The impact of public discourse on local representatives’ amalgamation preferences

ABSTRACT

Municipal amalgamation reforms have taken place across democratic regimes since the 1960s. While much focus has been devoted to such reforms’ causes and effects, less attention has been on public discourse surrounding territorial consolidation. This study analyses how local issue salience may impact local representatives’ stance on amalgamating their municipality with one or more others. We focus on the 2014–2020 Norwegian Local Government Reform and utilise a broad survey of local representatives. The analyses show that no one issue (local democracy, local belonging, local employment, municipal services and municipal finances) was perceived to have dominated more than others during the reform. And though certain issues, once prominent in local debates, affect the propensity of local representatives to support amalgamation, this relationship tends to differ between the largest and smallest municipalities, and is also conditioned on localisation issues in the (potentially) new municipality.

KEYWORDS:

• Municipal amalgamation
• preference formation
• public discourse
• territorial reform
• local representatives

Introduction

The search for an optimal fit between the scales of government institutions and the scales of the challenges faced by them has been a perennial quest for public administration reformers, scholars, and political thinkers since antiquity (Baldersheim and Rose 2010; Blom-Hansen, Houlberg, and Serritzlew 2014; Denters et al. 2014). Yet, it is questioned whether an optimal configuration of local jurisdictions within a domestic polity can even exist (Reingewertz and Serritzlew 2019). Nevertheless, the desires and rationales among national governments to reduce the number of local governments can be observed across time, countries, and governance systems (Erlingsson, Ödalen, and Wångmar 2020). They commonly rally economic, managerial, and democratic arguments to undertake local amalgamations: economies of scale to ensure lower unit costs of public services, higher quality services resulting from increased competences and specialisation, and larger democratic arenas to encourage greater competitiveness and responsiveness (Tavares 2018).

However, while such benefits may follow amalgamation, they may be accompanied by a potential trade-off. Indeed, the question of local government size is one that can be described as a dilemma between functionality on the one hand, and democratic qualities on the other (Reingewertz and Serritzlew 2019). Restructuring the jurisdiction of local communities may negatively impact democratic qualities such as participation and representativeness of the new municipality (De Vries and Sobis 2014; Zeedan 2017). Preferences for the territorial architecture of government may also be shaped by one’s feelings of community attachment and identity, which may kick in particularly strongly when the territorial boundaries of one’s polity is discussed (Zimmerbauer and Paase 2013).

The process of amalgamating municipalities may be state-directed or based on local initiatives (Strebel 2019). In the case of the latter, where there is an absence of coercion from above, amalgamation processes will depend on the desires and willingness of local representatives to amalgamate with one or more neighbouring units. From a democratic perspective, this willingness will (or should) be anchored in public support, thus ensuring democratic legitimacy for the decision (Denters et al. 2014, 11).

In this study, we explore the way public discourse correlates with local politicians’ support for amalgamation. Our research question asks how the issues related to municipal amalgamation, once salient in local public discourse, may influence local elected officials to support or oppose the decision to amalgamate their municipality. We focus on the 2014–2020 Norwegian Local Government Reform, in which, contrary to most other state-run reform processes, citizens’ opinions were heavily influential in its outcome (Klausen, Rose, and Winsvold 2021). We analyse the degree to which five topics associated with municipal amalgamation (municipal finances, municipal services, local belonging, local democracy, and local employment), once salient in public discourse, influenced local representatives’ amalgamation preferences. While our study is confined to a specific instance of reform, amalgamation reforms typically revolve around abovementioned functionalist and communitarian concerns and are largely independent of specific domestic contexts (Erlingsson, Ödalen, and Wångmar 2020; Tavares 2018). Our contribution to territorial reform literature, and its implications for generalisability, is thus twofold. First, at a general level, our study focuses on the influence of public discourse as a conduit for citizen engagement in shaping preferences of local decision-makers. Second, more specifically, we illuminate a relatively under-examined variant of municipal amalgamation reform driven by voluntary, argumentative processes (Klausen, Rose, and Winsvold 2021, Author Year; Oldervik 2022).

Analysing survey data of elected representatives of Norwegian municipalities, we note three key findings. First, the more a municipality’s finances were emphasised in local public discourse, the less likely the local representatives are to support amalgamation. Conversely, a greater emphasis on local belonging in public debates is associated with a higher likelihood of amalgamation support. These findings remain largely enduring irrespective of the type of municipality in which amalgamation was discussed, though differences can be observed between the largest and smallest, and the most rural and most urban municipalities. Second, amalgamation preferences are generally more positive in municipalities where political and administrative offices would be localised after an amalgamation. Depending on the specific topic under discussion in the municipality, this attitudinal difference is sometimes magnified, but never reduced. Third, while the topic of local belonging is associated with an increased support for amalgamation, this relationship is strongly dependent on whether local democracy was also discussed. This is interesting, as the two topics can be viewed as both pertaining to a communitarian view of local governance (Myksvoll 2021), and thus represent two sides of a similar coin. Our findings, however, indicate that local belonging and democracy are instead relatively independent coins when rallied as arguments in public discourse. These findings contribute to our understanding of how local territorial reforms may be viewed in public discourse, and how this in turn may influence local decision-makers, thereby providing a novel contribution to the complex dynamics involved to explain the success or failure of voluntary municipal amalgamations.

