,008 6,001 6,115 8,410 7,245 - Share in profit 9% 13% 11% 9% 10% Average share in profit in Euros 1,788 1,896 2,016 2,614 2,246-13 th month 20% 50% 14% 12% 20% Average 13 th month in Euros 2,914 3,061 2,656 2,923 2,933 - Lease car 63% 77% 92% 74% 78% - Laptop 61% 78% 81% 69% 73% - Mobile phone 61% 77% 64% 65% 66% Saving-as-you-earn plan 76% 95% 94% 76% 84% Day care 33% 78% 49% 23% 39% Training and education 52% 76% 64% 48% 57% Other industries (n=182) (n=204) (n=85) (n=176) (n=647) Gross annual wage including extras 45,722 51,099 45,068 51,029 48,775 - Bonus 20% 25% 41% 32% 28% Average bonus in Euros 4,075 3,479 3,955 8,896 5,399 - Share in profit 18% 17% 16% 14% 16% Average share in profit in Euros 2,124 2,659 2,093 2,155 2,300-13 th month 32% 44% 21% 18% 30% Average 13 th month in Euros 2,445 2,936 3,152 3,598 2,915 - Lease car 24% 18% 53% 55% 34% - Laptop 51% 52% 64% 57% 55% - Mobile phone 47% 51% 58% 56% 52% Savings wage 87% 93% 99% 77% 88% Day care 45% 65% 42% 30% 47% Training and education 69% 80% 75% 50% 68% 6.4.3 Preferences of IT Professionals in the IT Services Industry What then are the preferences of IT professionals in the IT services industry regarding the type of terms and conditions of employment? Table 6.12 shows that individual agreements, together with industry wide collective agreements, are the most preferred types of agreements. This is a pretty striking conclusion as table 6.11

A Survey Amongst IT Professionals 203 demonstrated that these can be regarded as the least favourable types of agreements when taking into account wage, bonuses, fringe benefits, and perks. Generally speaking there is somewhat of a paradox here. One would expect people to try to maximize their income. Table 6.12, however, shows that the type of agreement people prefer, is not always the most rational to pursue in this sense. Table 6.12 shows a few of the characteristics of IT professionals and their terms and conditions of employment of preference. First of all, the type of agreement people are familiar with is generally also the one most often preferred. The works council agreement is the only type for which this does not hold true. People working with a works council agreement seem to be more interested in an industry wide agreement. This could indicate that these employees have little faith in the bargaining powers of the works council (see also hereafter). As people get older they seem to prefer individual agreements over industry wide collective agreements. Also the preference for company level agreements seems to grow with age. There is a relative general dislike of works council agreements in all age categories. The highest attained level of education also seems to affect preference. Those who have received a higher vocational education and university education seem to favour individual agreements, while those who have received vocational education prefer an industry wide agreement. Company size again plays an important role as relatively many people who prefer individual agreements and industry wide collective agreements work in smaller companies. Job level seems to be one of the most important factors for preference. Sixtyfour percent of the directorate prefer individual negotiation, compared to 49 percent of managers and 40 percent of lower management. The lower the job level the higher the preference for an industry-wide and company-level collective agreement.

204 Developing Institutions Table 6.12 Preferences of IT professionals regarding the terms and conditions of employment in the Dutch IT services industry Industry wide CA Company level CA Works council agreement Individual agreement Total IT services industry 29% 20% 13% 38% Type of terms and conditions of employment Industry wide collective 61% 15% 11% 13% agreement Company level collective 22% 54% 9% 15% agreement Works council agreement 30% 19% 26% 25% Individual agreement 27% 9% 8% 56% Age 15-24 54% 14% 0% 31% 25-34 31% 18% 14% 38% 35-44 25% 23% 12% 39% 45-54 21% 20% 18% 41% 55-64 20% 30% 10% 40% Highest attained level of education University education 22% 19% 14% 45% Higher vocational education 29% 19% 15% 38% Vocational education 42% 23% 6% 29% Company size Less than 50 employees 35% 9% 7% 50% 50 or more employees 27% 25% 16% 33% Job level Directorate 14% 4% 18% 64% Management 20% 18% 13% 49% Coordinator, advisor, or chief 28% 20% 12% 40% Staff 35% 22% 12% 30% From table 6.12 two general conclusions can be drawn. Firstly, those preferring individual agreements generally seem to do so because they already have an individual agreement, but more importantly, have a higher position in the firm, which potentially increases their bargaining power. The opposite holds true for those preferring an industry wide collective agreement. With this information in mind the paradox that was spoken of earlier is solved as people still behave rationally and know that their chances to do well in individual negotiations is dependent on their relative position and bargaining power. Another conclusion is that IT professionals in the IT services industry overall seem to have little faith in the bargaining powers of the works council as

A Survey Amongst IT Professionals 205 works council agreements are least preferred of all types of agreement. Obviously this is an indirect conclusion. Table 6.13 shows a number of opinions of IT professional regarding contingent effects of some types of agreements and bargaining power of the works council, unions, and themselves. In table 6.13 those indicating they prefer works council agreements or individual agreements, seem to think collective agreements will only level their wages. These two groups also agree collective agreements are too rigid and will not meet their personal requirements. They also do not agree that unions have more expertise in matters of bargaining and think works councils are more adept at representing employees because of their direct relation to the firm within which they are operating. It is also interesting to note that those employees working in companies that do not us a collective agreement take the works councils more seriously than do employees in companies which do use a collective agreement. This may indicate that the strength and position of the works council within these latter companies is not so strong because of the position of the unions in these companies. An example of this is the former Philips works council described in the Atos Origin case in chapter five. A works council representative of Atos Origin made a remark on the difference in the level of autonomy between the works council of the former Philips, which was used to having the unions negotiate for them, and the works council of the former BSO, which had to negotiate with management without the assistance of the unions. The latter s autonomy and expertise was greater than that of the former Philips works council.

206 Developing Institutions Table 6.13 Preferences and opinions of IT professionals in the IT services industry regarding the type of agreement Industry wide CA Company level CA Works council agreement Individual agreement 1. Collective agreements are responsible for levelling of wages and do not account well enough for individual performances. (slightly) agrees 30% 34% 68% 67% Neutral 31% 29% 21% 19% (slightly) disagrees 39% 37% 12% 13% 2. I perform better than my colleagues and therefore I want to be able to individually negotiate with my employer. (slightly) agrees 65% 65% 75% 74% Neutral 23% 22% 19% 19% (slightly) disagrees 13% 13% 5% 6% 3. Collective agreements are too rigid and cannot cater individual terms and conditions of employment well enough. (slightly) agrees 27% 24% 69% 68% Neutral 33% 36% 19% 24% (slightly) disagrees 39% 40% 12% 8% 4. I have more faith in the expertise of unions than in that of the works council with regard to terms and conditions of employment. (slightly) agrees 44% 46% 18% 21% Neutral 35% 26% 25% 40% (slightly) disagrees 21% 27% 57% 39% 5. Works councils can better represent the employee than unions can because they are better informed about the wishes of employees within its firm. (slightly) agrees 35% 39% 66% 37% Neutral 36% 29% 27% 40% (slightly) disagrees 29% 32% 6% 23% 6. The works council is a management s tool. (slightly) agrees 47% 50% 18% 27% Neutral 31% 25% 34% 38% (slightly) disagrees 21% 26% 48% 35% Opinions four to six directly question the faith IT professionals have in the expertise of the works council with regard to the terms and conditions of employment in relation to the unions and management. Their preferences show that those working with a collective agreement, be it an industry wide or a company level collective agreement, think that the unions have more expertise in these matters than the works council. The opposite holds true for those falling under a works council agreement or individual agreement. On the whole IT professionals seem to favour the expertise of

A Survey Amongst IT Professionals 207 their principal representing party. Disregarding expertise and focusing at the independence of the works council in relation to the management, almost 50 percent of those falling under a collective agreement think the works council is a tool of the management. Those working with a works council agreement or individual agreement seem to have much more faith in their works council than those working with a collective agreement. The dichotomy in preferences regarding representation through the unions or the works council, between those who fall and do not fall under a collective agreement, is less visible when it comes to estimating how well a party is informed about the wishes of IT professionals. The general preference for the works council in this respect within all four types of agreements is interesting. Though those working with a collective agreement generally favour the expertise of the unions more than that of the works council, they generally think the works council represents them better than the union can. 6.5 Conclusion The Internet survey conducted within this study has corroborated many of the stereotypical assumptions about employees in the (Dutch) IT industry. IT professionals in the IT software and services industry are typically male (95 percent), higher educated (80 percent), and work more than 35 hours per week (94 percent). The survey also showed that not all of these assumptions are still the matter-ofcourse they used to be. Though an average age of 34.4 years still is much younger than in other industries where the average age is close to 43 years, the figure for IT professionals indicates they are not as young on average as they used to be. Twelve percent of them today are members of a union and 24 percent fall under the scope of a collective agreement. The relation between age and union membership should not be underestimated. The sector itself is almost half a century old. As such it is not surprising to see the overall age of its workers rise as well. It will be interesting to see what effect this will have on union membership in the future. Important information derived from the Internet survey for this study, besides union membership figures, were the preferences of IT professionals regarding the type of employment agreement they have to work with. IT professionals inclination towards the type of employment agreement with which they were already working was very interesting, as was an increased preference for individual agreements when their relative bargaining powers increased through their level of education and their position within the firm. The latter preference seems reasonable when you assume people seek to maximize their bargaining power. Having a strong individual negotiation position does not do a person a whole lot of good when he cannot negotiate himself. Similarly people without a strong negotiation position act rationally if they seek to increase their bargaining power through cooperation. In practice though not all IT professionals have a choice within their companies. No questions were asked in the survey with regard to the willingness to go to a different

208 Developing Institutions company in order to seek the bargaining power that would suit them best. The only information available is that in almost all cases, except in the case of a works council agreement, more than half of the IT professionals preferred the type of agreement present in the company they were working at. It is also difficult to ascertain whether respondents preferences have anything to do with corporate culture. IT companies often say they have a unique culture. The Internet survey, however, has asked no questions in relation to the company culture of respondents. It is very probable that the development of a corporate culture is a self re-enforcing process. Employees seek the type of company that best meets their preferences. So those that feel best with a company employing a collective agreement will stay and those that do not, will leave. A good example of this selection effect was the merger of Pink Elephant and Roccade described in chapter five. Though many of the former Pink Elephant employees adapted themselves to the new company, a number of them left as well. They simply did not feel comfortable with the new corporate culture, despite the efforts of PinkRoccade to offer both a works council agreement and a company level collective agreement. Corporate culture is something that needs to be taken into consideration as the preferences of IT professionals seem, to some degree, directly related to the type of collective agreement they are (used to) working with. Take for example Raet, an IT services company that had been employing a collective agreement throughout almost its entire existence. Since it became independent from Getronics in 2004, it stopped using a collective agreement because the new management felt it could do without. Employees, however, were not happy with the new situation at Raet and during 2004 a poll was conducted through which employees could voice their preferences. The outcome of this poll was in favour of a collective agreement. Summarizing the above, clear indications exist that professionals working in the IT industry are not as hostile towards unions and collective agreements as is often thought.

