Locality as a safe haven: place-based resentment and political trust in local and national institutions

ABSTRACT Recently, an emerging body of research has diagnosed a ‘crisis of trust’ in rural and peripheral areas. Yet, the majority of these works overlook local institutions as a potential alternative source of governance. Relying on original, nationally representative survey data from nine European countries, this article suggests that local institutions act as a safe haven for individuals with high levels of place-based resentment. Substantially, for these individuals, political trust in the local level significantly exceeds political trust in the national level. This dynamic is largely driven by concerns over representation. The stronger citizens feel that their place is overheard, the more they trust local relative to national institutions. This nuances the narrative of the ‘crisis of trust’ in an important way. Rather than conceiving it as a crisis haunting the political system in general, it seems to be better understood as a crisis of the national level specifically.


Introduction
In recent years, a number of studies have identified a 'geography of discontent' in Europe (McCann, 2020, for an overview, see Ejrnaes et al., 2023).This broader phenomenon is frequently linked to a 'crisis of trust' in rural and peripheral areas.On average, people living in these places are commonly found to exhibit significantly lower levels of trust in politics (e.g., Kenny & Luca, 2021;McKay et al., 2021;Mitsch et al., 2021;Stein et al., 2021), are less satisfied with the way how democracy works (Lago, 2022), and tend to be more supportive of authoritarian forms of government (Zumbrunn & Freitag, 2023).Similarly, several studies from the United States document an increasing resentment of rural Americans towards political elites, culminating in a perception that politics are fundamentally rigged against them (e.g., Cramer, 2016;Hochschild, 2016;Jacobs & Munis, 2023;Munis, 2022;Wuthnow, 2018).In this regard, a growing feeling of placebased resentment is often argued to undermine political trust, supposedly leaving people to turn their backs on the political system as a whole.
However, drawing such a conclusion might be premature.Crucially, the majority of existing studies overlook local institutions as an alternative form of governance, individuals with high levels of place-based resentment can turn to.Local councils and assemblies play an important role in citizens' daily lives, having gained more and more autonomy over the last couple of decades (Ladner et al., 2016(Ladner et al., , 2021)).Moreover, it is well established that citizens tend to trust local institutions substantially more than national institutions (for an overview, see Muñoz, 2017).Usually, these differences in trust are ascribed to local institutions delivering policies that match citizens' preferences more closely, and by enabling better opportunities for participation in politics (for overviews, see Mouritzen, 1989;Ostrom, 1972).Bringing these ideas together with the benchmarking literature on European Union (EU) support (for an overview, see Hobolt & De Vries, 2016), in this article, I suggest that these qualities should make local institutions a safe haven for individuals with high levels of place-based resentment.Building on works conceptualising support for European integration as rooted in a comparison between the merits of national and European governance (e.g., De Vries, 2018;Sánchez-Cuenca, 2000), I propose a similar benchmarking exercise, where individuals compare the virtues of local and national institutions.For people with high levels of place-based resentment, I then argue, that the result of this comparison should be particularly biased towards the local level.
I test this argument by relying on original, nationally representative survey data, fielded in nine European countries (Czech Republic, Denmark, France, Germany, Greece, Hungary, Italy, Poland, and Spain) among more than 9,000 respondents.In line with previous research, I also find a pronounced gap between trust in the local level and national trust, where the former considerably exceeds the latter.This gap, in turn, significantly widens with increasing levels of place-based resentment.My findings further suggest that the main channel underpinning this dynamic is connected to grievances relating to representation.The more individuals feel that their place does not get enough attention from politics, the more they trust the local relative to the national level.Furthermore, I find that this mechanism is largely restricted to individuals living in rural areas and somewhat confined to contexts that are characterised by stronger local institutions.Overall, the results of my analysis nuance the 'crisis of trust' narrative in an important way.Rather than viewing it as a crisis of the political system as a whole, it is better understood as a crisis of the national level specifically.
The remainder of this article proceeds as follows.I start by conceptualising place-based resentment and by reviewing the existing literature on geographic divides in political trust.I then develop my argument of how the local level may act as a safe haven by introducing the literature on EU benchmarking and discussing three possible channels through which place-based resentment may sway citizens' comparisons towards local institutions.After spelling out my hypotheses, I present my data, empirical strategy, and results.I conclude by outlining some broader implications of my findings, suggesting that besides place-sensitive economic policies, continuing to devolve political authority to local institutions might present another viable option to counteract the 'crisis of trust' in rural and peripheral places.

