Why do local governments support increased centralization, which may infringe their local autonomy? Local autonomy is a highly valued and promoted feature of European governance (Ladner et al. 2016), and decentralization has gained popularity in both scholarship and policy circles, along with world-wide trends of decentralization initiatives (Hutchcroft 2001). The field of immigrant integration is no exception, where local governance in the integration field has been promoted by politicians, EU institutions and integration scholars (Gebhardt 2016). In the integration literature, increased centralization is often criticized (Hernes 2021), and local autonomy is presented as a solution to ensure a holistic integration process in response to concrete problems (Scholten, Collett and Petrovic 2017; Jørgensen 2012; Dekker, Emilsson et al. 2015; Gebhardt 2016).

However, although it may sound counterintuitive at first glance, local governments themselves may not always prefer a high degree of local autonomy. In this article, using the integration field as a case, I investigate why local governments may support increased centralization. To explore this question, I conduct a comparative policy analysis of Norwegian and Swedish integration policies that specifically address central–local governance issues from 2010 to 2021. In studies of multi-level governance in the integration field, Zapata-Barrero et al. (2017) call for more theoretically relevant cases. Both Norway and Sweden are classified as countries where local autonomy has a particularly strong standing (Sellers and Lidström 2007), making them least likely cases for increased centralization. However, within the ‘Nordic family’ of managerial and institutional design, Sweden belongs to the East Nordic model characterized by having higher degree of institutional autonomy, while Norway belongs to the West Nordic model with more political steering (Öberg and Wockelberg 2016), which could also influence central-local dynamics. Thus, these two countries constitute particularly relevant cases to examine if and why municipalities may support increased centralization.

I start with a review of how central–local governance relations have been studied so far in the integration literature to highlight the research gap I address to fill. Building on the broader central-local governance literature, I present two opposing hypotheses for why municipalities may or may not support increased centralization. I continue with a presentation of the empirical data: a total of 190 municipal consultation responses from five different policy processes in Norway and Sweden from 2010 to 2021. In the empirical analysis, I examine whether or not recent changes in central–local governance in Norway and Sweden after 2010 were supported by the municipal community and synthesize the arguments the different municipalities used when justifying their position. In the conclusion, I discuss an analytical dilemma in the central– local literature by assessing a country’s degree of local autonomy (e.g., through indices): how to assess policy changes that involve increased central steering if such changes are actively supported by local governments themselves?

The study of national integration policies dominated the integration literature in its initial years, while the governance of integration drew less attention. From early 2000, however, the multi- level governance of integration policies received increased academic attention, including the central–local dimension. The exiting integration research has included different approaches to studying central–local governance. One strain of literature analyzes whether and how local governments – and particularly large cities – develop their own integration policies. Such studies show that local governments are not merely implementers of national integration policies, but increasingly formulate their own integration policies, which may diverge from the national integration policies and narratives of immigration and integration (Dekker et al. 2015: 2; Jørgensen 2014; Emilsson 2016; Penninx et al. 2017; Alexander 2003; Jørgensen 2012). A second strain of literature explores how local governments utilize the autonomy they have within national frameworks when implementing integration policies, focusing on e.g. implementation of policy tools such as meta-governance and network formation (Qvist 2016, 2017), cross-sector collaboration (Zapata and Compos 2023) and social procurement (Ek Österberg & Zapata 2023), and how local caseworkers negotiate tensions in national policies trying to combine aims of employment and economic self-sufficiency, and social inclusion and participation in a wider sense (Hagelund & Kavli 2011). The third approach is the integration mainstreaming literature, which includes descriptive analyses of whether or not countries have a state-centric (centralized) or poly-centric (decentralized) governance structure of the integration field (Scholten and van Breugel 2018; Scholten et al. 2017). These studies identify whether or not local governments have responsibilities for developing and implementing integration policies. The fourth approach goes one step further and analyses in depth how central governments regulate the leeway local governments have 1) to develop their own integration policies and 2) in their role as implementers of national policies. Emilsson (2015) studies how central governments can exercise power over local governments by examining how they use different instruments (coercive, remunerative, and normative) in their national integration policies to regulate the local leeway. Gebhardt (2016) takes a similar approach and shows how state-led civic integration programs across Europe may challenge and narrow the role of local governments in the integration process. Summarized, the constantly growing integration literature tackling central- local governance issues contributes with analysis of distribution of responsibilities for integration and different modes of national steering, and how local governments execute their autonomy within these national frameworks.