7 The Presence and Future of Collective Agreements This study started off with a fairly simple question: Why are there collective agreements in the Dutch IT software and services industry? The general answer that this study has provided is the following: Companies that are using a collective agreement in the Dutch IT industry, do so because of a direct and indirect hereditary effect of the dominant model of industrial relations in the Netherlands. The following sections of this chapter are geared towards specifying and expanding the previous answer, and will answer the rest of the original research questions. Table 7.1 repeats these original research questions. The first three questions are of a descriptive nature, question four then has a distinct explanatory nature, and question five has a prospective nature. Table 7.1 Original research questions 1 To what extent and in which particular ways do collective agreements emerge in the Dutch IT industry? 2 Which actors (firms, unions, employers associations and others) are involved in this process? What are their opinions on (various types of) collective agreements? What role have they played in enabling (or hampering) the emergence of collective agreements in this industry? What are the reasons for their preferences and actions? 3 What are the contents of these collective agreements in terms of flexibility and individuality? How do the stipulations in these agreements compare to prevailing arrangements on similar topics in similar IT firms that lack a collective agreement? 4 How can the extent and nature of these collective agreements in the Dutch IT industry be accounted for? 5 What are the prospects for collective agreements in the Dutch IT industry? Section 7.1 will start by answering question four, and explains the extent and nature of the collective agreements in the Dutch IT industry. Question one: To what extent and in which particular ways do collective agreements emerge in the Dutch IT industry?, is of a descriptive nature and has already

210 Developing Institutions been answered in the various chapters. In 2004, seven collective agreements existed in the Dutch IT industry: One industry wide collective agreement for the IT manufacturing industry and six company level collective agreements for the IT software and services industry. Because of a merger between two of these companies in 2005, 1 five of these company level collective agreements were left. Despite this merger, the number of people falling under the scope of a collective agreement in the Dutch IT industry remained the same. Section 7.2 will focus on question two regarding the actors (firms, unions, employers associations and others) that are involved in industrial relations in the Dutch IT industry and explain what their opinions are on employment agreements; what role have they played in enabling (or hampering) the emergence of collective agreements in this industry; and what the reasons are for their preferences and actions. Section 7.3 will then answer question three, centring on what the contents are of these collective agreements in terms of flexibility and individuality, and how the stipulations in these agreements compare to prevailing arrangements on similar topics in similar IT firms that lack a collective agreement. Finally, section 7.4 will answer question five and discuss the prospects for collective agreements in the Dutch IT industry. 7.1 Explaining the Presence of Collective Agreements in the Dutch IT Industry Can the presence of collective agreements in the Dutch software and IT services be explained by one or more hypotheses formulated in chapter one? The three (competing) hypotheses are actor centred, i.e. they assume that employers (and indirectly employees) have the freedom to select the type of regulation they want without these regulations being imposed on them by higher levels of decision making. The first hypothesis, the limited choice set hypothesis, takes into consideration the background of a company, its corporate culture and history, and exogenous factors, which can limit a company s choice set. Whatever the choice actors at one point make, it is a long-term choice and it will only be re-evaluated if one of the previously mentioned factors changes considerably. The second hypothesis, the anachronism hypothesis, suggests that the collective agreements present in the Dutch IT industry are remnants of an outdated bargaining structure and as such are anachronisms. In this view, collective agreements will disappear and actors will opt for a type of agreement that corresponds to the dominant type of bargaining power. The third hypothesis, the efficiency thesis, states that companies seek to reduce transaction costs involved in processes of collective and individual bargaining. Depending on their own estimations of the costs involved, companies will choose to negotiate over the terms and conditions of employment with either individuals, the 1 The takeover of PinkRoccade by Getronics.

The Presence and Future of Collective Agreements 211 works council, or the unions. If such alternatives do not exist, existing institutions will be changed into more efficient institutions, given that the group with the greatest bargaining power seeks to maximize efficiency, is strong enough to bring about these changes, and is able to come up with a more efficient alternative. The above three hypotheses will be called theoretical explanations from now on. This is to differentiate them from the empirical explanations that were mentioned in the conclusion of chapter five. Chapter five concluded with two basic empirical explanations and one further explanation that could link the two explanations. The first basic explanation was the lineage of a company. These lineages are an important feature of the Dutch IT industry and were already identified in chapter three. The first lineage is associated with the (public) mathematical and administrative centres that first started using computers in the Netherlands, and which later on were privatised. These privatised companies all started out using either a legacy collective agreement from the company they belonged to previously, or started using a collective agreement after some time. The point is that these companies have been using a collective agreement from the start and have been doing ever since. Companies belonging to the other two lineages, foreign companies and entrepreneurial firms, started out very differently. Foreign companies brought with them the industrial relations culture of their home countries and made adjustments where the Dutch legal system required them to do so. These companies refused to deal with unions and contributed to the non-union sentiments in the Dutch IT industry. Entrepreneurial firms had sentiments that were closely related to the foreign companies and preferred to regulate the terms and conditions of employment on their own, rather than together with the unions. Foreign firms and entrepreneurs have shaped industrial relations the Dutch IT industry and have therefore been responsible for the relative absence of collective agreements in this industry. Based on a comparative study of organizational culture differences, conducted by IRIC across 20 organizational units in Denmark and the Netherlands in the 1980s, Hofstede (2004) concludes that industries (and businesses) can indeed influence corporate cultures to some extent. Lineage, however, cannot explain all the cases where a collective agreement is present as there are two companies that have a collective agreement but stem from an entrepreneurial firm (Getronics) and a foreign firm (EDS). These companies made strategic choices based on their respective growth strategies, which was the second empirical explanation. 2 IT companies in the Netherlands can adapt three types of growth strategies: (1) autonomous growth, (2) growth through takeovers, or (3) growth through insourcing projects. Autonomous growth means companies grow predominantly through gradually attracting new people to their firm. This does not hold true for companies that grow through takeovers or insourcing projects. Both these latter growth strategies mean that companies grow through incorporating a large group of people in a short time. For these types of companies the decision to 2 Growth in this context refers to an increase the number of people employed.

212 Developing Institutions start using a collective agreement was a strategic choice that enabled them to reduce the resistance amongst the employees of these alien companies or departments. At the end of the conclusion of chapter five the concept of corporate culture was introduced as an overlapping empirical explanation that could cover both primary empirical explanations. Corporate culture, as a concept, showed that the companies who started using a collective agreement had been doing so because of a direct and indirect hereditary effect. Corporate culture can explain why companies stemming from a different background, but operating in the same environment and doing roughly the same things, are different. First lineage companies cultures have developed in the presence of a collective agreement and thus were directly affected by the corporate culture of the former company they belonged to. Second and third lineage companies have developed without the presence of a collective agreement. This does not imply that the corporate cultures of companies sharing the same origin are all the same. They do, however, share a common factor in their respective histories: the presence or absence of a collective agreement. The concept of corporate culture can also be applied to the second basic empirical explanation, that of the primary growth strategy. Autonomous growth allows a company to select new employees. It is also likely that potential employees partly select the company they want to work for based on its culture. In the case of takeovers or insourcing projects this is entirely different. Large numbers of employees can enter a firm that has a totally different culture than the one they left. Important in this respect is which corporate culture is dominant. Corporate culture thus can also have an indirect hereditary effect on a company s decision to opt for a collective agreement. Though they did not have any direct influence through a parent company, these companies were indirectly influenced by the companies they take over or source in. Corporate culture can thus have both a direct and an indirect effect on the decision to use or start using a collective agreement. As such corporate culture integrates the two basic empirical explanations into one single explanation. A corporate culture is shaped and influenced by a company s lineage. Corporate culture also influences and is influenced by the outside world. If a corporate culture is protective of its own culture, a company will prefer to grow autonomously. If a corporate culture is aggressive, it will more likely grow through taking over other companies. In doing so it has to deal with other company cultures, to which it can adapt if the other party s size and culture are dominant. It all depends on who is taking over whom, how many companies are taken over, and who is the stronger party. The ICK collective agreement is a different story and, as a case, does not fit the empirical explanations mentioned above. The ICK collective agreement is an industry wide collective agreement and not a company level collective agreement. Its roots lie within the office branch, which has ties to the IT industry, but these connections are limited and the ICK collective agreement covers only a part of the Dutch IT industry. In that sense the term industry wide collective agreement should be placed into perspective.

The Presence and Future of Collective Agreements 213 Section 7.1 will now continue to discuss the hypotheses and the extent to which these hypotheses fit the empirical data and explanations. 7.1.1 The Limited Choice Set Hypothesis The limited choice set hypothesis fits the empirical explanation best, as corporate culture, which plays a key role in the empirical explanation, limits the choices a company makes. Lineage and past choices can make it harder for a company to differ from conventions and ways things are regulated inside the company, e.g. terms and conditions of employment. Also when companies deal with the outside world, e.g. in the case of insourcing projects or takeovers, they have to deal with the cultures of other companies and groups of people, which can limit the number of outsourcing projects. An example of the way lineage limits the choices a company makes is Roccade. Roccade, which in 2004 existed in the form of PinkRoccade, was privatised by the Dutch government in 1990 and its employees had been working with a collective agreement for almost 40 years at that time. There is no doubt that their choice for a collective agreement was limited by its history. Because companies are always seeking to maximize their profits one might think that the hereditary effect involved with takeovers is more a matter of efficiency than a matter of choices being limited, but this is not true. The institutional theory North propagates is partly neoclassical and assumes that actors seek to maximize their profits. Each institutional system creates its own incentive structure that determines which actions are rewarded more than others. Because actors seek to maximize their profits, they are limited in their actions. An example of this would be EDS. EDS predominantly sources in IT departments of public organizations. The IT professionals working within these organizations are all accustomed to collective agreements. In order for EDS to become a more attractive outsourcing party for public organizations, EDS has chosen to use a collective agreement. Even today IT companies choose a collective agreement because of the embeddedness of collective agreements in Dutch system of industrial relations. 3 This directly influences the types of agreements IT departments within these companies are using. When IT companies seek to take over such IT departments through insourcing, 4 they are confronted with the way industrial relations are shaped in these departments, and thus, often, a collective agreement. The limited choice set hypothesis draws heavily on the notion of path dependency, which states that later choices depend on previous ones. Events can be narrowed 3 Table 6.6 shows 85 percent of the Dutch employees in 2004 fell under a collective agreement. In 2004 there were 195 industry wide collective agreements in the Netherlands covering 5,321,500 people, and 564 company level collective agreements covering 716,000 people (SZW 2004). 4 Insourcing and outsourcing are the same process. A takeover process is called insourcing from the viewpoint of the buyer, it is called outsourcing from the viewpoint of the seller.

214 Developing Institutions down to the choices made at the moments where the various companies started up, were privatized, or taken over. The choices that were made at these occasions have had effect on the contingent of future choices in this case for instance the type of agreement to regulate the terms and conditions of employment within the company. To give an example, Torex-Hiscom started out as an association of multiple experts from various Dutch hospitals. They all fell under the scope of the collective agreement for hospital personnel, but eventually started using their own, independent, collective agreement. The company was taken over twice and kept using a collective agreement. This is in large part due to the personnel of Torex-Hiscom, of which 40 percent is a union member. Cooperation with the unions is strongly embedded in the company culture. Management has attempted to introduce meritbased pay, but this plan was abandoned as the employees saw their activities as a collective enterprise, and not an individual one. Obviously the limited choice set hypothesis also holds true for companies not using a collective agreement. The hypothesis is not restricted to explaining the presence of collective agreements, but also the absence of collective agreements. Companies that have a different lineage from the privatized IT departments of state owned companies, will be geared towards different choices. Perfect examples of these types of companies are CMG, Capgemini, Centric, Ordina, and Baan. These companies in most cases regard unions as the odd ones out. The only tradition these employers share is that of unfamiliarity and lack of sympathy for unions. They hold the same aversion towards the collective agreement, which theoretically may be an option open to them, but to which they are simply not responsive. This does not mean, however, that companies with a collective agreement can never merge with companies not using a collective agreement. PinkRoccade, for instance, but also EDS and Getronics, have shown this not to be the case. While path dependency, on which the limited choice set theory is based, exemplifies the difficulty of institutional change, 5 it appears that institutional choice is, when possible, more likely to occur than institutional change. Of the ten cases, three have chosen to use a collective agreement after having used a different type of agreement. Empirical evidence presented by cases such as EDS and Getronics has shown that takeovers and insourcing processes can lead to choosing a different type of agreement. In the case of PinkRoccade, however, a compromise was made, and a best-of-both-worlds solution developed initially: one in which on a company level negotiations took place with the unions, and on a subsidiary/business unit level with the works council. The limited choice hypothesis can also be applied to employees. In chapter six, table 6.9 demonstrated that in most cases IT professionals preferred the type of agreement they were already working with. Over sixty percent of the employees covered by an industry-wide collective agreement preferred this type of agreement over other types of agreements. The same held true for those falling under a company level collective agreement (54 percent), and those solely working on the basis of an 5 All types of terms and condition of employment agreements are regarded as institutions. For a more elaborate explanation of institutions, institutional change and choice, see chapter one.