Territorial reforms: benefits and caveats

Since the turn of the century, and particularly since the time around the financial crisis, a wave of municipal amalgamation reforms has been observed both in Europe and globally, such as in Canada, Japan, and New Zealand (Hansen and Kjær 2020; Swianiewicz 2021).

Supporters of amalgamation see the case for increasing the size of local units as rewarding economies of scale (Dollery, Byrnes, and Crase 2007), wherein cost-savings, managerial professionalisation, and improved service quality can be seen as functional benefits of amalgamation (Myksvoll 2021). Although functional benefits of amalgamations are not guaranteed, and may depend on a host of other, sometimes external, factors (Dhimitri 2018; Tavares 2018), support for the functionalist view of governance and expectations of scale economic benefits and management improvements remain enduring motivators to undertake territorial consolidation at the local level (Erlingsson, Ödalen, and Wångmar 2020; Hansen and Kjær 2020; Myksvoll 2023). But local governments are not merely units of public service provision. They also form arenas and communities of democratic participation for their citizens. This communitarian perspective (Myksvoll 2023) focuses not so much on what governments do, but on who gets to exercise governance. Governance thus becomes an expression of sociality rather than functionality (Tatham, Hooghe, and Marks 2021), and, in the discussion of redrawing jurisdictional borders, invokes issues of identity, belonging and democratic self-rule (Myksvoll 2021). The post-functional theory of governance (Hooghe and Marks 2016) addresses the way the communitarian and the functionalist perspectives act as shapers of governance architecture preferences, favouring the former when it comes to territorial questions. In other words, while functionalist issues will shape one’s policy preferences, feelings of community and identity shape one’s polity preferences (Hooghe and Marks 2016, 2).

This is not to say that the functionalist view does not matter to the question of amalgamation. Rather that, as Strebel (2019) argues, while functional issues can act as a ‘push’ towards amalgamation, a ‘pull’ arises from the communal aspect and will even be the greater influential force. Indeed, the attachment or belonging to one’s local, territorially bounded community has been shown to increase aversion towards amalgamation among bureaucrats (Myksvoll, Tatham, and Fimreite 2021), citizens (Stein, Broderstad, and Bjørnå 2022), and local politicians (Myksvoll 2023) alike.

In studying voluntary amalgamation processes, the relevance will not be so much on objective measurements and amalgamation effects, however, but rather on the way amalgamation is presented and debated in public discourse (Strebel 2019). Government responsiveness, through which public discourse can act as a channel of citizen input and influence outside elections, is an integral part of representative democracy (Grimes and Esaiasson 2014; Pitkin 1967). In a democratic system, public discourse, or opinions formed and expressed in the public sphere, should therefore be expected to contribute to decision-makers’ preferences. Public opinion thus relates to the democratic norm of responsiveness, wherein elected officeholders may take note of and incorporate public opinion in public policy formulation and implementation (De Vries and Sobis 2014, 9; Mathisen 2023). Public discourse, through which the public’s opinions are channelled, thus forms an anchor of legitimacy from which local representatives make their decisions.

By seeing the issue of municipal amalgamation processes through the perspective of democratic responsiveness, public opinion formation, and public discourse, we may ask how various issues related to amalgamation, once salient in public debates, influence local representatives’ amalgamation stance. Derived from theoretical and empirical literature on municipal amalgamation – that is, the functional and communitarian perspectives described above, we identify five topics of relevance. These relate to factors that can push towards or pull against support for amalgamation: (1) municipal finances, (2) municipal services, (3) local democracy, (4) local belonging, and (5) local employment. We expect that these issues, once salient topics of public debate in a municipality, affect the degree to which local representatives find amalgamation a desirable outcome, by outlining the following hypothesis:

H₁:

Local politicians’ likelihood of supporting municipal amalgamation is affected by the saliency of functional and communitarian issues in local public discourse.

