Territorial reforms, mobilisation, and political trust: a case study from Norway

ABSTRACT The focus of this study is the political trust implications of territorial reforms, approaches to territorial reform, and the effects of the mobilisation of political-territorial collective identities. We focus on the political trust effects of political-territorial mobilisation grounded on territorial reforms, and of voluntary and forced structural reforms. The case examined is that of Norway, a country characterised by high levels of trust before a recent county reform. Utilising four survey waves from 2013 to 2019, we measure trust in national politicians both pre- and post-reform, giving us a quasi-experimental design. The findings indicate that political trust was not affected by whether the reform was forced on counties or they accepted it voluntarily. However, political trust was negatively affected by forced structural reforms in combination with regionalism, i.e., mobilisation of political-territorial collective identities. This finding provides new insight about how territorial reforms may affect political trust.


Introduction
The focal point of this paper is the trust implications of territorial reforms. As territorial reforms affect cultural, economic, and political interests, they are the most radical and contested measures in the reorganisation of local units (Ebinger, Kuhlmann, and Bogumil 2019). We are interested in how territorial reforms could have implications on political trust, as it is a prerequisite for a well-functioning society. Here, we define political trust as the public sentiment about the government and its political representatives (Norris 2011).
Peripheral political mobilisation against central political institutions is emerging as a new trend in many European countries, albeit with different political expressions (Ford and Jennings 2020;Jennings and Stoker 2017;Lee, Morris, and Kemeny 2018;Rodríguez-Pose 2018). The tension between central authorities and regions within a country is a well-known challenge, and political-territorial collective identities have been a focus of attention for political scholars (Keating 1998;Lipset and Rokkan 1967). The continuing presence of regionalism, understood as political-cultural processes in which a collectivity with a sense of territorial belonging takes action, or an elite mobilises and organises the cultural, economic, and political interests of a territory (Brunn 1995;Caciagli 2003;Nevola 2011), reveals the incompleteness of national integration and suggests regional lack of trust in the central political institutions.
Political-administrative regions within countries are constructions that can have been developed over a long time, more or less in accordance with a distinct culture and historic roots. They may also be comparatively new territorial administrative units designed locally or by the central government, but they are, like other bodies, likely to institutionalise their own practices, norms, interests, and become a political-territorial collective over time. The effects of territorial reforms on community identity have been widely studied in the literature, which concludes that structural reforms lead to social distrust and declining community identity, and negatively affects social cohesion, political, and social participation (Dahl and Tufte 1973;Ebinger, Kuhlmann, and Bogumil 2019;Nie, Verba, and Petrocik 2013;Putnam 2001). Moreover, a few empirical studies find a negative relationship between structural reforms and political trust (e.g., Denters 2002;Hansen 2012). The trust effects of political-territorial regionalism, manifesting in the form of mobilisation grounded on territorial reforms, however, appear unstudied, and the present article aims to contribute to filling this gap in the literature. Broader literature reviews on municipal amalgamations and their effects show that there is a dearth of research on the relationship between amalgamations and trust (Tavares 2018). More specifically, this article deals with relatively recently created regional (county) democratic units (around 50 years old), and different reform approaches: structural reforms requested by the central government but deemed 'voluntary' (voluntary reforms), and structural reforms designed and enforced by central government (enforced reforms).
First, we studied whether the reform was voluntary or enforced in relation to the effect on political trust. We test the hypothesis that political trust is linked to reform approaches, and that enforced reforms are accompanied by resistance (as suggested by Steiner, Kaiser, and Eythórsson (2016)), and consequently, declining political trust. Second, we suggest that regionalism, which manifests in the form of mobilisation on the grounds of political-territorial collective identities, is likely to lead to declining political trust in the affected territory. Understanding regionalism due to territorial reforms and its influence on political trust adds new knowledge to the literature on territorial reforms, on community identity and to the literature on trust.
