Implementing a diversity policy through public incentives: Diversity Plans in companies of the Brussels-Capital Region

ABSTRACT The development of diversity policies in the private sector is mostly a result of a voluntary engagement of companies. When interested, they decide about the ways and goals of such policies, implications for human resource management and the inclusion of staff with a migration background. This article explores a specific situation in which public authorities offered support and resources to companies in order to develop diversity policies. I present a first analysis of the implementation dynamics of the Brussels Diversity Plan, a policy instrument launched in 2007 by the Government of the city-region of Brussels. Building on the theoretical tools of the political sociology of public policies, I use qualitative material (interviews and administrative documents) to explore implementation processes in three private companies. I show that public authorities can play a decisive role in developing diversity policies in the private sector through incentives rather than constraint. Yet, the room for manoeuver offered to companies leads to important variations regarding how processes unfold and the results achieved.


I. Introduction
Over the past 15 years, diversity charters, awards and training programmes have spread widely across European countries (Boxenbaum 2006;Bereni 2009;Tatli et al. 2012), far beyond the United States were they had first been conceived in the 1980s (Dobbin 2009). One common feature of these initiatives is a reasoning based on the so-called 'business case for diversity'. According to this approach, employment discrimination is economically irrational (Becker 1957) as companies get competitive advantage from a proper management of diversity (Page 2010). Further, public policies based on moral and legal expectations are criticised for failing to reduce unfairness on the labour market (Thomas 1990). Another similarity of diversity policies is their voluntary character. When companies decide to take action in this domain, they also choose what kind of measures they will undertake (Bereni 2009;Dobbin 2009). To date, research has highlighted several factors favouring diversity instrument, namely workers and jobseekers with a migration backgroundor of foreign origin, to use the regional terminology. The first set of questions addresses preconditions for the adoption of diversity policies and the motivations of public authorities and companies at the various stages of the process. Second, this article investigates how these processes unfold and the extent to which they contribute to transforming these companies with regard to human resource management and the situation of workers with a migration background. After having specified my theoretical and methodological approach in the next section, I will outline characteristic features of the Diversity Plan: its context of emergence, its guidelines and the common approach to implementation. I will then examine three cases of such implementation, which differ both in terms of progress and effects. The final section discusses main results.

II. Theoretical and methodological approach
The Diversity Plan in Brussels can be seen as a policy instrument, that is 'a device (…) both technical and social, that organizes specific social relations between the state and those it is addressed to, according to the representations and meanings it carries' (Lascoumes and Le Galès 2007, 4). Policy instruments form a bridge between decision and implementation, and between the actors involved in both processes. After an instrument has been created, field agents use it and report on how implementation proceeds. If they are not satisfied with the results, policymakers (i.e. politicians and/or public officials) may modify some components of the instrument, or replace it by a new one. It is however worth noting that elaboration and implementation activities often take place at the same time and in parallel. They also unfold according to distinct logics and temporalities, which can lead to a disconnect (Jann and Wegrich 2007). Contrary to what political decision-makers may think or claim, policy instruments are not simply and immediately available to them. Rather, instruments have their own logics that may lead to unexpected effects. In this study, devoted to the implementation of Diversity Plans, I consider field practitioners to play a crucial role in this regard (Lipsky 2010), from the signature of a convention between the Region and the company to the achieved results two years later. Drawing on the historical sociology of Elias (1970), I approach every Diversity Plan as a configuration of interlinked and potentially conflicting actors, in this case representatives of public authorities and of the company concerned. Actors are caught up in reciprocal but often unequal dependency links, whose stake is to determine the content of the plan. All participants have room for manoeuver but they are constrained by their own resources, their understanding of the process and the action of other participants. Following Elias, I also see these plans as historical processes composed of successive steps. If each step can be seen as the product of previous ones, the outcome of the whole process is not determined in advance.
This article presents results of a larger study conducted from 2009 to 2012 in Brussels. It uses qualitative material from interviews with political and administrative actors, and also with employer and union representatives. In addition, I analysed a set of documents, including political speeches, administrative directives and reports, minutes of meetings and press articles. More specifically, the article draws on data collected during the first assessment of the Diversity Plan, a mission I undertook in 2011 on behalf of the regional government (Tandé, Desmarez, and Crosby 2012). I refer here to interviews conducted at the time with professionals involved in this policy instrument: three with consultants of the Diversity Cell that assists companies throughout their plan (collective interviews with four persons, length: three hours each), and one with a member of the office of the regional Minister of employment (length: one hour). My questions focused on the goals of the instrument, its impact on companies, the difficulties encountered and how they were resolved.