Lastly, many country and comparative studies that examine the policy development in the integration field sometimes also mention municipal views on policy proposals regulating central-local governance (Brochmann et al. 2010; Hernes 2017, 2021; Borevi & Bengtson 2015). However, there are no studies that investigate specifically how municipalities evaluate proposals of increased centralization, and their rationale for supporting or opposing such changes, a research gap this article seeks to address. Such knowledge is particularly relevant because the integration literature often portrays local autonomy as a mainly positive feature (Hernes 2021), which makes it important to broaden our understanding of the complexity and the dilemmas of the central-local governance questions from the perspective of the local level. This could be done by building on the broader central-local governance literature to develop theoretical hypotheses of why local governments may support or oppose policy proposals of increased centralization.

The central-local governance literature has provided important contributions on how to conceptualize and measure central-local relations. A large body of literature has dominantly focused on the constitutional and financial frameworks for local government, and on the different functions for which the latter are responsible (Hutchcroft 2001; Kjellberg 1995; Sellers and Lidström 2007; Pierre 1990; Goldsmith and Page 2010). But constitutional frameworks provide only a partial picture of local autonomy, and some scholars have argued that we also need to look beyond constitutional and financial frameworks and include more indirect control exercised through the regulation of policies, for example through policy instruments directed at local government (Fleurke and Willemse 2006; Goodman 2002). Further, as many studies concentrate on how to conceptualize and measure central-local local autonomy, there is a call for more studies with an ambition to explain why we may see cases and trends of decentralization or centralization (Blom-Hansen 1999; Ladner, et al. 2016), where municipal support or opposition may be one relevant factor.

Although decentralization has been dominantly promoted as a desired and popular feature (Hutchcroft 2001), central-local governance involves an inherent dilemma balancing different considerations. A paradox of central-local government relations is if the central government decentralizes politically important public services, the more it may be tempted to control local governments in their role as implementers of national policies (Haveri 2015). Kröger (2011, 150) calls this “the quarrelsome relatives’ paradox, where opposing values of local autonomy and central control are in conflict. Existing literature includes thorough discussions on the general dilemmas in central–local governance which expose justifications for both local autonomy and central steering (Hernes 2021; Kjellberg 1995; Ladner et al. 2019), which function as a good point of departure to develop hypotheses for why local governments may oppose or support proposals of increased centralization.

The point of departure – and null hypothesis – is that local governments would oppose increased central steering that would infringe local self-governance and autonomy, because local autonomy is regarded as an important democratic value and ideal in itself (Kjellberg 1995). The null hypothesis is based the assumption and rationale that subordinate units would prefer autonomy over detailed steering and control. More specifically, it presupposes that local governments would prefer having the autonomy to develop integration policies addressing specific needs in their local community without central interference and to adapt national policies to local conditions. Additionally, if the central government were to propose policies that would regulate the local integration process, local governments may oppose central steering regardless of whether or not they actually agreed with such policy changes. In the integration literature, several studies show an incongruence between national and local integration policies (Dekker et al. 2015: 2; Jørgensen 2014; Emilsson 2016; Penninx et al. 2017; Alexander 2003; Jørgensen 2012). Thus, if municipalities adopt an approach to integration that deviates from the national approach, they will naturally oppose central steering that would infringe their leeway to develop and adapt integration policies to local integration philosophies. However, even if municipalities agree with national policy and are prepared to implement them, they may still oppose such central steering because local autonomy is one of the most fundamental features or principles of local government (Ladner and Keuffer 2021).