The Presence and Future of Collective Agreements 215 individual agreement (56 percent). Interestingly, only those working under a works council agreement preferred other types of agreements. Only 26 percent actually preferred a works council agreement. The relative number of respondents preferring the type of agreement they are already working with is 49 percent of the IT professionals working in the IT services industry. This is just slightly less than half. 7.1.2 The Efficiency Hypothesis The efficiency hypothesis emphasizes the role of costs of the various types of agreements. Institutional theory suggests that institutions bring about incentive structures and reward certain actions more than others. Institutions also help reduce various costs involved. Williamson s theory of transaction cost economics for instance stresses that institutions can lower transaction costs. The efficiency hypothesis predicts that employers are aware of the fact that institutions may be efficient to different degrees and will seek out to use those agreements that best serve their goals. As such the collective agreement may have advantages over other types of agreements, or vice versa, other agreements over the collective agreement. One of the two basic empirical explanations for the presence of collective agreements in the Dutch IT software and services industry, the primary growth strategy, fits neatly in with the efficiency hypothesis. Companies seeking to take over or sourcing in companies that are using a collective agreement, predominantly EDS and Getronics, agree that their collective agreement, which they have (at least partly) adopted because of insourcing and takeover processes, has simplified these very processes. Takeover and insourcing processes are not the only reason why the efficiency hypothesis can say something meaningful about collective agreements. Companies like Atos Origin, PinkRoccade, Getronics, and EDS have all indicated that the collective agreement has helped them save costs. Atos Origin for instance had problems harmonizing the multitude of terms and conditions of employment amongst their employees. The collective agreement and, more importantly, the collaboration with the unions has helped them greatly in this process. In an evaluation of the collective agreement in 2003 they again had to agree that the collective agreement had many advantages, one of them being the reduction in negotiation costs. CMG was the only company that indicated that it thought collective agreements would increase the costs of negotiations, because it would mean introducing another party, i.e. the unions. At the time Getronics wanted to take over Roccade, it considered various alternatives to their works council agreement: (1) the ICK collective agreement; (2) a company level collective agreement; (3) a new, moderate sized, industry wide collective agreement; and (4) sticking to their works council agreement. One of the arguments for the collective agreement and the industry wide agreements was cost reduction because of economies of scale and the introduction of a relatively cheap

216 Developing Institutions expert, 6 i.e. the unions. On the other hand the introduction of the unions would also mean a relative loss of control, because now they would have to deal with a new party. At the time of the interviews, both EDS and PinkRoccade were trying to harmonize the works council agreements into their respective collective agreements. The costs of having to negotiate with both the unions and the works councils separately were becoming too high. 7.1.3 The Anachronism Hypothesis Can the collective agreements present in the Dutch IT industry be regarded as anachronisms? The answer here is negative. The collective agreements that currently exist in the Dutch IT software and services industry are not temporary solutions. This study has demonstrated that collective agreements were already present when the Dutch IT industry was in its infancy. The collective agreements that are still present today, and which came about during the growth of the industry, are still used to great satisfaction. Though the recession in the Dutch IT industry, which started during this study in 2001, had a negative impact on the relationship between the unions and the employers, collective agreements themselves have never been under discussion. At some point during the 2003 negotiations over their collective agreement, Atos Origin threatened to abandon the collective agreement. They, however, have re-evaluated the advantages the collective agreement offered them, and concluded it was still a valuable instrument to them. In 2004, 75 percent of the IT professionals working in the Dutch IT industry was not covered by a collective agreement. Large companies such as CMG and Capgemini, and relatively smaller ones like Centric and Ordina, are not inclined to accept union influences, nor any other outside pressure. The situation at Baan used to be a little bit different as employees were regarded as entrepreneurs. This also holds to some extent for Centric. Despite the fact that companies regard their workers as highly skilled employees, perfectly capable of negotiating their own terms and conditions of employment, some of these companies prefer the advantages that the collective agreement has to offer over other types of agreements, which do not involve the unions as a negotiation party. Employees preferences are also an important factor in determining under what type of agreement they fall. They show the same polarized view regarding unions as their employers do. Employers have learned that they cannot ignore the wishes of the employees. Without mentioning numbers, interviewees have indicated that some employees preferred to leave the company they had been working for after that company had been taken over by a different company. This is why having a collective agreement can have an advantage during takeover and insourcing processes, because in cases where the company that does the take over has the same type of agreement, it may reduce the number of people leaving the company. 6 It is more expensive for a company to hire such expertise on a permanent basis, than on a demand basis.

The Presence and Future of Collective Agreements 217 When looking at the preferences of IT professionals in the Dutch IT services industry, 38 percent of the respondents favours individual agreements over collective types of agreements. This preference is independent of age. This is not true of the highest attained level of education of employees. There is also a distinct relation between income and level of education. The higher their income, the more they favour individual agreements. The same goes for job level: the higher the job level, the more workers prefer to negotiate individually over the terms and conditions of employment. This is most likely the result of a rational evaluation, which leads them to believe that their chances of better terms and conditions of employment will improve through individual negotiations. Interestingly enough 70 percent of the IT professionals indicated that they thought they performed better than their colleagues and wished to negotiate individually over their performance. Fifty percent thought that collective agreements would only have a levelling effect on wages and do not take into account individual performance. As in the rest of the Dutch economy, older employees are more often union members than younger ones. The rise in union membership in the Dutch IT industry may thus be partly an effect of an ageing population of IT professionals. As the average age of the IT professional working in the Dutch IT industry is increasing, the number of union members may rise. Despite that possibility - this study was not specific enough to answer that question -, employees may want to increase their bargaining power through the unions. Especially in the less than optimistic market conditions since 2001, they have increasingly sought the help of the unions to improve or defend their position. The same thing happened during the first half of the 1990s when the Dutch IT industry also fell victim to a crisis. 7.1.4 Linking the Empirical Results Back to Institutional Theory One can wonder if this study provides additional insights for the development of institutional theory. The reason for conducting this study was not to test institutional theory, but rather an empirically driven question regarding the presence of collective agreements in an industry that is more or less foreign to the Dutch consultation model. Comparing the various theoretical explanations, one cannot underestimate the power of the institutionally based explanation. The limited choice set hypothesis offers the most convincing explanation that covers all of the empirical explanations. It also covers the largest number of cases. Table 7.2 presents an overview of the findings and how well the three theoretical explanations cover the various empirical sources, such as the empirical explanations from chapter five, the case studies of, again, chapter five, and finally, the preferences of IT professionals of chapter six.

218 Table 7.2 Developing Institutions The empirical data and empirical explanations for the presences of collective agreements versus the hypotheses Limited choice set Hypothesis Anachronism hypothesis Efficiency hypothesis Empirical explanations - Historical lineage of a company - Primary growth strategy - Corporate culture - No collective agreements because they are too rigid - Primary growth strategy - Corporate culture Cases - Torex-Hiscom - PinkRoccade - CMG - Getronics - EDS - Capgemini - Centric - Ordina - Baan - CMG - Capgemini - Centric - Ordina - Baan - Atos Origin - PinkRoccade - Getronics - EDS IT professionals - Preference for the type of agreement professionals are already working with. - Union membership and age - Preference and level of education - Preference and function level When defined as: a selection of multiple institutions that try to regulate the same actions, institutional theory says very little about institutional choice. This is a pity as it might give more insight into why multiple institutions develop to cover the same problem. Why do we have so many institutions in the Netherlands that serve the same goal? Take for instance the number of institutions in the Netherlands that regulate relations between people such as marriage and living together. Following Knight this would have to be explained from empowerment of different groups such as e.g. gay couples or people that do not want to get married yet still want to base their relation on a legal arrangement. Obviously having a choice is a good thing. What is surprising, though, is that when looking at the contents of most of these regulations, they are hardly any different. Consider e.g. the comparison of the various flexible and individual items in both collective agreements and works council agreements at the end of chapter five: there hardly is any difference at all. Section 7.3 will deal with this latter topic more thoroughly. The following sections will deal with answering the remaining questions of this study.

The Presence and Future of Collective Agreements 219 7.2 Explaining the (inter)actions of Actors This section will focus on the actors involved in the field of industrial relations, the employees, the works councils, the unions, the employers, and the employers associations. The two sub sections that will follow will compare the position of the works councils and unions in relation to the employers and employees. 7.2.1 Unions and Works Councils in the Dutch IT Industry From an Employees Perspective Employees in the Dutch IT industry are more diverse than is often assumed. Though IT professionals in the Dutch IT software and services industry are on average more highly skilled, predominantly male, and work full time more often than the average Dutch employee, this is where the differences end. This is not surprising as with the increasing age of an industry, its employees grow older on average as well. Most of the IT companies discussed in this study have been around since the 1970s. Employees who started out in their twenties in those days are now in their fifties. At the time of the survey two types of IT professionals can be observed that have diametrically different interests and demands with regard to what kind of terms and conditions of employment they want. The first group is that of the younger generation of IT professionals, which are still at the start of their career, usually have no children, and regard a company car and a good salary as the most important terms and conditions of employment. The second group consists of older IT professionals who are aged 45 and up, and who slowly start thinking of their retirement. It is among this latter group that most union members can be found. Obviously this division does not only take place in the IT industry, but can be witnessed everywhere. However, it is still worthy of note here, because this situation is relatively new in the Dutch IT industry. The younger generation of IT professionals does not take the unions seriously, because they think unions will not get them what they want. The older generation of IT professionals is much more receptive to the unions position. They hope to improve their situation together with the unions. Older IT professionals especially fear unemployment. Being more expensive than their younger counterparts, and usually less flexible geographically due to having families, they are more couch prone than others and hence seek means to strengthen their position. In this position, becoming a union member can be regarded as a rational solution. In some cases this has paid off, such as in the cases of Baan and Getronics. There, on one occasion, social plans were made that included extra benefits for union members over non-union members. One union, CNV, even goes as far as promoting individual advantages for its own members. One can question though to what extent these new members can really be called union members. It may just be that they are only temporary members until their positions improve when things get better in the IT industry. A surprising outcome of the Internet survey among IT professionals was their trust in the negotiation powers of the works council, yet the survey showed a significant lack

220 Developing Institutions of preference for a works council agreement. These outcomes do not coincide with the information retrieved from the ten case studies. This is most likely due to the fact that the information on the cases stems predominantly from interviews with management and members of the works council. Unfortunately the survey does not provide enough information as to why IT professionals dislike works council agreements so much in comparison to other types of agreements. In terms of primary and secondary terms and conditions of employment and fringe benefits, they are, on average, on par with company level collective agreements. The position of IT professionals towards works councils is rather ambiguous. Those professionals working under a(n industry wide) collective agreement predominantly seem to distrust the negotiation powers of the works council. This is much less so for IT professionals working under a works council agreement or an individual agreement. They seem to have more confidence in the power of the works council and regard them far less as a tool of management. This is, however, a result which should be examined further in future research. 7.2.2 Unions and Works Councils in the Dutch IT Industry: Employers Tools? The provocative title of this section is based on the author s belief that the collective agreements in the Dutch IT industry, which have appeared between 1996 and 1998, only came about because the companies themselves could use them to their advantage. Employers have tried to increase their bargaining power at the costs of the unions. Though the unions were very pleased with the collective agreements at EDS and Getronics, which were established in 1998 and 1997 respectively, and even considered the Getronics collective agreement as a personal victory, it was not their victory, but rather that of the respective companies. Both Getronics and EDS had been successfully keeping the unions at bay for years and only started using a collective agreement because they saw that it was a tool they could strategically use to their advantage. The collective agreement gave them the appearance of a company that did not take terms and conditions of employment lightly and an improved relationship with the unions would help them during takeovers and insourcing projects. The tightness of the IT labour market at the time must have played a role in their decision making. The Atos Origin case can be considered an example of the improved position of the unions in the Dutch IT industry. However, the Atos Origin case only proves that the unions, over the years, have found a way to manoeuvre themselves to the negotiation table by presenting themselves as experts, rather than as a traditional union. Atos Origin brought itself into a difficult position when it wanted to harmonize a variety of terms and conditions of employment. This is not to discredit the unions. The terms and conditions of employment the unions managed to negotiate together with the works council and management were relatively good compared to other Dutch IT companies at the time. During the negotiations over the second collective agreement, however, things changed radically and Atos Origin has been steadily increasing its bargaining power at the cost of the unions. In particular,

The Presence and Future of Collective Agreements 221 the uncoupling of the negotiations over wages from secondary terms and conditions of employment has been a disadvantage for the unions as they cannot negotiate and bargain over the primary and secondary conditions of employment at the same time. Though this may also be a disadvantage for the employer as well, the employer s position is much better: the employer can determine the wage budget that allows him to reduce the danger of estimating next years budget too optimistically; and it allows the employer to be more conservative and careful than he otherwise might have been. An aspect of collective agreement is that it only specifies that Atos Origin is required to talk to the unions and works council about the wages. 7 Should the parties not agree on the terms, Atos Origin is free to do whatever it wants. The recession in the IT industry has not had a positive effect on the unions position. Although it did help increase their membership numbers, the unions found it increasingly difficult to negotiate with the employers. Companies that do not use a collective agreement often use the works council as an excuse to keep the unions at bay. Union membership in these companies is often estimated as maximally five percent, whereas in companies using a collective agreement these estimates by the employer are in the order of ten percent. Though both types of companies, those with and without a collective agreement, regard their works councils as a valuable tool for measuring the preferences of the employees, employees evaluate the works council differently. In the previous section it was mentioned that in companies employing a collective agreement employees value the works council less than companies not using a collective agreement. Though this has not been researched extensively, future research may indicate that this is simply because unions take over many of the functions of the works council in these companies. In companies not using a collective agreement, the works council simply stands on its own and has to acquire the specific skills itself. At least in one of the cases, Atos Origin, this was mentioned specifically with regard to the works councils of BSO/Origin and Philips. Interviewees mention that the BSO/Origin works council had far more experience in negotiating over terms and conditions of employment than the former Philips works council. This was so because the former Philips works council had always depended on the unions doing the negotiations for them and the former BSO/Origin works council had to do this on their own. The ICK collective agreement case can also be regarded as an example for the relatively weak position of the unions in the Dutch IT industry. After reviewing its entire history, which is dominated by an uncooperative employers association, and a disheartening repetition of moves to get the collective agreement extended, it is surprising to see that the unions are even willing to invest more time in the ICK collective agreement. The employers association, which is mainly a lobby organization, only supports the ICK collective agreement because some of its members have a need for a cheap agreement that regulates the terms and conditions of employment for them. The original idea of creating a collective agreement which would be flexible on an individual level for both employers and employees, ended up 7 See chapter two. Employers in general are not compelled to negotiate over everything.

