Overcoming barriers to institutional integration in European second-tier urban regions

ABSTRACT Within the debates about the socio-economic advantages of cohesive urban regions, several barriers to institutional integration are said to exist, especially when a metropolitan government is absent and integration relies on inter-municipal cooperation. Some barriers are associated with different urban region structures, such as the asymmetric power relations and sociocultural contrasts between municipalities in systems with dominant core cities, or the lack of a leading city to overcome fragmentation and provide a shared identity in polycentric urban regions (PURs). This paper investigates whether urban regions formed around second-tier cities, whose features depart from both dominant core and PUR models, are able to mitigate these barriers when pursuing integration strategies. The analysis relies on interviews with municipal leaders in three representative European case studies, examining how they perceive the barriers to inter-municipal relations in second-tier urban regions. The findings show that perceptions vary not only between regions, with the three cases following different trajectories of integration, but also within regions, according to the geographical and socio-economic context of municipalities and the legacy of past relations. In general, barriers to integration are not minimized without explicit efforts to rebalance power relations, approach political cultures, mobilize core city leadership and develop a metropolitan identity.


Introduction
Many studies have been produced about the economic and social advantages of large and functionally diverse cities (Camagni, Diappi, & Leonardi, 1986;Jacobs, 1969;Melo, Graham, & Noland, 2009). Although evidence is mixed and highly dependent on context and indicators (Cox & Longlands, 2016;McCann, 2016), larger cities tend to be more productive, have a larger labour pool and ensure greater socio-economic diversity and interaction, and more opportunities for consumption and innovation. In parallel, cities hosting many and varied functions become more attractive for people and firms and sustain greater population growth (Glaeser, Kolko, & Saiz, 2001).
Today, many cities have coalesced into large and interconnected urban regions, meaning that some of these benefits can be achieved through integration with larger territorial scales. This has led policy-makers and researchers to advocate the formation of urban regions 1 as a way to build upon the potentials of larger population, functional mass and diversity, and stronger political influence in national and supranational contexts (BBSR, 2011;Otgaar, van den Berg, van den Meer, & Speller, 2008). Tighter functional and spatial integration is key to unlock the 'metropolitan potential' of urban regions (Lambregts, 2006), because a simple aggregate of nearby centres does not provide the same level of agglomeration benefits as a single large city of equivalent size (Lambooy, 1998;Meijers, 2008). To support this process, Ahrend, Farchy, Kaplanis, and Lembcke (2015) and others add institutional integration to the drivers of economic growth, arguing that less fragmented governance frameworks and collaboration in urban regions can contribute to harnessing the potential of size and diversity that emerges at that scale. Especially in contexts where local governments have sufficient authority and resources, they become a key vehicle of such integration, shaped either as a metropolitan government, municipal mergers, or local authorities working in cooperation.
However, earlier research has detected several barriers to institutional integration in urban regions. These are arguably more visible when a dedicated government level is absent and integration relies mainly on inter-municipal cooperation networks. Without a higher tier imposing metropolitan-wide decisions, such networks become vulnerable to the existing power relations and individual willingness to cooperate (Andersen & Pierre, 2010;Nelles, 2013), as well as to the incentives or disincentives to integration from national governments or supranational authorities. Some barriers can be associated with the different structures of urban regions, which define the shape of inter-municipal relations. For instance, several authors have analysed the difficulties of integration in urban regions formed around highly dominant core cities (for simplicity, here named 'primate urban regions'), due to the perception of asymmetric power relations and benefit distribution caused by the excessive socio-economic, cultural and demographic contrasts among jurisdictions (Feiock, 2007;Lefèvre, 1998;Nelles, 2009). On the other extreme, integration strategies in so-called polycentric urban regions (PURs), consisting of nearby urban centres of similar size, are hampered by the lack of a leading city bringing other stakeholders together, and of a sense of shared identity across the urban region (Dieleman & Faludi, 1998;Lambregts, 2006).
The question occupying the rest of this paper is whether urban regions departing from these two basic types also tend to overcome these barriers. The answers may give indications about not only how the spatial structure of urban regions affects their potential for institutional integration, but also what forms of inter-municipal collaboration are enabled in certain regions to mitigate potential barriers. This can become relevant for policy, even more so if parallel research is conducted on other dimensions of integration, such as the formal and informal cooperation networks carried by the multiplicity of actors participating in urban region governance (Sohn, Reitel, & Walther, 2009).
The next section will frame the question for the case of second-tier urban regions and describe the research approach. Section 3 specifies the existing barriers affecting intermunicipal collaboration in primate urban regions and PURs. The findings are presented in Section 4, based on interviews with municipal leaders about their perception of such barriers and whether they are overcome in their urban regions. Finally, Section 5 discusses the results and offers some conclusions.