The assessment mission allowed me to approach several companies involved in the Diversity Plan, and the analysis presented here is based on interviews with six company officials (Company 1: two human resources officers; Company 2: a human resource officer and a union delegate; Company 3: the director and a union delegate). The questions addressed the context in which each plan was launched, the interactions between regional officials and the company, as well as successes and failures of the process, notably relating to performance and the situation of minority workers after two years. I also use administrative documents referring to the Diversity Plan (operational and legal texts, activity reports of the Diversity Cell), and to the specific project of the three companies examined (initial draft, intermediary and final assessment). The latter were produced jointly by the company and the representatives of the Region.
The three companies examined here were part of a group of seven companies contacted during the assessment mission. Those companies had completed (or almost completed) a Diversity Plan; members of the company accepted to answer my questions (in each case: a manager and a staff member); companies were of different size (SMEs, large enterprises) and operated in different economic sectors. The assessment also aimed to include companies where the plan had produced significant effects and others where the impact had been limited. The three cases discussed in this article showed the most contrasting results.

Context and historical development
As the capital of the Kingdom of Belgium, Brussels hosts major state institutions as well as large private companies and organisations. For a long time, the city has also been a major site of tensions between the two main language communities of the country: the Dutchspeaking and the Francophone Belgians. Between 1970 and 1993, five successive reforms transformed the unitary Belgian state into a federal state. In this process, the Brussels-Capital Region and its institutions were created in 1989. Today, the region is composed of nineteen municipalities, and both regional and local administrations are bilingual (Mabille 2011). Parallel to these political and institutional transformations, significant changes also occurred regarding the composition of the local population. Originating from non-European countries, notably Morocco, Turkey and Congo, migrant workers concentrated in seven municipalities of the agglomeration where the price and quality of housing were the lowest. In these municipalities, the proportion of foreign residents in the population today varies between 30% and 50% (Delwit, Rea, and Swyngedouw 2007;Thys 2009). This situation is also due to the peri-urbanisation process that started in the 1960s, and further increased in the following decade. At the time, a large part of the Belgians with high incomes left the city to settle in the richer communes of the periphery. Finally, the current composition of the Brussels population also reflects the settlement of highly skilled Europeans, who came to work in international institutions, notably those of the European Union and NATO. Today 1.17 million people live in Brussels, just over one tenth of the Belgian population (Hermia 2015).
Over the past decades, Brussels has been one of the richest European regions in terms of gross domestic product, while it also has a very high unemployment rate (around 20%). Today, one third of the inhabitants of Brussels live with an income below the at-riskof-poverty line. This proportion is higher than in the other two Belgian regions (OSSB 2015). In Belgium, jurisdiction over employment and professional training is shared between the federal state, the three regions (Flanders, Wallonia and Brussels) and the three communities (the French-speaking, the Dutch-speaking and the German-speaking communities). From 1989 onwards, the authorities of the new Brussels-Capital Region have considered unemployment a constant challenge and they have taken action to decrease it through the intervention of the public employment agency (ORBEm, called Actiris today). Related to this agency, but partly autonomous, the Brussels-Capital Territorial Employment Pact was established as a policy forum in 1998 with assistance from the European Union. Created to foster economic growth and employment in the region, it gathers representatives of the regional minister of employment, the Brussels employment agency and other regional public bodies. Organisations of employers and workers were also systematically invited, in accordance with traditional Belgian practice in the field of labour and employment (Blaise 2010).
