Local Governments’ Response to Discrimination: A Temporal Framework to Analyze Local Anti-discrimination Policy and Actions

Abstract

Discrimination is widely studied, with extensive research measuring discrimination on the housing and labor markets. This study examines how local governments address this well-documented issue, by conducting content analysis on 45 policy documents and by performing semi-structured in-depth interviews with 24 alder(wo)men and diversity officers across nine Belgian cities. We introduce a temporal framework combining why, what, how, and when local anti-discrimination policy and actions are established. Such a framework is useful, as we do not approach policy as fixed, but pay attention to how actions evolve over time, even within one so-called anti-discrimination policy.This enables scholars and policymakers to identify decision-making patterns, predict changes over time, and understand contextual influences. Besides, unlike existing models rooted in integration or diversity policy, our framework captures the unique aspects of anti-discrimination policy, enabling a thorough understanding of the (non-)adoption of concrete anti-discrimination actions.

Introduction

Discrimination is a well-documented phenomenon, with an increasing amount of studies measuring discrimination in an objective way by means of discrimination tests on, among others, the housing (Auspurg et al. 2019; Flage 2018) and labor market (Zschirnt and Ruedin 2016; Heath and Di Stasio 2019; Quillian, Lee, and Honoré 2020). The mere perception of a name as belonging to an ethnic minority group appears to be sufficient to deny or diminish (housing) opportunities (Hogan and Berry 2011). This study emanates from the question of how local governments act on this highly documented fact and proposes a temporal framework to analyze local anti-discrimination policy and related actions. Drawing on the analysis of policy documents and semi-structured in-depth interviews, the temporal framework combines why, what, how, and when anti-discrimination policy and related actions are established. This type of framework is useful on multiple grounds. By bringing together different theoretical dimensions, we come to four types of what anti-discrimination actions can look like. By linking these types to why, how, and when anti-discrimination policies are established, this framework can be used to assess, understand, and predict different phases of policymaking. The latter also implies that we do not approach policy as fixed—by analyzing solely what type of actions emerge—but pay attention to how actions evolve, even within one so-called anti-discrimination policy. Besides, by considering cities with differing contexts in terms of dominant political color, broader policy discourse, and ethnic and socioeconomic composition, we can analyze to what extent these contextual elements matter for the (non-)adoption of anti-discrimination policy. Lastly, by applying the framework to different localities, these can be compared to each other, which could be used as a starting point for inter-government exchange on the side of policymakers or for inter-context comparison analyses on the side of researchers.

This study responds to four concrete research gaps: First and most important, existing frameworks do not pay attention to how policy and related actions evolve over time—even within one so-called policy paradigm or frame. Often, anti-discrimination—or wider integration or diversity—policy is approached as fixed. Hence, with the present study, we aim to contribute to the understanding of how policies evolve and/or change. Second, we do so by focusing on anti-discrimination policy and related actions. The existing frameworks are established in a broad setting, namely, integration or diversity policy (e.g. Adam 2013; Dekker et al. 2015; Flamant 2020). Few studies have focused on anti-discrimination policy (-actions) specifically, ¹ which is surprising given the extensive objective measurements of discrimination (Auspurg, Schneck, and Hinz 2019; Flage 2018; Heath and Di Stasio 2019). Third, our focus is on local-level policymaking. Integration and diversity policies were often studied at the national level. Until recently, the local level was generally considered to reflect decisions made at higher levels (Bullock, Wilk, and Lamb 2015; Schiller 2015; Borkert and Caponio 2010). Nonetheless, research has increasingly proclaimed the gravity of localities, characterized by “the local turn” (Bereni et al. 2020; Schiller 2017; Scholten 2013; Zapata-Barrero et al. 2017). Because ethnic minorities are part of the social landscape of cities, local policymakers are considered as being closer to social reality than higher levels counterparts and hence better suited to respond to their needs (Penninx et al. 2004; Borkert & Caponio 2010; Bullock et al. 2015). Lastly, to grasp the complexity and fluidity of anti-discrimination policy actions, we take multiple dimensions into account that can contribute to explaining why, what, how, and when these policy actions are established. Generally, only one or two dimensions are included in frameworks or typologies (e.g., Adam 2013; Escafré-Dublet et al. 2023; Scholten et al. 2017; Westerveen 2022), although more exist. Hence, until now, a more holistic approach combining more than two dimensions in one framework is missing.

Anti-discrimination Policy in Belgium

In order to address the aim of this study, nine Belgian cities—located on both sides of the linguistic border—are considered. Belgium poses an interesting case to analyze local-level anti-discrimination policymaking for at least two reasons. First, as Belgium has a longstanding migration history (Martiniello & Rea, 2012), its current population is superdiverse in ethnic, religious, and socioeconomic terms (Meissner and Vertovec 2015). Affecting the social landscape of cities and towns, ethnic minorities are confronted with various forms of discrimination in access to primordial services, like the housing and labor market (e.g., Baert 2018; Ghekiere et al. 2023). Second, being a federal state, certain competencies, whereunder integration policy—including anti-discrimination policy—were transferred to lower levels, leading to significant discourse differences between the Northern (Flemish) and Southern (Walloon) parts of the country (Adam 2011, 2013). Hence, the selection of Belgian cities situated in both regions adds strength to our temporal framework, as it indicates the robustness and generalizability of the framework in different contexts.