We avoid outlining a specific correlational direction between public discourse topics and amalgamation preferences because we do not expect this influence to necessarily remain constant across all types of municipalities. Local governments vary in their population size, their geographical extent, their financial situation, and as central or peripheral vis-à-vis other municipalities. As such, the contexts in which a potential reform is undertaken at the local level will vary: a city-municipality with hundreds of thousands of citizens has different challenges, goals, and priorities, than a peripheral municipality with a few thousand citizens (Keating 1995). Hence, local contexts can influence the degree to which various issues are discussed in public debates concerning amalgamation, and consequently, these issues’ impact on the desirability of amalgamation will vary (Leknes, Ridderstrøm, and Rommetvedt 2019; Strebel 2019, 2023). For example, Klausen, Rose, and Winsvold (2021) have studied citizens’ opinion formation and their assessments of the possible consequences of amalgamation. While the public generally perceived or expected no significant changes to the qualities of the municipalities’ functions or their communal/democratic features following a potential amalgamation, these views were found to depend on belonging to a municipality that represents a large or small share of a potentially new unit, indicating a possible fear of centralisation among citizens in smaller municipalities. We therefore expect the influence various issues have on amalgamation support to be dependent on the traits of the municipality, leading us to formulate our second hypothesis:

H₂:

The correlation between a topic related to municipal amalgamation in public discourse and local politicians’ amalgamation preferences is conditional on the type of municipality in which amalgamation is discussed.

The five distinct topics – municipal finances, services, local democracy, belonging, and employment – form our study’s main independent variables. However, we do not expect them to account for all variation in the local representatives’ amalgamation preferences. They are therefore subjected to a range of controls at the individual and the municipal level. At the individual level, we account for the representative’s position in the local political environment as well as their views on citizen participation in local decision-making processes. At the municipal level, we control for, and to test our second hypothesis, further test the possible interaction effects of various municipal characteristics. These include the municipality’s population, geographic, and economic size. We also include the municipality’s degree of centrality, as well as its position as a junior or senior partner in (potential) amalgamation agreements. The dependent and independent variables, as well as the controls, are described in further detail in the data and methods section.

The Norwegian local government reform

In 2013, the Norwegian government initiated the Local Government Reform. The government advocated ‘larger and more robust units’: larger municipalities were intended to enable ‘better welfare services, more sustainable community development and stronger local self-rule’ (Government manifesto 2013, 6;47).

The Minister for Local Government and Modernisation instructed all municipalities to commence consultations with neighbouring units. While it was obligatory for every municipality to investigate the potential for amalgamation, the central government relied mainly on local willingness and voluntary agreements to achieve amalgamations. The final authority nevertheless rested with the national parliament, and the central government did express willingness to coerce locally disputed amalgamations should it be deemed necessary to retain an overall territorial cohesion (Rec., 2014, p. 42).¹ Thus, while some municipalities were indeed compelled to amalgamate, the reform largely necessitated local willingness (Vabo, Fimreite, and Houlberg 2023).

Intermunicipal consultations, and those which in many cases were formalised into negotiated amalgamation agreements, became subject to local advisory referenda, opinion polling, or other forms of citizen input. Following this, the municipal councils made the decision on whether they desired amalgamation. In June 2017, the parliament voted on the new municipal structure, and on 1 January 2020, most amalgamations took effect.² The reform reduced the number of municipalities from 426 in 2014 to 356 in 2020. Table 1 summarises the main events of the reform.

Table 1. Timeline of the local government reform.

Stage	Event	Period
National	Central government initiates reform	Autumn 2013
	Expert committee lays out criteria for new municipal structure	Spring 2014
	Minister instructs municipalities to commence consultations and negotiations	Summer-Autumn 2014
Local	Municipalities consult and negotiate among potential amalgamation partners	Autumn 2014-Summer 2016
	Local advisory referenda are held	Spring-Summer 2016
	Municipal councils determine whether they want to amalgamate	Summer-Autumn 2016
Regional	County governors review municipal agreements and decisions, proposing a new municipal structure to the government	Autumn 2016
National	Government proposes a new municipal structure to the parliament	Spring 2017
	Parliament votes on amalgamations	Summer 2017
	Amalgamations are implemented, new municipal structure created	January 2020

As it was initiated and finally decided by the national government, but largely framed and formed locally, the processes of amalgamation in the Norwegian reform can be described somewhat paradoxically as a case of centrally dictated local autonomy (Nygård 2021). Indeed, Nygård (2021) finds that the voluntary nature of the reform provided opportunities for shaping the reform’s course and outcome locally. A central element of the reform therefore involved translating national arguments to local contexts (Klausen, Rose, and Winsvold 2021). In this sense, the reform was decidedly argumentative. Local consultations and negotiations, along with public opinion assessing the possible consequences of amalgamation, became essential elements of the reform (Klausen, Rose, and Winsvold 2021; Oldervik 2022).