The capacity for integration of second-tier urban regions
Recent research has positioned second-tier cities, or medium-sized cities in general, somewhere between the 'primate' and 'PUR' models above, in terms of how their urban regions emerged and are configured today. On one hand, they are not large and powerful enough to act as 'first-order' core cities, playing a historically dominant role over the surroundings, decentralizing activity hierarchically from core to periphery, and projecting their economic and political agenda over a relatively passive hinterland. On the other hand, they are not necessarily part of a PUR of similar centres, and can still be the 'leading'but not 'dominant'city of the region, acting as its main cultural and economic centre without overriding morphologies, development agendas and functional structures elsewhere. Cardoso and Meijers (2016) recently showed that second-tier urban regions have indeed a more balanced functional distribution than first-tier urban regions, which concentrate more functions in the core, leaving the remaining region underserved. Cardoso (2016) detected contrasts between the socio-economic sorting of population in some second-tier and first-tier urban regions, more fragmented and balanced across smaller centres in the former, and more hierarchic and spatially differentiated in the latter. These distinctions build upon earlier analyses contrasting opposing modes of urban region formation around large capitals and smaller cities (Hohenberg, 2004). An important contribution was Champion's triple typology of urban region development (2001), which associates the primate and PUR models to 'centrifugal' and 'fusion' types, but adds an intermediate 'incorporation mode', whose growth processes, functional structures and relations between centres reflect the description above.
So the question is whether second-tier urban regions closer to this intermediate model also enable a political landscape that reduces the barriers to integration specifically associated with the other two typologies. Compared to primate urban regions, can there be greater willingness for inter-municipal collaboration due to a less dominant core city, a greater sense of functional interdependence and fewer cultural and socio-economic contrasts between municipalities? And compared to PURs, can they profit from a leading city able to mobilize its greater resources and inspire change? Can such a city still have sufficient historical and cultural importance to cast its identity upon the urban region, anchoring the 'metropolitan idea' (Nelles, 2013) to a symbolic focus to which other actors are willing to adhere?
The broader question is whether second-tier urban regions have an added ability to pursue successful integration strategies. Integration is defined here as achieving governance capacity in issues that mobilize the urban region, by both coordinating territorial activity beyond individual jurisdictions and becoming a relevant actor in policy-making at higher levels of power. This is important for future policy directed at second-tier cities, as research suggests that harnessing the strengths of the urban region can bring them greater agglomeration benefits and political advantages than to first-tier cities. First, they can achieve a greater relative increase in population, compared with larger cities which already play a dominant role in the region (ESPON, 2005). Second, secondtier cities tend to have fewer urban functions than their population would imply (BBSR, 2011), making them likely to rely on functions located elsewhere to increase their mass and diversity; and their surroundings are indeed functionally better served than in firsttier urban regions (Cardoso & Meijers, 2016). Third, especially in countries dominated by large capitals, second-tiers tend to be weak interlocutors when dealing with higher levels of government, so joining forces to become a larger demographic and economic player may give them a stronger voice and avoid being overlooked by policy (Martin, Pike, Tyler, & Gardiner, 2015).

Research approach
These questions will be approached through a comparative account of three European case studies representing different spatial, historical and political contexts, as the answers are likely to be affected by place-specific conditions. The analysis is mainly based on interviews with municipal political leaders about the perceived contexts and the extent of inter-municipal collaboration.
The cases are Porto (Portugal), Antwerp (Belgium) and Bristol (the UK), whose interest relies on four features. First, with metropolitan populations between 1.1 and 2 million, they belong to the European 'second-tier city' typology (ESPON, 2012). Second, they are consistent with the description of second-tier urban regions departing from dominant core and PUR types: while leading core cities exist, these do not overconcentrate population and activities and are embedded in densely populated and polycentric urban areas, as Table 1 shows. Third, they lie in countries with important capitals, whose gap to second-tiers is significant 2 (ESPON, 2012), and may have an additional incentive to pursue integration for the 'upscaling' purposes discussed above. Fourth, they do not have a metropolitan government, making them privileged places to observe how governance capacity depends on the strength of inter-municipal collaboration. 2.1.1. Data sources and focus of analysis Inter-municipal collaboration can be explored by its outcomes, analysing changes in selected indicators and inferring causalities; by its outputs, describing the implementation of projects and plans; and by the perceptions and expectations of the stakeholders about the ongoing processes (Otgaar et al., 2008). The assessment of outcomes and outputs is relevant for future policy options, but does not tell us much about the balance of roles among actors and the underlying ethos of collaboration. Indeed, the question here is about the willingness of actors to engage in such collaboration, namely whether they recognize features in their urban region which have a positive effect on that willingness. This depends on their views about the existing power relations, the potential for joint gains and the future visions around which the incentives for integration congregate, recommending an approach based on evaluating perceptions, although they are subjective: 'perceptions of poweroften tied to historical identities and rolesprove to be more important than actual distributions of power based on measures such as population and economic clout' (Nelles, 2009, p. 301). The methodology to obtain primary data relies on open-ended interviews with key institutional actorsmayors or city councillorsas municipalities are assumed to be the main drivers of institutional integration strategies in the absence of a metropolitan government (Feiock, 2004). Elected politicians provide particularly relevant insights about the research question (more than 'disinterested' technocrats and civil servants, for instance) due to their effective decision-making powers, the potential spectrum of their mutual relations (from cooperation to rivalry) and their need to respond to the preferences of their electorate in the form of political action.
Of course, considering only this category restricts the conclusions, as institutional relations are only one of the determining factors of integration; further dimensions of cooperation, involving the civil society, firms, NGOs, etc., intervene to create a diffuse web of governance (Sohn et al., 2009). However, the specific barriers argued here as affecting specific types of urban region are primarily the outcome of factors shaping intermunicipal relations. Other barriers, affecting other dimensions of integration, may be typical of dominant cities or PURs, and their identification is a welcome future step of this research agenda. At this stage, the interviews cover the potential barriers presented above: the heterogeneity and balance of power relations, the political gap between core and peripheral locations, the role of the leading city and the existence of a metropolitan identity.
Nine municipal representatives were approached (three per case study). The strategy considers respondents from the core city and two smaller centres, one closer to the main agglomeration and one in a more remote area (respectively, the Morphological Urban Regions (MUA) and Functional Urban Region (FUA): see Table 1 and ESPON, 2007), due to the expectation of contrasting views according to geographical and socioeconomic contexts. This careful geographical distribution of respondents means that a significant part of the socio-spatial variety of each urban region is illustrated, partly compensating for the small sample. Also for this reason, an effort was made to make every interview as individualized and in-depth as possible, rather than summarizing the perspectives emerging from a larger set of conversations. Overall, this approach adds some explanatory value and representativeness to the analysis.
In the end, seven interviews out of nine intended were effectively conducted, between June and November 2015. The difficulties concerned Bristol, where only a councillor and member of the mayoral cabinet from the core city was available. Representatives from the other local authorities justified their unavailability with busy schedules. To substitute for the missing perspectives, an analyst from an urban policy think tank was interviewed about their recent research about city-region formation in the West of England. In the city of Porto, two councillors and members of the mayoral cabinet came to the interview, while in the MUA and FUA municipalities, the respondents were the mayors. In the city of Antwerp, the intended councillor was eventually replaced by a departmental director due to agenda incompatibilities (who nonetheless provided a political perspective), while the mayors were interviewed in the MUA and FUA. The interviews were conducted face to face (in one case via Skype), audio recorded and fully transcribed.
A respondent bias is possible: while the participants were not chosen according to the previous knowledge of their preferences, they were informed of the purpose of the research. Therefore, a preference for integration strategies is expected, or at least an awareness of their potential benefits. However, the point is not whether such strategies are welcomed in second-tier urban regions, but whether and how the barriers visible in other configurations are minimized. Another bias lies in misleading or incomplete answers by the respondents, who could be interested in exploring political rivalries or overstating the role of their municipality for political gain. This can be mitigated by triangulating inconsistent answers. Nonetheless, given the nature of the questions, the discrepancy between respondents is a valuable finding in itself: the level of subjectivity is revelatory of how different geographical and socio-economic contexts influence the perceptions about particular topics.

