Investigating institutional barriers and opportunities to an integrated approach for transport and spatial development: Mega urban transport development in a rapidly developing city, Seoul

ABSTRACT In response to rapid growth, megacities make large investments in mega urban transport projects. To ensure positive social outcomes, e.g., accessibility for all, an integrated approach to transport and spatial development is needed. This paper identifies the institutional barriers and opportunities to using an integrated approach, using Ostrom’s Institutional Analysis and Development Framework (IAD). We investigated a metro (subway) project in Seoul, South Korea, a rapidly-growing metropolis that has established an extensive transport network. Our analysis revealed that various rules have led to imbalances between transport and spatial planning, contributing to undesirable outcomes. Incongruence in scope between transport and spatial development, a lack of clear procedures for integrated planning, and institutional ambiguity regarding the responsibilities of the various actors affected social outcomes. We conclude that achieving desired social outcomes from mega urban transport projects requires institutional harmonization across sectors and levels, and local capacity to operationalize an integrated approach.


Introduction
In response to rapid urban growth, many cities around the world have made major investments in mega urban transport projects (MUTPs).The international urban agenda has increasingly recognized the role of urban transport development in enhancing economic growth and the quality of life of the urban population, especially in rapidly-growing megacities (ADBI, 2017;UN-Habitat, 2017).In practice, however, MUTPs rarely consider the broader, long-term consequences for society, and therefore they often fail to meet all of the many varied interests in project outcomes (Jones & Lucas, 2012;Stopher & Stanley, 2014).Because project benefits and negative impacts are spatially, temporally, and socially differentiated (Vanclay, 2002), many authors have argued that urban transport development alone cannot enhance the well-being of the whole population across a city (e.g., Martens, 2012;Van Wee & Geurs, 2011).
Many scholars have argued for integrated approaches to transport planning and spatial planning as a key to sustainable urban development (e.g., Bertolini, 2012;Lee, 2020;Sung & Oh, 2011).However, discrepancies between urban transport and spatial development still exist in practice, contributing to a failure to achieve a fair distribution of accessibility and quality of life across cities (Heeres et al., 2017;Martens, 2012).Discussions about integrated approaches have focused on the institutional arrangements for integration (Legacy et al., 2012) rather than on how to operationalize the approaches and actually achieve the desired outcomes (e.g., accessibility for all).Ostrom (2011) emphasized the need to analyze relevant actors, institutions and their interactions, and the outcomes at multiple levels within the institutional setting and socioeconomic situation.Critical investigation of the complex interplay between urban transport and spatial development at multiple levels is essential in order to comprehend how the intended social outcomes from MUTPs are determined and delivered.
The purpose of this paper is to identify the institutional barriers to and opportunities for an integrated approach to transport and spatial planning.We use Ostrom's (2005) Institutional Analysis and Development (IAD) Framework.To our knowledge, not many researchers have analyzed context in addressing the social outcomes of MUTPs.We applied the IAD Framework to investigate how an integrated approach could be operationalized to influence the social outcomes of MUTPs.By "social outcomes," we mean the broad societal consequences, which can be positive or negative, that occur as a result of the implementation of projects and from the spatial transformations that are induced by projects-e.g., changes in accessibility to opportunities, quality of life, and the extent to which benefits are fairly distributed across an urban population (Geurs et al., 2009;Jones & Lucas, 2012).We used a metro (subway) development project in Seoul, South Korea as our exemplar.Like Beijing, London, Paris, and Tokyo, Seoul is an example of a large metropolitan capital city that has experienced rapid economic and population growth, and in response has developed an extensive metro/subway system and policies for integrating transport and land use.