Data and methods

In this study we utilise a survey sent to elected officials of Norwegian municipalities. The survey includes a range of questions regarding their experiences of and attitudes towards the Local Government Reform. Every representative in every municipality received the survey. The data collection took place between October and November in 2018, placing it in the 2015–2019 local election term. Hence, the respondents were elected representatives of their municipality during the early and local stages of the reform (see Table 1), and experienced local debates and the decision on whether they would seek an amalgamation. After the initial survey distribution, recipients who had not responded were sent up to three reminders. In total, 8450 representatives were contacted, with 3387 recipients completing the survey in full, yielding a response rate of 40.1%.

The dependent variable (the respondent’s amalgamation stance) is coded as a dummy variable. The presence and prevalence of various topics associated with municipal amalgamation in public discourse was gauged through the local representatives’ perspectives. The respondents were asked the following: ‘If you consider the debate surrounding municipal amalgamation in your municipality, how much weight have you experienced was put on the following?’. All five topics were measured on an ordinal scale, where 1 = ‘Very little emphasised’ and 5 = ‘Very much emphasised’.

These five independent variables are subjected to a range of controls, at both individual and municipal scales. At the individual level, drawing on the notion of Miles’ Law (‘where you stand depends on where you sit’) (Miles 1978, 399), which has been found to apply to municipal decision-makers as well as civil servants (Henriksson and Christensen 2023), our individual-level controls include the respondent’s party membership status in terms of party affiliation, and their seniority within that party. We also include their rank within the municipal council, as well as their employment status. Finally, we also control for their views on citizen input in decision-making processes, both generally, as well as in contested issues specifically. As debates and attitudes surrounding amalgamation may be influenced by the type, or size, of the municipality (Strebel 2023), we include as municipal-level controls population, geographic, and financial variables. We also include the municipality’s degree of centrality,³ as well as the municipality’s position vis-à-vis its potential amalgamation partner(s).⁴ Table 2 displays summary statistics of all variables. Translated survey questions are included in the supplementary appendix (A1).

Table 2. Descriptive summary of dependent, explanatory, and control variables.

Explanatory factor	Variable	Source	Type	N	Min/Max	Mean (SD)
Dependent variable	Respondent’s amalgamation stance at the time of the municipal council’s decision	Survey	Dummy	3205	0/1	0.5 (0.5)
Amalgamation discourse topics	Local Belonging	Survey	Ordinal	3232	1/5	3.5 (1.2)
	Local Democracy	Survey	Ordinal	3232	1/5	3.3 (1.2)
	Municipal Services	Survey	Ordinal	3232	1/5	3.4 (1.1)
	Municipal Finances	Survey	Ordinal	3232	1/5	3.2 (1.1)
	Local Employment	Survey	Ordinal	3232	1/5	3.2 (1.1)
Controls	Rank in municipal council	Survey	Categorical	3197	1/4	1.6 (0.9)
	Party affiliation (opposition/government)	Survey	Dummy	3232	0/1	0.4 (0.5)
	Seeking re-election	Survey	Dummy	3175	0/1	1.5 (0.5)
	Terms served in local council	Survey	Discrete	3222	1/7	2.6 (1.7)
	Citizen Involvement: Preferences	Survey	Discrete	3117	1/10	6.4 (2.8)
	Citizen Involvement: Conflictual	Survey	Discrete	3121	1/10	7.3 (2.7)
	Employment status (full time employment)	Survey	Dummy	3226	0/1	0.7 (0.4)
	Population (in 1000)	SSB	Continuous	3232	0.21/673.47	2.1 (1.2)
	Geographic size (in 10 km^2)	SSB	Continuous	3232	0.6/896.9	3.6(1.0)
	Net budgetary result per capita (in 1000 NOK)	SSB	Continuous	3090	−6.44/44.31	1.0 (0.7)
	Centrality	SSB	Discrete	3232	1/6	3.9 (1.3)
	Junior/Senior	CCRD*	Dummy	2350	0/1	0.5 (0.5)

The descriptive statistics of municipal-level variables are presented in their original form. In our analyses they are transformed to a logarithmic scale due to their skewed distribution.

*CCRD = The Centre of Competence on Rural Development (Distrikssenteret).

The data is of a hierarchical nature, and we identify two cases of individual clustering of theoretical interest. First, the clustering of individual respondents within political parties can be expected to account for some of their amalgamation preference variation, driven by strategic considerations (Toubeau and Massetti 2013), differing ideological views on governance (Toubeau and Wagner 2015) or the institutional organisation of the public sector (Heywood 2015). Second, we also cluster our respondents in the municipalities where they are elected. We have theoretical priors to believe that there is significant heterogeneity between municipalities when it comes to the question of amalgamation (Blom-Hansen 2010, 51). We also find empirical variation in the data, suggesting that there are both empirical (ICC = .26) and theoretical reasons to cluster the data in a three-level mode (Leckie 2013; Sommet and Morselli 2017). Namely, respondents nested in parties, nested in municipalities. We empirically test these assumptions by observing the degree of intraclass correlation (ICC) of the party affiliation and the municipality variables on the dependent variable. In the empty models, an ICC of 0.26 is observed for the former, while an ICC of 0.29 is observed for the latter. In other words, roughly 26% of variance of the respondents’ amalgamation preference is attributable to party affiliation, while 29% of variation is due to municipal clustering among the respondents. Following recommendations of Leckie (2013) and Sommet and Morselli (2017), we therefore employ a three-level multilevel (logistic) regression model to analyse the individual local politicians’ amalgamation stance. Finally, several of the variables chosen for this study can be theoretically expected to covariate. Empirically testing multicollinearity reveal max (4.32) and mean (1.66) VIF values below discouraging levels in the proposed models.