Factors of inter-municipal collaboration in urban regions
In contexts where inter-municipal collaboration is the main driver of integration strategies, issues involving the formation of voluntary networks of interdependent actors become more prevalent than elsewhere. This is broadly the preference of the 'new regionalist' approach to governance (Kantor, 2008;Nelles, 2013). This perspective posits that a variety of networks with different interests, policy goals and geographical scopes can emerge as temporary collaborations, without a definitive institutional consolidation. The approach opposes both the 'metropolitan model', insisting on a single political authority for the urban region, with legitimacy emerging from direct elections and autonomy from other levels of government, and the 'public choice' approach, which advocates that fragmentation into small units of government will enhance competition, allow meaningful locational decisions by citizens and firms, and generally improve the delivery of services. From the opposition between the two not very communicative models (Feiock, 2004), new regionalism emerged as a synthesis, as it does not demand a rigid metropolitan government, but privileges a progression through cooperation, coordination and policy integration stages (Stead & Meijers, 2009) over fragmentation and competition. The barriers discussed below mainly affect this type of collaboration model.

Barriers in primate urban regions: asymmetric relations and political cultures
The decision to cooperate is a rational choice, in which actors compare the transaction costs with the potential benefits (Feiock, 2007). In general, municipalities collaborate more when they perceive symmetrical mutual relations, anticipate joint gains emerging from collaboration and recognize their interdependence in problems mobilizing the larger scale. However, Feiock (2007) adds, information is imperfect and preferences are inconsistent, leading to choices influenced by political, geographical and thematic contexts. Contexts change perceptions, which in turn influence the willingness to take risks and cooperate. Contexts can be shaped by the mutual relations of individual actors, and Feiock provides an example: economic, demographic and social homogeneity is positively related to the likelihood of collaboration. In urban regions with greater homogeneity among municipalities, interests are more likely to be uniform and nobody is in a dominant position to either impose the distribution of benefits or exit the negotiation. Excessive heterogeneity hampers integration in regions with highly dominant core cities, and 'municipalities in the position of dominant central city often face particular barriers' (Rayle & Zegras, 2013, p. 871). Asymmetrical power relations reduce the will to cooperate, especially in the absence of a metropolitan government mediating the territorial competition between jurisdictions (O'Brien & Pike, 2015).
Empirical research supports these views. In an early survey of metropolitan governance in Europe, Lefevre writes about the failure of 'technocratic projects dominated by the central cities ' (1998, p. 21). He notes that the most successful projects were led, or strongly supported, by the central cities, but only when they were ready to make concessions. These included accepting roles in the collaborative structure which did not fully reflect their demographic or economic weight in the region, and reshaping their boundaries to reduce gaps between jurisdictions. Interestingly, this 'first among equals' approach has been often implemented in second-tier urban regions: Lefevre gives the examples of Bologna and Stuttgart, to which the more recent cases of Manchester or Turin could be added. Boudreau, Hamel, Jouve, and Keil (2007) add a political culture dimension to the problem with the example of how the mindsets of the 'old downtown' and the 'metropolitan region' in Canadian urban regions are opposed. This gap is reflected, among other factors, in the political preferences of the population, often associated with greater conservatism in the 'suburbs' and left-leaning tendencies in socially mixed core cities (Hoffmann-Martinot & Sellers, 2005). This affects the profile of local authorities answering to specific electorates and further impedes integrated governance.