The interplay between territorial reforms, regionalism, and political trust is clearly of general European interest, but here we narrow the focus to a study of county mergers in Norway. Norway was chosen because it has recently experienced reform of its regions with both 'voluntary' and enforced county structural reforms. The context is harmonious: Norway, like the other Scandinavian countries, is a well-functioning unitary state, devolving the same sets of powers and discretions to all its county units, with top scores in international rankings on the best country to live in, and top scores on trust in politicians (Dalton 2005). Hence, Norway appears to be a rich and harmonious country. However, Norway is a vast territory and around 50 years ago was known for having regions with different political-territorial collective identities (Lipset and Rokkan 1967). In Norway, we find the most distinct peripheral area, the most recent reforms, and the longest traditions and broad political acceptance of policies that preserve a dispersed settlement, especially in the north (Stein 2019).
Our study focuses on the latest structural reforms in Norway. In 2017, the national government implemented a series of local government boundary changes that affected 15 out of the 19 Norwegian counties. To explore this, we use a methodological framework composed of a quasi-experiment with survey data. By utilising four survey waves (DiFi) conducted from 2013 to 2019 with 33,000 respondents, we can measure trust in national politicians both pre-and post-merger. Based on these data, we can observe how the process, that is, the enforced changes, affected levels of trust in the members of parliament (MPs), with the unaffected counties and the counties that accepted the reforms voluntarily serving as a control group. Within the treatment group, we make a distinction between counties that experienced forced mergers with weak mobilisation and forced mergers with strong mobilisation. The findings suggest that the process significantly lowered the level of trust in those counties where mobilisation was high.
The remainder of this paper is structured as follows. First, we present a broad conceptual framework of political trust and link this to county mergers, followed by a brief presentation of county mergers in Norway. Second, we give details of data collection and the methods applied before presenting and discussing the main results. Finally, we discuss the implications for policymakers and make recommendations.

Perspectives on trust
Trust is a debated concept with competing and complementary explanations of its origin. The cultural perspective argues that trust is embedded in the links and networks between people in the community; that is, trust in political institutions is rooted in cultural norms and communicated through early socialisation (Almond and Verba 2015;Inglehart 1997;Mishler and Rose 2001;Putnam 2001). The institutional perspective, on the other hand, argues that trust originates as an outcome of successful policies and that it is a consequence of institutional performance. Institutions that perform well generate trust; low performing institutions generate scepticism and distrust (Dalton 2005;Hetherington 1998;North 1990;Rothstein 2011;Van Ryzin 2007). Hence, from an institutional performance perspective, political trust can be enhanced by providing good services as a positive evaluation of services spills over to a positive sentiment about the government and its political representatives. Moreover, trust is discussed in relation to democracy and good democratic procedures and standards (Dahl 1989;Held 1991;Newton and Norris 2000) and in some elaborations of Easton's system support theory (Easton 1975). The general idea of political trust is also discussed in relation to national identity in the system building tradition (Bartolini 2005;Berg and Hjerm 2010;Rokkan and Urwin 1983;Stein, Buck, and Bjørnå 2021), and this approach serves as the point of departure for our understanding of the relationship between territorial reforms, regionalism, and trust.
Rokkan's most important contribution to political analysis was the addition of an independent territorial dimension to politics: the centre periphery axis linking the institutional architecture of a nation-state to its territorial structure, that is, it's given political and spatial characteristics (Rokkan and Urwin 1983). Moreover, differing historiographies linked to the process of nation building may create territorially different political interests, and any collective distinction may serve as an underpinning for political mobilisation (Sartori 2005) i.e., regionalism. Initially, regionalism seemed to be the ideology of the periphery: marginalised backward regions exploited by the 'centre' (Nevola 2011), an ideology that is well illustrated today in the works of Jennings (Ford and Jennings 2020;Jennings and Stoker 2017;McKay, Jennings, and Stoker 2021). They focus on the divide between citizens residing in locations strongly connected to global growth and those in locations that are not and find this to be a divide that exists between those from densely populated urban centres with an emerging knowledge economy and people who live in suburban communities, coastal areas, and post-industrial towns (Jennings and Stoker 2017, 4-5). The latter group has different values and feels 'left behind' in political visions and strategies; they protest and are resentful of the political establishment.