The Territorial Pact was entrusted with the task of preparing a regional anti-discrimination policy (Ringelheim and Van der Plancke 2010), after a study had revealed that jobseekers with a foreign background were regularly exposed to discrimination on the private labour market (Arrijn, Feld, and Nayer 1998). Between 1999 and 2004, several seminars targeted employers' representatives and unions of local economic sectors. Participants were sensitised to the problem of ethno-racial discrimination and the existence of a legal framework prohibiting such practices. In addition, employees of public and private employment agencies received training on psychological and social processes leading to discrimination, practical advice on how to prevent such processes and information on the penalties prescribed by the law. A few years later, however, another study outlined the persistent nature of discrimination in the private labour market (Martens and Ouali 2005). For participants attending the meetings of the Territorial Pact, these results revealed the failure of the measures undertaken in the previous years, and regional policy moved in a different direction. Considered to be insufficient to transform practices of employersif not counterproductive, references to non-discrimination and the legal framework became more discreet. Instead, the regional Minister, the representatives of the employment agency and the Territorial Pact, as well as the social partners decided to focus on voluntary participation of employers, motivated by self-interest, rather than on fear of criminal prosecution. Their new public discourse underlined the commercial benefits of a proper management of diversity, and the regional slogan became: 'Diversity, a real asset for your company'. References to diversity became ubiquitous: all discussion groups, job titles and policy instruments of the Territorial Pact were renamed to reflect this new orientation. And the same applied to the instruments created at the time: the Diversity Charter (2005), the Diversity Plan (2007) and the Diversity Label (2010). Resources allocated to this policy field increased, both in financial and human terms, and efforts were made to increase its public visibility.

The Diversity Plan
The Diversity Plan was launched in 2007. Its objectives and operating modes were defined in a working document, called the Operational framework for Diversity Plans in the Brussels-Capital Region. The management board of the regional employment agency adopted the document at the beginning of the year 2007, and it was then accepted by the regional minister of employment. An ordinance passed by the Brussels Parliament in September 2008 formalised the Diversity Plan. In May 2009, the regional Government adopted an order specifying implementation modalities of the instrument.
As one important feature, the Diversity Plan relies on incentives and persuasion. Private companies located in the Region are invited to engage in the process, and are not obliged. To a large extent, they are free to decide which measures they want to include in their project. Policy documents insist on the flexible dimension of the instrument, which has to be adapted to the situation and the expectations of every company. A mandatory dimension of the instrument relates to the involvement of workers. For a plan to be validated, the company's management must agree with the staff representatives or union delegates to create a structure monitoring the whole process. This support structure must be composed of both management and staff representatives. To motivate companies to engage, the regional government offers fundinga maximum of 10,000 eurosfor each project. To obtain this contribution, companies have to invest the same amount of money themselves. Another incentive is the assistance of a public diversity consultant. In 2005, a team of public consultants was created and integrated into the Territorial Pact. Their task was to implement the new regional instrument, and in particular to advise companies engaging in a Diversity Plan.
Over the two years projected for the realisation of a plan, firms are encouraged to carry out measures aimed at strengthening or adapting their operating modes with regard to diversity. A Diversity Plan may lead to the analysis and improvement of internal recruitment policies, to the development of new recruitment channels, and to the development of co-operational relationships with public and private employment agencies. Other measures aim to clarify job descriptions and to better integrate new recruits. Companies are generally advised to provide trainings in anti-discrimination and diversity management to their staff, and to share experiences, methods and tools in public events organised by the Diversity Cell to encourage other organisations to follow suit.
The Diversity Plan was presented as a way of increasing and promoting diversity among the staff with regard to four categories, referred to as the target groups of the policy. These groups were defined through the criteria of foreign origin, gender, age (younger and older workers) and disability. In the initial Operational framework, the term 'people of foreign origin' referred to every person who was not a national of an EU member state or who had one parent, or two grandparents, who were not citizens of an EU country. However, when the regional parliament and government legislated the Diversity Plan in 2008 and 2009, they dropped the criterion of foreign origin. This can be interpreted as a reaffirmation of the universalistic ideology which characterises most of political discourses and public policies in Brussels (Rea 2007). From then on, six criteria were considered: foreign citizenship, gender, age (distinguishing between younger and older workers), disability and low level of qualification. Several actors meet and work together during elaboration and implementation processes of the Diversity Plans. On the side of the Brussels-Capital Region, these are first the consultants and the coordination assistant of the Diversity Cell. They play a major role at all stages of the implementation process, from the preliminary discussions to the final evaluation of the project. Other actors are those who serve on examination and evaluation bodies, namely the Diversity Committee of the Territorial Pact and the management committee of the Brussels employment agency. These include representatives of regional public bodies involved in the employment field, of the federal anti-discrimination agency (Centre for Equal Opportunities and Opposition to Racism, called Unia today), and also representatives of larger employers' (BECI, UCM, Federgon) and workers' organisations (FGTB, CSC, CGSLB). Draft plans are first submitted to the Diversity Committee, which is expected to prepare the discussions that take place in the employment agency. In any case, it is ultimately the regional Minister of Employment who decides or refuses to sign a convention with a candidate company. On the business side, different actors may be involved in the preparation of the draft and the implementation process. According to how companies are organised, and depending on the interest generated by the project, these actors may be directors, officers of the human resource department, line managers, union representatives and staff members. More occasionally, other external actors take part in the implementation process, notably private consultants recruited for giving training to the staff.