Theoretical background: An integrated approach to urban transport and spatial development
Many scholars have argued that most MUTPs tend to focus on achieving project objectives in the short-term and they neglect the longer-term, broader outcomes from projects (Jones & Lucas, 2012;Stanley et al., 2017).Many desirable outcomes (e.g., enhanced accessibility and livelihood opportunities across a city, fair distribution of benefits among urban population) are rarely considered or addressed (Geurs et al., 2009;Lee, 2018;Switzer et al., 2013).
Achieving long-term benefits from MUTPs for the whole population requires attention, not only in relation to improved transport systems, but also in terms of appropriate changes to the spatial structure of cities, patterns of land use, and to the local environment (Bertolini, 2012;Geurs & van Wee, 2004;Mottee et al., 2020;Rodrigue, 2017).The social outcomes from urban transport development are closely related to the consequences of the spatial changes that occur at multiple scales (Geurs & van Wee, 2004;Jones & Lucas, 2012;Lee et al., 2020).Macro-scale spatial changes (e.g., polycentric urban development) and micro-scale changes (e.g., mixed-use land development and enhanced infrastructure and services around the nodes) jointly influence accessibility to opportunities across a city (Hall & Pain, 2006;Heeres et al., 2017;Loo & Verle, 2016).A lack of attention to the consequences of such changes will likely lead to a failure to address the fair and equitable distribution of desirable outcomes (Lee, 2020;Martens, 2012).
Many scholars (e.g., Banister, 2008;Cervero & Murakami, 2009;Straatemeier & Bertolini, 2020;Switzer et al., 2013) have advocated an integrated approach to urban transport and spatial development as being key to achieving desirable outcomes, such as environmental quality and the social wellbeing of urban populations.Heeres et al. (2016) and Lee et al. (2020) have argued for an integrated approach to transport and spatial planning at multiple levels, especially to ensure accessibility gains for all and a fair and equitable distribution of benefits.For example, at the metropolitan or city level, urban policies should focus on establishing transport networks that address the varying patterns of land use across a city (Martens, 2012;Wegener, 2004).At the local neighborhood level, integrating infrastructure development into adjacent land uses (e.g., station area development) and assessing the social consequences of spatial changes are essential (Rydin, 2010;Sung & Oh, 2011).The failure to consider multi-level integration will likely lead to unsuccessful and unintended social outcomes from MUTPs (Lee, 2020;Lee et al., 2021).
In practice, transport planning and spatial planning still occur in separate silos, each with its own paradigm, objectives, procedures, and priorities (Curtis & James, 2004;Legacy et al., 2012).Therefore, many potential benefits that would arise from the integration of urban transport and spatial planning are being missed (Van Geet et al., 2021).Earlier research has discussed the institutional barriers to an integrated approach, including organizational disintegration (e.g., Curtis & James, 2004), the lack of integrated planning instruments (e.g., Silva et al., 2017), and the lack of policy integration (e.g., Hull, 2008).Legacy et al. (2012) emphasized that simply establishing an integrated vision, strategic plan, or organizational structure will not necessarily lead to the intended outcomes.Achieving positive social outcomes in practice requires that an integrated approach be operationalized by actors and institutions at multiple levels (Heeres et al., 2016), which is what we focus on in this paper.An example of such an operationalization is the joint actions and networked coordination that should be established during policy formation and implementation being based on the mutual interests of the varied actors (Stead & Meijers, 2009).Curtis and James (2004) emphasized that the failure to deliver intended outcomes related to the disjuncture between strategic policymaking and the operationalization of an integrated approach (Van Geet et al., 2019).It is clear that facilitating desirable social outcomes from MUTPs requires operationalizing an integrated approach at macro and micro levels, and establishing institutional arrangements that support such an approach, from policy formulation to implementation.
To fully understand how outcomes are produced, the interactions between actors in the institutional setting of a project should be investigated (Alexander, 2005;Healey, 2003;Ostrom, 2005;Spijkerboer et al., 2017).Policy formulation and implementation are usually negotiated among actors at multiple levels in a planning process (Hooghe & Marks, 2003;McGinnis, 2011;Veeneman, 2018), in which various rules are established, taken up and used (Healey, 2003).Ostrom (2005) stressed that institutions-which include the formal and informal rules-shape a decision-making process by defining what actions and outcomes are required or prohibited, by whom, and what information should be exchanged among actors across all levels.These rules influence their interaction patterns and the flows of information between entities and across levels, thereby affecting project outcomes (Van Geet et al., 2019).The various rules concerning transport development and spatial development structure the interactions between actors, and influence the extent to which integration (of transport development and spatial development) is achieved.Ostrom (2007) also emphasized the importance of examining the broader contexts that structure planning processes (Cole & McGinnis, 2017).Many contextual factors, e.g., culture, politics, and the distribution of agency (power) influence MUTP development (Rode, 2019;Stead & Meijers, 2009;Veeneman, 2018).
Overall, critical investigation is needed into how an integrated planning approach can be operationalized so that it contributes to desirable social outcomes that arise from MUTPs.Such an investigation requires examining the dynamic urban transport planning and spatial planning processes at multiple levels, and considering the actors and rules in each specific context.