Barriers in PURs: leadership and identity
PURs consisting of proximate urban centres of similar size and importance are demographically and economically more homogeneous, reducing the risk of core city dominance and political-cultural gaps. But they also experience barriers to integration, emerging from the lack of leadership and the absence of a clear metropolitan identity. Post (2004) writes that strong political leaders act as a unifying force among heterogeneous populations and, by extension, other political leaders. They work as linking mechanisms between actors, acting as a 'broker with sufficient power and connections to bring stakeholders together and introduce the idea of collaboration' (Rayle & Zegras, 2013, p. 871). They may have privileged access to policy-makers at higher levels of government and thereby persuade them to obtain more benefits to their urban region. The fact that this type of leadership emerges more often in a core city is not unusual. Larger municipalities have greater human, financial and technical resources, which they mobilize to launch initiatives and develop projects that smaller centres cannot lead (Lefèvre, 1998). They usually elect higher profile mayors with important positions in national political parties and close ties to higher levels of government. This leading role by a city that stands out from the rest is usually absent in a PUR system.
The second aspect is identity. Dieleman and Faludi (1998) argue that styles of urban governance in the different cities of PURs are 'worlds apart', and cities profile themselves separately rather than as part of a greater whole. This restricts the scale of place attachment of the population and the organizing capacity of PURs compared to large centres with a single administration: Lambregts (2006) asks how 'metropolitan qualities' could be added to the Randstad, as they appear much lower than in 'real metropolises such as Paris, London, Madrid and Milan […] turning the Randstad into a "potential" metropolis at best or a "powerless", disjointed collection of middlesized cities at worst ' (2006, p. 119).
Shared cultural, historical and geographical references in the urban region may contribute to add such 'metropolitan qualities'. Nelles writes that, alongside the institutional system and the opportunities for collaboration, what matters for integration is 'the strength of the metropolitan idea' (Nelles, 2013(Nelles, , p. 1359, carried by the networks that emerge across the urban region. Networks act as conduits to transmit ideas and actions, linking policy areas and territorial scales and compelling municipalities to join in to remain relevant in the policy decisions. Therefore, the incentive to engage in inter-municipal cooperation is partly influenced by the number and quality of the networks, embodied by the concept of 'civic capital'. Urban regions 'with greater civic capital are more likely to produce stronger governance partnerships' (Nelles, 2013(Nelles, , p. 1360.
Low civic capital may result from the absence of a common identity in PURs, as competing conceptions of problems and priorities emerge from different understandings of the urban region. In contrast, stronger and more inclusive networks may appear in places sharing important references, as common objectives and interests are made evident across the urban region and more easily agreed upon, building bonds across territories. The resulting track record of collaboration, in turn, encourages more networks and further integration. This is likely to be facilitated by the presence of an important core city casting a shared historical and cultural identity over the larger region, allowing different places to recognize a common context and develop a 'script' about the future that holds the network together: Hajer, van't Klooster, and Grijzen (2010) describe how the implementation of this concept in an urban region was fuelled by the imaginations around the core city of Amsterdam. Place names illustrate this well: Euroregion, Mitteldeutchland and Randstad are policy names distant from the public imaginary, unlike urban regions (non-PURs) which kept the name of their core city, like Manchester, Milan or Lyon. As Nelles writes (2013, p. 1353, '[c]apacity cannot be attributed to an abstraction'.

Inter-municipal relations in second-tier urban regions
The hypothesis emerging from this discussion is that some second-tier urban regions may have characteristics that mitigate some barriers to integration affecting primate urban regions and PURs. On one hand, the absence of a highly dominant core city and the potentially more balanced distribution of functions and socio-economic groups may reduce asymmetric power relations and political-cultural gaps. On the other hand, the presence of an economically, culturally and historically leading city may ensure resources and leadership and cast a strong identity over the urban region, to which other actors are willing to adhere. This will be explored in the cases of Porto, Bristol and Antwerp. The next section describes the institutional context supporting inter-municipal collaboration. Then, the perceptions about the symmetry of power relations and the gap in political cultures are assessed, to verify the presence of 'primate city' barriers. Finally, the perceptions about the type of leadership of the core city and the existence of a metropolitan identity are examined, in order to detect the presence of 'PUR' barriers.

Institutional frameworks for inter-municipal collaboration
In Porto, the respondents criticized the existing model of metropolitan governance, a free association of municipalities with no additional political authority and a tight budget, admitting that inter-municipal collaboration is the main driver of urban region projects. But they lack the budgetary or juridical mechanisms to override place-blind central government decisions or the parochialism of individual municipalities, and the metropolitan power void makes municipalities pursue smaller, informal and purpose-built partnerships. They appreciate the freedom to choose partners and purposes, and some successful projects have emerged, but see those networks as fragile, temporary and unable to address the desired scale, compared to the advantages of a metropolitan authority: becoming a stronger player when dealing with the central government or the EU, supporting institutions which were formally created but never implemented (such as the metropolitan transport authority) and providing long-term stability to decisions which now can be reversed by a change in municipal cabinets. They stressed that the central state explicitly hampers metropolitan integration, by restricting the available juridical and financial mechanisms, and then removing important regional policies from their power with the argument that the municipalities are not managing them properly.
The long history of entrenched municipal autonomy in Flanders makes the concept of an Antwerp 'urban region' problematic. The Flemish government and the Provinces advocate the formation of 'collaborative clusters' and municipal mergers, but, unlike Porto, the respondents did not give it special importance and seemed to think about the territory in terms of autonomous municipalities, cooperating as and when needed. Despite its restricted power and resources, many cooperation areas and objectives covering larger scales are defined by the Province (although its delimitation does not correspond to the functional and morphological spaces proposed by researchers), reducing the incentive for municipalities to form their own partnerships. They build networks for pragmatic needs and do not seem to yearn for a more developed integration model, although they recognize the present ineffectiveness and redundancy. Both smaller municipalities oppose integration in a metropolitan authority led by Antwerp.
Following a history of changing boundaries around Bristol, the West of England Partnership, later expanded to the West of England Local Enterprise Partnership (LEP) to include businesses, was created as a joint committee of four local authorities, lacking an elected mayor or a combined authority with encompassing powers. That partnership eventually embarked, like other English cities, in the first wave of City Deals in 2012. Inter-municipal collaboration now happens within this framework, designed for cityregions to obtain further funding and policy-making powers and define their own economic strategies. This model enables decentralization, forces city-regions to work together and mobilizes local knowledge rather than centrally imposing sectoral and place-blind decisions (O'Brien & Pike, 2015). However, the government tends to favour urban regions with a track record of collaboration, and preferably a combined authority, which can be seen as a consistent player when presenting their claims. Therefore, unlike the negative impact of central government in Porto and the apparent indifference of political leaders in Antwerp, Bristol suffers strong pressures for integration, which, at the time of writing, have not yet been addressed by the creation of an institution to capture the devolved powers.