The general idea of the 'left behind' is also reflected in Cramer's (2016) concept of 'rural consciousness'. Rural consciousness includes a sense that decision makers routinely ignore rural places and fail to give rural communities their fair share of resources or a sense of relative deprivation, and a belief that rural citizens are fundamentally different from urbanites in terms of lifestyles, values, and work ethic. However, regionalism, over the years, has also been understood as an ideology of the richest regions in countries (such as Catalonia). Hence, regionalism is found both in rich and in marginalised regions and manifests in the form of mobilisation and movements based on territorial interests and identity that advance cultural, historical, political, and economic claims (Nevola 2011) and are built on a deep resentment and distrust of central government and its political representatives. It has even been suggested that such bitterness and political distrust are linked to distance from the centre: that as a general rule, greater distance from the centre has a negative effect on trust in politicians (Stein, Buck, and Bjørnå 2021). Territorial identities are powerful sources of discontent, and in extreme cases, such movements can claim not only cultural and administrative recognition, but also political autonomy.

Territorial reforms, identities, and reform approaches
An optimal jurisdiction size is a cornerstone of government design and has been on the reform agenda in many European countries for decades. Structural reforms have several motivations, the main one being that mergers can create economies of scale and reduce the public cost of administration and bureaucracy. A second motivation is that structural and functional changes in the population pattern can require a new structure (Storper 2014). A third issue that often arises in sparsely-populated countries such as Norway is the need for larger units to have the competence and knowledge required to deliver sufficient quality of government (Boyne 1995). 1 Hence, the most important drivers for territorial reforms are the objective of achieving improved local finance and human resources, improved quality and quantity of public services, exactness in legal decisions, and improvement for local autonomy and local democracy. These are all crucial characteristics of local government performance (Steiner, Kaiser, and Eythórsson 2016). Consequently, a structural reform, if successfully based on these motivations, would lead to increased trust from the institutional performance perspective.
The literature on territorial amalgamation reforms is vast and can mainly be divided into three focus areas. First, there is the research on the impact of territorial reforms on local democracy and elections. In Denmark, there is some evidence that the increase in municipal jurisdiction size has led to positive effects on voter turnout (Bhatti and Hansen 2019); however, it has led to decreased voter turnout in other countries (Lapointe, Saarimaa, and Tukiainen 2018;Koch and Rochat 2017;Blesse and Roesel 2019). Some researchers have found that amalgamations have detrimental effects on local democracy because of the reduction in internal political efficacy, that is, individual citizens' beliefs about their competence related to understanding and participating in politics (e.g., Lassen and Serritzlew 2011;De Ceuninck et al. 2010).
Second, the desired economic and administrative effects of amalgamation reforms are also contested. Some researchers argue that quasi-experimental studies which offer a causal interpretation do not show that mergers reduce total expenditures, and that potential savings in administrative costs, for example, can be offset by opposite effects for other domains (Blom- Hansen, Morton, and Serritzlew 2015). On the other hand, there is a body of literature substantiating the existence of economies of scale in most municipal services and showing that expenses can be reduced through economies of scale (e.g., Blesse and Baskaran 2016;Reingewertz 2012). In a more recent literature review, Gendźwiłł, Kurniewicz, and Swianiewicz (2021) conclude that saving on administrative spending is perhaps the only clearly confirmed gain of territorial amalgamation reforms. According to them, this finding should redirect the attention of researchers and policymakers to the varying institutional contexts, territorial organisation prior to the reform, strength of local identities, and political agency of the reforms' proponents and opponents, as important moderating variables.
Third, the effects of territorial reforms on identity have been widely studied in the literature, which implies that structural reforms tend to lead to social distrust and a decline in community identity (Dahl 1989;Ebinger, Kuhlmann, and Bogumil 2019;Nie, Verba, and Petrocik 2013;Putnam 2001). Existing communities are part of people's identities, understanding the term 'identities' as referring to a feeling of similarity and solidarity (Capello 2018). Political-territorial collective identities, then, have common and shared values and a solidarity that rests on feelings of attachment and anchorage to the local area. Sources of similarity are identified in history, tradition, culture, and language, which together participate in the creation of borders with others, and have the potential to serve as grounds for mobilisation (Keating 2008;Lipset and Rokkan 1967;Rokkan and Urwin 1983;Sartori 2005). Structural reforms destroy the symbolism of being a recognised community (Denters et al. 2014) and consequently are likely to lead to decreased trust in the government and its political representatives within the framework of the community identity perspective, especially if areas with clear politicalterritorial collective identities are affected. At the very least, it is likely to take time before residents will accept the newly formed entity as a community rather than just an administrative unit (Allers et al. 2021).