Where the European Union adopted the Racial Equality Directive in 2000, Belgium established three federal laws: the anti-discrimination law, the anti-racism law, and the gender equality law. In 2008, the Flemish government inaugurated an equal opportunity and equal treatment decree, while the Walloon government introduced a decree against certain forms of discrimination. Both prohibit discrimination on the basis of 20 different characteristics, whereunder race, skin color, gender, and age. ² The focus of Belgian local governments in terms of anti-discrimination is often either formulated very generally or is specifically directed to the housing market. The latter can be explained by (1) the fact that access to housing is seen as primordial for participation in society and (2) the awareness around discrimination in this domain is higher because of objective measurements of discrimination that are strongly mediatized. Nevertheless, the topic of anti-discrimination in Belgium is strongly interwoven with integration policy, being handled by the department of integration and civic integration (“inburgering”), with a strong focus on integrating newcomers and their offspring and promoting social cohesion in general.

Although the above holds for both sides of the linguistic border, both differ in their broader policy discourse. The Flemish community replaced an interventionist multiculturalist approach by interventionist assimilationism, yet keeping successful multicultural policies in place. Opposingly, the French community was characterized by laissez-faire assimilationism. As the competencies were transferred back to the Walloon region in 1993, this policy model is abandoned for the idea that equal opportunities and integration is a problem of Flanders and Brussels, consequently lacking a clear policy frame (Adam 2011, 2013; Martiniello 2012). It is within these differing policy discourses that Westerveen (2022) observed that ethnic inequalities were given different explanations on both sides of the linguistic border: on the Flemish side, individual shortcomings of ethnic minorities (e.g., language skills, educational level, or cultural capital) are viewed as the main reason. This, hence, translates into a non-redistributive, yet color-conscious policy frame (e.g., providing language classes, ethnic entrepreneurship). On the Walloon side, then, socioeconomic inequalities are put forward as the underlying trigger for ethnic inequalities, translating into a redistributive, yet color-blind policy frame (generic policies for example to enhance the competencies of all citizens).

It follows that, although the present study’s focus is on ethnic discrimination, also policies directed at other grounds of discrimination are considered. First, if we would solely focus on policies directed at ethnic discrimination, many actions and policies would be overlooked. Since anti-discrimination is often included in broader diversity policy (Saeys et al. 2019; Schiller 2015, 2017), certain policies and actions are applicable to multiple grounds of discrimination, or one policy might comprise actions directed toward different grounds. Second, and related to the first point, policies and actions might be formulated in very generic terms (Vermeulen and Stotijn 2010; Schiller 2015)—as is the case in Wallonia, targeting the population as a whole. This makes it increasingly difficult to single out one ground of discrimination.

Theoretical Foundation

Studies have already analyzed integration and diversity policies of European cities (Zapata-Barrero, Caponio, and Scholten 2017), but few have focused on anti-discrimination policies specifically. To systematically analyze local anti-discrimination policy and actions, we set out from four big questions generally used to describe integration and diversity policy and apply them to anti-discrimination policy and actions: why are anti-discrimination policy actions implemented, what are anti-discrimination actions, how are these actions implemented, and when are they rolled out in respect to the occurrence of discrimination.

Why Are Anti-discrimination Actions Implemented?

The local turn led to increasing focus on the local level to draw its agenda (Zapata-Barrero, Caponio, and Scholten 2017). Although no research explains why anti-discrimination policy specifically is (not) established, previous research has analyzed the reasons for formulating integration policies in general. Scholten (2013) described three types of agendas that lead to policy attention for a specific topic. First, problem agenda refers to awareness and thus the increasing attention through the occurrence of a specific event. The latter can be level (e.g., local) and context (e.g., a specific city) dependent. Hence, the notion of the occurrence of discrimination in a given context can impact the establishment of policies to counter it. However, the occurrence of other events, like, for instance, the terrorist attacks of 2015 in France, led to a replacement of the anti-discrimination policies by deradicalization policies in that context (Escafré-Dublet, Guiraudon, and Talpin 2023). Second, the political agenda reflects the power distribution in a government, defining the willingness to address a certain problem (Scholten 2013). Although the relevance of political context is contested (Schammann et al. 2021), left-wing governments are found to adopt more targeted integration policies (de Graauw and Vermeulen 2016). In France, which is known to overall follow a color-blind ideology, left-leaning local parties are found to be primordial for the adoption of diversity policy measures (Flamant 2020). Third, every level of government has its own policy responsibilities, labeled as policy agenda (Scholten 2013).

What Anti-discrimination Actions Can Be Put in Place?

First, legal studies often distinguish between compliance instruments, aiming at convincing people to behave desirably through training, information, or raising awareness on the one hand and deterrence instruments, applying a repressive approach through enforcement or punishment on the other (Baldwin & Cave 1999). Hence, compliance instruments can be seen as soft actions , including campaigns, information brochures, and academic or sensibilizing discrimination tests (Verstraete et al. 2017). ³ Opposingly, deterrence policies have a hard nature, such as the development of a complaint procedure. However, only a minority of cases are reported. Except for blatant discrimination, it is often difficult to identify (Abravanel 2002). Besides, even if discrimination is identified, gathering proof poses another obstacle (De Schutter 2003). Also, the belief might exist that filing a complaint is useless or might cause revenge actions (Silver and Danielowski 2019). Therefore, complaint procedures are sometimes combined with the implementation of juridical discrimination tests to gather proof to sanction discrimination (Verhaeghe 2022). Importantly though, deterrence is not limited to sanctioning but also includes enforcement through positive action. One type of action is more effective than the other depending on the actor. Where soft actions have a bigger impact on individuals who discriminate due to a lack of information, hard actions are necessary for actors who are acquainted with the law but discriminate nonetheless (Verstraete et al. 2017).