A potential limitation of the study concerns the nature of the main independent variables. The variables measure the respondents’ own perceptions of how much each topic dominated in public debates. Moreover, the survey questionnaire did not differentiate qualitatively between these issues as having had a ‘positive’ or ‘negative’ connotation in relation to amalgamation, but rather on the degree to which they were seen as having been prevalent. As such, a focus on for example local belonging, as a measure of local attachment and identity, can be understood both as the public’s concerns for declining local identity in the face of territorial reform (Stein, Broderstad, and Bjørnå 2022), but may also take the form of establishing a new, secondary identity based on the larger regional territory (Terlouw 2016).

We approach these limitations by structuring the analysis of our findings around them, but also by refraining from too strongly determining specific causal directionalities between the independent and the dependent variables. Rather, it is our interest to observe the presence (or not) of such correlations: in other words, we are primarily interested in whether public discourse can impact representatives’ territorial reform preferences. Secondarily we analyse the directionality of potential correlations and discuss possible explanations.

Results

Figure 1 displays the degree to which the respondents perceived the five issues to be prevalent in the municipality’s debate on the amalgamation issue (on a 5-point scale, ranging from 1 = None (Very little emphasised), to 5 = Strong (Very much emphasised). As we hold the assumption that their successive categories are equally spaced, for the purposes of our analysis, we treat these independent variables as continuous even though they are ordinal (Williams 2020)⁵

Figure 1. Issue salience in debate prior to amalgamation decision.

Note: Kernel density distribution of all respondents (N = 3205) of the degree they perceived different issues to have been prevalent in public discourse in their municipality during the LGR.

The respondents did not perceive any one consideration to significantly trump others in the public amalgamation debates, and most centre around the mid-point (‘neither little nor strong’). However, there was a somewhat greater focus on local belonging (mean = 3,50/5) and somewhat a less focus on municipal finances (mean = 3.19/5). Excepting local economy and local employment, the distribution of the respondents’ answers is not significantly normally distributed around the mean, however, and there is a visible skewed distribution. This suggests that, while on average the respondents perceived a somewhat moderate emphasis on these topics, there is a significant portion who noted that the issues of local services, democracy, and belonging were strongly emphasised.

In Table 3, we present the results from four separate multilevel logistic regression models. The dependent variable measures whether respondents want to merge the municipality were they are elected (0 = ‘No, I did not want to merge’; 1 = ‘Yes, I wanted to merge’). The first model (1) includes only our covariates of interest. In the second model (2), we include individual-level covariates. In the third model (3), we remove the individual-level covariates and add municipal-level covariates. In the fourth model (4), our preferred model, we include both individual and municipal-level covariates.

Table 3. Multilevel logistic regression.

	pr(Merge municipalities)
	(1)	(2)	(3)	(4)
Debate: Local Economy	−0.140*** (0.048)	−0.128*** (0.049)	−0.139*** (0.001)	−0.124** (0.049)
Debate: Local Services	−0.023 (0.053)	−0.037 (0.054)	−0.038*** (0.001)	−0.062 (0.055)
Debate: Local Democracy	0.079 (0.052)	0.077 (0.054)	0.060*** (0.001)	0.060 (0.054)
Debate: Local Belonging	0.211*** (0.052)	0.205*** (0.054)	0.222*** (0.001)	0.218*** (0.054)
Debate: Local Employment	−0.084* (0.049)	−0.071 (0.051)	−0.067***(0.001)	−0.052 (0.052)
Involvement: Preferences		−0.068*** (0.018)		−0.074*** (0.019)
Involvement: Conflict		0.013 (0.019)		0.006 (0.019)
Member: 0-4y		−2.241*** (0.419)		−2.234*** (0.442)
Member: 5-10y		0.041 (0.373)		−0.055 (0.366)
Member: <10y		0.197 (0.367)		0.031 (0.359)
Position: Council member		0.009 (0.363)		−0.115 (0.355)
Position: Deputy mayor		0.127 (0.111)		0.125 (0.113)
Position: Mayor		0.167 (0.217)		0.227 (0.219)
Work: Full time		0.220 (0.202)		0.284 (0.205)
Opposition		0.145 (0.113)		0.154 (0.113)
log(population)/1000			−0.085*** (0.001)	−0.088 (0.141)
log(income)/1000			0.130*** (0.001)	0.148 (0.103)
log(km2)			0.243*** (0.001)	0.237*** (0.089)
Centrality level			−0.225*** (0.001)	−0.236* (0.129)
Senior partner			1.243*** (0.001)	1.254*** (0.229)
Constant	−0.279 (0.481)	1.239** (0.557)	−0.759*** (0.001)	1.031 (0.906)
Observations	3,205	3,059	3,064	2,923
Log Likelihood	−1,713.589	−1,627.906	−1,566.276	−1,481.237
Akaike Inf. Crit.	3,443.177	3,291.812	3,162.552	3,012.474
Bayesian Inf. Crit.	3,491.757	3,400.277	3,252.964	3,161.983