The opposing hypothesis is that local governments may support increased central steering infringing their self-governance and autonomy, the rationale being that local governments may have a self-interest in supporting increased centralization in certain situations. First, central steering may be a tool to ensure a redistribution of resources and responsibilities among municipalities (Kjellberg 1995). In such cases, central steering may delimit local autonomy for some local governments, while others may increase self-governance through central steering (Hernes 2021). If a municipality believes it would benefit from such a redistribution of responsibilities and resources, it may support increased centralization. Second, while the central government (in unitary states) may mandate local governments to implement national policies and provide specific services, in most countries the central government must finance such imposed responsibilities. Consequently, if local governments have already developed particular policies based on local needs and financed them through local funding – or if they have not implemented certain policies due to lack of funding – they may support centrally managed policies if they entail increased funding for municipal service provision. Third, central governance may involve questions not only about whether local governments are responsible for particular service provisions, but also about whether local service provision is regulated in such a way as to infringe local governments’ leeway for local adaptation (Goldsmith 2010; Hernes 2017). Still, local sectoral interests may welcome increased central regulation to gain more leverage in their local or multi-level struggles to ensure sufficient resources and attention for their own sectoral field. For example, local departments working on integration may support increased regulation of the delivery of local services to refugees (e.g., introduction programs) because statutory regulation may facilitate funding for integration measures locally (Fiva et al. 2014). Finally, in Ladner et al.’s (2019) cross-national measurement of local autonomy, a country’s local autonomy index rises with the number of policy fields for which local governments are responsible. According to the logic in the null hypothesis, one could assume that local governments would want responsibility for as many policy fields as possible in order to achieve more power. However, there are certain tasks they may not want responsibility for, either because they think that other governmental levels are better suited or to avoid the blame game in a potentially contested policy field.

Even though municipal actors may generally advocate for local autonomy as an important principle, the different rationales presented above demonstrate how local governments may have good reason to oppose or to support increased centralization.

To study why municipalities may support centralization, I analyze the municipalities’ position and arguments in five integration policy processes in Norway and Sweden that specifically address questions of central–local governance. I conduct document analyses of the municipalities’ public consultation responses. See Table 1 for a list of the five policy processes included in the empirical analysis, their overall topic and the number of municipal consultation responses. The empirical analysis includes all 190 municipal consultation responses in the five policy processes. Table 1

List and description of analyzed policy processes and number of municipal consultation responses in each process

The municipal responses to the consultation process in these policy processes are highly relevant for the exact purpose of this study (investigating why local governments may support or oppose proposals of increased centralization). First, the two sub-policies of integration policies investigated – settlement policies and integration programs – explicitly involve questions of central–local governance and how different actors prioritize between instruments promoting central coercion as opposed to local autonomy (Hernes 2017; Gebhart 2016). Further, in the five selected policy processes, central-local governance is at the heart of the policy proposal. As many other policy processes that regulate central-local governance in the integration field often also involve a shift towards more restrictive (or liberal) integration polices (Hernes 2021), these dual purposes may complicate analysis of the consultation process and argumentation. However, the policy proposals in this study are mainly about the central-local governance and distribution of responsibilities between public actors, which implies that these topics are – at least among – the main subjects in the public consultations process (and municipal responses). Third, in unitary states (such as Norway and Sweden), the central government may mandate and regulate the implementation of tasks by lower governmental levels. Sweden and Norway both have extensive welfare policies where the local governments have wide responsibility and autonomy for developing and implementing public policies. On local autonomy indices, both countries are classified as countries where local autonomy has a particularly strong standing (Sellers and Lidström, 2007). Thus, increased centralization may be presumed to face stronger opposition in these countries and should be thoroughly and explicitly justified in the policy process. In Scandinavia, public consultations constitute an arena for actors – including municipalities and their interest organizations – to state their opinion about proposed national policy changes. Accordingly, public consultation responses constitute an original and formal source to identify support for and opposition to centralizing policy instruments among the municipalities and the rationales behind different positions.

The analyzed municipal consultation responses are between one and 15 pages long.¹ In Norway, all public consultation responses for the selected policy processes are publicly available on government websites. The same applies in Sweden for the 2016 public consultation process. For the 2010 and 2014 processes, the municipal consultations were obtained after a formal request was made to the Department of Government Archives (Arkiv-och registratursektionen, Regeringskansliet).