Second, established policy actions can differ in terms of targeted audience. A distinction can be made between targeted policies—which target specific groups within the population, like ethnic minorities, and generic policies—which are directed to a whole population or area (Alexander 2004). A recurrent critique of multiculturalism is that ethnic differences are highlighted by proclaiming recognition and representation of minorities (De Zwart 2005). However, ethnic groups were targeted in an attempt to close social inequality gaps. This leads to a paradox between efficiency-aimed policies by targeting groups on the one hand and reinforcing social differences on the other. This “dilemma of recognition” can be answered in three ways (De Zwart 2005; Borkert and Caponio 2010; Vermeulen and Stotijn 2010). Accommodation recognizes and targets specific groups who should benefit from the implemented policy. This color-conscious strategy is adopted by multicultural policies. Denial ignores ethnic categories, despite existing social inequalities, aligning with color-blindness, and is found in diversity policies (Schiller 2015). Replacement is an in-between strategy whereby policymakers create new categories to avoid underscoring existing ethnic categories while targeting specific groups nonetheless (De Zwart 2005; Vermeulen and Stotijn 2010).

Nevertheless, this distinction should be nuanced. Certain policymakers feel compelled to adjust policies to the increased complexity following migration and diversity becoming more widespread in countries and areas of social life (Scholten, 2020). Mainstreaming policies tend to be directed to society as a whole (van Breugel & Scholten 2017; Zapata-Barrero & Cantle 2018), yet while establishing rather than ignoring a general awareness of diversity (Scholten et al. 2017; Westerveen & Adam 2019). Besides, diversity is incorporated in different domains rather than being handled as a stand-alone topic (Westerveen and Adam 2019; Scholten 2020). Hence, no predefined actors or departments work exclusively on diversity (Cianetti 2020; Scholten et al. 2017; Westerveen & Adam 2019). Whether mainstreaming applies in practice or if it is a symbolic approach covering for replacement strategies, however, can be questioned (De Zwart, 2005; Scholten et al. 2017).

The above reflects the difficulty to delineate anti-discrimination policies. Both targeted and generic approaches have in common that anti-discrimination is the primary goal, and the actions are implemented to directly address this goal. Hence, both are direct anti-discrimination policy approaches. Opposingly, mainstreaming entails the idea that diversity and followingly discrimination is deeply embedded in social life, leading to the need to incorporate these topics in other domains rather than addressing them in a distinct way (Cianetti 2020; Scholten et al. 2017; Westerveen & Adam 2019). This leads to less focus on diversity and discrimination as such, by including these topics in, for example, the general housing market policy. Consequently, the underlying idea is that by primarily focusing on general housing market (and other) policies, anti-discrimination is also addressed. Hence, this can be seen as a form of indirect anti-discrimination policy approach.

How Are Anti-discrimination Actions Implemented?

Questioning how these actions are implemented, a top-down approach refers to the government implementing policies in the field, whereas a bottom-up approach refers to the field co-shaping policies with the government (Vermeulen & Stotijn 2010). This distinction relates to the level of state-society cooperation (Zapata-Barrero, Caponio, and Scholten 2017) and hence to governance, whereby the role of civil society is relevant both in decision-making and implementation (Caponio 2010). The form of governance that localities apply depends on the power distribution between the local government and civil society, as it portrays the level of freedom and input civil society has in policymaking (Mahning 2004; Schiller 2018). In some cases, the so-called “crowding out effect” occurs, whereby responsibilities are delegated to civil society (Borkert & Caponio 2010; Cianetti 2020). This form of laissez-faire is opposed to interventionism, where the local government actively takes up a central role (Cianetti 2020; Adam 2013). Crowding out or laissez-faire in general can be a deliberate choice. The most obvious reason is the neoliberal idea supporting free market forces (Cianetti 2020; Kymlicka 2010; Penninx et al. 2004). Nevertheless, the motive may come from the color-blind ideology, whereby the state does not want to favor individuals or groups to the detriment of others, despite existing social inequalities (Vermeulen and Stotijn 2010; De Zwart 2005). This reasoning might lead to the deliberate choice of doing nothing (Westerveen and Adam 2019). Besides cooperation with civil society, local-level policymakers also interact with higher levels or with other local governments (Dekker et al. 2015; Zapata-Barrero et al. 2017). An increasing number of European cities share their views on policymaking and their experiences through city networks and projects (Schiller 2015; Alexander 2004). The European Coalition of Cities Against Racism (ECCAR)—founded by UNESCO in 2004—is an example where different cities work together within the framework of an action plan consisting of 10 points aimed at reducing racism, discrimination, and xenophobia at the local level (ECCAR 2021).

When Are Anti-discrimination Policy Actions Rolled Out?

Having discussed why actions are rolled out and what these actions may look like, the logical next question is whether these actions are implemented proactively or reactively with respect to the occurrence of discrimination. Proactive actions are taken before the unwanted outcome occurs and can vary from soft actions, like information campaigns, to hard actions, like positive action (Cianetti 2020; Verstraete et al. 2017). Also, research-oriented correspondence tests, aimed at gathering information, are a form of proactive action (Freiberg and Squires 2015). Conversely, reactive actions are developed to address the occurrence of an unwanted event. Considering anti-discrimination policy, this translates into the establishment of a complaint procedure or organizations where targets of discrimination can file a complaint. In Belgium, individuals can turn to UNIA, ⁴ the interfederal center for equal opportunities. As mentioned before, only a minority of discrimination cases result in filing a complaint because it is challenging to prove discrimination (Abravanel 2002; Silver and Danielowski 2019; De Schutter 2003). Therefore, juridical discrimination tests could be used as a reactive policy tool. These tests make pursuing discriminatory actors easier, as documentation is gathered and unlawful behavior exposed (Verhaeghe 2022; Yinger 2015; Freiberg and Squires 2015; Boggs, Sellers, and Bendick 1993; Ahmed 2015). The “when” question is inherently related to the dimension of time and is hence central to our study that emanates from the question of how local governments act on the occurrence of discrimination and proposes a temporal framework to analyze local anti-discrimination policy and related actions.