However, the trust effects of a territorial reform and whether the reform sparks repercussions in the form of regionalism are likely related to the reform approach. Bottom-up reforms can emerge because residents and their politicians find it context-appropriate or where incentives, financial or otherwise, are offered. A bottom-up approach, where territorial reform is introduced incrementally, is likely to experience low levels of resistance (Kaiser 2014;Steiner, Kaiser, and Eythórsson 2016), and is less likely to lead to regionalism and diminished political trust. There may also be a range of middle-ground approaches, initially based on requests from the central government, but where it may be up to territorial units to 'find their partner' themselves. Such approaches may or may not be accompanied by incentives. In these cases, it may be difficult to define how 'voluntarily' the amalgamations are. As Swianiewicz (2010) noted, 'the difference between bottom-up voluntary change and reform implemented in a top-down manner is not very sharp'. Compulsory reforms usually leave some discretion for local governments; voluntary reforms frequently employ powerful incentives. However, in this study, we treat bottom-up and middle ground approaches as 'voluntary reforms'. However, as per the supporting material, we have conducted robustness checks with mergers that could be categorised as being somewhere in between forced and voluntarily. The results remain consistent with our main argument.
The third category is central government-designed and enforced structural reforms, 'forced reforms', which we assume are likely to meet resistance and create conflict between the lower tiers of government and the central government. This is because residents and local politicians do not like topdown decisions on matters that affect them (Steiner, Kaiser, and Eythórsson 2016) and because it disturbs established political-territorial collective identities. Forced reforms, we propose, may lead to regional opposition to central politicians, and may even lead to regionalism. Moreover, we propose that regionalism, as manifested by mobilisation on the grounds of politicalterritorial collective identities, is likely to lead to declining political trust.
Based on these assumptions about the effect of enforced structural reforms, we formulate the following hypotheses: H 1 : There is a decline in political trust in regions subject to forced structural reforms compared to regions that have engaged in voluntary structural reforms.
H 2 : There is a decline in political trust in regions that mobilise on the grounds of political-territorial collective identities, compared to regions without such mobilisation.
Our framework established in H 2 can be summarised in Figure 1

Empirical background: the case of enforced county mergers in Norway
After Norway's 2013 parliamentary elections, the government changed from a majority-centre-left government to a minority conservative-centre government, dependent on two centrist minor parties. One of the major priorities for the new government was to implement an overhaul of the structures of the Norwegian public sector. The most imminent was the mergers of municipalities (for more about the process and elections see Fitjar 2019; Folkestad et al. 2019;Stein et al. 2020). To get the two minor parties to support the total package of mergers of municipalities, the two largest parties in the government had to agree on reforming the structure of the counties. The main position of the largest government parties was to scrape the county level and split their portfolio (mainly upper secondary school, regional transport, and oral health) between municipalities and the national government. However, the smaller parties wanted to maintain the regional level and they believed that the county portfolio could be expanded with fewer counties.
As shown above, the structural reform of Norwegian municipalities and counties are interlinked, wherein the county amalgamations were an outcome of negotiations between the parties constituting the central government. The Norwegian counties today may be more of an administrative division than a 'true' regional level government: they are in charge of relatively few tasks, they have a relatively low score on the Regional Authority Index (a measure of the authority of regional governments developed by Marks, Hooghe, and Schakel (2008)), and they have no formal powers over the municipalities as both counties and municipalities are regulated under the same Act (Lov om kommuner og fylkeskommuner (kommuneloven)).