Theory
Place-based resentment and the 'crisis of trust' The importance of place in politics has been well established by landmark works in political geography (e.g., Agnew, 1987).One way in which places matter, is that they can act as markers of group membership, forming a central part of an individual's social identity (e.g., Proshansky, 1978;Tajfel, 1981).These place-based identities, in turn, can create powerful in-group biases, underpinning a degree of affective polarisation between people living in different areas (Lyons & Utych, 2023).Besides this, they also carry profound political consequences.For instance, numerous studies find that voters are responsive to place-based appeals, preferring political candidates from their own locality (e.g., Jacobs & Munis, 2019;Munis, 2021;Munis & Burke, 2023;Schulte-Cloos & Bauer, 2023).Furthermore, strong local attachments have been related to increased turnout (Borwein & Lucas, 2023b), voting behaviour along the cosmopolitan-nationalist divide (Hegewald & Schraff, 2022), and support for far right political parties (Fitzgerald, 2018).
Place-based resentment, by contrast, can be conceived as an extension of place-based identities.Following Cramer (2016), it describes a feeling, where 'an identity rooted in place … is infused with a sense of distributive injustice' (p.12).Place-based resentment, therefore, goes beyond a mere attachment to place, by combining place-based identities with a sense that one's own place is treated unfairly, in particular, concerning economic, representational, and cultural issues (Huijsmans, 2023;Munis, 2022).In this regard, place-based resentment comprises three major components, connected to perceptions that (a) one's place is getting less resources than just (economic component), (b) policy-makers pay too little attention to the concerns of people living in one's place (representation component), and (c) place-specific ways of life are looked down upon by people living in other places (cultural component) (Cramer, 2016;Huijsmans, 2023;Munis, 2022).
Overall, place-based resentment is argued to result in a deep-seated belief that the political system is fundamentally rigged (Cramer, 2016).In particular among rural residents in the United States, place-based resentment has been connected to anti-intellectualism (Lunz Trujillo, 2022), support for Donald Trump (Lunz Trujillo & Crowley, 2022), and voting for the Republican party (Jacobs & Munis, 2023).Ethnographic works further underline this.For instance, Hochschild (2016), documents a growing distrust of rural Americans in government, which they view as mostly benefiting elites living in cities.Likewise, Thompson (2023) finds that high levels of place-based resentment significantly predict lower levels of political trust among voters living in rural America.Studies from Europe paint a similar picture, suggesting a growing gap in political trust between people living in rural and peripheral areas, on the one hand, and individuals living in urban centres on the other (e.g., Kenny & Luca, 2021;McKay et al., 2021;Mitsch et al., 2021;Stein et al., 2021).For instance, studying urban-rural disparities in political trust over the last two decades, Mitsch et al. (2021) conclude that rural residents are increasingly losing faith in politics, and suggest that this might partly be explained by a general dissatisfaction with the economy and public service provision in these areas.In connection to this, Stein et al. (2021) investigate centre-periphery divides in political trust in Norway, finding that individuals living further away from the national capital exhibit considerably lower levels of trust in politicians.In addition, comparable trends are also reported with regard to related, more diffuse, measures of political support.In contrast to city dwellers, individuals living in the countryside are found to be more in favour of authoritarian regimes (Zumbrunn & Freitag, 2023) and tend to be less satisfied with the way how democracy works (Lago, 2022).
In light of these findings, many of these works tend to echo a narrative of a 'crisis of trust' that is already widespread in the more general literature on political support (see Van der Meer, 2017).Political scientists and politicians alike often share a 'deeply rooted concern with political trust', frequently claiming that trust in politics is deteriorating (Van der Meer & Zmerli, 2017, p. 1). 1 From this perspective, political trust is central to the survival and legitimacy of democratic governance (Miller, 1974, p. 951).Low levels of political trust are, therefore, commonly regarded as a warning sign for an erosion of democracy itself, and may thus fundamentally challenge the stability of contemporary democratic political systems (Dalton, 2004).In this regard, existing studies on geographic divides in political trust tend to be framed around these concerns, implying that place-based resentment essentially subverts the viability of democracy altogether. 2  However, drawing the conclusion that there is a general 'crisis of trust' might be premature.While the existing literature provides valuable insights into geographic divides in political trust regarding a number of different institutions, it tends to omit local institutions as a fully-fledged layer of democratic governance. 3Yet, in particular, institutions located at lower levels could provide an attractive alternative for individuals with high levels of levels of place-based resentment.Over the past few decades, the local level has gained considerable influence, with more and more authority being transferred to municipal councils and assemblies (Ladner et al., 2016(Ladner et al., , 2021)). 4In this respect, it has been well documented that political institutions located at lower tiers of governance consistently elicit higher levels of political trust than national-level institutions (e.g., Chang & Chu, 2008;Cole & Kincaid, 2000;Denters, 2002;Goot, 2002;Pharr et al., 1997;Uslaner, 2001, for an overview, see Muñoz, 2017).Commonly, these differences are explained by an inverse relationship between jurisdictional size and political trust, with smaller jurisdictions being regarded as closer to the people living in them.In this sense, local institutions are often argued to produce policy outcomes that are more in line with citizens' preferences and to foster better opportunities for citizen participation in political decision-making (for overviews, see Mouritzen, 1989;Ostrom, 1972). 5Precisely these qualities, I suggest, should make local institutions a safe haven for individuals with high levels of place-based resentment.