Data and Methodology

The Case(s)

For the present study, we selected nine cities in Belgium: five in Flanders (Gent, Aalst, Genk, Leuven, and Oostende) and four in Wallonia (Namur, Charleroi, Ottignies-Louvain-la-Neuve, and Arlon). Rural areas are not included, as these are more likely to be inactive on themes like integration policies (Schammann et al. 2021) and consequently also on anti-discrimination policies. This is due to the generally lower numbers of ethnic minority inhabitants and the limited resources of rural areas. The cities were chosen based on the political color of the local government and the alderman in charge of the thematic that is closest to (anti-)discrimination (Table 1). According to Schammann et al. (2021), the political context does not determine whether or not cities are active on themes like integration and, by extension, discrimination. In contrast, de Graauw and Vermeulen (2016) show that there is a tendency for left-wing governments to adopt more targeted integration policies. Additionally, we opted for cities with varying numbers of ethnic minority inhabitants. The latter is part of the local voters’ body, possibly impacting cities’ decision-making (de Graauw and Vermeulen 2016). Next to varying political contexts, the different cities vary in terms of ethnic and socioeconomic composition (Table 2). By including cities situated in both regions with different policy discourses (see above), as well as with varying local political contexts, and ethnic and socioeconomic compositions, we can investigate to what extent these contextual elements matter for the (non-)adoption of anti-discrimination policy. Besides, considering diverse contexts adds strength to our temporal framework, as it indicates the robustness and generalizability of the framework in different contexts.

Table 1. Political context of the selected Belgian cities

* The alderman who has a thematic related to discrimination in his/her competencies.

** In this city, the competencies are divided among two Aldermen. They both represent a different center-left or left party.

*** CPAS is the public center for social welfare and provides citizens in (socioeconomic and other) difficulties with assistance.

Table 2. Ethnic and socioeconomic composition of the selected Belgian cities

Source: Statbel 2021 (most recent data)

The Data

The data consist of two forms of primary data: official policy documents and semi-structured in-depth interviews. First, we analyzed official policy documents published by local governments on their websites for the legislature 2019–2024. This consists of policy agreements (“bestuursakkoord”) and policy statements (“bestuursnota”) related to equal opportunities, discrimination, diversity, housing, and employment. Concrete action plans dealing with equal opportunities, diversity, or discrimination as well as general policy declarations published on the website of the local government were also included. In Wallonia, local governments are obliged to establish a transversal strategic program (“Programme stratégique transversal”—PST) as well as a social cohesion plan (“Plan de cohesion sociale”) for every legislature. The latter is a regional tool aimed at safeguarding the access to fundamental rights in a context of increasing precariousness. Hence, the focus is on closing the socioeconomic inequality gap. We included these documents for the legislature of 2019–2024 of the four Walloon cities. Lastly, some cities conducted studies related to anti-discrimination, like correspondence tests on the housing market. If the results are published on the website of the local government or in the form of a research report, they were also added. A total of 45 documents were included.

We conducted content analysis on the policy documents by following an abductive approach (i.e., going back and forth between literature and data). This means that we used a codebook based on the theoretical dimensions described before. Passages that could not be attributed to one of the dimensions were open-coded in order to create a new dimension. We also coded whether the policy actions apply to ethnic discrimination or other grounds of discrimination. The former is described as policies directed at discrimination based on a person’s ethnic or national origin, nationality, religion, culture, or language. The latter was defined as policies directed at discrimination based on a person’s gender, disability, sexual orientation, or other grounds of discrimination.

Second, we conducted semi-structured in-depth interviews with alder(wo)men or cabinet officers active in the cities and working on discrimination or a related topic. Besides, we interviewed diversity officers or officers working for equal opportunities/diversity or related departments. By interviewing both profiles, we include viewpoints of both politicians who are potentially replaced every legislation and civil servants who are generally in service over different legislations and are expected not to be politically driven. A total of 24 respondents were interviewed in the first half of 2023. After transcribing the interviews, we applied open, axial, and selective coding. Again here an abductive approach was followed, whereby axial codes were derived both from the literature and from the interviews.

The interviews were conducted by the first author of this article. Being a young white female researcher with an Italian surname merits some reflection. Overall, having an Italian background did not seem impactful on the tone or content of the conversations. Probably more influential is that certain respondents were familiar with the authors’ work on the topic of (anti)discrimination. This led to a feeling of safety and ease to talk about this sensitive topic in center- or left-wing cities and to a feeling of tension in right-wing cities. The latter was explicitly mentioned by a respondent, saying that “Although we will not necessarily be on the same page on all points, I always find it inspiring to hear ideas and also force myself to formulate my ideas clearly.”

Results

Going back and forth between the literature and the fieldwork, we come to the temporal framework presented in Figure 1, which is a logical sequence that visualizes the relation between why, what, when, and how local anti-discrimination policy and actions are established. We see this as a temporal process in which these dimensions at different points in time can influence the (non-)adoption of different actions.

Figure 1. Temporal framework of local anti-discrimination policies and actions.

Awareness, Willingness, and Capability

When looking at why anti-discrimination policy actions are rolled out, three elements are crucial: the awareness, willingness, and capability of local governments. Every development of anti-discrimination policy starts with an awareness among the local government about the occurrence of discrimination (Table 3). This awareness is raised by information upflow from civil society (like Unia, a human rights organization in Belgium), evidence from discrimination tests in the locality conducted by independent academic researchers or during previous legislations, and the so-called gut feelings of local policymakers. In most cities, alder(wo)men and diversity officers cite that Unia shares (in)formal complaints regarding discrimination with the local administration. This, often combined with their own sense of the problem, leads to increasing awareness about discrimination as a problem in the locality. However, the awareness creation appears to differ between regional, political, and socioeconomic contexts. Although most cities refer to information upflow from civil society, only two cities took the initiative and investigated whether discrimination is a problem in their city. Gent has been conducting discrimination tests on the housing and labor markets since 2015, and Leuven conducted a society and city monitor to identify local issues. Both cities happen to be left-wing and located in the Flemish region. Besides, with, respectively, 38.3% and 49.4% of people who have a higher education degree (Table 2), Gent and Leuven are the two Flemish cities with the highest educational level. Surprisingly, the presence of inhabitants of non-Belgian origin seems less influential in the awareness creation of discrimination.