222 Developing Institutions in nothing short of a nightmare for the unions. When asked why they still supported the agreement at all, one of the unions responded that some employees would be even worse off without the ICK collective agreement. 8 7.3 Individuality and Flexibility in Collective Agreements: A Paradox? The third research question, mentioned in table 7.1, asks to what extent the contents of these collective agreements differ in terms of flexibility and individuality from other types of agreements? How do the stipulations in these agreements compare to prevailing arrangements on similar topics in similar IT firms that lack a collective agreement? This question was answered in chapter five in section 5.4. The conclusion was that there is no difference in terms of flexibility. This is not surprising as the level of competition in the Dutch IT industry is so high that companies carefully watch what competitors are doing. This was especially the case during 1998-2000 when the Dutch IT industry was booming. Recalling the case of Atos Origin and its first collective agreement, which in retrospect was considered too good by Atos Origin, it had the function of attracting employees. The collective agreement of Atos Origin proved for some time to be the reference for terms and conditions of employment for other companies. The only difference in terms of flexibility can be noted between software companies and service companies. The latter thrive much more on individual performances than the former. Performance related pay makes much less sense to a programmer who is often part of a team than to a person who is seconded and to whom performances and result can be ascribed individually. Though attempts were made to introduce a pay for performance system at Torex-Hiscom, the employees simply refused such a system. A particular concern pertained to the method by which they would be judged, which was rather vague and experimental. The failure to introduce a pay for performance at Torex-Hiscom can be considered an example of company culture limiting choices. Performance related pay, however, is becoming part of the general culture of companies in the Dutch IT industry and as such will to some extend influence companies internal cultures and decisions. In the case of Torex-Hiscom it is a matter of power and because the company is relatively small and less diverse in terms of activities than other companies, the employees have a strong cultural unity. 8 One of the union leaders antipathy towards the collective agreement was so great that he mentioned he would become physically ill if he would have to negotiate for the ICK collective agreement again in the future.

The Presence and Future of Collective Agreements 223 7.4 The Future of Collective Agreements in the Dutch IT Industry. So what can be said about the future of collective agreements in the Dutch IT industry? Based on the hypotheses one can imagine a number of scenarios. The first scenario based on the limited choice hypothesis, is that everything will remain the same. Companies will not change the nature of their regulations and will continue to use what they have done in the past. This however is not a very likely scenario as the Dutch IT industry is in constant flux. Instead a second scenario based on the limited choice hypothesis, which encompasses the movements within the industry, is much more likely to occur. This second scenario predicts that there will be a decline in the total number of collective agreements, but a growing number of employees falling under the scope of a collective agreement. The reason for this is the consolidation process in the Dutch IT industry. The consolidation process is a part of this industry and will remain a strong centralizing force in the industry until a few large players remain. When two companies use a collective agreement and one takes over the other, the total number of collective agreements decreases. This scenario is more than hypothetical as in 2005 Getronics successfully took over PinkRoccade. 9 This means that although the actual number of collective agreements in the Dutch IT industry decreases, the number of employees falling under a collective agreement will remain roughly the same. This does not necessarily harm the position of the collective agreement in the Dutch IT industry. This second scenario, however, could in theory turn out to be a sour grape for the future of the collective agreement in the Dutch IT industry, for there are more factors at play here. If a company that does not use a collective agreement takes over a company that does and has a stronger culture than the party that is taken over, it is uncertain whether they will start using a collective agreement after the takeover. As an effect the number of collective agreements may decline as well as the number of workers falling under its scope. History has, however, shown that in practice actually the opposite is more likely to occur in these cases, and companies will consider the option of using a collective agreement themselves in order to be come a more attractive party to the company that is taken over. Outsourcing projects also will continue to play an important role in this respect. Outsourcing projects often have the effect that IT professionals generally not categorized previously as IT professionals working in the Dutch IT industry are now regarded as IT professionals working in the IT industry. This means an increase in the total number of IT professionals working in the Dutch IT industry. Working outside the Dutch IT industry often means that these people are more accustomed to working with a collective agreement than without. During interviews a few firms now working with a works council agreement have indicated that outsourcing projects may be the primary reason for them to start using a collective agreement in the future. 9 Something which Getronics already had planned on doing back in 1996/1997.

224 Developing Institutions This means that presently, the only foreseeable means for an increase in the number of collective agreements and increase in number of IT professionals falling under a collective agreement will stem from the indirect hereditary effects of outsourcing projects and consolidation processes. A scenario for the future of the collective agreement based on the efficiency hypothesis would look quite similar to the second scenario based on the limited choice hypothesis mentioned above. The Dutch system of industrial relations and the dominant position of the collective agreement within it, create incentives towards using a collective agreement for companies seeking to source in IT departments or taking over other firms. Though several companies such as Getronics, who have started using a collective agreement of their own, indicated that collective agreements have reduced certain costs for them, it does not seem likely that other companies will start using a collective agreement because of this. Change costs time and money and other reasons must be present to make a transformation like this. As such efficiency will only play an additional role. Although one that should not be underestimated as the Atos Origin case demonstrates. A scenario based on the anachronism hypothesis would predict a decline in the number of collective agreements because they would have no place in an industry such as the IT industry. IT professionals have no need for rigid labour contracts and are perfectly capable of negotiating over the terms and conditions of employment themselves. Though indeed a large segment of IT professionals does not think very highly of unions and collective agreements, empirical evidence presented in chapters five and six, clearly demonstrates that collective agreements are far from the archaic rigid structures they are accused of being. The question however is, will the collective agreement meet the demands of the IT industry and its workers in the future? The answer to that question would be positive. IT professionals unfamiliar with individual negations will prefer having unions negotiate for them in the future under the condition that they will negotiate as successfully in the future as they have in the past. Empirical evidence in chapter six has indicated that it is likely that, on average, IT professionals that fall under the scope of a works council agreement or collective agreement have a better position than those who have to negotiate individually. Though some people that work at a company that uses a collective agreement may indeed be more successful at improving their terms and conditions of employment more than the unions, these individual and their colleagues, on average, are better off than their colleagues that have to negotiate individually. So although from an individual point of view the anachronism thesis may play an important role for some people, it is in essence company culture and efficiency that are decisive for the future of the collective agreement in the Dutch IT industry.

The Presence and Future of Collective Agreements 225 7.5 Some Final Remarks This study makes clear that employers in the Dutch IT industry will use institutions, works councils and unions strategically and to their advantage. An example of such strategic behaviour is e.g. the situation in which a company has to lay off a great number of people. In most cases employers prefer to draw up a social plan with the unions instead of negotiating with their works councils, whom the employers often think are too much involved to be objective. The cases of Atos Origin, EDS, and Getronics prove that the unions have managed to position themselves in the Dutch IT industry. They, however, could only do so because the companies in question chose to allow them to do so. Without the proper incentives for these employers, the unions would never have attained the position they have now. This is not to say they would have had no place in the Dutch IT industry at all; that would stretch the point too far. Unions have managed to mould themselves into a servile role of experts. They had to because that was the only way for them to get in. Though union membership numbers have increased slightly over the years, and may continue to do so until the figure reflects that of the rest of the Netherlands, they will still find it hard to convince employers that unions are representing their employees. If the unions truly want to gain more than a foothold in the rest of the IT industry, they should do a number of things. Unions should try to win the hearts of the employees. This seems close to impossible as a dominant segment of the IT professionals believes that they perform better than their colleagues and therefore should earn more. As the unions represent the collective, this will not be the first place individuals will look to improve their position. This is an interesting paradox as the best collective results are best gained collectively. This belief that a better position can be achieved through individual negotiations works to the advantage of the employer as it generally empowers him. Individuals see no need to organize as they think they are better equipped than their colleagues and therefore want to be individually rewarded. Perhaps the unions will one day find a way to break through this paradox and convince individuals that they are better off together than alone. One can question whether IT professionals are in a real need of unions. Many of them have incomes far higher than the average employee in the rest of the economy. In that respect the unions would do better to focus on gaining members amongst companies using the ICK collective agreement, and amongst companies using neither a collective agreement, nor a works council agreement, for that is where IT professionals that could best use the help and support of the unions and decent terms and conditions of employment may be found. Moreover, employees lack enough trust in the unions. This is to a large extend a matter of culture and probably unions will be more successful with employers. Cases such as Getronics, EDS, and Atos Origin have shown that employers using the unions and collective agreements as a tool, give the unions more credit than employees do, in trusting them as a partner that has shown itself to be an expert in the fields of industrial relations and working conditions, and which is a valuable asset to companies seeking to restructure and save costs on their terms and conditions of employment.

226 Developing Institutions Unions should be more proactive during insourcing and takeover procedures of IT companies as these are the best times for them to play their cards. The IT industry is increasingly becoming part of the rest of the economy because productivity becomes more and more dependent on information technology. IT companies will be increasingly confronted with the dominant system of industrial relations in the Netherlands. It is at these key moments that unions can help IT companies. In a way this will mean that unions will continue to play a role as a tool for the employers, but there is nothing wrong with that as it would mean that they would help the employees working within these companies just as much. This study is titled Developing Institutions: Collective Agreements in the Dutch IT Industry. Collective agreements were originally introduced in the Netherlands to balance the power between the two power groups on the labour market: employers and employees. It is strange to find collective agreements in the Dutch IT industry because collective agreements reflect the power of the collective employee. It is strange because both employers and employees in the Dutch IT industry proclaim that they are evenly matched and can work things out between themselves well enough without the intervention of unions. This study has demonstrated that collective agreements in the IT industry have mainly originated through changes in the cultural setting of companies. Institutions themselves are notoriously rigid and not easy to change. Theorists often claim it takes a revolution, a radical shift in the power structure, to bring about institutional change. Power structures express the division of power between various power groups. Institutions reflect the formal and informal rules of a society, and power groups are more successful at strategically using these rules to their own advantage. Institutions and culture resonate. Institutions develop within a specific culture. Cultures are unique and therefore so are institutions. Despite their sui generis character, institutions across different cultures often provide solutions to similar problems. So when power balances change, groups that gain more power are able to introduce and impose their institutional solutions upon others. They can do so under two conditions: (1) if they are powerful enough to create new institutions; and (2), if they cannot use existing institutions to their own advantage. If they are not powerful enough and/or are not willing to impose their own rules upon others, they will create institutions that will coexist with the ones that were already in place. Institutional change, and ultimately social change, takes place through the parallel development and selection of institutions by power groups. From the above follows that if power groups share a more or less equal amount of power, a multitude of institutions can exist. Institutions development means that there are more possibilities for social change, which in effect makes societies and their economies more flexible and increases changes for a more efficient institutional system. Hall (1985) explained that power groups, if too strong, can become capstones to societies. These capstone powers lock societies by blocking institutional and social development. Without institutional development, there is no social change, and

The Presence and Future of Collective Agreements 227 without social change there is ultimately no economic development, because institutions will only benefit one group and not society as a whole.