Benchmarking approaches to support for European integration
My argument largely draws on the benchmarking literature on support for European integration.Essentially, this approach to EU support revolves around citizens using the national context they live in as a benchmark to evaluate the merits of European integration (for an overview, see Hobolt & De Vries, 2016, pp. 421-423).The origins of this argument can be traced back to the work by Sánchez-Cuenca (2000) who suggests that citizens living in countries with poorly performing political institutions 'have little to lose by opting for "more" Europe' (p.148).Building on this, Harteveld et al. (2013) make the argument that citizens who face bad performance at home, 'might actually start trusting the European level because it provides them with an alternative source ofpotentially bettergovernance' (p.549, emphasis in original).Similarly, Muñoz et al. (2011) advance the notion of a 'trustworthiness standard', where citizens use their trust in domestic institutions as a reference point when deciding whether to trust institutions at the European level (for similar arguments, see Kritzinger, 2003;Rohrschneider, 2002).More recently, De Vries (2018) has further developed these arguments, proposing that national performance evaluations inform a comparison between the benefits associated with the status quo of EU membership and the benefits associated with a hypothetical exit from the EU.Following this, individuals are argued to support European integration when ill-performing institutions at home render a scenario outside of EU less attractive.
However, these benchmarking approaches are contested.Various studies suggest that trust in the EU might simply be viewed as an 'extrapolation' of trust in national-level institutions (e.g., Ares et al., 2017;Armingeon & Ceka, 2014;Harteveld et al., 2013;Lipps & Schraff, 2021;Muñoz et al., 2011;Talving & Vasilopoulou, 2021;Torcal & Christmann, 2019).This idea is based on the assumption that citizens' knowledge of political affairs at the European level is too limited to allow for an independent evaluation of the EU, forcing citizens to rely on national-level cues to compensate for this information deficit (Anderson, 1998).In this respect, these studies propose that trust in national-level institutions 'spills over' to the European level, resulting in congruent levels of trust across different layers of governance (Ares et al., 2017).
In terms of empirical evidence, which logic prevails appears to depend on the level of analysis employed.At the individual level, studies usually find extrapolation effects, while at the country level a benchmarking dynamic tends to be present (e.g., Lipps & Schraff, 2021;Muñoz et al., 2011;Torcal & Christmann, 2019).Yet, when it comes to the local level, simply reducing trust in local institutions to a mere extrapolation of national trust seems to be less appropriate.While individuals may struggle to independently evaluate institutions at the European level, this is less likely when it comes to local institutions which are arguably much closer to citizens' daily lives.Studies showing that local performance assessments matter for citizens' trust in local institutions further underline this (Fitzgerald & Wolak, 2016;Steenvoorden & Van der Meer, 2021).In addition, recent works analysing individuals' ability to differentiate between different levels of governance, document that citizens are capable to separately evaluate the trustworthiness of institutions located at different levels (Proszowska et al., 2022(Proszowska et al., , 2023;;Schneider, 2017).
Against this backdrop, I propose that trust in the local and national level are underpinned by a similar benchmarking exercise, where individuals compare the benefits of local and national institutions.Here, the benefits of local institutions should play an especially important role for individuals with high levels of place-based resentment.Essentially, individuals who feel that their locality is short-changed of its fair share of resources, attention, and respect might specifically blame national politics for these injustices, while they rely on local institutions to make up for these deficiencies.In this regard, I suggest that this dynamic may operate through three distinct channels each relating to one of the three components of place-based resentment outlined above.