Table 3. Why do localities implement or not anti-discrimination actions

Once the local government is aware that discrimination is an issue, they must be willing and capable to tackle the problem for an anti-discrimination policy to be formulated. Awareness alone is insufficient for the establishment of a clear policy and actions. Willingness appears to be affected by the problem definition, whether it is a politically sensitive topic and priority, and the perceived responsibilities. Here, the two regions differ substantially. In Wallonia, economic inequalities are generally defined as the main problem, potentially leading to discrimination. Consequently, it is often not a political priority to act on discrimination since available resources are dedicated to economic challenges. Peculiarly, Namur and Charleroi—the Walloon cities most engaged in tackling discrimination—are also the cities with the lowest socioeconomic status, with below national average incomes and above national average unemployed (Table 2). Conversely, discrimination based on several grounds is defined in Flanders. However, depending on the political coalition in the city, discrimination does not always find its way to the political agenda. In the most right-wing case, the political priority is the native population feeling alienated, consequently putting no effort into countering discrimination. Although political color appears relevant, the percentage of inhabitants of non-Belgian origin appears to play only a minor role in our investigated cities.

Additionally, capability reflects the economic and human resources a local government has to engage in anti-discrimination policy. Capabilities are strongly related to willingness, as cities with more concern to act on discrimination will have clear competency descriptions for the employees, a sufficiently large team of employees and diversity officers, and dedicate internal local financial resources to anti-discrimination actions. One element, however, is more ambiguous. In Flanders, the regional government proposes “Plan Living Together” ⁵ (Plan Samenleven), providing funding for 24 concrete actions directed at successfully living together in a diverse society. Since this plan reduces the costs—financial, intellectual, and time-wise—for local governments, it is likely to impact their policies. However, this regional plan only proposes four actions directly targeting discrimination, translating into approximately 10% of the total budget. Consequently, anti-discrimination may end up lower on the political agenda than actions related to integration, like Dutch language courses. This ambiguity is expressed by a department employee of a Flemish city:

“You also notice with the resources that have come from Plan Living Together, that certain actions such as learning opportunities in Dutch, which are less polarizing or less sensitive, receive a much larger budget than actions related to discrimination. And that’s a pity of course.” (Employee at a Flemish local government department)

Types of Local Anti-discrimination Policy

When there are awareness, willingness, and capability to tackle discrimination, concrete actions are established. But what do these actions entail? Our results show that a first distinction can be made between direct and indirect actions, reflecting whether anti-discrimination is the primary or secondary focus. The latter comprises actions directed to inclusion and participation (e.g., buddy projects, language classes) or improving housing supply (e.g., community land trust, taxes on vacant housing). Although these actions are primarily aimed at social inclusion or housing, reducing discrimination is a secondary side-product, as expressed by an alderwoman in a Flemish city:

“By actually implementing the broad housing policy, we want to be able to change something on the supply side that will hopefully change something about discrimination.” (Alderwoman in a Flemish city)

Also, actions directed at creating a cohesive society, intending to foster encounter and respect (e.g., human library, promoting intercultural and interreligious encounters), can be seen as indirect actions, although these are often included in the anti-discrimination or equal opportunity plans. On the one hand, this practice might be seen as intentional ambiguity as discrimination is a sensitive topic, hence addressing discrimination in a fuzzy way by applying less sensitive actions. On the other hand, it may be a form of mainstreaming, whereby discrimination is addressed in a broader sense than topic-specific actions. In contrast, direct actions whereby discrimination is the primary focus are either directed at ethnic discrimination (e.g., skin color pencil collection to sensibilize kids) or other grounds of discrimination (e.g., rainbow flags as sensibilization regarding the LGBTQIA+ community). Nonetheless, the most recurring actions are directed at discrimination in general and can be applied to different grounds (e.g., bystander training, discrimination tests). Interestingly, more left-leaning governed localities use both direct and indirect actions, whereas right-leaning governments tend to prioritize indirect actions. One exception in this regard is Genk, where both indirect and direct actions are deployed. This could potentially be explained by their higher number of inhabitants of non-Belgian origin (46.25%). There is not one city implementing solely direct actions.

When further analyzing direct actions, we come to types of anti-discrimination actions, depending on whether they are soft or hard on the one hand and generic or targeted on the other hand. For the latter, our results show that an additional subdivision is needed, differentiating actions directly targeting potential perpetrators from victims of discrimination. By combining both dimensions, we come to four types of anti-discrimination measures presented in Figure 2. First, generic awareness actions are directed to the whole population and are aimed to raise awareness about the occurrence of discrimination and its unlawfulness. Almost all Flemish cities provide bystander trainings, aimed at teaching potential witnesses (bystanders) of a discriminatory event to react correctly. Second, targeted awareness actions are also soft by focusing on raising awareness on the topic, but do so by targeting potential perpetrators of discrimination with the actions. Both Walloon and Flemish cities consider housing discrimination to be the most urgent issue, leading them to organize training sessions for realtors to remind the latter of the difference between legitimate selection and discrimination. Third, targeted support actions also follow a soft approach, but by targeting potential victims of discrimination. The most frequent examples are campaigns to increase the willingness to file a complaint or having a local contact point where individuals can turn to for support. Fourth, targeted enforcement actions are different from the three other types in their hard nature, focused on deterring unwanted behavior. Here, only two actions come to the fore. The local contact point that receives complaints related to discrimination can follow up on those and sanction the perpetrator. However, as official complaints remain scarce, one city decided to conduct juridical discrimination tests, whereby proof of discrimination is gathered. This makes pursuing perpetrators easier as the discriminatory behavior is exposed. However, this method is only applicable to the housing and labor markets.