Note:*p < 0.1; **p < 0.05; ***p < 0.01.

By inspecting the results across all four models, we can see that the variable measuring the importance of local economy and of local belonging in the debate remain significant across all computations. The former, however, only reaches significance at p < 0.05 and should be interpreted with some caution. While all debate topics display significant correlations in the third model (when municipal characteristics are introduced), in the final model (4), only local belonging and local economy remain correlated with wanting to merge the municipality. Our first hypothesis is thus supported only in two of five cases. The prevalence of local economy and local belonging show robust correlations with the dependent variable. They also differ in their correlational directionality. While an increased focus on local economy reduces the likelihood of local representatives’ amalgamation support, an increased focus on local belonging increases said likelihood. It is important to remember that these explanatories do not contain or measure the qualitative nature of the topics as they arose in public debates. As such, different explanations for the observed correlations are viable.

Regarding the impact of local economy, a focus on this topic in public discourse could stem from a critical view of intended economic benefits of an amalgamation. Amalgamating two or more municipalities may entail a cost in the form of fiscal equalisation mechanisms reflecting the new jurisdictional map. From an economic perspective, citizens will therefore seek to avoid amalgamation, should it be assumed that the post-amalgamation redistribution of public expenditures favour other parts of the new jurisdiction (Brink 2004; Hanes, Wikström, and Wångmar 2012). The positive correlation shown by local belonging may be an indication that, by focusing on this aspect of amalgamation processes, it is simultaneously consciously acknowledged and ‘dealt with’, so as to avoid the loss of belonging citizens may feel following an amalgamation. Identity discourse can in this way promote a secondary identity, covering a greater area than the individual municipality itself (Terluow 2018), thus promoting amalgamation efforts.

Of the control variables, many of the individual-level covariates do not reach statistical significance in Model 4. However, representatives’ beliefs on citizens’ input in decision-making processes (Involvement: Preferences) shows a negative correlation. Representatives who believe that citizens’ input should always be considered in decision-making processes are less likely to support amalgamations as those who believe local governments should be more selective with citizen input. Interestingly, this (or indeed any) correlation is not observed regarding the incorporation of citizens’ preferences in conflictual policy decisions. ‘Freshmen’ party members are more negatively disposed towards mergers compared to non-party members. Additionally, local politicians who are members of a party which is in government at the national level are more likely to support mergers than those affiliated with an opposition party. This effect is, however, not statistically significant in the final model.

Some of the municipal level covariates in the final model do indeed reach a 5% level of statistical significance. Most notably, the status of the municipality the politician represents as a junior or senior (potential) merging partner has a positive effect on their willingness to merge. In other words, local politicians in municipalities in which central services and political offices would be headquartered are more inclined to support mergers. The population and income variables are not significant in the final model. Geographical size is significant and positive, while the municipality’s centrality is significant and negative. As such, representatives in geographically larger municipalities are more positive towards mergers, but those representing more rural municipalities are more averse towards mergers. The degree of centrality of the municipality is only significant at the 10% level in the final model (Model 4).

Table 3 displays the direct correlations between the dependent variable and the independents. Although they shed light on several interesting relationships, the results represent data aggregated. In other words, the findings observed in Table 3 may be obscuring specific local contexts. We therefore explore potential interacting dynamics by running a series of interaction terms, in which we explore the probability of amalgamation support when the various topics are emphasised, across the various values of our municipal-level covariates. Specifically, we test this with two explanatory factors: First, the characteristics of the municipality (population, geography, economy, centrality), and second, junior/senior merging partner status.