Locality as a safe haven: place-based resentment and political trust in national and local institutions
One of the most basic understandings of trust views the concept as the evaluation of a relationship, where 'A trusts B to do x' (Hardin, 2000, p. 26).From this perspective, an evaluative conceptualisation of political trust implies that citizens of a political community somewhat rationally assess the performance of political institutions (see Easton, 1975, p. 449).While these performance evaluations may relate to a wide variety of policy outputs, much of the existing literature on political trust focuses on economic performance as a critical source of trust in politics (e.g., Miller & Listhaug, 1999;Van der Meer & Hakhverdian, 2017, for an overview, see Van der Meer, 2018).In relation to this, the economic component of place-based resentment may represent a first channel biasing citizens' evaluations of national and local institutions towards the latter.Research on the political effects of inequality has demonstrated that an unequal distribution of available resources considerably undermines political trust and engagement (e.g., Anderson & Singer, 2008;Solt, 2008).Connected to this, persistent regional inequalities have been found to substantially erode citizens' trust in national institutions (Dellmuth, 2023;Lipps & Schraff, 2021), suggesting that failure to remedy these inequalities by means of increased regional redistribution might be perceived as a policy failure specific to the national level.Schraff and Pontusson (2023) further substantiate this link by finding a strong association between regions lagging behind economically and far right voting in national elections.Thus, individuals with high levels of economic place-based resentment might perceive national institutions as wasteful, complacent and ineffective at solving the issues they face in the place where they live.In contrast, local institutions may be viewed in a more favourable light, delivering policy outcomes that are closely tailored to local preferences (see Mouritzen, 1989, pp. 663-664).
Besides economic performance, the quality of democratic procedures also matters for political trust (Van der Meer & Hakhverdian, 2017).In this regard, a second possible channel could be connected to the representation component of place-based resentment.Individuals who feel that policy-makers pay too little attention to the place where they live, might perceive local democratic processes, and their ability to foster citizen engagement, as a better venue to voice their concerns than distant institutions at the national level (see Dahl & Tufte, 1974).This perspective aligns with recent research by McKay et al. (2023) who document a prevailing belief that government tends to favour capital cities, implying that national institutions are often perceived as representing areas that are already located at the locus of political power.Connected to this, a number of studies looking at municipal mergers, resulting in an increase in size, empirically underpin this phenomenon.For instance, research on the aggregate level, commonly finds that municipal amalgamations depress voter turnout (e.g., Allers et al., 2021;Heinisch et al., 2018;Lapointe et al., 2018).Similarly, individual-level studies point towards the same direction, showing that increases in size as a consequence of municipal mergers have a negative effect on levels of political efficacy, satisfaction with government, and trust (e.g., Hansen, 2013Hansen, , 2015;;Lassen & Serritzlew, 2011).
In addition to evaluations of economic performance and democratic procedures, political trust can also originate from a person's 'long-standing and deeply seeded beliefs about people that are rooted in cultural norms' (Mishler & Rose, 2001, p. 31).Departing from this understanding of political trust, the cultural component of place-based resentment may also sway citizens away from national institutions and more towards the local level.When feeling that people from other places disrespect the cultural peculiarities of one's own place, local political institutions might be perceived as champions of these place-specific values and norms defending them against political developments at the national level (see Rokkan & Urwin, 1983).

Hypotheses
In light these considerations, the local level may then potentially be regarded as an alternative source of governance, or safe haven, capable of filling a void left by the national level.In this sense, when comparing the merits of both local and national governance, local institutions should become relatively more attractive to national institutions with increasing levels of placebased resentment.Although place-based resentment can be viewed as a general sentiment that the area where one lives is treated in an unfair way (see De Lange et al., 2023;Huijsmans, 2023), it is commonly conceptualised in reference to the urbanrural divide (Cramer, 2016;Munis, 2022).Following this, place-based resentment is somewhat unidirectional, with rural residents resenting urban elites, rather than the other way around.In line with this, current research finds that placed-based resentment and its effects for political attitudes and behaviour tend to be more pronounced among ruralites than urbanites (Borwein & Lucas, 2023a;Jacobs & Munis, 2019;Munis, 2022).One reason for this could be that political and urban elites are often equated with one another (see Cramer, 2016;Hochschild, 2016;Wuthnow, 2018), making urbanites a clearer political out-group than ruralites.This, in turn, implies that ruralites should be more biased towards the local level, while urbanites, often already wielding the levers of power, should have a lower incentive to turn towards the local level in relative terms.
H 2 With increasing levels of place-based resentment, the increase in local trust relative to national trust should be more pronounced for rural residents than urban residents.
Lastly, whether the local level can indeed act as an attractive alternative to the national level, should also depend on the degree of authority devolved to local institutions in a given context.After all, if local institutions are rather weak to begin with, they are less likely to be regarded as capable of delivering better policy outcomes, providing a more suitable venue for citizen engagement, and as a defender of place-specific values and norms.Comparative research already hints at this.Using data from 15 European countries, Fitzgerald and Wolak (2016) find that the gap in trust between local and national institutions is larger in federal than in unitary systems, concluding that citizens are more inclined to trust local over national institutions 'when local governments hold political decision-making power' (p.136).Furthermore, when local institutions are weaker, they might also be less visible to citizens.This would make an independent evaluation of their trustworthiness more difficult and a benchmarking exercise, where citizens compare the relative merits of each level, less likely.
H 3 With increasing levels of place-based resentment, the increase in local trust relative to national trust should be more pronounced in contexts with higher levels of local autonomy than in contexts with lower levels of local autonomy.