Figure 2. Types of direct anti-discrimination actions.

Considering when what type of direct anti-discrimination action is favored over another, we find that local governments always start their anti-discrimination policy by implementing a set of soft actions directed at increasing awareness or providing support (types 1–3). Omitting the localities applying no actions whatsoever, general awareness actions are found in all cities, regardless of their social, economic, or political context (note that the most right-governed cities are the ones without actions). The picture changes with respect to targeted awareness actions: only politically (center-)left cities deploy actions like training sessions for realtors, directly targeting plausible perpetrators. The most recurrent and valued targeted support action in the respondents’ view is the local permanence the interfederal equal opportunities organization has in most cities. Those are present in as good as all Belgian cities, irrespective of the local political coalition as this stems from decisions made on higher levels of policymaking.

This first set of measures (varying from type 1 to 3) is generally followed by a second set of soft actions, being aware that the first set is insufficient. All cities express a preference for soft actions, which can be explained by the underlying political mindset whereby policymakers aim to convince as many people as possible, which is more likely when no sanctions are imposed:

“My objective as a politician, I’m not an activist, I’m a politician, so I want to make sure that I get as many people as possible on board with positive policies. And if you have resistance, then you lose any possible connection to bring people into that narrative.” (Alderwoman of a Flemish city)

A second reason lies in the belief of local governments that they have no enforcement power. Only one Flemish city claims the power of enforcement by conducting juridical discrimination tests on the rental housing market, whereby discriminatory landlords risk legal prosecution. This was applied after having deployed soft actions, which were insufficient to tackle the problem. This city has a longstanding socialist history, already performing actions against discrimination in previous legislatures. The locality has a relatively high socioeconomic status and is multicultural, although it is not the city with the highest percentage inhabitants of non-Belgian descent. This willingness to go a step further is seldomly present in other cities, where they were either no longer willing or were lacking the resources:

“We always say that money is the sinews of war, and this is unfortunately the case. What’s more, we’re in a disgusting economic situation, which means we’ve been able to carry out this study [academic discrimination tests on the housing market], but we can’t do it in 2023. So you have financial limits.” (Cabinet officer in a Walloon city)

Additionally, in most other cities, juridical discrimination tests remain a sensitive topic that is generally left off the table. The left-governed cities in Wallonia refrain from considering enforcement actions at this time, which we can explain by two related elements: (1) the Walloon region has a broader discourse focusing on socioeconomic inequalities, rather than discrimination, leading to (2) only a recent development of anti-discrimination actions by a slowly raising awareness that discrimination also touches this region. As is the case in the Flemish region, local governments start their anti-discrimination policy with soft—awareness focused—actions, before considering going a step further to enforcement. Both left-governed Walloon cities also have another socioeconomic context, with below national average income levels and higher unemployment rates.

Reactive Prevention

The choice for different types of anti-discrimination actions can be linked to when the actions are rolled out respective to the occurrence of discrimination. The preference for a soft approach relates to the preventive nature of the actions. This is as opposed to reactive actions, which are implemented after a discriminatory event. We call them preventive rather than proactive, as the actions are only implemented once the local government is aware that discrimination is an issue in the locality. Yet, the actions are deployed before and to avoid specific events to occur. In a locality, this was illustrated by the term reactive prevention, which the cabinet officer of a Walloon city explained as follows:

“It’s a bit of both [proactive and reactive]. I’d say we put a lot of emphasis on prevention, but that when it’’s already there (…) it’s true that if there wasn’t discrimination somewhere, there wouldn’t necessarily be any need for prevention.” (Policy advisor in a Walloon city)

Considering political, social, and economic differences between the localities, two elements emerge. First, three cities adopt no actions whatsoever, and two of these cities (Aalst and Arlon) are right-wing governed. Even the alder(wo)man working on the topic closest related to diversity and/or discrimination is affiliated with a right-wing party (Table 2). The third city (Ottignies-Louvain-La-Neuve) is a center-governed city, as opposed to the other Walloon cities that are more clearly leaning to the political left. Both Walloon cities adopting no strategies (Arlon and Ottiginies-Louvain-La-Neuve) have rather high socioeconomic status as compared to the other analyzed Walloon cities. Secondly, although all cities—except the three mentioned before—adopt proactive actions, only one (Gent) in the Flemish region also implemented reactive actions.

Top-down or Bottom-up

So far, we discussed why, what, and when anti-discrimination actions are implemented, but not how. Although in most cases the local government intervenes, the three cases mentioned above can be defined as laissez-faire. However, not as a neoliberal policy idea but because their problem definition focuses on economic inequalities (in the Walloon region) or they do not think of discrimination as an issue that needs a reaction (right-wing case in Flanders). Regarding both cities in the Walloon region, this may sound ironic, given that both cities defining the problem in economic terms are also the ones with the highest socioeconomic status but are politically right or center governed. In the other cases, the state intervenes with varying levels of cooperation with—mostly—civil society. State-society cooperation varies from a top-down to a bottom-up approach. In the former, the local government takes up a leading role as opposed to civil society assuming an executive role:

“On all competencies and policies, we very often work together with civil society, in which that the city has a directing role.” (Alderwoman in a Flemish city)

“(…) So yes, inevitably, there comes a time when the city, well, we’re going to take the lead because… I’m not even sure there’s a big explanation behind it, but because the lead has been taken naturally and that’s that.” (Cabinet officer in a Walloon city)