This process came into effect in June 2017. As shown in the map in Figure 2, four counties (Oslo, Rogaland, Møre-Romsdal, and Nordland) did not merge. Ten counties merged voluntarily, albeit some more enthusiastically than others. 2 Five counties were subjected to enforced mergers: Østfold, Akershus, and Buskerud into the new county 'Viken', and Troms and Finnmark into the new county 'Troms-Finnmark'. Although there was local opposition in Østfold, Akershus, and Buskerud to their absorption into Viken, the opposition was much stronger in Troms, and even more so in Finnmark. The county administration in Finnmark refused to participate in formal negotiation talks about the merger and passed multiple resolutions in the county parliament opposing the merger. The Finnmark administration also organised a 'referendum' on the merger in May 2018. Even though the 'referendum' fell far short of meeting the criteria of the Venice Commission's code on referendums, it showed the level of resistance from political leaders and people with eighty-seven per cent voting against the merger (Finnmark 2018). This result is similar to the results found in opinion polls around the same time. The opponents of the territorial reform in Troms and Finnmark also set up two organisations to cooperate in opposing the merger: For Finnmark and For Troms.
The northern region is considered special in the Norwegian context. There is relatively broad academic literature discussing the region's uniqueness, as it is a sparsely populated territory with vast distances and natural resources, local traditions and adaptions, special communities, challenges, and possibilities (e.g., Brox 1966;Jentoft, Nergård, and Røvik 2013). Some researchers have urged the preservation of its distinctive artefacts, stigmata, and culture (e.g., Drivenes, Hauan, and Wold 1994), while others have found relationships between forging collective identities and successful developmental policies within the region (Bjørnå and Aarsaether 2009). In one contribution, the issue of identities within the region has been found to represent exceptionally fertile soil for mobilisation against amalgamation reforms (Bjørnå 2013). Previous studies have also shown that the trust in politicians in Norway is lower among the people living furthest away from the capital (Stein et al. 2021). Thus, we found grounds for categorising the county as having strong regional mobilisation.
Even though the three counties now constituting Viken opposed the enforced structural reforms the resistance was much less fierce, both among politicians and in the population. The three counties merged to form Viken (Østfold, Akershus, and Buskerud) are all situated in the region around the capital (Oslo) (see Figure 2). An opinion poll in June 2017 showed that fifty-one per cent of the respondents were against the creation of Viken as a new unit, while twenty-six per cent were in favour. 3 In contrast to Troms-Finnmark, opposition to Viken did not manage to mobilise political-territorial collective identities. The differences in these experiences make it possible to carry out a quasi-experiment, where some respondents have not experienced a merger, some have experienced a voluntary merger, and some have experienced an enforced merger. Moreover, some respondents experienced a forced merger accompanied by regionalism, while others did not.

Methods and data
Our methodological framework is a quasi-experiment that uses survey data. Finding natural or quasi-experiments has long been a part of programme evaluation in psychology (e.g., Campbell 1969;Cook and Campbell 1979) and economics (Angrist and Krueger 2001;Meyer 1995), but such an approach is also emerging in political science (Lassen 2005) and in political trust research (e.g., Ares and Hernández 2017;Hansen 2012). Blom-Hansen, Morton, and Serritzlew (2015) argue that experimental methods are underused in public management research. They introduced five different experimental designs in public management research: lab, survey, field, natural, and quasiexperiments. Quasi-experiments share the characteristic with natural experiments that the intervention comes from the outside. The process not controlled and manipulated by the researcher but is created by nature or the political-administrative system. In contrast to other experiments, however, assignment to experimental and control groups is not randomised in quasiexperiments (Blom-Hansen, Morton, and Serritzlew 2015). Therefore, the two groups may differ in terms of their exposure to the experimental intervention (Shadish, Cook, and Campbell 2002). Quasi-experimenters thus face the challenge of interpreting whether differences in results are caused by the experimental intervention or by initial differences between the groups. In this study the regions in question may differ in terms of social, political and economic characteristics, and those differences may explain differences in outcome. To address these crucial challenges, we run a placebo test, a propensity score matching (PSM), and a nearest neighbour algorithm to gauge potential endogeneity issues. These are the robustness checks advocated by Gendźwiłł, Kurniewicz, and Swianiewicz (2021) (see Appendix C). Both tests show that there is no systematic difference between the treated and untreated units on a set of socio-economic characteristics (education, income, immigration, age, party preference and marital status). This gives an indication that the non-random randomisation process has been successful. Even though there still can be potential unobserved endogenity issues, quasiexperiments have considerable potential for new studies (Blom- Hansen, Morton, and Serritzlew 2015). They are considered to be stronger designs than traditional observational studies because they address endogeneity problems more directly. Further, quasi-experiments, such as natural experiments, render it possible to investigate the effects of phenomena that are difficult for researchers to manipulate. Consequently, if randomisation cannot be sacrificed, many research questions remain unanswered (Blom- Hansen, Morton, and Serritzlew 2015). Therefore, we believe that a quasi-experiment is appropriate in this case, even though it comes with a caveat that the treatment is not randomised.