The empirical analysis documents the municipalities’ position on specific government proposals and whether they 1) support, 2) oppose or 3) do not refer to them or are unclear about their position. Furthermore, based on the theoretical assumptions underlying the two hypotheses, the actors’ arguments for their position were mapped. Although municipal consultation responses constitute a reliable and original source, it is not always straightforward to determine whether a municipality supports or opposes specific policy changes. In some processes, there is only one concrete central–local regulation up for evaluation, while other policy processes involve various policy proposals. In the latter processes it may be challenging to determine whether a municipality is for or against the overall proposal, since it may support certain elements but oppose others. Thus, in the empirical analysis we document whether the actors explicitly express their position for each change that regulates central–local governance relations. In larger legislative processes involving multiple policy changes, the consultation response does not always include a response to all the proposed elements, and there are examples where the municipality discusses a policy change but does not express its position. In the two latter cases, the municipalities’ responses were classified as “unclear”.

The empirical analysis of policy processes in Sweden and Norway produces two main findings: municipal actors often support measures of increased centralization, and the municipal community is often divided. How do municipal actors justify their support of, and opposition to, proposals for increased centralization? In the following section, I first synthesize the arguments for and against 1) increased centralization of responsibilities and authority, and 2) central steering through closer regulation of municipal service provision.

First, the empirical analyses show that many municipal actors support a transfer of responsibilities and authority to higher levels of government. In the Swedish processes for refugee settlement, supporting municipal actors emphasized that central allocation was necessary to ensure equal distribution of and responsibility for refugees among the municipalities. Additionally, many municipal actors supported a transfer of certain responsibilities from local to either national or regional level because it would make higher governmental levels responsible and accountable for certain parts of the service provision within their sector. In both the Swedish and the Norwegian cases, the municipalities have depended on these higher governmental levels to provide certain integration measures, but often without receiving them. For example, in Norway the county level had already been responsible for the provision of upper secondary education for this group, and in Sweden the Swedish Public Employment Service was mainly responsible for employment measures. In these cases, the municipalities argued that a transfer of responsibilities would be positive because it would make these higher levels of government legally accountable for the respective parts of the service provision in the integration process. Thus, since the integration process is a multi-level process including multiple actors at different governmental levels, the local sectoral interests may welcome higher levels of government being assigned more formal responsibilities and being made accountable for delivering measures within their fields of expertise. These actors also often point to equal service provision as supporting arguments for increased centralization.

Still, several municipal actors do oppose transferring the respective responsibilities for the integration field to higher government levels. These actors argue that a transfer of responsibilities would not be effective because it is the municipalities that possess the experience and competence to create such programs and are therefore best suited to ensure a holistic integration process.

Second, what arguments do municipal actors put forward for and against increased central steering through closer regulation of the municipal service provision? The empirical analysis shows that municipal opponents of increased central regulation emphasize the need for local leeway to adapt national policies to local conditions. For example, in the Norwegian consultation, municipal actors argue that the proposal for closer regulation of the program content would give them less leeway for individual and local adaptation and warn that this may have a negative impact on the program quality. These actors also argue that reducing local leeway may result in fewer opportunities for user participation, emphasizing the refugees’ rights as an argument for maintaining local autonomy. This latter point is in direct opposition to many of the arguments in favor of increased centralization, where municipal actors supporting such measures emphasize that increased central regulation is necessary to ensure the refugees’ legal safeguards by ensuring equal service provision and minimum standards for program quality across all municipalities.

At first glance, and in line with the null hypothesis, it may sound counterintuitive for local governments to support policies that reduce or infringe local autonomy. However, the empirical analysis of policies regulating the central–local governance of integration policies in Norway and Sweden finds (partial) support for the opposing hypothesis, by documenting how some policies of centralization were (at least partially) supported by the municipalities, and that most changes were supported by the majority of municipalities and their national interest organizations.

As local autonomy has a particularly strong standing in Norway and Sweden compared to other countries (Sellers and Lidström, 2007), this finding may be surprising, however, the analysis of the local governments’ argumentation shed light on their rational. Overall, the analysis shows that municipal support for increased centralization may be motivated by municipalities’ self-interest, e.g., either by (re)distributing responsibilities within the municipal community or by making higher levels of government legally accountable. An additional argument used by municipal actors is that increased central steering is necessary to ensure an equal service provision and uphold the target group’s (refugees) legal safeguards against potentially poor local service provision.