The Institutional Analysis and Development Framework
We used Ostrom's (2005) IAD Framework to explore the complex interplay between urban transport and spatial development processes at macro and micro levels in Seoul.We considered how, within specific institutional and socioeconomic contexts, an integrated approach to MUTP can be enacted and would likely influence the broader social outcomes from projects.The IAD Framework is a tool for analyzing complex situations that involve many actors and multitudes of interests (Spijkerboer et al., 2017;Van Geet et al., 2019).The framework has been used to investigate the ways decisions are made and implemented by actors in interactions with institutions at multiple levels (McGinnis, 2011).It offers a coherent theoretical approach that can be used to study the roles of stakeholders and institutions, and their interactions in varying contexts (Cole & McGinnis, 2017).
The IAD Framework comprises action situations, contexts, and outcomes (see, Figure 1a).Ostrom (2005, p. 14) defined action situations as social spaces where "participants with diverse preferences interact, exchange goods and services, solve problems, dominate one another, or fight."The Framework focusses on investigating how individual and collective stakeholders select actions, engage in patterns of interaction, and realize outcomes (McGinnis, 2011;Ostrom, 2011).Each action situation is shaped by the specific context in which it is embedded, such as the characteristics of a community, as well as by the formal and informal institutional rules at play.What actions are possible, what outcomes can be produced, and what is contained in the information available to the actors are all affected by the broader context and by the rules-in-use that support prescriptions concerning permitted or prohibited actions (Ostrom, 2005).Ostrom (2011) emphasized that there can be situations where rules do not exist, or where there are institutional ambiguities.
There are seven types of rules that structure the action situations (Figure 1b).First, position rules specify the positions that actors could hold in a decision-making process.Second, boundary rules specify who may enter or exit a position and how they do this.Third, choice rules specify the sets of action that can, may be, or must not be taken at specific points in time.Fourth, information rules specify the amount and type of information that should be available to participants and how this information is used and shared among actors.Fifth, aggregation rules specify how decisions are made, e.g., by an individual actor or in collaboration with others.Sixth, pay-off rules specify the costs and benefits that derive from particular actions or outcomes.Seventh, scope rules specify sets of possible outcomes, as well as who has jurisdiction.
Drawing on the literature relating to an integrated approach to transport and spatial planning, we developed a heuristic diagram to illustrate the potential action situations in a MUTP planning process.Figure 2 provides a schema for analyzing the action situations and institutions that govern the actions  of key stakeholders and influence project outcomes (such as greater accessibility for all and fair distribution of benefits) within a collective action arrangement relating to transport and spatial planning.As depicted in Figure 2, we examined the interplay between spatial planning and transport systems, at macro and micro levels, from policy formulation to implementation.The action situations were established and investigated using Ostrom's rules-in-use in a MUTP setting (the metro system in Seoul).Consistent with the IAD Framework, the planning processes were analyzed in terms of how institutions and actors selected actions, interacted, and realized social outcomes.We also identified the institutional barriers to and opportunities for integrated planning processes, and we considered how the MUTP planning process in Seoul affected the social outcomes from the metro project.
Our application of the IAD Framework in the context of the Seoul metro project involved three steps.First, we established the action situations and identified the variables (i.e., rules) of these action situations using action verbs (Ostrom, 2011; see Table 1).Second, we analyzed the planning processes, focusing on the key institutional issues in implementing an integrated approach to urban transport and spatial development at macro and micro levels.Third, we identified the barriers to and opportunities for enhancing the social outcomes from urban transport projects.