Although there are compelling arguments for combining different indicators when measuring political trust (e.g., Marien and Hooghe 2011), there are also important similarities between trust in political actors, liberal democratic institutions, and courts and police (Denters, Gabriel, and Torcal 2007). In the current study, we used a single-item variable, trust in politicians, as an indicator of the concept of political trust (other studies with single-item indicators include Hetherington and Rudolph 2008;Newton 2001;Rudolph and Evans 2005;; Van der Meer 2010). The respondents were asked to give their answers on a 7-point scale, ranging from −3 to 3, where −3 indicates very low trust and 3 indicates very high trust. Trust in national politicians was measured by the following question: 'How much or little trust do you have that politicians in Stortinget (the national parliament) are working for citizens' best interests?' Our data were obtained from the Norwegian Agency for Public Management and eGovernment citizen surveys (DiFi).
We use a difference-in-difference (DiD) model and test whether counties affected by the merger display a significant change in trust compared with the control group, which is composed of counties unaffected by the merger. As we do not have repeated observations for the same individuals over time, we have chosen to aggregate individual-level data to the county level (individual level models and the number of respondents per county are listed in the Appendix C). Running the analysis on individual-level data would produce biased estimates, because we do not have pre-treatment observations on the individual that received the treatment. We aggregate an average trust indicator per county-year in the DiFi survey, which consists of 33,851 respondents over four survey waves (2013,2015,2017,2019) on 18 counties. 4 Below, we present the results from the different model specifications. First, we used a simple DiD model in which the treatment and time of treatment interact. The model takes the following form: where D i is the effect of the treatment, that is, forced mergers, and Y i is the outcome, that is, the change in level of trust. α i is the county fixed effects and λ t is the time fixed effects. The focus of our interest is the following: where treat i is a treatment group dummy and time i is a dummy for the posttreatment period. To test our hypotheses, we are interested in the DiD estimator ρ ðtreat iÞ � time i ð Þ À � .

Results
Before testing our hypotheses, we ensured the validity of our model by using a visual inspection of the parallel trends-assumption, a placebo test, a PSM, and nearest neighbour matching (cf., Abadie 2005; Gendźwiłł, Kurniewicz, and Swianiewicz 2021). In Figure 3, we plot the average trust in MPs over time.
The field year of each survey wave is denoted on the x-axis. The grey lines represent the local polynomial regression (loess) for each county, and the black lines show the average of the treatment and control groups. To fulfil this assumption, the difference between the treatment and control groups was constant over time in the pre-treatment condition. To our knowledge, there is no statistical test for this assumption, yet a visual inspection can help us ensure that the requirements of the assumption are met. From the plot, we can infer that the pretreatment trends are parallel. The solid line, which represents our control group, is parallel to the two treatment conditions represented by the two dotted lines. Thus, we can conclude that this assumption holds true. To further ensure the validity of our design, we used a placebo treatment test. In this model (see Table F in the Appendix), we administer a placebo treatment in 2015 which is the halfway point in our observations over time, and four years before the actual treatment occurs.
The results showed no significant effect of treatment. This, in combination with the visual inspection of Figure 3, the PSM, and nearest neighbour matching (Appendix C, Figures 1-4), leads us to conclude that our model is internally valid and robust to potential observable endogeneity issues.