Diversity plans can be seen as processes composed of successive stages. Usually, it is the Diversity Cell that takes the initiative and contacts company officers. Regional consultants inform them of the existence of the instrument, of the resources they may benefit from and of the benefits they can expect from such a project. The main issue at this stage is to know whether the company wants to carry out a Diversity Plan. In many cases, prospecting does not lead to an agreement with the company, at least in the short term. When the firm is interested, a regional consultant helps analyse the staff composition and form a support structure, which will be in charge of monitoring the project. The main issue of this second step is to know what measures the company wants to take and whether they will meet the expectations of the regional examination bodies. The support structure or, when appropriate, the works council of the company, have to formally accept the content of the project. This content is then forwarded to the Diversity Committee of the Territorial Pact, and then to the management committee of the regional employment agency. In some cases, these bodies may ask the company to specify or amend a part of the draft. The regional minister of employment is the last to accept or reject the application. Once the Minister, the regional employment agency and the company have signed an agreement, a new stage begins, during which the planned measures will have to be carried out. After one year, and again after two years, the company will be invited to assess its achievements. The order in which these stages occur as well as their length may vary from company to company.

IV. Three cases of implementation
Company 1: a successful process with limited impact Company 1 is a medium-sized Belgian company in the fast food sector. Selling quality products, its public communication emphasises social and environmental commitment. In 2007, the Diversity Cell contacted its human resource (HR) department, and several meetings were held during the year. A regional diversity consultant and an HR officer of the company discussed the relevance of a Diversity Plan, as well as its potential scope, objectives and timing. They also addressed the ways to gather data regarding the composition of the staff, and which employees would join the support structure. Interactions between these two actors were peaceful and constructive. Officials of Company 1 paid tribute to the work done by the consultant to coordinate this preparatory phase, and they progressively expressed their willingness to engage in the project. Worker representatives were absent during these discussions. The company had neither a trade union delegation nor a works council. An informal personnel committee existed, but its members were not involved. In October 2007, both parties agreed on the content of a draft, which was submitted to the regional Diversity Committee. The latter examined the proposed measures, followed by the management board of the regional employment agency. Both bodies validated the proposal without criticism or any particular suggestion. The regional Minister of employment signed an agreement with the company and the plan officially started. During the two following years, some of the actions were completed, largely due to the perseverance of the regional diversity consultant. Some others were abandoned, mostly due to other priorities and disagreement within the company regarding the pertinence of the project, notably between the headquarters and some franchised restaurants. In November 2009, and despite limitations in terms of actions and effects, the final evaluation reflected the satisfaction of both sides.
With regard to human resources management, the measures aimed mostly at enhancing the existing situation, and did not consider substantial changes. Concerning selection and recruitment, it was decided to provide standardised application and CV forms on the website. The action was completed, but did not become a systematic recruitment tool, since other channels continued to exist. Another measure was to set up a partnership with human resource consulting firms, specialised in the issues of disabled people and children with behaviour disorders. Some partnerships were established, while some others did not work. Regarding staff supervision, the project planned to standardise assessment interviews and to hold a training session on neutral recruitment. Due to other priorities, training sessions were finally postponed after the end of the two years devoted to the plan. As regards internal communication, the handbook for managers was adapted to mention, within a paragraph, the importance of equally welcoming all new employees. An internal newsletter addressed diversity issues, but was not followed by similar initiatives during the two years. It was also planned to address the theme of diversity at internal meetings, which finally did not occur. Regarding external communication, Company 1 reported on its website that it had signed the regional Diversity Charter. It also participated in several public events to express its commitment to diversity. This was done during the plan, but no systematic communication in this domain was developed, and these interventions ceased at the end of the project.