In a bottom-up approach, the local government listens to concerns addressed by civil society and provides them a seat at the decision-making table. This approach was only applied in the two cities with more than 30% inhabitants of non-Belgian origin. These cities also happen to be two cities with a rather low socioeconomic status. Here, a co-creation between government and society is central, such that the implemented actions meet the needs of the field:

“(…) It’s also for us to identify the false good ideas. Because there’s a lot of stuff, well it’s thought higher up. But the thing is that higher up, when they think it, they do it at higher levels, and then it’s too bad for the field.” (Alderwoman in a Walloon city)

Aside from state-society cooperation, a limited level of state-state cooperation was addressed. Besides the vertical cooperation in Flanders between the regional level providing subsidies (through Plan Living Together) to localities, different localities mentioned involvement in the ECCAR as a form of intercity—horizontal—cooperation. ECCAR assemblies are mostly perceived as a formal platform to exchange good practices. Overall, a hesitancy is expressed regarding whether cities can draw much from these assemblies as it takes place on a large scale and is consequently far removed from the specific contexts of cities. Hence, it is mainly seen as a label indicating a city is committed to the issue of discrimination:

“(…) Very meeting-like too. You can’t really do much with that at the local level. I sat there once and didn’t know what I was doing there.” (Policy advisor in a Flemish city)

“It does still seem important to maintain that network. Also because the ECCAR label is not unimportant. ECCAR is also linked to UNESCO, so it is a quality label somewhere.” (Policy advisor in a Flemish city)

Despite their doubts about the effectivity, a consensus seems to exist regarding the positive impact of the obligation of ECCAR members to translate a 10 points plan against discrimination to each local context.

To a Temporal Framework

Combining these results leads to the framework in Figure 1, which is a logical sequence that visualizes the relation between why, what, when, and how local anti-discrimination policy and actions are implemented at the local level. It is a temporal process in which these dimensions at different points in time can influence the (non-)adoption of different actions.

When a local government is aware of discrimination as a problem in the locality and is willing and able to act on it (why), a first set of actions is implemented (what). These actions are general awareness, targeted awareness, or targeted support actions (types 1–3) and have their soft nature in common. They can be developed by following a top-down or bottom-up approach (how). Generally, this is followed by a second set of actions, again ranging from type 1 to 3 (what). Both sets of actions are established with the main focus on prevention (when). After the deployment of a set of actions, an awareness re-evaluation takes place, where the local government analyzes whether the implemented actions were sufficient. If they conclude that the problem is still present, and they are still willing and capable to act on it (why), local governments go over to reactive (when) enforcement actions (what) that are implemented in a top-down way (how). Hence, here action type 4, namely, targeted enforcement actions, is put into place.

The framework can be applied to localities over different contexts (see Appendix Figures II–X). We illustrate this with examples from four cities. In the Walloon city of Namur—discrimination was brought to their attention through information upflow from Unia and academic correspondence tests conducted by an independent researcher. Defining discrimination as a problem and making it a political priority led the local government to implement soft actions in the form of academic and sensibilizing correspondence testing in a top-down way. The idea was to prevent future discriminatory events from occurring. Although the local government evaluated the soft actions as having a positive yet insufficient effect, they argue to have no financial resources left for further actions. The city Arlon, conversely, also refers to information upflow from civil society but defines the problem in socioeconomic terms. Consequently, discrimination is not a political priority, there is no personnel working on the thematic, and no financial resources are allocated. Here, no anti-discrimination actions are deployed. The city of Gent is the only city going to the reactive phase, implementing juridical correspondence testing on the housing market to sanction discriminatory behavior. The city of Genk defines discrimination as a problem following information upflow from Unia. Their main strategy entails engaging with citizens through bottom-up participatory moments to define aspects contributing to successfully living together in a diverse society. However, their approach is limited to soft preventive actions, without considering reactive or hard actions.

The political color of the local government and of the alderman working on diversity related topics seem crucial for the local approach to anti-discrimination policy, both for its (non-)adoption as for the choice to envisage enforcement besides soft actions. The number of inhabitants of non-Belgian origin seems, although to a lesser extent, also an indicator. In a context with a high percentage (>30%) ethnic minorities, a bottom-up approach to shaping soft actions appears more likely. Lastly, also the socioeconomic status of the city might impact whether the problem is defined in economic terms or as discrimination. However, more influential in this regard is the regional discourse, as both high and low socioeconomic status Walloon cities focus on socioeconomic inequalities. Also, a difference is perceived between the Flemish and Walloon region, rather than between high and low socioeconomic status cities over both cities.

Conclusion

Discrimination is a well-documented phenomenon, with an increasing amount of studies measuring discrimination in an objective way by means of discrimination tests on, among others, the housing (Auspurg et al. 2019; Flage 2018) and labor market (Zschirnt and Ruedin 2016; Heath and Di Stasio 2019; Quillian, Lee, and Honoré 2020). However, few studies focused on anti-discrimination policy specifically. The present study departs from the question how local governments act on this highly documented fact and proposes a temporal framework to analyze local anti-discrimination policy and related actions. We applied an abductive approach, going back and forth between the literature and fieldwork. For the latter, nine Belgian localities were considered with varying political and social contexts. Anti-discrimination policy and actions in Belgium are often either formulated very generally or are specifically directed to the housing market. The latter is because access to housing is seen as primordial for participation in society, and the awareness around discrimination is higher in this domain as objective measurements of discrimination are strongly mediatized. Also, in terms of targeted grounds of discrimination, policies are generally framed very broadly. Besides, the topic of anti-discrimination in Belgium is strongly interwoven with integration policy, being handled by the department of integration and civic integration (“inburgering”).