After establishing that the model is internally valid, we proceed to the DiD model to test our hypotheses. Figure 4 displays the coefficients from Table 1. Our dependent variable is estimated using a simple DiD model with a multivalued treatment indicator. In model 1, we can observe a significant drop in trust starting from the 2017 survey wave to 2019 in all counties. The drop of −.325 points indicates a general decline in trust in national politicians in Norway, albeit from a high initial level. In model 1 we also test our H 1 about there being a stronger decline in trust in the counties where territorial  reforms were forced. We find that although there is a lower level of trust in the 'forced' counties after the reform compared to the other counties, the effect is not significant (β = −.066). In model 2, Troms-Finnmark and Viken (created as the result of forced mergers) are separated based on regional mobilisation as seen in H 2 , and then compared with the control condition. In Figure 4, we see that the level of trust in MPs is lower, in general, in the northernmost county (Troms-Finnmark) compared with the control group (−.361 points). In Viken the trust is slightly lower, but only −.163 points, which is only significant at a statistical significance value level of 90%. When analysing the effects of the reform, the DiD estimator, we see that there is a significant drop in trust in national MPs in Troms-Finnmark at −.234 points. For Viken, we observe a positive but insignificant estimator. This tells us that the treatment had an effect in Troms-Finnmark and not in Viken.

Discussion
These results indicate that a forced merger does not necessarily lead to a decline in trust in politicians. As seen in Table 1, the level of political trust remains the same in nearly all counties, with no significant decline in trust in the merged counties compared to the non-merged counties (with one exception). This suggests that structural reforms can become administrative affairs that are largely accepted. Hence, not all regions subject to territorial reforms have an identity element that leads to decreased trust. This finding adds to the literature on territorial reforms on community identity (Dahl and Tufte 1973;Ebinger, Kuhlmann, and Bogumil 2019;Nie, Verba, and Petrocik 2013;Putnam 2001). We hypothesised a lower level of political trust in regions subject to forced territorial mergers than in regions that have done so voluntarily. The two forced territorial mergers were in Viken and Troms-Finnmark. The models in Table 1 show no decline in trust in Viken, a finding that contradicts arguments in the structural reform literature (Kaiser 2014;Steiner, Kaiser, and Eythórsson 2016), where it is asserted that forced mergers are likely to meet with local resistance and result in conflicts between lower tiers and the central government. This is not always the case, and we can reject H 1 . H 2 posited a lower level of political trust in regions with mobilisation on the grounds of political-territorial collective identities than in regions without such mobilisation. As shown above, voluntarily merged regions and nonmerged regions displayed similar trust levels, and there were no significant protests voiced against voluntary mergers. Hence, the instances of forced mergers found in Viken and Troms-Finnmark, where various degrees of protests were evident, becomes particularly interesting in the discussion of H 2 . In Viken, however, there was no strong regional mobilisation. This could indicate that the tensions between the centre and periphery are less pertinent in regions located around the capital (Oslo). It also indicates that that political-territorial collective identities matter for political trust: The Viken region is not made up of regions with the kind of deep-seated and historically based collective distinctions that could underpin political-territorial collective identities. Consequently, it would be more difficult to mobilise on such grounds against a forced territorial merger. Other grounds for mobilisation, such as a fear of a lower quality of public services because of a merger, would lead to a decline in trust in all the counties affected by the territorial reform. Table 1 shows the same level of political trust in the merged counties, including Viken, compared to the non-merged counties.
The exception is the northernmost region, Troms-Finnmark. Table 1 shows a significant decline in trust in politicians in the region affected by a forced merger with strong regional mobilisation. This indicates that mobilisation is a necessary condition for a decline in trust, even after a forced merger, further suggesting that regionalism, not the forced merger itself, is the cause of the decline in trust in politicians. In a classical centre-periphery relationship, a territorial merger is an example of a central authority using force to implement structural changes in the regions. When residents can mobilise the latent structural tension between the centre and periphery (Rokkan and Urwin 1983), a decline in trust in national politicians is to be expected. Although this particular experiment does not include the rhetoric used in the mobilisation processes, it has been clearly documented that the two organisations that opposed the Troms-Finnmark territorial reform used the rhetorical power inherited in territorial collective identities, 5 that is, that the organisations mobilised on some form of 'rural consciousness' (Cramer 2016) and the cultural, economic, and political interests of the territory. Hence, it is likely that mobilisation on the grounds of political-territorial collective identity is relevant in relation to trust in politicians. Hence, this suggests that political territorial identities matter for the acceptance of territorial reforms and that local mobilisation affects political trust. This indicates support for H 2 .