The various interactions between municipal characteristics (population, geographic, economic size as well as degree of centrality) fail to display any consistent moderating influences. In other words, although we observe some direct effects of municipal characteristic and the local representative’s amalgamation stance with varying level of statistical significance (see Table 3), the type of municipality, be it large or small in terms of its population, landmass, or financial robustness, does not affect the influence of any of the debate topics on the representative’s merger stance. However, we do observe a statistical difference between the maximum and minimum values on our measure of the centrality of the municipality and when local democracy is debated (see supplementary appendix A2 for regression output and predicted probabilities). Thus, we fail to find consistent and comprehensive support for hypothesis 2.

This is an interesting finding in and of itself: it indicates that the observed relationships between the five amalgamation topics and the dependent variable are retained regardless of the type of municipality considered. We do, however, find a pattern regarding the municipality’s position as a junior or senior partner in relation to its (potential) merger partner(s) of interest to further explore. Figure 2 displays the results of interaction analyses with these variables.

Figure 2. Interaction terms with junior/senior partner.

Note: Interaction regression modelling of independent variable Senior partner (X) on Merge municipalities (Y) given different values of the five explanatories (Z). N = 2,923. The model includes the same controls as Model 4 above. Full regression tables are reported in the accompanying appendix (A3).

By splitting the respondents’ amalgamation preferences on the localisation variable, we find that senior amalgamation partners are generally more positively inclined towards amalgamation than their potential junior partners. These attitudinal differences mostly remain unaffected by increased focus on local economy and belonging – the two topics which have been shown to have a consistent effect on merger preferences (see Table 3), leading us to believe that these two topics are relatively robust explanatory variables. However, when local democracy, employment, and services are increasingly discussed, the difference in amalgamation preferences between junior and senior municipalities’ representatives magnify.

As such, no single discourse topic, when more prominent in local debates, serves to reduce the merger attitude differences between central and peripheral amalgamation partners. Instead, their differences are either unaffected or increased. Discussing local democracy seems to divide junior and senior amalgamation partners. While their differences in amalgamation preferences are non-significant when local democracy is not discussed, increasing focus on it in public discourse serves to create a wedge between the two groups of representatives. This happens primarily in senior partner municipalities, where support for amalgamation is increased, while the preferences among junior amalgamation partners remain stable. For peripheral communities, in other words, it does not matter how much or little one discusses local democracy in an amalgamation process, whereas attitudes towards amalgamation are increasingly supportive in central municipalities. The same picture is observed in the case of local employment.

This indicates that, for peripheral communities, amalgamation represents a potential threat to retaining employment. This lends support to the findings of Klausen, Rose, and Winsvold (2021) who found a potential fear of centralisation among citizens in smaller municipalities. A similar logic can be applied to the question of local services: while the representatives’ amalgamation preferences are within confidence intervals and as such not statistically dissimilar when municipal services are not discussed in public discourse, the more this topic is prevalent, the more their preferences diverge – again with localisation ‘winners’ increasingly in support of amalgamation, while junior partners support it less. Issues of locality thus seem to relate both to local democracy, proximity of services as well as employment in the local community, and attitudinal differences to the question of amalgamation are only ever magnified, rather than contracted, once these topics are made prevalent in public discourse relating to the amalgamation process.

This is interesting when considering the voluntary nature of the Norwegian reform, as they depended on local mutual agreement to see amalgamations through. Our findings indicate that no amalgamation topic serves to unify amalgamation partners when considering the central or peripheral status the old municipalities would have in a new municipality.

So far, our analyses have been produced with the assumption that only one topic varies in its prevalence at any given time, while the others are held constant. But municipal amalgamation is complex and multifaceted. Hence, it should be expected that public discourse relating to amalgamation is, too (Leknes, Ridderstrøm, and Rommetvedt 2019). We therefore expect that multiple topics were debated simultaneously, and that the representative’s merger stance may be affected by various combinations of the five topics. Table 3 shows that Local Economy and Local Belonging remain robust and significant (p < 0.05) influential variables on the representatives’ amalgamation stance. We therefore limit our tests to these, to observe whether their effects on the dependent variable (amalgamation stance) remain unaffected, or is changed, depending on the values of one of the other topics. Like our testing of municipal characteristics as interacting dynamics, we find that the interactions between various topics to produce mostly non-significant results (see supplementary appendix A4 for full regression outputs). We find, however, one notable exception in the interaction between local democracy and local belonging. As displayed in Figure 3, when local democracy is not emphasised in public discourse, there is a significant difference in the probability of supporting amalgamation between those who perceived local belonging to have been greatly emphasised, and those who did not. When neither local belonging nor local democracy is emphasised in local discourse, the predicted probability of supporting amalgamation is roughly 40%.

Figure 3. Interaction between local democracy and belonging.

Note: Interaction regression modelling of independent variable Debate: Local Belonging (X) on Merge municipalities (Y) given different values of Local Democracy (Z). N = 2,923. The model includes the same controls as Model 4 above. Full regression tables are reported in the accompanying appendix (A4).