Methods
We used the Second Phase Subway Development (2PSD) in Seoul as the exemplar to discuss the potential of an integrated approach to enhancing the social outcomes from MUTPs.There were several reasons for choosing Seoul and the 2PSD.First, Seoul is a rapidly growing metropolis that has experienced rapid population increase and economic growth.Second, Seoul has developed an extensive urban transport network in response to environmental pressures and the socioeconomic problems arising from rapid urbanization.Third, the subway has become the most used mode of transport in Seoul, due to its expansion across the metropolitan area.Fourth, since the time the 2PSD was initially planned, Seoul's urban policy has focused on the need to integrate transport and spatial development (SMG, 2003).Fifth, much previous research has been conducted on the Seoul subway system, which has established that there are various issues relating to the social outcomes from the subway developments.These issues include that it was difficult to meet all objectives (e.g., balanced development across a city by facilitating station area development), and that there has been a differential distribution of accessibility and quality of life across the metropolitan area (Jang et al., 2017;Jin & Jin, 2015).
Research methods included document analysis and in-depth interviews with key stakeholders who had been involved in the planning processes related to the 2PSD at national, city, and neighborhood (local district) levels.Because rules-in-use are often developed over a long time and usually are implicitly understood by participants rather than being written down explicitly (Ostrom, 2008), to identify and comprehend the rules-in-use, we cross-corroborated legal and policy document analysis with our in-depth interviews.
We analyzed documents concerning transport and spatial planning and implementation relating to the Seoul metro system (Appendix C provides the list of documents used).To identify the action situations, we considered the various steps in the planning process, as well as the defined project goals, the actors, and their interests.We did this by examining a range of relevant document, including all Table 1.Rules and action verbs used for institutional analysis (adapted from Ostrom, 2005Ostrom, , 2011)) project documents relating to the 2PSD, consultation reports, policies and guidelines on urban rail development, and spatial and transport policies and plans at multiple levels.To identify the influential rules for each action situation, we examined laws and regulations on urban rail development, land use, and station area development at different periods.To identify the broader contextual factors and project outcomes, we studied the ex-ante impact assessment and ex-post project evaluation reports, and empirical research about the outcomes from the 2PSD.The document analysis provided input for our in-depth interviews with key stakeholders.
A total of 33 in-depth semi-structured interviews took place between May 2018 and May 2019 (see, Appendix B for the list of interviewees).Interviewees included key stakeholders (i.e., planners, policymakers, and experts) from various levels (national, regional, city, and neighborhood levels), sectors (transport, land use, or both), and organizations (policy institutes, governments and public corporations).We also interviewed private sector professionals involved with station-oriented development planning and implementation.The 33 interviewees had key roles during the strategic and/or operational stages of the 2PSD.Starting with people in obvious government positions (i.e., key planners), we used a snowball sampling approach to identify the people to be interviewed.Our collection of interview data ceased when we considered that additional interviews led to recurring viewpoints, indicating that saturation had been achieved (Hennink et al., 2010).The lead author conducted interviews in a manner consistent with ethical social research (Vanclay et al., 2013), with informed consent being given verbally.The interviewer recorded the interviews, which lasted between 60 to 90 minutes.We carried out all interviews in Korean and transcribed them for analysis.
The interviews had three parts.
We started with open-ended questions, which provoked discussion about the key issues related to the social outcomes from the 2PSD, and about the broader socioeconomic context.Then, we asked semi-structured questions to investigate the action situations according to the variables specified in Table 1.Interviewees were asked to describe the rules in the planning process generally.Then, they were asked which of these rules they thought were especially important in the 2PSD planning process.They were also asked how these rules affected the action situations, and what rules they did not apply.Finally, the interviewees were asked to discuss how the social outcomes from the 2PSD were affected by the planning process.To consider the different contexts of each neighborhood, we asked the interviewees about the socioeconomic conditions of the neighborhoods in which they had worked.
The research process was iterative in that the determination of the rules-in-use started by examining official documents, and was further informed by the in-depth interviews with key stakeholders.
The key points that emerged from the analysis of the documents and the initial interviews were validated by a small number of second-round interviews with five key planners.
The transcripts of the interviews and documents, all of which were in Korean, were coded and analyzed using ATLAS.ti8. We assigned codes based on latent as well as manifest content to capture the underlying meanings of the data (Babbie, 2010).Interview passages that pertained to each action situation of the 2PSD were coded deductively, using the seven types of rules from the IAD Framework.Because action verbs (e.g., to be, enter or leave, do, jointly affect, send or receive, pay or receive, and occur) are important identifiers of the different rules (Ostrom, 2005), we analyzed the transcripts for their presence (see, Table 1).We note that the grammatical concept of action verbs also exists in the Korean language and that identifying them in the transcripts was relatively easy to do.We also identified statements about the broader context, as well as statements about social outcomes.We inductively discovered the key issues affecting integration.

An institutional analysis of the second phase subway development in Seoul
In this section, we use Ostrom's (2005) rules to describe the action situation of the 2PSD and to analyze the key institutional issues for an integrated approach to subway development.Below, we indicate the specific rules that apply to each action situation in brackets (e.g., Sc1 = Scope Rule 1).We provide the lists of rules, respondents, and data sources in Appendices A-C respectively.