228 Developing Institutions

Samenvatting (Summary in Dutch) De informatietechnologie sector, meestal aangeduid als IT sector, wordt algemeen gezien als een vreemde eend in de bijt van het Nederlandse systeem van arbeidsverhoudingen. Vaak wordt aangenomen dat zowel werkgevers als werknemers in de Nederlandse IT sector wars zijn van vakbonden en collectieve regelingen. De organisatiegraad, voor zover deze objectief in beeld gebracht kan worden, is traditioneel laag. Het aantal collectieve arbeidsovereenkomsten is mogelijkerwijs nog lager. Naar de oorzaak van deze on-polderse situatie in de Nederlandse IT sector is nooit grondig onderzoek gedaan. De reden hiervoor is onduidelijk. Het ontstaan van een drietal CAO s bij toonaangevende IT bedrijven als EDS, Getronics, en Origin in de jaren 1997 en 1998 is de aanleiding geweest voor het onderhavige onderzoek. De totstandkoming van deze drie CAO s was opmerkelijk te noemen, omdat ze waren ontstaan bij bedrijven die zich voornamelijk op dienstverlening richtten; diensten welke voornamelijk uitgevoerd werden door hoog opgeleide professionals. Het type bedrijf en werknemer waarvan wordt aangenomen dat ze model staan voor een nieuw stelsel van arbeidsverhoudingen. Een stelsel, waar de werknemer zonder tussenkomst van een vakbond onderhandelt over zijn of haar arbeidsvoorwaarden. Juist daarom is het ontstaan van deze drie CAO s een opmerkelijk fenomeen te noemen, dat meer aandacht verdient. Dit onderzoek heeft dan ook als hoofddoelstelling te verklaren waarom er CAO s zijn ontstaan in de Nederlandse IT sector. Subdoelstellingen zijn: het in kaart brengen van de ontwikkeling van arbeidsverhoudingen in deze sector, de sector zelf inzichtelijk maken en het belichten van de belangrijkste actoren (werkgevers, werknemers en hun collectieven). De opbouw van dit boek bestaat uit een zevental hoofdstukken. Het eerste hoofdstuk behandelt de probleemstelling en het theoretische kader. Het tweede hoofdstuk bespreekt het Nederlandse stelsel van arbeidsverhoudingen en de instituties die in Nederland bestaan voor het reguleren van arbeidsvoorwaarden. In hoofdstuk drie wordt de Nederlandse IT sector besproken en gedefinieerd. Hoofdstuk vier belicht vervolgens de werkgevers en werknemers organisaties actief in de Nederlandse IT sector en wordt de enige in de sector voorkomende sectorale regeling, de ICK CAO, als case besproken. Hoofdstuk vijf bestaat uit een tiental case studies van bedrijven met en zonder CAO. Aan het eind van dit hoofdstuk worden de vijf CAO s en vijf ondernemingsovereenkomsten inhoudelijk vergeleken op basis van hun flexibiliteit en

230 Developing Institutions individuele keuzemogelijkheden. Hoofdstuk zes, het laatste empirische hoofdstuk, geeft een beeld van de (voorkeuren van) werknemers in de Nederlandse IT sector. Het zevende, concluderende, hoofdstuk ten slotte, beantwoordt de onderzoeksvragen. Naast de probleemstelling zoals die hierboven is besproken, geeft hoofdstuk één de theoretische onderbouwing voor deze studie. Het theoretisch kader voor het beantwoorden van de onderzoeksvragen is de institutionele theorie, zoals die is vormgegeven door North (1991a, 1991b, 1992, 1993, 1994) en Knight (1992). De belangrijkste invalshoek van North binnen deze theoretische stroming is het idee dat instituties zogenaamde prikkelstructuren definiëren. Simpel gezegd zorgen instituties er niet alleen voor dat handelen mogelijk gemaakt wordt door regels, maar dat instituties bepaalde handelingen beter belonen dan andere handelingen. North trekt, zeker in later werk, de efficiëntie van instituties en rationaliteit van actoren steeds meer in twijfel. Waar North de efficiëntie van instituties laat afhangen van onderhandelingen tussen verschillende partijen, benadrukt Knight juist de verschillen in macht tussen deze partijen en de vaak negatieve gevolgen hiervan op de efficiëntie van instituties. Instituties, zo stelt Knight, drukken voornamelijk de belangen uit van de groep met de grootste macht. Op basis van deze theoretische inzichten is een drietal hypothesen geformuleerd. De eerste hypothese, de beperkte keuze -hypothese, benadrukt dat actoren beperkt worden in hun keuzevrijheid door keuzes die zijn gemaakt in het verleden. Keuzes zijn padafhankelijk en de keuze voor een bepaald type arbeidsvoorwaardenregeling kan invloed hebben op een latere keuze. De tweede hypothese, de efficiëntie hypothese, geeft juist aan dat de keuze voor een bepaald type arbeidsvoorwaardenregeling gebaseerd is op basis van efficiëntie overwegingen. De laatste hypothese, de anachronisme hypothese, stelt dat de keuze voor een bepaald type arbeidsvoorwaardenregeling afhankelijk is van de voorkeur van individuen die hun onderhandelingsmacht willen vergroten en kiezen voor instituties die die onderhandelingsmacht het beste vertegenwoordigen. Deze voorkeur kan beïnvloed worden door een collectieve dan wel individualistische cultuur of tijdgeest. Als zodanig kunnen instituties als een anachronisme worden gezien wanneer ze niet binnen een bepaalde cultuur of tijdgeest passen. Hoofdstuk twee bevat een beschrijving van het Nederlandse systeem van arbeidsverhoudingen en toont hoe een instituut als de CAO tot stand is gekomen en is ingebed in het Nederlandse stelsel van arbeidsverhoudingen. Het eerste empirische hoofdstuk, hoofdstuk drie, dat de Nederlandse IT sector definieert en beschrijft, laat zien dat de oorsprong van de IT sector in Nederland ligt bij de door de overheid opgerichte rekencentra. De groei en ontwikkeling van de sector was echter niet in handen van de overheid. Vooral in de jaren 70 en 80 van de vorige eeuw ontstond er een veelheid aan kleine IT bedrijfjes die werden opgezet door zelfstandigen die vanwege de toenemende vraag naar hun producten en diensten enorm konden groeien. Daarbij kwamen de buitenlandse IT bedrijven (uit vooral Engeland en de Verenigde Staten) die zich in Nederland vestigden. In een tijdsbestek

Samenvatting (Summary in Dutch) 231 van ongeveer 15 jaar kwam zo een sector tot stand die volledig wars was van bestaande Nederlandse tradities en voornamelijk gevormd werd door een ondernemersmentaliteit en door management praktijken overgenomen uit het land van herkomst. Hoofdstuk vier belicht de geschiedenis van arbeidsverhoudingen in de Nederlandse IT sector aan de hand van het handelen van de bonden en werkgeversverenigingen. Toen de bonden in de beginjaren 90 van de vorige eeuw zich rijkelijk laat begonnen te richten op de IT sector, was het eigenlijk al te laat voor hen en was er een sector ontstaan waar geen plaats meer was voor vakbonden, CAO s en collectief overleg. IT ers waren feitelijk zelfstandige ondernemers die prima over hun eigen arbeidsvoorwaarden konden onderhandelen. Bedrijven als bijvoorbeeld IBM, hielden er in die tijd een actief antivakbondsbeleid op na en verboden hun medewerkers in sommige gevallen te spreken met vakbondsbestuurders. De leden die de bonden wisten te werven in deze tijd kwamen binnen doordat de sector zich sinds het begin van de jaren 90 van de vorige eeuw in een crisis bevond een fenomeen dat zich iedere zeven tot tien jaar in de sector voor lijkt te doen. Evenals de bonden hebben ook de werkgeversverenigingen weinig vat op de sector gehad. In het begin kwam dit vooral omdat er bijna net zoveel werkgeversverenigingen waren als deelgebieden. Echter later, zelfs later nadat zich een aantal harmonisatierondes hebben voorgedaan en het aantal werkgeversverenigingen is afgenomen, beperkt zich de voornaamste functie van deze clubs tot het ondersteunen en tot dienst zijn van werkgevers. Werkgeversverenigingen zijn niet in staat een bepaald beleid op te leggen aan de sector als geheel. Ze zijn lobbyclubs die de werkgevers en sector vertegenwoordigen, regelgeving beïnvloeden en moeten zorgen dat de overheid investeert in de sector. De casestudy over de ICK CAO, de enige sectorale regeling in de Nederlandse IT sector, geeft goed aan hoe de verhoudingen in de sector liggen. Wat in 1995 begonnen was als een ambitieus plan om een oude kantoormachinebranche CAO op te waarderen naar een moderne CAO voor een zo mogelijk nog modernere sector, loopt uit op een teleurstelling voor de vakbonden. Een volgens de vakbonden magere CAO is het resultaat in 1997. In eerste instantie was er de belofte deze uit te breiden en verder in te vullen in toekomstige versies, maar in plaats van de initiële CAO afspraken in de jaren erna uit te bouwen via een ge-avv de CAO, zien de bonden het aantal regelingen alsmaar magerder worden omdat de werkgevers niet gebonden willen worden. Ook de beloofde avv komt nooit tot stand vanwege problemen met aanpalende sectoren als de grafiemedia waar men wel in staat is een avv voor elkaar te krijgen. De casestudy geeft een goed beeld van de positie van de bonden in het hardwaresegment van de IT sector. Hoofdstuk vijf laat zien dat de positie van de bonden in het hogere segment van de IT sector over het algemeen niet veel beter is. Er heerst vijandigheid en met bedrijven waar de vakbonden geen voet tussen de deur hebben weten te krijgen wordt voornamelijk via de pers gecommuniceerd. In dit onderzoek zijn tien bedrijven in de vorm van casestudies nader onderzocht. Vijf van de tien bedrijven gebruiken een

232 Developing Institutions CAO (Atos Origin, EDS, Getronics, PinkRoccade en Torex-Hiscom) en vijf bedrijven gebruiken een ondernemingsovereenkomst (CMG, Capgemini, Ordina, Baan en Centric). De vijf wel-cao-bedrijven representeren vrijwel alle ondernemingscao s die er in het software- en dienstengedeelte van de sector te vinden zijn. De niet-cao bedrijven zijn geselecteerd op basis van hun overeenkomst met de wel-cao bedrijven in termen van nationaliteit, omvang, markt en producten en diensten. De casestudies maken duidelijk waarom bedrijven wel of geen CAO gebruiken. Bij de wel-cao bedrijven kan onderscheid gemaakt worden tussen bedrijven die sinds hun ontstaan een CAO hebben (PinkRoccade en Torex Hiscom) en bedrijven die pas in een later stadium een CAO zijn gaan hanteren (Atos Origin, EDS en Getronics). Afkomst, in dit onderzoek lineage genoemd, kan voor een deel de aanwezigheid van CAO s verklaren. Bij de bedrijven waar men pas in een later stadium CAO s is gaan hanteren moet de verklaring ergens anders gezocht worden. De verklaring kan worden gevonden in de activiteiten die de bedrijven zijn gaan ondernemen in relatie tot bedrijven waar wél met een CAO wordt gewerkt. EDS is ooit in Nederland begonnen met de NS als grote klant. Later, in het midden van de jaren 90 van de vorige eeuw, is het bedrijf zich meer gaan richten op het insourcen van IT afdelingen van bedrijven uit de publieke sector; bedrijven waar de medewerkers allemaal gewend waren te werken met een CAO. EDS wilde de weerstand die vaak bestond bij deze werknemers wegnemen door zelf ook een CAO te nemen en daarmee te bewijzen dat men in de IT sector ook arbeidsvoorwaarden serieus neemt. Voor EDS was de keuze voor een CAO dus van strategische aard. Hetzelfde geldt voor Getronics, dat in de jaren 90 van vorige eeuw het geprivatiseerde staatsbedrijf Roccade wilde overnemen. Wilde men op de shortlist komen als potentiële overnemer, dan moest men een CAO hebben. De keuze voor een CAO bij Origin, zoals Atos Origin heette voordat het met het Franse Atos fuseerde, was uiteindelijk ook een strategische. Origin was een bedrijf dat was ontstaan uit een fusie tussen de voormalige IT ontwikkelafdeling van Philips en BSO/Origin. Origin had een flink aantal overnames gedaan en het aantal verschillende arbeidsvoorwaarden waarmee men werkte, rees de pan uit en men wilde deze gaan harmoniseren. Het harmonisatieproces bleek echter dusdanig complex dat de ondernemingsraden de hulp van de bonden hebben ingeroepen. Bedrijfscultuur speelt bij alle cases een grote rol en lijkt voor een groot deel te bepalen of een bedrijf wel of geen CAO heeft. Bedrijven die autonoom groeien en geen grote overnames doen, blijven vasthouden aan het type arbeidsvoorwaardenregeling dat ze van oudsher gebruiken. Bedrijven die juist wel veel aan overnames doen, dan wel groeien door middel van insourcing, zijn eerder geneigd zich aan te passen. Bedrijven kiezen in deze gevallen voor een CAO als hen dat strategische voordelen biedt. De voorkeuren van werknemers kunnen een rol spelen bij de keuze voor een bepaald type arbeidsvoorwaardenregeling. Ten slotte zijn in het laatste empirische hoofdstuk, hoofdstuk zes, ook de individuele werknemers in de IT sector ondervraagd door middel van een Internetenquête. Veel van de stereotype assumpties ten aanzien van deze werknemers werden bevestigd. De meeste werknemers zijn man (95 procent), hoger geschoold (80 procent) en werkt