Beyond these measures of general scope, increasing the share of workers of foreign origin in the company was not an aim of the original project. Neither the representatives of the Brussels Region nor those of the company requested to include such goals in the project, as both parties had agreed that the existing staff composition was satisfactory. Absence of persons with a foreign background at the management level was noticed, but did not lead to any particular proposal. After two years, the number and the hierarchical position of workers with a foreign background remained the same. Questioned on this issue in 2011, company officials explained: 'We were already good. Europe, outside Europe, in any direction, that is ok.' The issue of discrimination was not addressed in internal training sessions. When questioned on this point, the same actors argued that their firm had a long-standing commitment in this domain, which seemed to immunise it against such practice: 'No, we didn't talk much about it, because that's very much part of us. (…) Attitude will always be the most important thing' (Project Officer, Company 1, August 2011).
From the beginning, the Diversity Cell was the main promoter of the project. At the time, the Diversity Plan had just been created and the priority was to find private companies willing to test it. Developing the skills of regional consultants was also an urgent necessity, as some of the new recruits were not familiar with non-discrimination and diversity issues. Company 1 appeared to be a good candidate because of its positive image as dynamic and open-minded, and also because of its experience in managing a diverse workforce. Seeing this firm as a model, the regional authorities easily accepted its project. For their part, officials of Company 1 initially had doubts regarding the relevance of the project. They considered their firm already diversified, and did not identify any problem with respect to the target groups mentioned by public authorities: 'What did [the plan] bring? Above all, it made us realise that we were doing this diversity naturally. (…) For us, it is deeply engrained it the culture of the company' (HR Manager, Company 1, August 2011). Their company was also expanding, with a growing number of employees in Belgium and more than ten restaurants in the region. To convince them, the Diversity Cell highlighted the funding and the assistance that would be provided by a consultant throughout the project. In addition, the company already had positive experiences with management of workers with a disability, which had regularly been expressed in public. The Diversity Plan was thus presented as a new step in this perspective. The company was finally convinced by the possibility to participate in public events organised by the Diversity Cell, which could help to promote its commercial offers and to develop a network of partners. Yet, the regional consultant played a crucial role in the implementation of the plan: 'The administrative constraints, honestly, it was ok. [The consultant] wrote everything. We gave her the raw material, and she put it in a suitable form' (Project Officer, Company 1, August 2011). Several measures contained in the plan were pushed to the background by other considerations, and the commercial arguments associated with diversity did not receive strong support. The interest of the company was initially limited and remained so.

Company 2: a deadlock in the implementation
Company 2 is a multinational business in the retail sector. At the end of the 2000s, it owned hundreds of supermarkets and local stores, and employed more than 16,000 people in Belgium. At the time, the company was facing a serious internal crisis. The management launched restructuring processes aimed at closing stores and dismissing hundreds of employees. The unions replied with strikes and demonstrations. In 2006, the Diversity Cell solicited Company 2 to carry out a plan, on the grounds that it had been one of the first to sign the regional Diversity Charter. Regular meetings took place in the following years between a diversity consultant and a project officer of the company.
In March 2009, Company 2 presented a project to the Diversity Committee of the Territorial Pact, and then to the management board of the regional employment agency. The two bodies validated the project and the regional Minister of employment signed a convention shortly after. The plan started, but a new phase of social conflict hampered its implementation. A new restructuring plan envisaged closing several hypermarkets, to modify the branch agreement for the remaining staff and to freeze wages for three years. Unions unanimously rejected this new strategy and spontaneous strikes took place in several stores. In reaction to these new tensions, the regional Minister of employment withdrew support of the Diversity Plan and stopped it 12 months after its start.
The plan that started in early 2009 had modest ambitions and most of the measures were not completed. Regarding recruitment, it was planned to diversify recruitment channels to reach out to job seekers with disabilities or a foreign background. First attempts to establish partnerships with external consulting firms were, however, unfruitful. The plan aimed at training store and HR officers on discrimination and diversity issues. This was the main measure completed in the first months, and it was well received. Yet, the officer in charge of the plan within Company 2 doubted its ability to modify recruitment and management modes: 'How do I convince a store manager to hire a disabled worker? Because he no longer pays a portion of the salary! For some of [the managers], there are values, but the pressure on earnings is such that this is not the priority' (Project Officer, Company 2, August 2011). Moreover, the introduction of a module on non-discrimination and diversity into the internal training programme was abandoned for budgetary reasons. Other completed measures concerned internal and external communication. A special issue of the internal newsletter gathered views of employees on diversity management. Company 2 also published local initiatives and the firm's commitment in this domain on its website. In contrast, modifying the composition of the staff with regard to the target groups defined by the instrument was not a goal. References to the issue of workers with a foreign background led to an outcry among union delegates. From the start, they rejected any analysis of the staff composition from this perspective. In the initial draft, it was only mentioned that the company already had a diverse workforce and therefore no problem in this respect. The few months devoted to the plan did not bring any change on this point. The regional Diversity Plan proved to be too weak as an instrument to transform the norms and practices of Company 2, and it did not survive the conflict.