Symmetry of power relations
As Table 1 shows, Porto, Antwerp and Bristol are not demographically dominant core cities and are part of larger, densely populated urban regions. However, Antwerp has fewer secondary centres (only Mechelen counts as a secondary MUA) and the core city is much larger than the second largest municipality (493,517 against 81,927), with an overall high number of small municipalities. The distribution in Porto and Bristol suggests a greater inter-municipal balance: in Porto, the core city is not even the most populated municipality and there are several centres with over 100,000 inhabitants. In Bristol, there are four large local authorities in the partnership, all with important centres.
Heterogeneity affects the perception of symmetrical relations in Antwerp. While the core city respondent stressed their commitment to bring other actors into a collaborative structure, they are clearly the dominant partner able to impose the matters to discuss: 'If one community takes some measures, they have an effect on the neighbours. This is true for Antwerp, because if we really push something through, the effect will be quite large' (officer at Antwerp municipality, interview, 2015). This is seen by the other partners as a dominance ambition. Mayors in the core MUA and FUA municipalities described Antwerp as having an effective power over smaller places and the last word in common decisions. They criticize the approach of Antwerp in joint debates as 'shut up and learn', but seem to consider that a 'fact of life' deriving from the 'natural' location of power in the region (mayor of FUA municipality, interview, 2015). Peripheral areas tend to develop networks between them, expanding the scale of their activities while bypassing the core city.
In a region known by the spatial dispersion of its population, functions and economic activities (De Meulder, Schreurs, Cock, & Notteboom, 1999), the fact that such dominance is seen as normal and unavoidable suggests that the 'urban region' idea is not a political priority (although arguably part of the political debate), and relations occur under an embedded core-periphery mindset. While this is not consistent with recent spatial evidence (Labo XX, 2012), historical factors and unchanged power relations are likely to conceal the actual spatial structure of the urban region and distort its recognition.
Porto and Bristol seem more aware of the interdependence of the urban region. In line with Cardoso and Meijers (2016), respondents in both cities argued that many of their important urban functions lie beyond their borders and that they need the larger scale to prosper. They gave examples of their commitment to build symmetrical relations with partners. Councillors in Porto stressed their reliance on the regionthe airport, the port, the trade fairs, industry and logistics are all outside the city. The new executive (elected in 2013) described the quick progression towards the 'integrated policy' stage of collaboration (Stead & Meijers, 2009) in several projects, arguing that waiving leadership and choosing partnership roles below their economic and symbolic weight were crucial for the perception of symmetrical relations between actors, as suggested by Lefèvre (1998). The mayor of a smaller neighbouring municipality recognizes the 'intelligence' of this approach, anticipating a tendency for increasingly balanced relations. Both respondents have attempted a flattering comparison to Lisbon, where smaller centres, in their view, need to confront core city dominance and thereby strive for as much autonomy as possible.
The Bristol city councillor similarly stressed how much of the region's economic power sits outside their cityin advanced industry, higher education or tourismand how interdependence with Bath and other centres is essential for joint growth. To mobilize that scale, authorities in Bristol argue that they 'want to be an equal partner […] we are very keen not to dominate, but to have a kind of sub-region that works well together' (Bristol city councillor, interview, 2015). The councillor agreed that capital cities may have greater difficulty in achieving this balance, as they are culturally more distinct from their wider region, and, being the seat of governmental and financial power, there is an incentive to centralize decisions and resources.
Collaboration also seems to be facilitated by spatial contiguity. The Bristol councillor believes that working together under symmetrical relations is more efficient across an uninterrupted urban fabric, despite municipal boundaries, and this arguably affects the pace of development in the different areas covered by the LEP. The same happens in Porto, where spatial contiguity is argued as the trigger of several inter-municipal projects. The respondents in the core city and the MUA municipality agree that an urban continuum eventually dissolves asymmetrical power and balances the gains of collaborations among actors. In their view, this partly mitigates the lack of a metropolitan authority, which would be important to find common logics with the more remote places of the urban region, unclear about the benefits of collaboration. Indeed, the mayor of the small FUA town developed this view, stressing the weak relations between the core agglomeration and the rest. In contrast to examples given by other respondents, he believes that only urban functions 'expelled' by the core can locate in other places, and that Porto disregards cooperation attempts: 'everyone understands its leadership as the most relevant centre. It is not, regardless, a city which had the ability to move beyond its administrative boundaries. In my opinion, Porto is a small Lisbon' (mayor of FUA municipality, interview, 2015).