Samenvatting (Summary in Dutch) 233 meer dan 35 uur per week (94 procent). Hoewel de gemiddelde leeftijd inmiddels gestegen is tot 34.4 jaar, is dat nog steeds lager dan het gemiddelde in de economie als geheel, dat ongeveer 43 bedraagt. Een andere opvallende uitkomst was de organisatiegraad. Indien men de respondenten in de subgroep werkzaam in de IT software en dienstensector bekijkt, dan is 12 procent van hen lid van een vakbond. Voor de overige respondenten (voornamelijk werkzaam in IT functies in andere sectoren) is dat percentage 16 procent. Een ander opvallend resultaat van de enquête was de voorkeur van werknemers voor het type arbeidsvoorwaardenregeling dat men heeft. De voorkeur voor individueel onderhandelen neemt pas toe als de onderhandelingspositie van een werknemer sterker is door bijvoorbeeld opleiding en functieniveau. De enquête schetst een beeld dat werknemers in de IT sector minder vijandig tegenover de vakbonden staan dan soms gedacht wordt en dat men niet negatief staat tegenover het idee dat anderen (ondernemingsraad of vakbonden) voor hen onderhandelen. Conclusie Het blijkt dat de hypothesen redelijk tot goed in staat zijn de cases te verklaren. Het is met name de beperkte keuze -hypothese die het meest weet te overtuigen en kan verklaren waarom bedrijven vasthouden aan de keuzes die ze gemaakt hebben. Immers de beperkte keuze -hypothese is deels gebaseerd op de notie van padafhankelijk die zegt dat keuzes tot op zekere hoogte afhankelijk zijn van keuzes die in het verleden zijn gemaakt. Bedrijven die zich weinig mengen met andere bedrijven en voornamelijk autonoom groeien, ontwikkelen een sterke bedrijfscultuur, waartoe ook het onderhandelen over arbeidsvoorwaarden behoort. Bedrijven die al vanaf het begin een CAO hanteren, of juist niet, blijven dat in de meeste gevallen ook doen. Overnames zijn een moment waarop bedrijven kunnen losbreken van de keuzes uit hun verleden. Men kan als het ware met een schone lei beginnen, ware het niet dat voor beide partijen in een overname de vraag geldt wie de dominante partij is en hoeveel waarde men hecht aan de overname. Zo is men bij EDS en Getronics overstag gegaan en een CAO gaan hanteren en zijn bedrijven als Raet en Origin uiteindelijk toch weer een CAO gaan gebruiken nadat ze deze in eerste instantie overboord hadden gegooid. Ook vanuit de efficiency-hypothese kunnen de keuzes die bedrijven hebben gemaakt voor een CAO verklaard worden. Deze hypothese stelt dat bedrijven, indien mogelijk, die instituties zullen selecteren die hun transactie- en onderhandelingskosten kunnen verlagen. Naast het feit dat bedrijven als EDS, Getronics en Getronics hebben aangegeven dat ze met een CAO beter kunnen concurreren met andere bedrijven zoals hierboven al is beschreven, levert een CAO hen ook kostenbesparingen op in de vorm van schaalvoordelen. Van alle respondenten reageerde enkel LogicaCMG met een vermoeden dat vakbonden, als derde partij, de onderhandelingskosten zouden vergroten. Getronics heeft zelfs aangegeven dat men een CAO samen met een aantal andere bedrijven heeft overwogen, omdat dit de schaalvoordelen nog meer zou vergroten. Een andere interessante observatie binnen de efficiency context waren de harmonisatietrajecten

234 Developing Institutions die waren opgestart binnen Getronics en EDS. Deze waren er op gericht om de onderhandelingskosten terug te brengen. Het hanteren van zowel een ondernemingsovereenkomst als een collectieve arbeidsovereenkomst was te duur en men was bezig onderwerpen vanuit de ondernemingsovereenkomst over te hevelen naar de CAO. De derde hypothese stelde in feite de vraag of de CAO niet als anachronisme beschouwd kan worden. Het antwoord hierop luidt negatief. De huidige CAO s in de Nederlandse IT sector zijn geen tijdelijke oplossingen. CAO s bestaan al sinds het ontstaan van de sector. De bedrijven die al sinds hun ontstaan een CAO gebruiken geven aan dat ze er tevreden over zijn en niet van plan zijn om het overleg met de vakbonden stop te zetten. De crisis in de IT sector na het oplossen van de Millennium bug en het instorten van de New Economy heeft er wel voor gezorgd dat de relatie tussen werkgevers en vakbonden onder druk kwam te staan. Atos Origin heeft toen nog de CAO voor zichzelf geëvalueerd en is tot de conclusie gekomen dat het hanteren van een CAO uiteindelijk meer voordelen dan nadelen heeft. Als men de ervaringen van de bonden in de IT sector met de ICK CAO vergelijkt met de redenen waarom bedrijven een CAO zijn gaan hanteren, dan komt er een beeld naar voren dat de vakbonden en de CAO niet meer dan een instrument voor werkgevers zijn. Hetzelfde geldt voor de ondernemingsraden die vaak als schild gebruikt worden door bedrijven om de vakbonden buiten te houden. Diezelfde ondernemingsraden worden tijdens massaontslagonderhandelingen vaak gepasseerd omdat men dan weer bang is dat de partij niet objectief genoeg kan zijn en prefereert men te onderhandelen met de vakbonden. In dit onderzoek is tevens bekeken of er verschillen bestaan tussen CAO s en ondernemingsovereenkomsten op het gebied van flexibiliteit en individuele keuzemogelijkheden. De conclusie hier is dat die verschillen er eenvoudigweg niet zijn. Een verklaring hiervoor kan zijn dat bedrijven elkaar nauwgezet volgen; ook op het gebied van arbeidsvoorwaarden. Dat geldt zeker voor de IT sector waar bedrijven altijd in felle concurrentie met elkaar zijn. Wat betreft de toekomst van CAO s in de IT sector is het meest voor de hand liggende scenario dat het aantal CAO s in de sector zal afnemen, terwijl het aantal mensen dat valt onder een CAO zal toenemen. De reden hiervoor zijn overnames en insourcing. Een goed voorbeeld hiervan is de overname van PinkRoccade door Getronics begin 2005. Beide bedrijven hadden een CAO, zijn samengegaan, waardoor het aantal CAO s is afgenomen, terwijl de hoeveelheid mensen die valt onder een CAO gelijk is gebleven. Het ligt niet voor de hand dat bedrijven die op dit moment geen CAO hanteren dat op korte termijn wel zullen gaan doen, tenzij ze worden overgenomen door een grotere partij. Op dat moment lijkt het verhaal op dat van de overname van Pink Elephant door Roccade in de jaren 90 van de vorige eeuw. Hoewel Capgemini heeft aangegeven dat het inziet dat een CAO voordelen kan bieden bij insourcingprojecten denkt men toch dat de cultuursprong die men moet maken te groot zal zijn. Het zijn echter de vakbonden die tijdens overnames hun kans moeten grijpen. De geschiedenis tot dusverre heeft laten zien dat ze op dit soort momenten het meest succesvol kunnen zijn.

Samenvatting (Summary in Dutch) 235 De CAO s die zijn ontstaan in de Nederlandse IT sector zijn niet van tijdelijke aard. Deze studie heeft laten zien dat CAO s in het Nederlandse systeem van arbeidsverhoudingen een belangrijke positie innemen en dat, naarmate de IT sector meer en meer verstrengeld raakt met de rest van de economie, de CAO een steeds belangrijkere rol voor hen zal spelen als strategisch instrument bij overnames en insourcing activiteiten.

236 Developing Institutions

List of Interviewed People Interviews with union representatives Organization Job Title Date FNV Bondgenoten Ans van Uffelen Mariska Stuivenberg Gerard van Hees Union leader IT industry Union leader Services Union leader Services February 23, 1999 November 5, 2002 November 8, 2002 January 17, 2003 Various email exchanges between 2003-2004 CNV Dienstenbond Maurice Limmen Director IT industry April 7, 2003 Various email exchanges between 2003-2004 Various email exchanges 2004 Siegbert van der Velde Union leader De Unie Babet Hendriks Union leader March 31, 2002 Various email exchanges 2004 Interviews with HRM representatives of IT companies Organization Job Title Date Atos Origin Pascalle Jansen Pascalle Jansen HRM representative HRM manager February 22, 1999 May 27, 2003 Baan Marco-Jan Bax Manager Human Resources June 10, 2003 Capgemini Henk Wesselo Manager People & Professions June 6, 2003 CMG Ellen Bekke HRM representative May 19. 2003 Getronics Lars Kok Theo Kimmel Manager Compensation & Benefits Manager HRM June 12, 2003 September 16, 2004 Raet Arjan de Jong Manager HRM Oktober 2004 (via telephone) Various email exchanges 2004 PinkRoccade Hank Kuit Manager Corporate Human Resources July 8, 2003 Ordina Diederik A. Taselaar Manager HRM May 21, 2003

238 Developing Institutions Interviews with works council members of IT companies Organization Job Title Date Getronics Ernst Kuntz Paul van Bavel Chairman works council Member works council May 20, 2003 Was also present at May 20, 2003 Centric Anonymous Lidwien Meijers CMG Sonja Meijerink Arie van den Berg Torex-Hiscom Louis Dubbeldam Björn Vosters Capgemini Bert Wolthuis Michel Hagenus Member works council Member works council Chairwoman works council Member works council Chairman works council Member works council Chairman works council Member works council 2003 August 5, 2004 (via telephone) PinkRoccade Jan Kees de Jong Member central works council March 20, 2003 Baan Marc Teutelink Member central works council June 10, 2003 Atos Origin Willem van Haasteren Herman Strik EDS Herman Jacobs Reinier de Smidt Rob Korevaar Chairman works council Member works council Member works council Member works council Member works council Ordina Robbert Kuit Member works council May 13, 2003 Interviews with employers associations May 14, 2003 Various email exchanges 2004 July 2004 (via telephone) Various email exchanges 2004 May 12, 2003 Was also present at May 12, 2003 June 3, 2003 Was also present at June 3, 2003 May 15, 2003 Was also present at May 15, 2003 March 19, 2003 Was also present at March 19, 2003 Was also present at March 19, 2003 Organization Job Title Date AWVN Alex Verheijden Manager Business & Societal December 11, 2002 Services Group Werkgeversvereniging ICT Jo Lindelauf Manager February 3, 2003 Other relevant interviews Organization Job Title Date IBS Consist Harrie Gooskens Managing Director (Interviewed as expert of the Dutch IT Industry) Various email exchanges between 2002 and 2004 SOI Sijbe Dijkstra Secretary January 23, 2003 Various email exchanges 2004

References Index 239 Automatiseringsjaarboek. 1981-82. 2 vols. Nijmegen: International Information Research. Automatiseringsjaarboek. 1983-86. 4 vols. Dordrecht/Nijmegen: Kluwer / International Information Research. Automatiseringsjaarboek. 1987-90. 4 vols. Amsterdam: VNU Business Publications. Automatiseringsjaarboek. 1991-94. 4 vols. Breda: EDP publications. Bakels, H.L. 2000. Schets van het Nederlands arbeidsrecht, (16th edition). Worked on by Asscher-Vonk, I.P and W.J.P.M. Fase. Deventer: Kluwer. Beltzer, R.M. 2003. Uitzendrelaties galore. SMA. Januari 2003, 58, 1. Pp. 3-5. Berenschot. 2003. Internet survey beloningen 2003. Bijlsma, A.G.M. and R.H. Van het Kaar. 1999. Ondernemingsraad en arbeidsvoorwaarden. Looise, J.C. and R. van het Kaar (eds.). De volwassen ondernemingsraad. Alphen a/d Rijn: Samsom, pp. 151-175. Blauwhof, Getrud and Irene van der Krol. 2001. Twee werelden, één toekomst: De geschiedenis van PinkRoccade 1950-2000. PinkRoccade. Boes, Andreas and Andrea Baukrowitz. 2002. Arbeitsbeziehungen in der IT-Industrie Erosion oder Innovation der Mitbestimmung? Berlin: Edition Sigma. Brinton, Mary C. and Victor Nee (eds.). 1998. New institutionalism in sociology, New York: Russel Sage Foundation. Bush, Paul D. 1987. The theory of institutional change. The journal of economic issues. Vol. XXI, No. 3, September 1987. Chandler, A.D. 1977. The visible hand: The managerial revolution in American business. Harvard: Harvard University Press. Coase, Ronald. 1937. The Nature of The Firm. Economica, November 1937, pp. 386-495. Reprinted in Williamson, O. and Winter, S., eds., The Nature of the Firm: Origins, Evolution and Development, Oxford University Press, 1991. CBS. 2003. De Digitale Economie 2003. Voorburg: Centraal Bureau voor de Statistiek. CPB. 1997. Challenging neighbours: Rethinking German and Dutch institutions, Berlin: Springer. De Korte, Ton. 2000. De CAO van de toekomst. Tijdschrift voor HRM, Winter 2000, 4, pp. 9-33. Eggertsson, Thráinn. 1990. Economic behaviour and institutions, Cambridge: Cambridge University Press. Elster, Jon. 1983. Explaining Technical Change, Cambridge University Press, Cambridge. Elster, Jon. 1986. Introduction. Rational Choice, edited by Jon Elster, Oxford: pp. 1-33.