The Diversity Cell had been particularly interested in involving Company 2 because such a plan would have reached a very large number of workers as the company had more than 3000 employees. At the time, the Diversity Cell, the Diversity Committee and the management board of the employment agency did not perceive the social conflict as a threat for the implementation of a Diversity Plan. On the contrary, this plan was seen as a first step towards the implementation of a diversity policy at the national level. The decision of the regional Minister of employment to stop the plan was therefore a total surprise for themas well as for representatives of Company 2: 'The disruption of the plan strongly discouraged us. The reaction of everyone here was to say: "We didn't think it was that political". (…) It was a cold shower' (Project Officer, Company 2, August 2011). For the Minister, however, there was apparently no other choice: 'In the eyes of the public, how do you explain this? The Minister signed and gave 10,000 euros, and on the other side people are laid off? There is a communication problem that is clearly difficult to manage' (Member of the Minister's office, October 2011).
On the side of Company 2, several HR and communication managers initially saw the plan as a way to reconcile ethical and commercial motivations in their company, where profit logics dominated. In addition, these actors wanted to reassert their influence within the organisation, seeing the plan as a way to re-centralise and harmonise management practices that differed from one store to another. If the project officer praised the work provided by the regional consultant, she also criticised the difficulty of adapting the Brussels instrument to the complexity of the organisation and its conflicts. Furthermore, a gap existed between the project supporters on the one hand, and the directors and trade unions on the other hand. Managers of the company expressed scepticism on the stated benefits of the instrument. They regularly argued that the measures proposed by the Diversity Cell were insufficiently concrete. They finally accepted to engage in the process to take advantage of the funding and communication opportunities. Compared to the annual budget of the company, the amounts proposed were modest but could help to set up training sessions for store managers and HR staff. Taking part in public events was also interesting at a time when the company appeared in the media only on the occasion of strikes and police intervention. The unions also showed limited interest in the regional instrument, that they perceived as a diversion tactic in a context where stores were going to be closed and jobs eliminated. The unions then tried to use the plan to settle disagreements which had little relevance to diversity issues, for instance reinstating workers who had left the company because of sickness. To obtain their agreement, this measure was added to the draft, but without provision for specific funding, which indicated that it did not match the expectations of the regional authorities.

Company 3: reconciling business and justice expectations
Company 3 is the Belgian branch of an international firm in the catering sector. The company supplies food on-board high-speed trains, and transports luggage between airports and train stations. In 2007, when discussions started with the Diversity Cell, Company 3 was facing a serious conflict. Staff members clashed with the management, and employees were also divided along ethnic and political lines. A first group was composed of Brussels inhabitants with a Moroccan background. Recently recruited, they claimed to be discriminated against by a second group, mostly Dutch-speaking Belgians, who had been working in the company for a longer time. While the first were members of the Socialist union FGTB, the others were affiliated with the Flemish and Christian union ACV. Frequent strikes were threatening the survival of the company. In this context, Company 3 officials started negotiating the feasibility of a plan with the Diversity Cell. Spurred on by the management and employees supporting the project, a first draft was presented in July 2008. The Diversity Committee of the Territorial Pact and the management board of the regional employment agency hesitated at first to implement a Diversity Plan in such a context. They imposed several conditions: training the entire staff on nondiscrimination and diversity issues, creating a support structure gathering managers and workers, and analysing the staff composition from the viewpoint of the national origin. After these measures had been carried out, the Diversity Committee and the management board validated a new draft. Following the signature of a convention with the regional Minister of employment, new measures were carried out. Between 2009 and 2011, the plan provided a way out of the conflict in which the company had been engaged for years.