Proximity of political cultures
The gap in political cultures between core cities and their urban regions is often expressed by the contrasting political preferences of the population, and their approximation is seen as a positive influence on inter-municipal collaboration. Antwerp, Porto and Bristol illustrate the different shapes this can take, but the political leaders from the core city and smaller centres had different views about its importance for integration.
Antwerp is a case of increasing uniformity across the urban region. The historical contrast between a conservative periphery and a left-wing core city had made collaboration difficult, with mayors from different parties answering to contrasting electorates and life ambitions. This partly changed in the 2012 elections, when the conservative New Flemish Alliance (N-VA) party won in most municipalities, as well as in the province. Forty-two out of 50 municipalities in the urban region now have centre-right mayors.
Porto shows increasing political proximity to the core agglomeration, not followed by the remaining urban region. The Portuguese municipal map changed in 2013 with the surprising victory of independent candidates in several cities, including Porto and its neighbouring municipality Matosinhos, while the dominance of the traditional political parties was kept elsewhere. Interestingly, the first new inter-municipal network formed after the elections, the Atlantic Front, involves precisely Porto and Matosinhos, as well as the contiguous city of Vila Nova de Gaia, where independents came second.
Bristol represents a more pronounced gap between core and periphery, in the sense that even the institutions governing the different local authorities are different. Bristol is one of the few English cities with a directly elected mayor, while the other three authorities are governed by an assembly of councillors. Furthermore, the mayor of Bristol (at the time of the interviews) was an independent, while the surrounding authorities are dominated by centre-right forces.
The radar charts show the distribution of the political spectrum in the different parts of the three urban regions (Figure 1).
The perceived importance of political proximity was stronger in Antwerp than elsewhere and the core city official credited it for the first steps of an integrated response to issues like housing and infrastructure, enabling collaboration which would not happen otherwise. This view is not shared by the smaller municipalities, which argue that proximity is not producing greater willingness to collaborate due to the political dominance of the core, enhanced by the fact that the mayor of Antwerp is also the national leader of the N-VA party. Note that many smaller municipal cabinets formed mayoral coalitions with the N-VA but depend on local political parties, with semi-autonomous agendas.
It is unclear whether the contrast between the renovation trend led by independents in and near the core and the preservation of the political status quo elsewhere amounts to two political cultures in the Porto urban region. The respondents did not explicitly admit that, although they acknowledged how both tendencies did cluster geographically. Similarly to Antwerp, what emerges from this clustering are different perceptions in the core and the periphery. Core city councillors stress that political renovation has brought mayors together and allowed blocked collaboration efforts to resurface. The FUA mayor criticizes the narrow scope of that trend. Initiatives like the Atlantic Front, mentioned above, are seen by some media as 'Porto-centralist' and dismissive of the urban region as a whole (Mendes, 2013), creating new fault lines in the urban region.
The case of Bristol, with different structures governing that city and the other local authorities, presents a deeper challenge. The figure of a mayor is associated with personal leadership and charismatic qualities, illustrated by the idea of Bristol electing an independent who embodies the city's 'sense of exceptionalism' (Byrne, 2014). By contrast, the Council, popular in England, is associated with collegial decision-making and less symbolic capital. This gap may be difficult to surmount. The Bristol city councillor mentioned the confusion around the proposal of a West of England 'strategic mayor', due to the suspicion by the local authorities that the mayor of Bristol, being the only 'mayor' around, would eventually occupy that position, igniting fears of core city dominance.

Core city leadership
There is a fine balance between using core city leadership to provide an impulse to partnerships and imposing an undesired hierarchy over the urban region. Core city officials in all three cases stressed their willingness to mobilize substantial technical, financial and human resources to integrate partners and achieve results. However, they do this in different ways, and create different perceptions about their role.
Porto city councillors described their approach through the 'theory of the natural leader someone who we just recognise as a leader even if not chosen or wanting to be one. If we apply this to cities, we find the same logic' (Porto city councillor, interview, 2015). This is a 'first among equals' position, in which the leading city gets to choose their role, but opts for greater equality: they establish the networks, mobilize the resources to support and finance them, but then retreat to a secondary position in the implementation stage (Lefèvre, 1998). In Antwerp, as anticipated, core city leadership is strong, and the respondent admits that the city can pursue an agenda for the urban region due to their much larger resources. Integrating other partners is preferable, but they are ultimately able to take unilateral decisions, an approach regretted by the neighbouring municipalities. Bristol has decided to waive some leadership pretensions, according to the city councillor interviewed, to reduce fears of their neighbours about their role in a future combined authority. Moreover, the region is composed only of four large local authorities, which avoids that a single actor has much greater resources than others, although some differences persist.
Such differences may be embodied by the charismatic core city mayors, who are, in all cases, quite high profile and vocal. The mayors of Porto and Bristol at the time of the interviews shared important features: they were elected as independents and managed to build cohesive cabinets with people from different political forces; they had a well-regarded professional background rather than being full-time politicians and they appeared to embody the personal characteristics that make their city special, such as independence, cosmopolitanism and innovation (Byrne, 2014;Minder, 2013). All these features arguably matter for a city-regional leader, but these mayors have no formal powers outside their jurisdictions, which restricts their capacity to lead integration. That capacity is informally present in Antwerp, whose mayor is also 'the most important man in Belgium' (mayor of MUA municipality, interview, 2015). However, his potential influence over higher tiers of policy-making and unofficial authority over other actors in the region are hampered by the strongly parochial politics and not expected by others to extend their benefits beyond the core.