Context
By the 1990s, Seoul had become a global metropolis with a population exceeding 10 million.Seoul had been going through rapid economic and population growth since the 1950s (SMG, 2015).Between 1950 and 1990, the city built much large-scale urban infrastructure, resulting in economic growth, a continuously-rising population, and traffic congestion.The population went from 1 million in 1950 to over 10 million in 1990, an average annual growth rate of 6% (SMG, 2015).During the 1980s, the Seoul Metropolitan Government (SMG) focused on the urgent issues stemming from this rapid development (SMG, 2016).Because of road congestion, the shortage of public transport, and the need to support urban expansion, development of a subway system was increasingly urgent.
Construction of the First Phase of subway development (i.e., Lines 2, 3, and 4, totaling 186 km) started in the 1970s and it came into operation in the early 1980s (Figure 3).In the 1990s, the network was extended with the Second Phase Subway Development (2PSD) (Line 5, 6, 7, and 8 totaling 145 km).The stated key aims of the 2PSD were to: support the employment growth that was concentrated in the center and sub-centers (i.e., Young Dong and Yeouido); increase access to these centers from a wider area; and support the populations living in new satellite cities outside the SMG's boundaries (SMG, 2003; Figure 3).The 2PSD project also aimed to facilitate the development of local socioeconomic hubs at nodes and to enhance the living environment and quality of life of people in neighborhoods, in order to contribute to balanced development and social equity across Seoul (SMG, 1997(SMG, , 2003)).
Several studies have identified that the subway development contributed to ongoing changes in spatial structure, which led to Seoul becoming a polycentric metropolitan area (Jin & Jin, 2015;Kim & Suh, 2016).The 2PSD project enhanced access to the major hubs in Seoul and extended the subway network to a wider area (Choi et al., 2012).However, it is not clear if the project fully realized its broader social goals, such as facilitating balanced development across the metropolitan area and triggering station-oriented development at local hubs (Lee et al., 2021).In the center and subcenters, accessibility to opportunities was enhanced due to the increasing concentration of jobs and services at these centers and because of the improved transport network.In contrast, many other districts, especially those in peripheral areas, experienced only limited spatial change and little increase in access to opportunities (Jin & Jin, 2015).In peripheral locations, station areas only played a minor role as local hubs because of their limited potential for commercial development (Choi et al., 2012;Lee & Kim, 2009).Moreover, the positive consequences for local living environments were also limited due to the poor integration of station areas into neighborhoods and the poor state of pedestrian pathways (Go, 2008;C. Lee et al., 2015).Furthermore, there was a socially-differential distribution of outcomes.In particular, the districts with lowest incomes (e.g., peripheral areas in the Northeast of Seoul) had the highest proportion of people with poor access to public transport, jobs, and services (KOSIS, 2010(KOSIS, , 2015;;KRA, 2012;W. Lee et al., 2011).

Action situations
We identified and analyzed the multiple action situations of the 2PSD (see, Figure 4).The first stage of the 2PSD included project decision-making and developing a proposal.The second stage involved detailed planning and design at the macro level (i.e., planning and design of an overall route across Seoul) and at the micro level (planning and design of stations and adjacent developments).The third stage included implementation of station area development, which often took place several years after the subway commenced operation.The action situation of each stage is illustrated in Figure 4.

Stage 1: Decision-making and developing a proposal
In Stage 1, the Seoul Metropolitan Government (SMG) had the lead role in developing the initial proposal.Upon receiving an order from the President and the Office of the Prime Minister (Po1, i.e., Position Rule 1, see, Table 1 and Appendix A), the SMG developed an overall route plan and a financing plan (Pa1-2, Po1), mainly with experts from the central government's Transport Policy Institute (Ag2, Bo1).The experts developed a number of project options, reflecting a varying combination of assumptions, including projected demand, which were specified in the key macro-scale urban policy (i.e., Seoul Urban Plan) (Ch2).Because of the central government's concerns about rapid implementation and the financial and economic feasibility of projects, it ultimately approved the 2PSD proposal (Ag1, Ch4, Po2).During this first stage, district governments and communities were involved only to a limited extent, primarily by attending public hearings about initial options that were facilitated by the SMG (Bo2-3, Ch3, In1).