Data and methods
All three hypotheses are tested by relying on data from an original online survey among 9,114 respondents fielded in nine European countries (Czech Republic, Denmark, France, Germany, Greece, Hungary, Italy, Poland, and Spain).The countries were selected to provide for a representative picture of the EU and differ on a variety of dimensions, including geographic region within Europe, economic performance, population density, and degree of local autonomy as shown in Table 1.Data were collected via online access panels administered by the survey company Bilendi.Nationally representative quotas for age, gender, education, and NUTS-2 region were applied. 6 My dependent variable is operationalised following De Vries (2018), forming a local trust differential by subtracting respondents' trust in their national parliament from respondents' trust in their local council.Positive values indicate that local trust exceeds national trust, while negative values mean the opposite. 7In this regard, this variable measures the extent to which individuals prefer local-level institutions relative to national-level institutions, which is precisely what my hypotheses stipulate.In plain words, if the local level indeed serves as a safe haven, then the local trust differential should become more and more positive with increasing levels of placebased resentment.
My main independent variable, respondents' place-based resentment, is measured by means of a scale adapted from Munis (2022), using the five survey items shown in Table 2. Respondents answered each item on a 5-point Likert scale ranging from 'strongly disagree' to 'strongly agree'.The items were designed in such a way that they correspond to the economic, representation, and cultural components of place-based resentment.All of the items take respondents' respective geographical in-groups and out-groups along the urban-rural divide into account.In this regard, I follow the established literature on place-based resentment, operationalising the concept as a phenomenon that primarily unfolds as a conflict between cities and the countryside (Cramer, 2016;Hochschild, 2016;Munis, 2022;Wuthnow, 2018).
Respondents' geographic in-groups and out-groups were classified by preceding the place-based resentment questions with a survey item asking respondents whether they think they live in a 'very rural', 'rather rural', 'rather urban' or 'very urban' place. 8Using these self-classifications, respondents' in-groups were defined as rural if they described themselves as living in a 'very rural' or 'rather rural' place, while their out-groups were defined as urban.In contrast, if respondents described themselves as living in a 'very urban' or 'rather urban' place, their in-groups were shown as  urban, whereas their out-groups appeared as rural.I rely on respondents' selfclassifications instead of more objective measurements of urban-rural location, as this more closely approximates the place-based identity component inherent in place-based resentment (see Nemerever & Rogers, 2021).All hypotheses are tested by estimating ordinary least square (OLS) regressions with fixed effects at the country level.H 1 is tested by regressing the local trust differential on a scale combining all five place-based resentment items after conducting a factor analysis (Cronbach's α = 0.83). 9To test H 2 , I interact the place-based resentment scale with a dummy variable indicating urban or rural residence based on respondents' self-classifications introduced above. 10H 3 is tested by first splitting the sample into two subsamples comprising countries with lower and higher levels of local autonomy respectively.I classify countries into either sample according to their score on the Local Autonomy Index (LAI) in 2020 (Ladner et al., 2021) as shown in Table 1.Countries are classified as part of the lower or higher group if their respective level of local autonomy is below or above the sample mean (M = 63.42). 11Since I expect the relative increase in local trust to be more pronounced for rural residents (H 2 ), I further split each of these sub-samples into an urban and rural sample.This leaves me with four sub-samples comprising rural residents in either higher or lower local autonomy contexts and urban residents living in countries with stronger or weaker local institutions.I then estimate separate OLS regressions, modelling the relationship between place-based resentment and the local trust differential, for each of these four subsamples.All models employed in the analysis control for a range of potential confounders including standard demographic variables such as gender, age, education, and income as well as social trust, respondents' left-right selfplacement, place-based identity, and government-opposition voting.