We found that local resistance and diminishing trust are linked to whether or not regionalism is present, that is, whether a political-cultural process where a collectivity or a group representing a territorial identity mobilises and organises the cultural, economic, and political interests of a territory (see Brunn 1995;Caciagli 2003;Nevola 2011). Although the general levels of political trust declined all over Norway over time, the decline in trust in politicians was steeper in the northernmost region. Hence, the current study's findings support the Rokkan-inspired hypothesis about the impact of distance: a greater distance from the centre can negatively affect regional citizen trust in politicians (Stein et al. 2019). We suggest that the central state faces a continuous struggle to integrate regions, but that the conception of 'us' in the region and 'the elites' in the centre may never be removed. Rural areas, particularly the distant periphery, require constant attention and responses. Here, we are particularly likely to find latent conflicts that could be subject to regionalism.

Conclusions
Our study suggests that territorial reforms may have negative consequences for trust in central government politicians. In the current study, the negative consequences of territorial reforms for political trust are linked to organised mobilisation on the grounds of political-territorial collective identities. After the reform, trust in politicians was most diminished in the northernmost region, where there was a strong regional mobilisation based on a historic centre-periphery conflict. This is an explanation for mobilisation in line with basic elements in a Rokkan centre-periphery framework, where politicalterritorial collective identities can be mobilised and affirm an inherent tension between the centre and regions.
The present study's design as a quasi-experiment provides a foundation for indicating causation between different types of mergers, regionalism, and trust in politicians, and adds to the literature on county and municipal mergers. For political trust, whether it is a forced merger or not, does not necessarily matter the most. Moreover, establishing that territorial reforms and the way they are introduced may or may not affect trust in politicians; the decisive factor is the presence of political-territorial collective identities that can serve as the focus for mobilisation. This is relevant for policymakers when considering local mergers. This study is also relevant to the literature on political trust, as it adds to our understanding and the growing literature on the relationship between political and territorial collective identities and political trust (McKay, Jennings, and Stoker 2021).
As a limitation of this study, we acknowledge that it assesses the trust effects of mergers in a relatively short time period. We were not able to assess whether this would have long-term consequences, the assessment of which would require the conducting of new surveys. Moreover, our analysis is a lesson from a quasi-experiment in Norway for which the treatment groups were not randomly selected. Hence, we could not confirm the external validity of this experiment. However, we find it plausible that similar territorial reforms combined with regionalism may have negative effects on national integration. Policymakers should consider this before implementing territorial reforms. More information on the relationship between territorial reforms and political trust from similar experiments in other countries would be of both interest and use to scholars in the field, and we highly recommend the implementation of further research focusing on the relationship between regionalism and political trust. Notes 1. To this, it should be added that the status of the provincial tier of government in Norway has been highly contested since it was reformed from an assembly of the municipal mayors to a council of directly elected representatives in 1975. Some of the most prominent political parties have called for its abolition ever since. 2. E.g., the merger of Hedmark and Oppland to form Innlandet was formally voluntary, but only after the final bill passed by the national parliament (Stortinget) stated that the counties would be forced to merge if they did not do so voluntarily. For robustness checks, we have run models categorising Innlandet as forced merged (see Appendix C). 3. https://www.klassekampen.no/article/20170607/ARTICLE/170609976 4. For Trøndelag, out data does not allow us to separate the respondents living in the old counties, Nord-and Sør-Trøndelag. We therefore treat respondents from these counties (voluntary merger) as one territorial unit. This has no consequence for our analysis as we do not hypothesise a difference in treatment effects between respondents living in the northern and southern region of Trøndelag. 5. https://forfinnmark.no/