Stage 2: Detailed planning and design
In Stage 2, the Seoul Metro Development Division (SMDD), which was established within the SMG to plan the project, held a central position in the planning process at city and local district levels (Po4).At the city level, the SMDD developed a detailed route plan, mainly with transportation experts (Ag2, Bo1).To decide on the locations of stations, the SMDD conducted a thorough investigation of local contexts along the routes (Ch6), examining the location of local markets, the availability of public land, local development plans, and geographical and topographical conditions.Urban planning experts, policymakers, and other sectoral experts participated in finalizing the detailed route plan only when consultation was facilitated by SMDD (Ag3).The plan was continually adjusted until agreement between all experts and policymakers was achieved (Bo2, Ch3, In2).At the local district level, the SMDD developed a plan and design for stations and adjacent areas, including station facilities, pedestrian paths, and mixed-use development around the stations (Ch10-11, Po7, Sc5).They were also responsible for land acquisition and for determining compensation amounts for any negative impacts on private land (Ch8, Po5, Pa3).The SMDD negotiated details of the development plan with key stakeholders, including: other divisions in the SMG (e.g., land development division) (Po5); district governments (Bo4-5, Ch12); private landowners (Bo4-5, Ch12); and the owners of adjacent commercial buildings (Pa4).Transport experts investigated the impacts on local mobility and identified mitigation measures using Transport Impact Assessment (Ch13; CUTR [Center for Urban Transport Research], 2000).
After the detailed plans were reviewed by expert committees of the SMG, local residents and business communities were consulted on the route plan and station design (Bo3, In3, Po6).In addition to the general consultation process, through the district government, the local communities could officially request modification of the plans (e.g., changing the location of station exits or the route of subway lines) and could demand mitigation measures for potential negative impacts (Ch14).After meetings with district governments and experts, the SMDD determined what was feasible and beneficial for the majority of stakeholders (Ag3, Ch3, Ch15).

Stage 3: Station area development
After the subway development was completed, new actors would often become involved in the operational stage.For example, private developers might develop areas adjacent to stations that were designated as being available for development by the SMG (Bo6, Ch16, Po8).Consistent with guidelines issued by the district government, developers had to prepare a specific land use plan, details of building design, and traffic management plans (Ch17).The SMG and district governments reviewed the development proposals, focusing on a few criteria, especially the floor area of buildings (Ch19).District governments and the SMG discussed the details of the development proposal with the initiating developers, and provided recommendations for improvement (e.g., integrating subway exits into private buildings and pedestrian-friendly urban design on the ground floor; Bo7, Ch20, Po9, Sc7).Developers would usually be given permission to increase development density (e.g., the height of buildings and floor area ratio) if they donated land to the SMG for public amenities or pedestrian paths, or if they implemented designs that were suggested in the District Plan (Ch18, Pa5).

Analyzing the key institutional issues for an integrated approach to subway development
Using the conceptual framework in Figure 2, we identified the key institutional barriers to and opportunities for an integrated approach to transport and spatial development at multiple levels.

Stage 1: Strategic decision-making stage
The strategic decision-making stage involved only a limited number of actors, primarily the central government, the SMG, and transport experts.A primary issue was that, although the SMG had an integrated goal (i.e., effective land use around key local stations to facilitate balanced development across the city), no one was in a position to be able translate this goal into specific targets in the 2PSD proposal (Po1).As the lead actors, the SMG and transport experts focused on developing a proposal to reduce congestion and expand transport networks for the rapidly increasing population (Po1, Pa1, Sc2).The central government played a limited role during this stage (Po2-3).A transport planner (Res1) emphasized that: "It was an emergency.The speed of urban development was too fast, and the problem of congestion was out of control.SMG focused on building more subway lines as soon as possible."Transport planners and urban policy experts (Res3,5,8,11) emphasized that the fundamental issue was that most experts did not understand the relation between subway development and effective spatial development.As such, there was no procedure for joint decision-making that would integrate spatial and transport development during the strategic stage.
The second finding was that the limited scope of transport policy (SC4) and the limited involvement of district governments and communities during the strategic stage (Bo3) created a situation in which, during the development of the initial proposal, no consideration was given to the varying needs and conditions of the different districts.Transportation planners at the national and city levels (Res1, 10) and urban spatial policymakers (Res11, 13) identified that: "SMG mainly thought about enhancing overall accessibility to the city center.Any issue related to development at the district level was rarely discussed at any time during the strategic planning stage" (Res13).Interviewees identified that the limited involvement of local actors was related to the fact that the decentralization policy had not yet issued (Res11, 13).There was no mechanism by which local actors could get involved in developing the transport plans and the spatial plans of Seoul.
The third issue related to unclear roles and responsibilities for developing and implementing the regional infrastructure plans.Although the new satellite cities were developed to relieve the concentrated development of Seoul (Sc3), no agency was in charge of establishing a regional transport plan.The SMG was only responsible for developing subway networks within Seoul using its own budget (Pa1, Po1).An urban spatial policymaker (Res11) and a transport planner at the national level (Res 3) both argued that the central government did not understand the importance of having a strategic overview for the metropolitan area overall, and it did not take responsibility for providing such direction.Some interviewees (Res4,11,14) mentioned that the central government did not facilitate collaboration between SMG and other cities and that local governments had to be responsible for their own development (Po4).Consequently, the SMG and other cities individually developed their own urban transportation development plans.