240 Developing Institutions Fase, W.J.P.M. 1980. Vijfendertig jaar loonbeleid in Nederland: terugblik en perspectief, Alphen aan den Rijn: Samsom uitgeverij. FNV, CNV, and MHP. 2004. Letter to the Department of Social Affairs titled: Wijzinging avv-beleid, August 4, 2004. Groenewegen, John, Frans Kerstholt and Ad Nagelkerke. 1995. On Integrating New and Old Institutionalism: Douglass North Building Bridges. Journal of Economic Issues, June 1995, 29, 2, pp. 467-476. Heerwegh, Dirk. 2001. Survey-onderzoek middels het Internet: een exploratie van het terrein, Leuven: Onderzoeksverslag van het Departement Sociologie, Afdeling Dataverzameling en analyse van de Katholieke Universiteit Leuven. Hofstede, Geert. 2004. Cultures and Organizations: Software of the Mind. New York: McGraw-Hill. Huizinga, Frederik and Peter Broer. 2004. Wage moderation and labour productivity. CPB Discussion Paper, No 28, March, 2004. Jacobs, A.T.J.M. 1986. Het recht op collectief onderhandelen: in rechtsvergelijkend en Europees perspectief, Alphen aan de Rijn/Brussel: Samsom H.D. Tjeenk Willink. Kahmann, Marcus. 2003. Trade Unions and the Growth of the Information Economy. ETUI/ICFTU. DWP 2003.02.01. Kamerbrieven. 2004. Kiezen voor groei, 15-10-2004. Kamerstukken I. 2000-2001. No. 9. Kerstholt, Frans. 1994. Van een neoklassieke naar een neo-institutionele benadering van historische economische groei: kreeg Douglass North de eerste Nobelprijs voor historische sociologie? Sociale Wetenschappen, 37, 1994, 1, p. 30-56. Kerstholt, Frans, Jan Van Wezel & Adriaan van Liempt. 1999. Institutions, Figurations and Performance: A neo-functionalist approach of labour markets, Utrecht: Lemma. Kleinknecht, Alfred and C.W.M. Naastepad. 2002. Is loonmatiging goed voor de export? Economisch Statistische Berichten 87 (4373), p. 624-626. Knegt, Robert (ed.). 1995. Instituties van de arbeidsmarkt, OSA-voorstudie V43, Den Haag: OSA. Knight, J. 1992. Institutions and Social Conflict, Cambridge: Cambridge University Press. Koning, G.E. 1998. Nieuwe technologieën en werknemersinspraak, Lelystad: Koninklijke Vermande. Korevaar, Kees. 2000. De netwerk-cao. Tijdschrift voor HRM, Winter 2000, 4, pp. 35-53. Kreps, D. M. 1990. Corporate Culture and Economic Theory. Perspectives on Positive Political Economy, ed. by J. E. Alt, and K. A. Shepsle, pp. 90-143. Cambridge University Press, Cambridge. Lambregts, J.W.P., Niehoff, P.E., Poortman, J. & Suijkens, M. 1998. Rapportage onderzoek: bedrijfsgroep IT-ET, Utrecht: Basis & Beleid.

Index 241 Leisink, Peter. 2002. The Information Industry new employment relations? Paper presented at the IREC Conference 2002, 29-31 August, Aalsborg, Denmark. Lievense, Peter. 1998. Samen sta je sterk. Computer Partner, nummer 8. Nagelkerke, Ad & Ton Wilthagen. 2000. Op weg naar een institutioneel mozaïek: de Nederlandse arbeidsverhoudingen aan het begin van de 21 e eeuw. SMA, april 2000, 55, 4, pp. 157-165. Mahony, Michael S. 2002. Software: The Self-Programming Machine. Atsushi Akera and Frederik Nebeker (eds.) From 0 to 1: An Authoritative History of Modern Computing. New York: Oxford University Press, pp. 91-100. Norberg, Arthur L. 2002. The modern Computer Business. Atsushi Akera and Frederik Nebeker (eds.) From 0 to 1: An Authoritative History of Modern Computing. New York: Oxford University Press, pp. 163-176. North, Douglass C. 1991a. Institutions. Journal of economic perspectives, Volume 5, Number 1, pp. 97-112. North, Douglass C. 1991b. Towards a theory of institutional change. Quarterly review of economics and business, volume 31, No. 4, winter 1991, pp. 3-11. North, Douglass C. 1992. Institutions, ideology, and economic performance. Cato journal, volume 11, No. 3, winter 1992, pp. 477-488. North, Douglass C. 1993a. What do we Mean by Rationality? Public Choice, 77: pp. 159-162. North, Douglass C. 1993b. Institutions and Credible Commitment. Journal of Institutional and Theoretical Economics, 1: pp. 11-23. North, Douglass C. 1994. Economic Performance Through Time. Nobel prize lecture. December 19, 1993. Denzau, Arthur T & North, Douglass C. 1994. Shared Mental Models: Ideologies and Institutions. Kyklos. Blackwell Publishing, vol. 47(1), pages 3-31. Ordina. 1993. Ordina, 20 jaar een balans. Ordina. OSA. 1995. Instituties van de arbeidsmarkt, OSA-voorstudie V43. Den Haag: OSA. OSA. 1996. Soepel geregeld: Instituties en efficiëntie van de arbeidsmarkt. OSA-rapport 24. Den Haag: OSA. Post, Henk A. 1996. Ongoing innovation The way we built Baan. Barneveld: Baan Business B.V. Rood, M.G. 2000. Introductie in het sociaal recht. Deventer: Gouda Quint. Rojer, Maurice. 2000. De wondere wereld van de CAO. Tijdschrift voor HRM. Winter 2000, 4, pp. 87-113. Rojer, M.F.P. 2002. De betekenis van de CAO en het Algemeen Verbindend Verklaren van CAO s. Ministerie van Sociale Zaken en Werkgelegenheid, Directie Arbeidsverhoudingen, Afdeling Collectieve Arbeidsvoorwaardenvorming en Medezeggenschap, december 2002. Sandler, Todd. 1992. Collective Action: Theory and Applications. New York: Harvester, Wheatsheaf. Simon, Herbert A. 1961. Administrative behavior. 2 nd edition. New York: Macmillan. Slangen, Roger. 2004. Outsourcemarkt wordt volwassen: Dienstverleners staan onder zware Druk. Computable Business review 3 (Q4 2004): 14-19.

242 Developing Institutions Smelser, Neil J. & Richard Swedberg (eds.) 1994. The handbook of economic sociology, Princeton, N.J.: Princeton University Press. Smit, Evert, Keimpe Schilstra, & Jaap Pauwe. 1995. Belangenbehartiging van werknemers: een toekomstverkenning. Rotterdam: Erasmus universiteit, Human Resource Systems Research Unit. Schotter, Andrew R. 1994. Microeconomics: A Modern Approach. Harper Collins College Publishers, New York. Schutte, C.E.M. 1995. Overzicht van het CAO-recht. Nijmegen: Ars Aequi Libri. Schwartz, J. 1993. Functional explanation and metaphysical individualism. Philosophy of Science, 60, pp. 278-301. Steijn, Bram. 2001. Werken in de informatiesamenleving. Assen: Koninklijke Van Gorcum. STZ advies & onderzoek. 1994. Net werk Over werken in de automatisering, Amsterdam. SZW. 2004. Voorjaarsrapportage CAO afspraken 2004. SZW. Juni 2004. Taylor, Michael. 1990. Cooperation and Rationality: Notes on the Collective Active Problem and its Solution. The Limits of Rationality. Edited by K. Cook and M. Levi, Chicago: University of Chicago Press, pp. 222-240. Tijdens, Kea & Carolien van den Brekel. 2000. Werken in de ICT sector: onder welke voorwaarden? Amsterdam: Stichting FNV Pers, Maart 2000. Tros, Frank. 2000a. Schuivende marges: trends in arbeidsvoorwaardenvorming en arbeidsverhoudingen, Alphen a/d Rijn: Samsom/NVP. Tros, Frank. 2000b. CAO: spelverdeler en grensrechter. Tijdschrift voor HRM, Winter 2000, 4, pp. 57-85. Tros, Frank. 2001. Decentraliserende arbeidsverhoudingen: de casus arbeidstijden. SMA, 57, 1, januari 2002, pp.57-70. Tros, Frank. 2003. Werkgevers- en werknemersorganisaties in Nederland. Compendium politiek en samenleving, 43, april 2003. Van Dael, Ruud. 2001. Iets met computers: over beroepsvorming van de informaticus. Delft: Eburon. Van den Brekel, Caroline and Kea Tijdens. 2000. Keuzemogelijkheden in CAO s. AIAS research report nr 3, Juni 2000, Amsterdam: AIAS/UvA. Van der Heijden, P.F. and A.C.B.W. Doup. 1991. Medezeggenschap per convenant, Sinzheimer Cahier 3, Amsterdam: Hugo Sinzheimer Instituut. Van der Meer, Marc and Evert Smit. 2000a. De revitalisering van een aloud instrument: de vernieuwing van CAO s op ondernemingsniveau. Van der Meer, Marc & Evert Smit. Innovatie of Imitatie?: CAO-vernieuwing op ondernemingsniveau. s-gravenhage: Elsevier, pp. 9-23. Van der Meer, Marc and Evert Smit. 2000b. Conclusie: CAO-innovatie en de maatwerkparadox. Van der Meer, Marc & Evert Smit. Innovatie of Imitatie?: CAO-vernieuwing op ondernemingsniveau. s-gravenhage: Elsevier, pp. 121-133.

Index 243 Van Liempt, Adriaan. 1997. Towards an Institutional Explanation of the West: Explanations on the Border Regions of Sociology, the Political Sciences, and Economics. Tilburg University: Master s thesis Sociology. Van Liempt, Adriaan. 2000. Origin: de modernisering van arbeidsvoorwaarden via een CAO. Van der Meer, Marc and Evert Smit (eds.). Innovatie of Imitatie?: CAO-vernieuwing op ondernemingsniveau. s-gravenhage: Elsevier, pp. 101-111. Van Liempt, Adriaan. 2002a. Collectieve contracten in een individualistische ICTsector? Paradox, traditie of anachronisme? Arbeid en ICT in onderzoek, Utrecht: Lemma. Van Liempt, Adriaan. 2002b. The Origin BV Collective Agreement. Van der Meer (ed.). The Trade-Off between Competitiveness and Employment in Collective Bargaining: The National Consultation Process and Four Cases of Company Bargaining in the Netherlands. Working paper No. 10. July 2002. Geneva: International Labour Office. Van Liempt, Adriaan and Ans van Uffelen. 2000. Arbeidsverhoudingen in ontwikkeling: het ambivalente karakter van arbeidsvoorwaardenregelingen in de ICT-sector. SMA, juni 2000, 55, 6, pp. 244-252. Van Liempt, Adriaan and Martijn van Velzen. 2003. Industrial Relations in the Dutch and U.S. IT Industries: Two Systems Moving Apart Together? Paper presented at the 15 th Annual Meeting on Socio-Economics, Aix-en-Provence, June 26-28, 2003. Van Oost, E., G. Alberts, J. Van den Ende and H.W. Lintsen. 1998. De opkomst van de informatietechnologie in Nederland. Stichting Historie der Techniek, Den Haag: Ten Hagen Stam. Van Peijpe, T. 1985. De ontwikkeling van het loonvormingsrecht, Nijmegen: Ars Aequi Libri. Van Velzen, Martijn. 2004. Learning on the fly: Worker Training in the Age of Employment flexibility. PhD study. Veenhuizen, Jan Paul. 2001. PinkRoccade: van twee werelden naar één. Zeggenschap, 12 e jaargang, juli 2001, nr. 2, pp. 29-31. Visser, Jelle. 2000. CAO-vernieuwing bezien vanuit de vakbeweging. Van der Meer, Marc and Evert Smit. Innovatie of Imitatie?: CAO-vernieuwing op ondernemingsniveau. s-gravenhage: Elsevier, pp. 141-149. VNO-NCW. 2004. Nederland moet actiever: Werk(en) in de kenniseconomie 2010. Den Haag: VNO-NCW. Whitley, R. 1999. Divergent capitalism: the social structuring and change of business systems. Oxford: Oxford University Press. Williamson, Oliver E. 1994. Transaction cost economics and organizational theory. Smelser, Neil J. & Richard Swedberg (eds.). The handbook of economic sociology. Princeton, N.J.: Princeton University Press, pp. 77-107. Wilthagen, Ton. 1998. Flexicurity: A New paradigm for Labour market Policy reform? Berlin: Wissenschaftszentrum Berlin Discussion Paper no. FS I 98-202, March 1998.