Certain sidenotes should be made when proposing a framework. First, and implicitly addressed above, anti-discrimination policy seems difficult to delineate. We note a distinction between direct actions, where addressing discrimination is the primary goal, and indirect actions, where discrimination is the side-product. In the latter, the main goal is directed toward inclusion and participation in general, improving housing or creating a cohesive society based on encounter and respect. This complexity may be intentional ambiguity, given the sensitive nature of discrimination against the background of increased polarization and the rise of radical-right in Europe (Bergh and Kärnä 2022; Corrochano, Mata López, and Ruiz Rodríguez 2023; Carral, Tuñón, and Elías 2023). However, it may also be an attempt at mainstreaming anti-discrimination policy. More precisely, given the multiple discrimination grounds and the broader contextualization of discrimination (e.g., tight housing market), local governments may opt to implement actions directed to benefit society as a whole (e.g., broad housing policy). These are in turn hoped to reduce discrimination (Scholten 2020; van Breugel & Scholten 2017). For policies to fall under mainstreaming rather than color-blindness, a consciousness regarding this diversity should be upheld (van Breugel and Scholten 2017). However, in practice, mainstreaming can be used as strategic color-blindness (Westerveen and Adam 2019). More precisely, such an attempt at mainstreaming might fall under intentional ambiguity avoiding the topic of discrimination in a context where it is sensitive and politized. Hence, local governments might want to avoid the population to think they are putting significant effort into reducing (ethnic) inequalities.

Second, the impact of the local context cannot be underestimated. As opposed to what Schamman et al. (2021) found, political color matters for the presence or absence of an anti-discrimination policy. In right-wing cities, where discrimination is not defined as problematic, no actions are deployed. However, when looking at the cases that did implement anti-discrimination actions, a consensus exists about the preference for soft over hard actions, regardless of the dominant political color, partially reconciling with Schamman et al.’s (2021) findings. However, only left-wing governed cities are willing to go a step further, to hard actions sanctioning discriminatory behavior. This tendency adds to the finding of De Graauw and Vermeulen (2016) that left-wing governments are more likely to adopt targeted as opposed to general actions. Hence, localities being (center-)left governed are primordial for the establishment of anti-discrimination actions, but they can differ in terms of focus or approach (Flamant 2020). Besides political color, the presence of ethnic minorities is also relevant to the way in which anti-discrimination policy is shaped. In the two cities with the highest percentage of inhabitants of non-Belgian origin (both more than 30%), a bottom-up approach by means of participatory moments with citizens and civil society was prioritized over top-down approaches, as in most other cities. Also, the socioeconomic status of the city might impact whether the problem is defined in economic terms or as discrimination. However, more influential in this regard is the regional discourse, as both socioeconomic advantaged and disadvantaged Walloon cities focus on socioeconomic inequalities. Also, a difference is perceived between the Flemish and Walloon regions, rather than between high and low socioeconomic status cities over both cities. Nevertheless, the cities with the most elaborate anti-discrimination policy tend to have a high socioeconomic status in Flanders and a low socioeconomic status in Wallonia. Although attention was paid to the political, social, and economic context of the localities, further research could assess the impact of context in a quantitative way to unveil statistical relationships.

Besides, the regional (or national) level contributes to what happens locally, leading to significant regional differences in local anti-discrimination policy. The regional government impacts the local level by providing subsidies for specific actions, like Plan Living Together in Flanders. Besides, the regional political vision influences local governments. This is clear in Wallonia, where the regional government defines socioeconomic inequalities as the main concern, obliging local governments to formulate social cohesion plans that aim at closing the socioeconomic inequality gap. As a consequence, discrimination remains mostly off the table. Hence, the localities generally follow a color-blind approach inspired by the regional level (Adam 2011, 2013). The latter finding resembles the situation of color-blind France, where ethnic minorities are never directly targeted by the actions because the main focus lies on socioeconomic inequalities (Escafré-Dublet, Guiraudon, and Talpin 2023). The regional government being more (center-)right in the Flemish region leads to discrimination being a sensitive topic (Adam 2011, 2013). This translates into ambiguous anti-discrimination actions in localities and a stronger focus on integration and diversity policies. Similar findings were also found in France, where a strong focus lies on living together rather than on tackling discrimination directly (Escafré-Dublet and Hamidi 2023). Notwithstanding the relation between regional and local levels, certain cities move away from the regional vision and apply stronger anti-discrimination actions. This process can be seen as a sort of governance decoupling and happens mostly when the locality’s dominant political color is opposed to that of the regional level (Scholten 2013).

Having addressed these elements, our framework is innovative for different reasons. First, by bringing together why, what, how, and when local anti-discrimination policy and actions are implemented, we do not approach policy as fixed but pay attention to how actions evolve over time even within one so-called anti-discrimination policy. This enables scholars and policymakers to identify decision-making patterns and predict how decisions will evolve over time and how they might be influenced by the context. Our framework enables a thorough understanding of the (non-)adoption of concrete anti-discrimination actions. Second, and building further on the previous point, by having brought together different theoretical dimensions, we came to four types of what anti-discrimination actions can look like, differentiating between general awareness, targeted awareness, targeted support, and targeted enforcement actions. By linking these types to why, how, and when anti-discrimination policies are established, this framework can be used to assess, understand, and predict different phases of policymaking. Hence, it can lead to a more thorough, yet simplified, understanding of the course of anti-discrimination policy. Third, our framework applies to anti-discrimination specifically and separates it from integration or diversity policy. Existing frameworks might be argued to be applicable to anti-discrimination policy but were constructed considering integration or diversity policy and hence cannot grasp the specificity of anti-discrimination policy and actions (e.g., Adam 2013; Dekker et al. 2015; Flamant 2020). Lastly, by applying the framework to different localities, these can be compared to each other (as demonstrated in the appendix), which could be used as a starting point for inter-government exchange on the side of policymakers or for inter-context comparison analyses on the side of researchers.