Stage 2: Developing detailed plans at macro and micro levels
Our research identified that choice rules and scope rules led to settings in which decisions about the overall route were based only on a few criteria related to cost-benefit ratio (e.g., expected passenger numbers), rather than by considering the broader social outcomes of the project.With a primary concern about the economic efficiency of the project (Sc4), the SMDD and transport experts prioritized providing a service to areas with high demand (Ch5, 7).This was identified by several interviewees (15)(16).A SMDD transport planner emphasized that: "When planning the overall route, the key concern was providing services to the center and sub-centers where demand was already high.Areas with low demand or an area which did not have a concrete development plan were not considered" (Res9).Urban spatial planners at the city level (Res11, 13, 31) argued that transport planners and SMDD only considered current demand, rather than thinking of the long-term wider goals of subway development: "SMDD hardly included lines that connected peripheral centers and newly developed areas in the final plan due to the low benefit-cost ratio, even though the initial proposal aimed to provide better services to peripheral areas and trigger balanced spatial development."Furthermore, the spatial planners emphasized that without a joint decision-making process between urban and transport planners (Bo2, In2), they did not have much chance to stress the need to promote the wider outcomes of the subway development.
Regarding the development of a detailed plan at the micro level (i.e., for station and adjacent area development), our research showed that the pay-off rules and position rules contributed to the limited scope of station development, i.e., the lack of development of a large proportion of the adjacent areas.The SMG had an urgent need for transport network expansion and had to finance station area development as well as the construction of the subway lines (Po4, 7, Pa1, 3).To meet its targets, the SMDD developed a cost-effective and time-saving plan for station area development (Sc4).This strategy resulted in the SMDD only focusing on developing a limited number and range of facilities within stations, even though they wanted to operationalize the station-oriented spatial development policy (Res9,16,17).A SMDD planner emphasized that "we ended up focusing on a narrow scope due to the time-consuming process [of the integrated approach] and the greater cost involved in land acquisition, compensation, and construction of buildings etc."He indicated that the key barriers to an integrated approach included the complicated process of negotiating with all relevant stakeholders (Bo5, Ch8, Ch11-12, Po5, In3), which caused delays and increased costs: "we had to hold thousands of meetings.It often took too long to negotiate with all those affected communities and private landowners." Our research identified that the limited scope of station area development and the fragmented responsibilities of the different divisions led to a lack of joint decision-making process within the SMG.The Land Development Unit within the SMG was in charge of all the processes involved in land development around stations, and the SMDD had no authority to intervene in the decision-making of that unit (Ag4, Po10).A SMDD planner (Res16) emphasized that "without any joint decision-making procedure and clearly defined roles, we just had to accept the decisions made the Land Development Unit.If the unit decided to stop the process [of land development around a station] due to the complexity of the project or other priorities, we had to give up planning station area development [at that specific location]." Another dominant issue was confusion regarding roles and responsibilities in planning local-level accessibility during the operational stage.Our study identified that no one had any commitment to enhancing connectivity between stations and the rest of their district, or to enhancing the pedestrian environments around the stations.Many interviewees (Res8,19,31,33) stated that a key barrier was a lack of clarity regarding who should pay for and how to implement actions to enhance local accessibility.A SMDD planner admitted that the SMDD focused mainly on the technical and financial feasibility of the project (Ch7, SC4): "we mainly cared about mitigating impacts on private land.We thought that works, such as enhancing local environments or access to stations, were not within our scope."Furthermore, urban spatial planners (Res20-22, 31) stated that the only role of district governments was to assist the SMG's land acquisition process and to collect complaints from local communities (Ch8).Mitigation measures suggested during the Transport Impact Assessment, such as improving pedestrian networks, were rarely considered due to their technical complexity or high cost.