The initial draft did not foresee fundamental transformations of recruitment modes, but some actions were completed in this domain. Vacancies were first advertised internally. A number of employees with a disadvantageous status were integrated into its regular staff. Other actions were abandoned, like posting a standardised CV form on the website of the company. With regard to internal communication, the number of staff briefings increased, as well as the information flow between the management, the works council and the employees. Regarding external communication, the company demonstrated its commitment to diversity on its website and through all communication tools, including post and e-mails. Representatives of the company took part in several seminars organised by the Diversity Cell. Furthermore, the key point of the Diversity Plan was to transform management norms and practices. It is in this area that most actions were completed and most effects achieved. The entire staff was trained on issues of discrimination and diversity, notably the psychological dynamics favouring discrimination and the legal framework prohibiting these practices. Two employees were designated as contact persons to solicit in the case of any suspected unfair treatment. At the end of the plan, workers with a migrant background considered their situation to have improved, both in terms of hierarchical positionpromotions were granted where they had been blocked for yearsand recognition of cultural legitimacy. These workers were given the opportunity to share traditional dishes with colleagues at corporate parties (end-of-year party, Muslim holidays).
Among the 43 firms who started a Diversity Plan between 2007 and 2010, Company 3 was one of the few to directly contact the Diversity Cell with the idea of engaging in the process. The board of directors and some of the employees supported the project, which they saw as a way to save their firm. The major aim of the management was to reduce the conflict and to get economic activity back on track. This implied restoring a serene work climate and reinstating the managers' authority. In this regard, diversity consultants and other regional interlocutors were expected to play a mediation role. To show from the start his willingness to get involved in the process, the director of the company signed the Brussels Diversity Charter in October 2007. The support of the company remained strong during the entire process and the management was finally satisfied: 'Many clichés have fallen. Frustrations have been expressed in both sides. Everyone has learned that there are differences, that we must work together, make concessions and efforts … Gradually, it has revitalized the morale of the people' (Director, Company 3, September 2011).
On the side of the FGTB delegates, the survival of the company was also a priority. Yet, these delegates, some of them with a migrant background, were also demanding better recognition of their competences and qualifications, and more equitable management practices. They considered management modes to be fuzzy, discretionary and unfair. They supported the implementation of all measures and welcomed the results achieved: 'The management acknowledged that there had been discriminations in the past. (…) Following the Diversity Plan, there have been major changes in the methodology of selection and recruitment, but also in promotions, where previously it would have been impossible' (Union delegate, Company 3, September 2011). It should be noted that employees initially accused of discrimination, although opposed to the project, did not actively mobilise against it and accepted the following evolutions.
The attitude of the Brussels authorities was surprising in this case. As a general rule, implementing a growing number of Diversity Plans each year is a major concern. Yet, the high level of conflict within Company 3 led them to question the relevance of the project. The general manager of the employment agency openly expressed his doubts during the preliminary discussions. Employers' and unions' representatives also considered the instrument premature and the diversity consultants insufficiently experienced to face this situation. After having decided to impose a test period, they were progressively pleased by the achievements, and the final evaluation was warmly welcomed. In December 2011, the Minister of Employment awarded the Diversity Label to Company 3.

V. Discussion and conclusion
In the United States, the rise of diversity management in the 1980s was analysed as a defensive strategy of human resource professionals against the attacks of Ronald Reagan's conservative administration (Dobbin 2009). By presenting gender and racial characteristics of workers as a resource for the economic development of firms, rather than an issue of recognition and reparation for historical wrongs, these professionals partially saved Equal Opportunity and Affirmative Action programmes introduced since the 1960s, putting 'old wine in new wineskins' (Kelly and Dobbin 1998). In France, the development of diversity policies in private companies in the early 2000s appeared to reflect an inverse logic: human resource professionals and business clubs mobilised against the new anti-discrimination legislation derived from the European Union law (Bereni 2009). In both cases, private actors mobilised against the state, either to preserve existing equality policies or, on the contrary, to avoid having them imposed. By studying creation and implementation processes of the Diversity Plan in three private companies located in the Brussels Region, I explored a scenario still insufficiently addressed in studies on diversity policies: public authorities can play a decisive role in developing such policies in private companies through incentives rather than constraint and in an intensive interaction of public and private actors.
To date, most of the research on diversity policies in organisations focuses on the private sector, while interest in the public sector is more recent. The interweaving of public and private actors in the development of such policies has also received limited attention, while it has become a prominent characteristic of governance and policy instruments in Europe (Lascoumes and Le Galès 2007). The Brussels context has specific features, such as the institutionalised participation of state (here in its regional component), employers and trade union representatives in the field of employment. But it takes place in a wider environment, where diversity charters and labels as well as anti-discrimination trainings spread across European countries as the result of joint actions of public and private actors, notably spurred on by the EU Commission and EU legislation (Ringelheim and Van der Plancke 2010). What lessons can be drawn from the Brussels experience?