Results
Before testing each hypothesis separately, Figure 1 presents some first descriptive insights.Panel A illustrates the result of a paired t-test between local and national trust, while panel B shows the distribution of the local trust differential.Both Panel A and panel B rely on the full sample.Panel C and panel D show the results of two paired t-tests between local and national trust for samples restricted to urban and rural residents respectively.There are two main takeaways from this figure.First, the trust gap between the national and the local level, as documented in other studies (e.g., Chang & Chu, 2008;Cole & Kincaid, 2000;Denters, 2002;Goot, 2002;Pharr et al., 1997;Uslaner, 2001), also exists in the data at hand.As shown in panel A, average trust in the local level (M(local) = 4.96) is considerably higher than average trust in the national level (M(national) = 3.83), t(9,113) = 40.94,p < 2.2e-16.This is also reflected in the distribution of the local trust differential shown in panel B. While a large share of respondents trusts the local level and the national level to equal parts (i.e., the local trust differential is at 0), the majority of respondents tends to trust the local level more than the national level (i.e., the local trust differential is positive).This is also evidenced by the slightly positive mean of the local trust differential (M = 1.13). 12Second, the trust gap is larger among ruralites than it is among urbanites.While the paired t-tests in panel C and panel D show that local trust is higher than national trust in both the urban (t(6,395) = 29.97,p < 2.2e-16) and rural sample (t(2,717) = 28.96,p < 2.2e-16), the gap in trust between institutions tends to be more pronounced for rural (M(local) -M(national) = 5.21 − 3.70 = 1.51) than urban residents (M(local) − M(national) = 4.85 − 3.89 = 0.96).Table 3 presents the results of three OLS regressions, testing H 1 by regressing the local trust differential on the place-based resentment scale.In Model 1, the coefficient for place-based resentment is positive and statistically significant, indicating that with stronger levels of place-based resentment, the local trust differential becomes more positive.Put differently, the more respondents feel that their place is cheated in one or another way, the wider the gap between local and national trust.This finding remains robust when control variables are added in Model 2 as well as in Model 3 and counts as evidence in favour of H 1 .
To test H 2 , Figure 2 plots the predicted values of the local trust differential by place-based resentment conditional on respondents' urban and rural residence.Full model results can be found in Table A5 in Appendix A. 13 The local trust differential strongly increases with growing place-based resentment among rural residents, while it remains virtually constant for urbanites.In light of this, the increase in the local trust differential appears to be primarily confined to individuals from rural areas, thereby supporting H 2 . 14 The OLS regressions presented in Figure 3 further investigate through which of the three channels theorised above the increase in the local trust differential operates.Since the increase in the differential seems to mostly concern rural residents, the sample is split according to respondents' answers on the urban-rural self-classification item, estimating regressions   Note: OLS regression coefficients with country fixed effects.Thick and thin lines are 95% and 99% confidence intervals, respectively.Models control for gender, age, education, and income as well as social trust, left-right self-placement, place-based identity, and government-opposition voting.For full model results see Table A7 in Appendix A.
for an urban and rural sub-sample respectively.Instead of regressing the local trust differential on the full place-based resentment scale, an independent variable capturing each component of place-based resentment is entered into each model separately. 15Results for the full models are presented in Table A7 in Appendix A. Notably, the main channel that seems to underpin the widening of the trust gap between local and national institutions, appears to relate to the representation component of place-based resentment.Among rural residents, increasing feelings of lacking representation are associated with an increase in local trust, relative to the national level, while the coefficients for the economic and cultural components are not statistically significant at conventional levels.For urban residents, findings tend to be more mixed.Whereas the economic component correlates positively with the local trust differential, concerns over representation seem to narrow the gap between local and national trust.This might explain the somewhat constant slope observed for urban residents presented in Figure 2 as the effects for the economic and representation components at least partially cancel each other out.
Figure 4 plots the results of four OLS regressions testing H 3 , splitting the sample into contexts with higher and lower local autonomy and with regard to urban as well as rural residence.Full model results can be found in Table A8 in Appendix A. In line with H 3 , the increase in the local trust differential seems to be restricted to rural residents living in contexts characterised Note: OLS regression coefficients with country fixed effects.Thick and thin lines are 95% and 99% confidence intervals, respectively.Models control for gender, age, education, and income as well as social trust, left-right self-placement, place-based identity, and government-opposition voting.For full model results see Table A8 in Appendix A. Denmark, France, Germany, and Spain are classified as higher local autonomy contexts.Countries with lower local autonomy are the Czech Republic, Hungary, Greece, Italy, and Poland.by higher local autonomy.Here, trust in the local level increases relative to national trust, the higher respondents' place-based resentment.By contrast, the opposite appears to be the case for urban residents, where the gap between local and national trust appears to narrow down ever so slightly as place-based resentment increases.Furthermore, as expected by H 3 , both coefficients for urban and rural residents in lower autonomy contexts do not reach conventional levels of statistical significance.
Investigating the relationship between place-based resentment and the local trust differential for each country separately reveals some interesting exceptions to this more general pattern.As shown in Table A9 in Appendix A, among the countries with higher local autonomy, no statistically significant relationship between place-based resentment and the local trust differential can be detected in Spain.Furthermore, among the group of countries with lower local autonomy, as shown in Table A10 in Appendix A, the Czech Republic stands out, with rural residents' local trust differentials strongly increasing with growing levels of place-based resentment. 16Given these divergent country patterns, these results only count as partial evidence in favour of H 3 .
In summary, the analysis has shown that for respondents with high levels of place-based resentment, the local level acts as a safe haven.When feeling that their place is not getting its fair share, individuals tend to trust the local level significantly more than the national level.This mechanism seems to primarily work through a channel connected to the representation component of place-based resentment.Individuals who perceive that their place is not listened to, turn away from the national level and invest their trust in the local level instead.Furthermore, this pattern appears to be largely restricted to individuals from rural places, and somewhat more pronounced for respondents living in contexts with higher levels of local autonomy.