In a larger comparative perspective, Norway and Sweden share many cultural, political, and institutional similarities as unitary social-democratic welfare states in Northern Europe, and we do not know whether similar arguments would be used by local governments in other countries with different institutional contexts or central-local governmental relations. Nevertheless, the arguments used align with the different values hypothesized and outlined by Kjellberg (1995) as more generic values for central steering. Looking beyond the Nordic similarities, Öberg and Wockelberg (2016) points out that within the Nordic model, Norway and Sweden are characterized by different institutional heritages when it comes to vertical autonomy and political steering. Nevertheless, despite these differences, the empirical analysis shows that Swedish and Norwegian local governments use the same type of arguments in favor of increased centralization. Thus, as these arguments reflect more general considerations favoring central steering over local autonomy applicable in both institutional contexts, it is plausible to assume that these arguments will also be used in other countries. To test the potential broader (bounded) generalizability of these findings, similar studies in countries that differ in their central-local relations – along with different political and institutional context – is encouraged.

Further, it may be plausible to assume that the policy area under study may also affect the relatively large municipal support and the arguments are used. The case of immigrant integration policy is a highly politicized topic, which may affect the municipal positions and argumentation. For example, the argument of making higher levels of government legally accountable, and thereby supporting the transfer of responsibility from the local to the regional or central level, may perhaps be more commonly used in relation to more politicized policy areas than for policies areas of less complexity or politicization. However, this hypothesis should be tested in future analyses.

The finding that many municipalities actually support increased centralization, raises a methodological question for the general central-local governance literature. In this literature, many scholars attempt to measure the degree of local autonomy cross-nationally (Sellers and Lidström 2007; Ivanyna and Shah 2014; Do Vale 2015; Ladner et al. 2016; Goldsmith 2002). Such indices are essential to investigate broader trends, but it is a complex task. Ladner and Keuffner (2021) provide an exemplary discussion of the challenges involved in creating such indices, and how the theoretical and methodological choices made by researchers may directly affect countries’ scores and classification based on these indices. Furthermore, such indices are descriptive tools of a country’s attributes and do not document or analyze how a country ended up with those attributes; for example, the process leading up to a change in local autonomy score (e.g., that a local government becomes responsible for a new policy field).

In this regard, the findings from this study exemplify the analytical difference between descriptive classifications of central–local policies and structures, and analyses of the processes driving these policies and structures. The empirical analysis shows several examples of increased centralization within the integration field, e.g., policies under which responsibilities are transferred from the local to the regional or central level, or policies delimiting local autonomy in service provision. However, these changes have come about with the partial support of the local governments themselves. This finding poses a challenge for how to use central–local indices when the aim is to analyze (changes in) a country’s central–local governance relations: How should we analytically assess policy changes that involve increased central steering if such changes are supported by local governments themselves? For example, a general principle or criterium in several central–local indices is that local responsibility for a policy field implies a higher “autonomy score” (Sellers and Lidström 2007; Goldsmith 2002; Ladner et al. 2016). As Ladner and Keuffer (2021) write:

According to our understanding of local autonomy, the deciding element of local autonomy is the question of whether the provision of important tasks and services falls to local governments rather than to higher levels of the state. The more it does, the more important they are. (p220)

However, in the case of the Swedish Establishment Act, the majority of municipalities supported the transfer of the overall responsibility of the introduction program from the municipalities to a national agency. Although the local autonomy indices do not cover all policy areas and detailed changes in the countries’ central-local relations, this Swedish empirical case raises a question about the general principle in these indices: that they emphasize that local responsibility for a policy field is a deciding indicator of local autonomy. Ironically, according to this general principle, the Swedish transfer of responsibility would imply that Sweden’s “local autonomy score” would decrease, even though it was supported by the majority of municipalities. To use a hypothetical example to highlight the point: if a unified municipal community initiated and politically pressured the central government to transfer responsibility for a specific policy issue from the local to the national level, this would (ironically) be classified as a decrease in the municipalities’ local autonomy score according to this principle in existing indices. These examples highlight why more studies of whether increased centralization happens with the municipals’ support or opposition (or both) is crucial to make a broader assessment of how to measure developments or trends in local autonomy.

I would like to thank Åsne Danielsen for valuable research assistance, and participants in the NORKOM conference 2021 for constructive comments on earlier drafts. This research has received no external funding.