First, interventions of the state that use incentives may help initiate diversity policies in companies where such policies do not exist. As it appears in the three examined cases, the government of the Brussels-Capital Region influenced the behaviour of private companies, at least as regards the implementation of a plan, by offering them financial and human resources, and by working actively to convince them to take part -notably through arguments about the benefits they could expect from it. This approach allowed for a number of measures these companies had not undertaken so far. Firms were not reacting to the threat of sanctions based on anti-discrimination legislation, even if the latter exists in the Brussels and Belgian context. Brussels officialsdiversity consultants in particulartook care to distinguish Diversity Plans from judicial procedures and their own work from that of the labour inspectorate. References to discrimination largely remained discrete. As distinct from the United States (Dobbin, Kim, and Kalev 2011) the external influence of industry peers could not be identified as a crucial factor motivating businesses to implement diversity programmes. In Brussels, external influence also played a crucial role but this influence was exerted by the regional state.
Second, the Diversity Plans originating from a tripartite diversity policy are very similar to policies going back to solely business initiative. The main peculiarity of the Brussels Diversity Plan is its procedural framework, which partially guides and restricts the choice of the businesses involved. As a consequence, the most homogeneous element of different cases of implementation is the compliance with the administrative procedure, from the signature of a convention to the final evaluation. Otherwise, however, the Brussels Region promotes a diversity policy that closely resembles those born in the private sector. Public authorities use the arguments of the 'business case' instead of references to morality and respect for the law. The Diversity Plan is based on the commitment of private companies, since the Brussels authorities decided not to compel or sanction companies that refused to engage in the process. It mostly involves convincing companies that they can produce and sell goods and services even better with a new state of mind. Here, the state as well as employers' and trade union organisations associated with the regional policy play the same role as human resources experts who contributed to the rise of diversity management in the United States.
Third, and paradoxically, diversity policies involving public and private actors share another feature with those of the private sector: their achievements vary widely (Doytcheva and Caradec 2008). In other words, diversity policies are characterised by their own diversity. Precisely because they are based on a fluctuating conceptual frame and provide room for manoeuvre, diversity policies rely fundamentally on the commitment of actors, especially on what they expect from them (or, reversely, do not). Careful observation of the ways in which Diversity Plans were elaborated and implemented revealed significant differences between the cases. Some measures proposed in the initial programmes were carried out and then integrated into the day-to-day operation of the company. But others were completed only once, without changing practices and procedures in place before the start of the plan. Moreover, even when companies initially committed to implementing some actions, they did not always do so. Public authorities then only had the option of withdrawing the funds originally earmarked for these actions. The impact on human resources management also differed. If practices and procedures were informal and disputed, the plan introduced new ways of acting. But the impact was more limited in big companies where human resource management was already structured. Regarding the situation of workers of foreign origin, when this was closely examined in the analysis of the staff composition and during trainings provided to the staff, it helped to end discrimination practices when they had been reported as an existing problem. Yet, although this category of workers was considered a target group of the policy instrument, it was not always central. Some managers celebrated diversity as a distinctive dimension of their firm, claiming it was fully integrated into management modes, but did little more than produce this rather general discourse. In the event of a labour dispute, union delegates also refused to engage with migration-related inequalities if they perceived the Diversity Plan (and the notion of diversity itself) as an attempt to divide the employees.
The analysis of the Brussels Diversity Plan illustrates that organisations respond to diversity, and diversity policies, but in different ways. As already observed (Ahmed 2007b), organisations may respond by aiming to conserve, and not to transform, operational modes or the staff composition. Thus, it was not throughout a serious aim to change the position of workers with a foreign background in any substantial way. Main actors had other priorities: the public authorities aimed to multiply Diversity Plans in the region and private companies often aimed to access public resources. Not entirely surprisingly, measures benefitting those of immigrant backgrounds may depend on the involvement of minority workers in the definition and implementation of such policies: when they mobilise, their situation in the organisation is likely to improve. Yet, due to the multiple paths that diversity policies are likely to follow, it surely remains necessary to continue observing what actors do and how they engage in such processes.