Discussion and conclusion
In rural and peripheral places, political trust is often argued to be in crisis.Rooted in a feeling of place-based resentment (e.g., Cramer, 2016;Hochschild, 2016;Jacobs & Munis, 2023;Munis, 2022;Wuthnow, 2018), individuals living in these areas tend to exhibit considerably lower levels of trust in political institutions (e.g., Kenny & Luca, 2021;McKay et al., 2021;Mitsch et al., 2021;Stein et al., 2021), are less satisfied with democracy (Lago, 2022), and seem to prefer authoritarian forms of government more frequently (Zumbrunn & Freitag, 2023).In this regard, existing research on geographic divides in political trust often echoes concerns over a growing number of individuals turning their backs on the political system as a whole.
Yet, this conclusion has been largely based on studies overlooking local institutions as an alternative layer of governance.Responding to this shortcoming, in this article, I have pointed out an important qualification to the narrative of the 'crisis of trust' prevalent in the existing literature.Drawing on works conceiving support for European integration as grounded in a benchmarking exercise (e.g., De Vries, 2018;Sánchez-Cuenca, 2000), I have argued that the local level acts as safe haven for individuals with high levels of placebased resentment.In line with this argument, I document a substantial gap between local and national trust, which is particularly large for individuals who feel that their place is treated in an unjust way.This dynamic appears to operate mostly through a feeling of lacking representation, with those who perceive their place not to have a voice in politics, trusting the local level much more relative to the national level.In turn, this pattern seems to be a primarily rural phenomenon, and somewhat confined to contexts characterised by higher levels of local autonomy.In light of this, the 'crisis of trust' in rural and peripheral areas should not be understood as a crisis haunting the political system as a whole, but rather as a crisis looming over the national level specifically.
In terms of implications for policy-making, these findings reinforce the importance to continue devolving authority to local institutions (for a similar argument, see McCann, 2020).Besides place-sensitive economic policies, redistributing resources from richer to poorer areas (see Lang et al., 2022;Vergioglou, 2023), giving local communities a greater say in their own affairs, might be another viable option to counteract the 'crisis of trust' in rural and peripheral places.After all, when the national level is perceived as not paying enough attention, providing institutions that are closer to people's daily lives with more tools to decide over their own fate could be a suitable way forward.
Importantly, some limitations remain to this study.First, the correlational nature of the article's analysis does not imply causal relationships.In this sense, I cannot preclude that political trust is itself also a driver of placebased resentment.While the causal arrows might indeed go both ways, from a theoretical perspective, it seems more likely that place-based resentment influences political trust and not the other way around.Second, the increase in the local trust differential seems to be absent among urban residents.This, however, might be a result of the specific operationalisation of place-based resentment employed here.Crucially, the out-group for urban residents is defined as people living in rural areas.Yet, this might not be the relevant out-group for urbanites in the context of this dynamic.Future research should further scrutinise this possibility by looking at alternative out-groups and by employing different operationalisations of place-based resentment.In particular, De Lange et al.'s (2023) measurement of regional resentment might be a good starting point in this regard.Lastly, the rather small number of country cases surveyed only allows for a somewhat limited analysis of possibly moderating institutional variables at the country level.A more systematic investigation of these factors, including local autonomy, but also electoral institutions, would be a welcome addition to the literature.

Notes
1. Whether trust in politics is indeed declining is fiercely debated in the existing literature (for overviews, see Thomassen et al., 2017;Van der Meer, 2017).2. Often, existing studies make this point implicitly.Nevertheless, some works explicitly suggest that there is a 'crisis of trust' in rural and peripheral areas (Mitsch et al., 2021, p. 1), with low levels of trust potentially leading to 'a breakdown in democratic institutions' (Thompson, 2023, p. 1) Besides looking at centre-periphery divides in political trust in national politicians, they also investigate whether people living in the periphery exhibit higher trust in local politicians than national politicians.Yet, they only find weak evidence in favour of this argument, concluding that citizens in the periphery tend to hold equally low levels of trust in both national and local politicians.They further speculate that this might be explained by local politicians being 'perceived as insignificant or as brokers for the central government' (Stein et al., 2021, p. 50). 4.This trend fits well with the more general development of authority migration away from the national level towards supranational and subnational layers of governance (see Hooghe & Marks, 2003;Hooghe et al., 2016;Hooghe & Marks, 2016;Hooghe et al., 2017). 5. Notably, some works propose the opposite, suggesting that political trust is higher in larger jurisdictions.One argument for this is that politics in smaller jurisdictions are characterised by closer interpersonal relationships and are, therefore, believed to be more prone to corruption (Dahl & Tufte, 1974).Furthermore, due to possible economies of scale, larger jurisdictions are presumed to deliver public goods more efficiently, which, in turn, is said to bolster political trust (Alesina & Spolaore, 1997).6.Data were collected from 07 February 2023 until 13 April 2023.Around 1,000 respondents were sampled in each country.For descriptive statistics of all variables employed in the analysis and data collection periods per country, please see Tables A1 and A2 in Appendix A. For the fullquestionnaire of the survey please consult Appendix B. 7. Details on the measurement of all variables employed in the analysis can be found in Table A3 in Appendix A. 8. Importantly, this self-classification variable may have primed respondents to their place-based identity, biasing the overall level of place-based resentment upwards.9. Please see Figure A1  indicates that this solution fits the data much better than a 1 factor solution (BIC = 3684.34).This underlines that political trust and place-based resentment are different latent constructs.For the factor loadings of this analysis please see Figures A2 and A3 in Appendix A. 10.Respondents are coded as rural, when they indicated that they live in a 'very rural' or 'rather rural' place, and are coded as urban when they have indicated otherwise.As a robustness test, I re-estimate all models using the full urbanrural self-classification item instead (see Table A4 and Figure A4 in Appendix A).
The substantive findings remain the same.11.Following this, I classify Denmark, France, Germany and Spain as higher local autonomy contexts and the Czech Republic, Greece, Hungary, Italy, and Poland as lower local autonomy contexts.Figure A5 in Appendix A shows the distribution of the LAI in 2020 for all 27 EU member states.Note, that the mean for the EU (M = 60.11) is slightly lower than the sample mean.In this regard, the average level of local autonomy is higher in the countries surveyed than in the EU as a whole.Furthermore, Greece, Italy, and Poland should be considered borderline cases, since their overall level of local autonomy can still be regarded as rather high (see Ladner et al., 2021, p. 71).12.Given that both trust questions were measured in the same battery, which has been shown to increase the correlation between different trust items (see Brosius et al., 2020), the local trust differential presented, might even underestimate the gap between trust in the local and the national level.13.While place-based resentment is commonly conceptualised with regard to the urban-rural divide (Cramer, 2016;Hochschild, 2016;Munis, 2022;Wuthnow, 2018), evidence from the Netherlands (De Lange et al., 2023) suggests that it can also be understood in connection to the centre-periphery cleavage.I further investigate this possibility by interacting the place-based resentment scale with a measure of respondents' perceived positions on the centre-periphery divide.As shown in Table A6 in Appendix A, none of these interaction terms reach conventional levels of statistical significance.Likewise, when including both an interaction term between place-based resentment and urban-rural residence as well as place-based resentment and centre-periphery residence, the interaction terms for urban-rural residence remain statistically significant, while the statistical significance for the interaction terms for centre-periphery residence seem to depend on the specification of the regression model.Interestingly, however, when specifying a three-way interaction term between the place-based resentment scale, centre-periphery residence and urban-rural residence, it appears that the increase in the local trust differential is driven by rural residents who simultaneously live in the periphery of their country (see Figure A6 in Appendix A).This pairs nicely with De Lange et al. (2023) who find that place-based resentment is most pronounced in rural areas that are located at the periphery.14.Looking at the constituent parts of the local trust differential among rural residents is also informative in this regard.As shown in Figure A7 in Appendix A, the increase in the local trust differential among rural residents mostly comes down to a decrease in national trust, while local trust remains constant across different levels of place-based resentment.
15.To operationalise the economic component, I simply rely on respondents' answers on the economic place-based resentment item.To capture the representation and cultural components, I calculate an average representation (Cronbach's α = 0.81) and cultural scale (Cronbach's α = 0.67) from the two items corresponding to each of these components.To ease interpretation, all of these three variables are standardised.16.For reasons of low statistical power due to small sample sizes, Tables A9 and  A10 in Appendix A only show the results of bivariate regressions between the local trust differential and place-based resentment.