The strength of the metropolitan idea
Inspiring a 'metropolitan idea' is arguably the greatest strength of Porto, seen by the respondents as a symbol of regional identity increasing their willingness for integration. The city has successfully developed its image and name into a 'brand', which is useful for firms in the region (presenting themselves as from 'Porto') and for smaller municipalities (presenting their amenities as components of the Porto offer). A variety of networks are championing this urban region agenda, mostly building upon the 'Porto' designation. Examples include the regional tourism strategy, industrial and commerce chambers, and associations of universities and businesses.
This sense of identity is not limited to firms and institutional networks. The respondents noted that the population is generally oriented to Porto as a cultural focus and that mobility 'puts us all in the same city' (mayor of MUA municipality, interview, 2015). Again in comparison to Lisbon, core city councillors argue that moving across the region is not only practical, but also bidirectional: unlike the capital, where peripheral inhabitants travel daily to the centre, but those working and living in the core hardly need to leave, in Porto both peripheral and core populations can cross several municipalities in daily commuting. This arrangement increases socio-economic diversity in smaller centres, reduces city hierarchies and increases the cultural proximity between places, developing 'an emerging state of mind oriented to integration' (Porto city councillor, interview, 2015). The respondents described the urban region partly as a 'single large city': the mayor of the FUA municipality stated that 'we want to be the industrial district of Porto ' (interview, 2015), in reference to the former industrial area, historically inside the city. The mayor of the MUA municipality compared the urban region to Greater London: a set of adjacent 'boroughs' with their own identities building a large city in need of joint governance.
Like in other aspects, perceptions change from core to periphery, and the mayor of the FUA municipality does not feel that this 'state of mind' necessarily brings practical benefits for his constituents, such as better health services or transport. He would welcome a metropolitan authority to distribute those benefits, and sees its absence as the greatest obstacle to the transformation of the strong metropolitan identity into a tool of territorial development.
A comparable notion of identity seems absent in Antwerp. The mayors interviewed do not see their municipalities in the position to integrate a 'greater city' vision, and showed greater interest in keeping local specificities and attending to everyday issues than profiling their territories as part of Antwerp, alleging too distinct economic and cultural profiles. The core city alone is the key economic player and they hope to profit mainly by attracting residents. Although the region is overall very urbanized, networks championing a joint economic agenda have not developed, and investment agencies either promote solely Antwerp or the whole of Flanders, with no intermediate scale. According to the Flemish government, only up to 25% of all associations (municipal, NGOs, etc.) in Antwerp have inter-municipal scope, making it the most isolated municipality in the urban region. Smaller municipalities tend to have more inter-municipal networks, but usually bypass the core city (Agentschap Binnenlands Bestuur, 2014).
This mirrors the sense of identity of the population. The politicians interviewed openly referred to the typical views of the core city by peripheral populations as 'dangerous, dirty, ethnically tense and poor'. The feeling seems mutual, as core city inhabitants 'see themselves as a metropolis' and are aloof about the surroundings (mayor of MUA municipality, interview, 2015). As described by the respondents, these contrasts are caricatural generalizations, but they reveal embedded core-periphery ideologies that restrict any attachment to an 'urban region' identity. Nonetheless, this perception may change as the movement of population and firms becomes more diverse and complex, and there were also references to Antwerp as the focus of culture and leisure, and to people across the urban region appreciating that fact over traditional rivalries.
The 'metropolitan idea' in Bristol is stronger than in Antwerp. But it is not, like Porto, a case of the core city projecting its identity across the urban region: integration is conceived under the 'West of England' designation or the 'Bristol-Bath' bipole. According to the core city councillor, there is a sense of interdependence between places, but not a cultural orientation towards Bristol. The historical reliance on autonomous urban centres, he adds, made cities like Bath retain a strong sense of individuality. Despite 'cross-fertilization ' (interview, 2015), there is no willingness to waive individual identities or profile smaller centres as 'part of' Bristol, like in Porto. In this sense, the approach to integration seems closer to a 'PUR' model, which arguably explains the tensions associated with the proposed combined mayoral model, when seen from outside the city.
Under that framework, several networks have formed to champion the urban region agenda, in tourism, investment agencies and transport, although they may not yet possess the longevity and leadership components seen by Nelles (2013) as necessary to carry the metropolitan idea further. Due to the failure to interview representatives from the surrounding local authorities, their perceptions about an urban region identity cannot be discussed. However, Boddy and Hickman write that the four partners 'have rarely, if ever, been natural allies ' (2013, p. 751). They illustrate the existing tensions with the opposing views about urban growth and housing delivery, as Bristol is surrounded by a green belt and additional growth necessarily falls on the surrounding authorities. They suggest a contrast between urban and 'hinterland' mindsets which may lower the potential to develop a common identity.