244 Developing Institutions Wilthagen, Ton, et al. Unpublished. Beyond international competition: Strategies of flexibilization and participation in multinational companies. Yeager, Timothy J. 1999. Institutions, Transition Economies, and Economic development, Westview press.

Index 3COM, 62 ABN AMRO, 135, 137, 139, 140, 141 Agridata, 90, 136, 137 Aino, 89 AINO, 89, 158, 166 Alberts, G., 243 AMD, 59, 62 Apple, 48, 62, 197 Arpanet, 50, 54 ASM Lithography. See ASML ASML, 59 ASP, 64 Atos Origin, 60, 65, 76, 77, 99, 121, 123, 141, 146, 147, 174, 179, 180, 205, 215, 216, 218, 220, 221, 222, 224, 225, 237, 238 ATW (Working Hours Act), 32 Avv, 21, 22, 23, 24, 27, 28, 30, 31, 34, 240 AWGB (equal treatment act), 33 AWV. See AWVN AWVN, 35, 76, 77, 96, 101, 117, 129, 238 Baan, 48, 57, 61, 122, 123, 154, 170, 171, 172, 175, 179, 214, 216, 218, 219, 237, 238, 241 Bakels, H., 239 Baukrowitz, A., 15, 239 BAZIS, 153, 154 BBA, 24 Beltzer, R., 158, 239 Berenschot, 95, 183, 184, 185, 186, 187, 239 Bijlsma, A., 36, 98, 239 Blauwhof, G., 148, 150, 239 Boes, A., 15, 239 BRAVO, 150, 151, 152, 153, 176, 177 Brinton, M., 4, 239 Broer, P., 30, 240 BSO, 53, 91, 99, 141, 142, 147, 148, 167, 180, 205, 221 BTP, 147 Bull, 47, 76, 90, 91, 99, 121 Burroughs, 45, 47 Bush, P., 3, 239 Cap Gemini. See Capgemini Capgemini, 58, 61, 65, 82, 105, 122, 123, 163, 164, 165, 166, 167, 174, 179, 214, 216, 218, 237, 238 CBS, 15, 65, 69, 70, 71, 72, 73, 86, 96, 137, 185, 186, 187, 189, 193, 196, 197, 239 Centric, 61, 122, 123, 124, 167, 168, 169, 176, 177, 178, 179, 214, 216, 218, 238 Chandler, A., 239 Cisco, 55, 59, 62, 119 CMG, 47, 61, 97, 98, 99, 105, 122, 156, 157, 158, 159, 160, 176, 214, 215, 216, 218, 237, 238 CNV, 30, 34, 85, 86, 88, 89, 92, 95, 97, 99, 100, 103, 106, 113, 128, 132, 133, 134, 138, 143, 154, 155, 158, 166, 173, 174, 176, 177, 178, 192, 193, 219, 237, 240 Coalitions, ban on, 18 Coase, C., 9, 239 Collective action, 4, 7, 8, 9, 150 Collective agreement, 123 Collective Agreement Act, 34 Collective bargaining, 2, 13, 22, 23, 24, 31, 39, 86, 87, 153

246 Developing Institutions Compaq, 62, 99 Control Data, 45 Corporate culture, 12, 180, 208, 212, 218 COSSO, 80, 81 CPB, 239, 240 CRM, 63 Data General, 47 Datelnet, 91 Datex, 99, 126 DAVO, 150, 151, 152, 153 De Korte, T., 6, 7, 11, 239 De Unie, 86, 99, 100, 106, 107, 128, 131, 132, 133, 143, 163, 173, 174, 176, 177, 237 Debis, 90 Dell, 62 Denzau, A., 12, 180, 241 Doup, A., 41, 242 Durkheim, E., 4 EAI, 63 EDS, 61, 90, 91, 99, 106, 121, 123, 124, 135, 136, 137, 138, 139, 140, 149, 173, 178, 179, 180, 195, 211, 213, 214, 215, 216, 218, 220, 225, 238 Eggertsson, T., 239 Elster, J., 8, 239 Employers organizations, 5, 35, 75 Employment contract, 30, 32, 33, 34, 41 ENIAC, 44 ERP, 57, 63, 123, 171 ESC (European Social Charter), 31 Fase, W., 21, 23, 37, 239, 240 FENIT, 80, 81, 82, 85, 105, 108, 128 FHI, 84, 108 FNV, 30, 31, 34, 85, 86, 89, 91, 92, 93, 94, 95, 97, 98, 99, 100, 103, 106, 107, 109, 110, 111, 113, 128, 132, 133, 143, 145, 150, 154, 155, 163, 173, 174, 175, 176, 177, 178, 193, 237, 240, 242 Framework agreement, 104, 114, 128, 143, 151 FTU, 70, 71 Getronics, 48, 53, 58, 65, 76, 87, 90, 91, 97, 99, 105, 106, 121, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135, 137, 138, 144, 149, 150, 157, 167, 176, 179, 180, 195, 208, 210, 211, 214, 215, 218, 219, 220, 223, 224, 225, 237, 238 Geveke Electronics, 125 Government, the retreating, 17 Groenewegen, J., 240 HCS, 53, 88 Heerwegh, D., 184, 240 Heidemij, 87, 90 Hewlett-Packard, 47, 48, 99 HIO, 51 Hofstede, G., 211, 240 Honeywell, 45, 47 Huizinga, F., 30, 240 IBM, 44, 45, 47, 48, 51, 61, 62, 63, 70, 76, 88, 99, 106, 128, 149, 164, 197 ICK collective agreement, 75, 77, 80, 81, 82, 83, 84, 101, 102, 104, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 121, 127, 128, 130, 138, 157, 168, 196, 197, 212, 215, 221, 222, 225 ICT Telecom, 83, 118 Iiyama, 62 ILO, 27, 31, 141 Incentive structure, 4, 7, 213 Industrial relations, Dutch system of, 5 Infotheek, 88 Institutional change, 7, 8, 11, 13, 214, 241 Institutional matrix, 3, 5, 7 Institutional theory, 3, 4, 9, 213, 217, 218 Institutions, 1, 2, 3, 4, 5, 6, 7, 8, 9, 11, 12, 13, 30, 153, 211, 214, 215, 218, 225, 226, 227, 239 Collective agreements as, 5 efficiency in, 12 Institutions, definition of, 3, 4

Index 247 Intel, 59, 62 Inter bellum, 17 IOP, 145 IPO, 59 ISIC, 67 isoft, 155 ISP, 64 ISV, 48 ITOP, 99 Jacobs, A., 18, 19, 238, 240 Kahmann, M., 15, 240 Kerstholt, F., 240 Kleinknecht, A., 29, 240 Knegt, R., 240 Knight, J., 2, 3, 7, 8, 9, 13, 218, 240 Koning, G., 240 Korevaar, K., 238, 240 Kreps, D., 12, 180, 240 Krol, I., 148, 150, 239 Leisink, P., 15, 241 Lievense, P., 80, 81, 85, 241 Lintsen, H., 243 LogicaCMG, 123, 155, 156, 158, 159, 160, 161, 176, 179 Lucent, 59, 121, 141 Mahony, 46, 241 MAI Systems Corporation, 47 Mainframe, 46, 47, 52, 136, 153 Maxtor, 62 MHP, 30, 34, 86, 89, 95, 103, 178, 193, 240 Microsoft, 53, 63, 153, 155 Millennium bug, 56, 58 Minicomputers, 47, 49, 52, 53 Motorola, 59 Naastepad, C., 29, 240 Nagelkerke, A., 6, 240, 241 Nederland-ICT, 78, 82, 83, 84, 110, 111, 118 Nee, V., 4, 239 New economy, 59 Nixdorf, 47 Norberg, 46, 241 Nortel, 59 North, D., 2, 3, 4, 5, 7, 9, 11, 12, 180, 213, 240, 241 OEM, 48 Olson, M., 4, 7, 8, 9 Oracle, 48, 99, 102 Ordina, 61, 99, 122, 123, 126, 161, 162, 163, 173, 179, 214, 216, 218, 237, 238, 241 OS/400, 63 OSA, 20, 21, 22, 23, 24, 25, 26, 28, 29, 240, 241 Part time work, 27, 92, 107, 132, 186, 188 Pauwe, J., 79, 80, 242 PBO, 27 Philips, 47, 51, 62, 91, 141, 142, 144, 168, 180, 205, 221 Pink Elephant, 148, 149, 150, 180, 208 PinkRoccade, 60, 76, 87, 121, 123, 126, 137, 148, 149, 150, 151, 152, 153, 176, 177, 178, 179, 180, 195, 208, 210, 213, 214, 215, 216, 218, 223, 237, 238, 239, 243 Raet, 53, 87, 88, 90, 99, 103, 126, 127, 128, 130, 133, 135, 154, 208, 237 RCC, 76, 90, 99, 129, 148, 149 Remington Rand. See Unisys RMA, 87, 90, 148 Rojer, M., 14, 241 Rood, M., 241 Sandler, T., 7, 241 SAP, 48 Schilstra, K., 79, 80, 242 Schotter, A., 242 Schutte, C., 38, 242 Schwartz, J., 242 Seagate, 62 SER, 10, 25, 26, 27, 28, 98 Shepsle, K., 240 Simon, H., 11, 241 Slangen, R., 53, 241 Smelser, N., 4, 242, 243 Smit, E., 6, 15, 61, 79, 80, 99, 192, 242, 243

248 Developing Institutions Social plan, 134, 152, 159, 161, 163, 225 SOI, 97, 98, 103, 104, 106, 238 Sony, 62 SOP, 99 STAR (Labour Foundation), 10, 24, 25, 26, 27, 28, 40 Steijn, B., 6, 242 Sun, 62, 63 Swedberg, R., 4, 242, 243 SZW, 1, 18, 36, 137, 197, 213, 242 Taylor, M., 7, 242 Taylorism, Post-, 7 Tijdens, K., 15, 94, 242 Torex-Hiscom, 121, 123, 124, 153, 155, 170, 175, 178, 179, 195, 214, 218, 222, 238 Transaction costs, 9, 13, 210, 215 Tros, F., 6, 10, 27, 76, 85, 86, 98, 242 T-Systems, 90, 121 Union members, 22, 86, 87, 89, 93, 94, 98, 99, 100, 106, 127, 133, 134, 143, 147, 151, 153, 159, 161, 167, 184, 193, 194, 217, 219 Unisys, 45, 99 Van Dael, R., 50, 51, 52, 87, 242 Van den Brekel, C., 94, 242 Van den Ende, J., 243 Van der Heijden, P.F., 41, 242 Van der Meer, M., 6, 242, 243 Van het Kaar, R., 29, 36, 98, 239 Van Oost, E., 43, 44, 45, 243 Van Peijpe, T., 19, 243 Van Uffelen, A., 15, 36, 192, 237, 243 Van Velzen, M., 15, 117, 243 Van Wezel, J., 240 VAR, 48 Veenhuizen, J.P., 150, 243 Vereniging ICT Nederland, 78, 81, 82 VERMOCON, 79 VHP, 89, 95, 99, 103, 106, 107, 113, 128, 143, 151, 177 VIFKA, 77, 78, 79, 80, 81, 82, 102, 103, 104, 105, 106, 107, 108, 110, 116, 127, 128 Vifkantec, 82, 83 Visser, J., 243 VNO-NCW, 29, 35, 76, 96, 243 Volmac, 53, 99, 128, 148, 164 VONIT, 81, 82 VVP, 147 Wang, 47, 99, 126, 127, 133, 135 Wassenaar Agreement, 27 WAZ (labour and care Act), 32 WCAO, 20, 23, 36, 37, 38, 42, 140 Werkgeversvereniging ICT, 76, 77, 82, 83, 84, 101, 111, 112, 113, 118, 119, 121, 168, 197, 238 Western Digital, 62 Whitley, R., 243 Williamson, O., 215, 239, 243 Wilthagen, T., 6, 241, 243, 244 WLV (Wage Formation Act), 17, 26, 27 WMM (minimum wage and minimum holiday allowance Act), 32 WOR (Works Council Act), 35, 36, 41, 96, 147, 174, 178 WSOI, 103 Yeager, T., 9, 244