Notes on contributor
Sven Hegewald is a PhD candidate at the Center for Comparative and International Studies, ETH Zurich.

H 1
The stronger individuals' place-based resentment, the more they trust local institutions relative to national institutions.

Figure 1 .
Figure 1.Local trust and national trust compared.Note: Panel A illustrates the result of a paired t-test between local and national trust for the full sample.Panel B shows the distribution of the local trust differential for the full sample.Positive values indicate that local trust exceeds national trust, while negative values signal the opposite.The dashed line indicates the mean value of the local trust differential.Panel C and panel D illustrate the results of two paired t-tests between local and national trust for the urban and rural samples respectively.The red dots represent the means with 95% confidence intervals.

Figure 2 .
Figure 2. Predicted values of local trust differential by place-based resentment conditional on urban-rural residence.Note: Based on OLS regression with country fixed effects.95% confidence intervals displayed.Model controls for gender, age, education, and income as well as social trust, left-right self-placement, place-based identity, and government-opposition voting.For full model results see TableA5in Appendix A.

Figure 3 .
Figure 3. OLS regression results: Local trust differential on place-based resentment components, by urban-rural residence.

Figure 4 .
Figure 4. OLS regression results: Local trust differential on place-based resentment scale, by lower and higher local autonomy and urban-rural residence.
in Appendix A for the factor loadings.Following De Lange et al. (2023), I have also conducted an additional validation analysis testing to what extent political trust and place-based resentment are different latent constructs.When comparing a 1 factor with a 2 factor solution, the lower BIC (Bayesian information criterion) of the 2 factor solution (BIC = 394.73)

Table 1 .
Overview of countries included in online survey.Data for GDP per capita and population density are measured in 2022 and stem from Eurostat.Data were accessed in November 2023.Local autonomy is measured by the Local Autonomy Index (LAI) in 2020(Ladner et al., 2021).
People in [in] work harder than people in [out], because in [in] it is harder to make ends meet.Note: Adapted from Munis (2022).If respondents see themselves as living in a very rural/rather rural place: [in] = rural areas; [out] = urban areas; If respondents see themselves as living in a very urban/ rather urban place: [in] = urban areas; [out] = rural areas.
Stein et al. (2021)sagree with the argument that a 'reservoir of political trust' is necessary to 'preserve fundamental democratic achievements' (Van der Meer & Zmerli, 2017, p. 1), a general 'crisis of trust' would imply that place-based resentment undermines political trust per se, regardless of the institution in question.As I further argue below, however, there are good reasons to believe that local institutions are sheltered from the adverse effects of geographic grievances.3.Stein et al. (2021)are a notable exception to this general pattern in the literature.