Discussion and conclusions
The absence of a metropolitan government turns Porto, Antwerp and Bristol into privileged arenas to observe how integration is driven by the shape and strength of intermunicipal partnerships. With no higher tier structure imposing cooperation and masking tensions, it becomes more evident that tighter integration depends primarily on the willingness to collaborate of individual municipalities. The interviews have shown that such willingness depends on the perceptions by the actors about their mutual relations, but perceptions vary greatly between and within urban regions. This paper starts by putting together a systematized list of barriers to institutional integration and highlighting their association with 'primate' and 'polycentric' urban region types. As for the question whether second-tier urban regions departing from such types are able to minimize those barriers, first we must consider the limitations of the present findings. These result primarily from the size and the nature of the sample. Focusing only on municipal relations certainly neglects other types of urban region networks which play an important role in creating an integrated institutional space (for instance, shaping the 'metropolitan idea'). However, the specific barriers under analysis are very much a result of the different shapes of those municipal relations, which justifies the focus of this paper, while opening up pathways for research about other dimensions of integration.
On the other hand, the sample is admittedly small to create an exhaustive portrait of inter-municipal cooperation in each urban region: it might be that other municipalities develop different perceptions and approaches. However, as discussed earlier, the strategic geographic sorting of the individual interviews did produce important illustrations of what is likely to occur in other places. The perceptions of institutional actors in close-by contexts are interdependent and prone to influence each other and to cluster in recognizable patterns. Therefore, while the scope of the research provides incomplete stories about each case study, it highlights patterns likely to illustrate an important part of the attitudes towards integration in these and other second-tier urban regions.
Several conclusions can be drawn from this analysis. First, although all cases are second-tier cities embedded in larger and densely populated urban regions, Porto and Bristol are more consistent with the research hypothesis than Antwerp. The respondents in the former two cities believe they are distinct from capital cities, mentioning the greater sociocultural contrasts between city and surroundings in capitals, as well as their different spatial-functional structures, arguing that their urban regions have greater functional interdependence and weaker hierarchies between places, and that they need the strengths of the wider regions to prosper, with the core city offering the necessary leadership. They believe that the barriers to integration visible in capitals are thereby reduced. This reflects the contrasts between types of urban region argued by Cardoso and Meijers (2016) and others.
However, this narrative is typically told by the core cities and becomes more unfamiliar as the distance to the centre grows. Perceptions vary notably within urban regions, with core city respondents generally in line with the research hypothesis and existing spatial evidence (however thin), and more remote and spatially detached places perceiving greater imbalances between themselves and the 'core' and unsure about the distribution of benefits. As discussed, a factor likely to affect this variation is morphological contiguity, seen as dissolving asymmetries and increasing the sense of interdependence. Urban region policies advocating horizontal relations between partners must always consider the perspectives of individual localities about the benefits of integrating the urban region, as well as how an excessively large, spatially discontinuous and heterogeneous territory may restrict the functioning of an inter-municipal model.
The examination of existing barriers also stresses the differently perceived power relations in Antwerp, Porto and Bristol, supporting the arguments by Feiock (2007) about the role of excessive scalar contrasts between municipalities. This happens even if the core city does not dominate the second-tier urban region 'as a whole' (see Table 1): the perceptions of individual municipalities about their own relevance in comparison to others count more. The resulting power asymmetries between partners affect their willingness to collaborate more than the opportunities brought by the proximity of political colours or benevolent core city leadership, as the case of Antwerp suggests, providing clues about the different weight of the barriers under analysis. Following Nelles (2009, p. 301), both 'actual' and 'perceived' asymmetries restrict cooperative intensity.
Therefore, a second aspect distorting the perception of the conditions for integration is the way historical roles and identities and unchanged power relations, which are not finetuned with the territorial structure, can mask the actual spatial configurations of urban regions. It becomes clear that barriers to integration are not mitigated (just) by the spatial, functional or demographic structure of the urban region, but need explicit efforts by partners aimed at actively shaping their mutual perceptions. The cities of Porto and Bristol show greater commitment to build horizontal relations with others and waive pretensions of dominance in the interest of better collaboration, as argued by Lefèvre (1998). But the way different partners react to that approach stresses the complexity of inter-municipal relations, which depends not only on the core city but also on the overall territory, including the legacy of past urban growth and the type of governance in smaller centres. Nelles (2009) argues that regions with more centralized configurations reveal more intense perceptions of disparities, which this analysis confirms. But asymmetric relations also emerge in apparently more balanced contexts, when parts of the urban region integrate at different speeds: in Porto, the opportunities emerging from political proximity near the core risk creating new fault lines and sending a message of elitism to the remaining territory.
The findings about the formation of urban region identity are restricted to the focus of the research. However, and despite the limitations of the institutional system, Porto is arguably the case where such an idea is most developed in the sense given in this paper: a leading but not dominant core city acting as a symbol of territorial identity beyond its boundaries, which becomes attractive for other actors to adhere to; they see themselves as part of an 'extensive city', with impacts in both the profiling of firms and institutions and the general attitudes of the population. This arrangement is strengthened by daily mobility, inter-municipal and multidirectional, both for core city and 'hinterland' inhabitants, reflecting recent research about how non-hierarchical mobility across municipal borders is upscaling the sense of place attachment of populations, no longer confined to the city, but developing into a 'metropolitan identity' (Kubler, 2016). Such insights can guide transport or housing policies directed at intensifying this 'extensive city' idea.
In comparison, the sense of interdependence and complementarity in Bristol is not materialized by the projection of the core identity over the urban region, but by a new framework under which individual city identities are negotiated. This is tied to historical identities and roles and closer to a PUR approach to integration. Antwerp, regardless of how much its spatial structure and institutional system resemble the other cases (namely Porto; similarities between the configuration of urbanization in Northwestern Portugal and Flanders can be found in the accounts by Portas, Domingues, &Cabral, 2007 andDe Meulder et al., 1999, for instance), is characterized by a 'little capital' approach to integration. This points to a detachment between spatial and socio-economic changes happening on the ground and the time they eventually reach policy-makers and institutions in order to become operational and change historically embedded attitudes. Notes 1. In different countries also known as city-regions, metropolitan regions, etc. The term 'urban region' is used here to denote the type of territorial form generally associated with these definitions. 2. Antwerp is an ambiguous case, as different criteria provide different perspectives on the polycentric structure of the Belgian urban system. However, for the issue of political and economic importance considered here, the primacy of Brussels is clearly stated in the ESPON (2012, p. 28) report; see also Hall and Pain (2006).

Disclosure statement
No potential conflict of interest was reported by the author.

Funding
This work was supported by the Fundação para a Ciência e a Tecnologia (Portuguese Ministry for Science, Technology and Higher Education) [grant number SFRH/BD/80157/2011].