However, when local belonging is emphasised in public discourse, but democracy is not, the predicted probability of amalgamation support surges to around 80%. Once local democracy enters the discussion, however, the degree to which local belonging is discussed no longer matters. In between these extremes, it is also interesting to observe that as democracy is increasingly a topic of conversation, support for amalgamation decreases when local belonging is not discussed, while increasing when it is. Local democracy and local belonging both may be viewed as pertaining to the communitarian side of local governance (Myksvoll 2023). Here, however, they initially appear more independent forces than this would imply. Hence, when the former is not a part of public discourse, it matters whether the latter is: when the former is a part of public discourse, it no longer matters whether the latter is.

Concluding discussions

Explaining the occurrence of voluntary municipal amalgamations is a complex endeavour. Examples of literature that has sought to enlighten this draw on theories of government tools (Vabo, Fimreite, and Houlberg 2023), postfunctionalism (Strebel 2019) or economic incentives and voting behaviour (Miyazaki 2014; Yamada 2016). In this study, we contribute to this discussion by adding to it the perspective of public discourse, through which democratic legitimacy regarding amalgamation decisions may be put.

The political process to undertake a municipal amalgamation, especially where this is conducted voluntarily at the local level rather than being coerced from above, requires the willingness of local representatives to drastically alter the jurisdiction they represent. From a democratic perspective, such willingness can (or should) be shaped by public opinion, which, outside of elections, is expressed through public discourse.

This paper has analysed how local public discourse shape local representatives’ amalgamation willingness by focusing on five topics related to amalgamation: local economy, services, democracy, belonging, and employment. The analyses show that a public discourse which emphasises local finances is associated with an increased likelihood of representatives’ amalgamation support, while a greater focus on local belonging increases it. These correlations remain enduring across different types of municipalities, although they are conditioned by whether the municipality will retain or lose key political and administrative offices. Moreover, the positive correlation of local belonging is strongly moderated by whether local democracy was simultaneously discussed in the municipality.

These findings contribute to our understanding of why some municipalities may choose to amalgamate, why some do not, and how various issues associated with amalgamation, through public discourse, may influence this process. The data employed in our study does not specify the qualitative nature of the discourse. It would rather necessitate a more qualitative approach to individual municipalities as they dealt with the question of amalgamation. We nevertheless find the variables’ relationships (including those lacking any) to shed interesting lights on the dynamics at play at the local level during territorial reform, and encourage further research, particularly on the qualitative nature of various amalgamation-related topics in public discourse, to understand how these issues, given varied connotations (i.e., presented in a negative or positive light) influences local decision-makers.

The Norwegian Local Government Reform’s voluntary local processes, backed by threats of coercion from above (Vabo, Fimreite, and Houlberg 2023), represents a relevant and interesting case to study the impact of public discourse on amalgamation preferences. It also provides opportunities for generalisability beyond the specific Norwegian context. Voluntary amalgamation processes have taken place in other countries, such as Japan (Yamada 2016), Finland (Saarimaa & Tukiainen 2015), Sweden (Hanes & Wikström 2010), and Switzerland (Strebel 2019), to name a few. It has not been unusual for local referenda to be included in such processes. As referenda outcomes may be shaped by public discourse, such discourse becomes an important link in the causal chain leading to the decision to (not) amalgamate. Holding that functionalist and communitarian concerns are commonly raised during amalgamation processes (though their specific prevalence in individual municipalities may vary) (Erlingsson, Ödalen, and Wångmar 2020), our findings have a generalisability potential to such similar cases, and shed light on citizen engagement as a democratic channel of input to local decision-makers facing the amalgamation question. Future research on voluntary amalgamation processes are therefore encouraged to study the democratic side of such processes, by focusing on the intricate link between public opinion and decision-making.

Acknowledgments

The authors would like to acknowledge colleagues at NORCE Norwegian Research Centre for continuous feedback and helpful comments during the writing process, as well as feedback received from the 2023 Norwegian National Conference of Social Sciences.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplemental data

Supplemental data for this article can be accessed online at https://doi.org/10.1080/03003930.2024.2302604.

Notes

1. This resulted in 13 cases of coerced amalgamation.

2. One amalgamation took place in 2017, and four in 2018.

3. The municipality’s degree of centrality is based on proximity to the number of workplaces and how many different types of service functions (goods and services) those living within the individual jurisdiction can reach by car within 90 minutes (SSB 2017).

4. This is measured by whether the municipality would retain the political and administrative headquarters in the new municipality. The measure is sourced from intermunicipal negotiation agreements reached in most municipalities ahead of local referenda, available from The Centre of Competence on Rural Development (Distrikssenteret) (Myhr 2015).

5. Treating the variables as categorical in our regression models shows that the results remain similar to that of a continuous treatment.