Stage 3: Station area development
Prominent issues in station area development were the limited level of quality control over land use around stations and institutional ambiguity regarding an integrated approach to station area development.The planning regulation primarily focused on floor ratio and the density of development of the projects around stations, and the SMG had limited influence on the way land was developed by private developers (Ch19, Sc6).An urban planner working at the district level stated that: "Our urban planning system lacks rules and strategic planning directions [to developers] that would facilitate effective land use and quality in the living environment in the vicinity of transport nodes."Some interviewees (Res20, 28) also emphasized that there were few incentives to encourage developers to implement appropriate actions in their development proposals, e.g., to integrate station exits into private buildings, improving the public paths leading to stations, and having socially-desirable retail activities on the ground floor (Ch17-18).Given the lack of controls or methods to promote integrated development around nodes, private developers primarily focused on maximizing development density.The SMG and district governments could only "encourage" (as defined by Territorial Plan and Use Law) developers to consider an integrated approach (Ch20).A district official emphasized that "private developers only listened to us if they would expect profitability.We often tried to suggest improvements in the way they used and designed land around stations, but it did not work very well due to the limitations of the current law and district plans" (Res30).
We identified that the lack of incentives and controls in the application of an integrated approach (Ch17, 20, Po9) created a situation in which the financial capacity and vision of district governments played a key role in land-use and transport integration around nodes.This was mentioned by all interviewees who worked at the district level, as well as by the urban planners working for the SMG.Some districts at the periphery that were eager to quickly realize station area development (e.g., to increase their rate revenue) tended to accept developer proposals outright, rather than encourage the developer to minimize negative impacts.In contrast, other districts (e.g., districts located in the subcenters) that had greater financial independence often required that private sector proponents improve their plans by adjusting land use or site design.A district planner (Res21) mentioned that "the level of integration was closely related to the socioeconomic conditions of districts.Those that had enough money and negotiation power, being in primary locations such as sub-centers, strongly gave many recommendations to private companies in order to maximize the benefits at the local level."

Barriers and opportunities to an integrated approach to enhancing social outcomes
Figure 5 depicts a summary assessment of the extent to which the planning process for the 2PSD project in Seoul reflected an integrated approach.This figure depicts that there was an imbalance and incongruence between transport and spatial development at multiple levels, and from the strategic to the operational stages.It also depicts that there were missing action situations, and an absence of linkages between the action situations.The planning process influenced the social outcomes, specifically: an unbalanced distribution of benefits across the city; a limited level of positive impacts at the local level; and differential outcomes among social groups.Furthermore, the macro level context (e.g., stage of urban development) and the local socioeconomic contexts together influence the extent to which the MUTP planning led to the social outcomes.Based on our analysis, we discuss explanations for the ongoing difficulties and potential opportunities to enhancing social outcomes below.

Social outcomes and integration at the macro level
Our analysis of action situations illustrates that during strategic policymaking, institutional incongruence in scope between transport and spatial development is an obstacle to addressing the broader social outcomes from MUTPs (e.g., balanced development across a city).Such incongruence is closely related to the lack of a joint decision-making process between transport planners and spatial planners (i.e., absence of aggregation rules) that would potentially translate an integrated vision into a strategic plan and targets (Heeres et al., 2016).During the strategic stage, the Seoul case showed that the macrolevel planning process, which only involved transport planners, resulted in decision-making about the overall route that disregarded the overarching long-term goals of urban policy (especially facilitating effective land use around stations at local hubs and balanced development across the city).The operational stage became driven by technocratic planning, which prioritized the serving of highlypopulated areas in the center over peripheral areas.Because the plans to extend the transport network to a wider area were delayed by 10 to 15 years, people living in such areas (especially those with poor socioeconomic conditions) suffered from limited accessibility to public transport and opportunities (KRA, 2012;KOSIS, 2015;Res3, 8, 9, 31).
In addition to institutional barriers, our analysis reveals that contextual factors (Cole & McGinnis, 2017;Ostrom, 2011)-e.g., knowledge of actors about the relations between MUTPs and spatial changes at a macro level-play an important role in applying an integrated approach to MUTP planning processes.The Seoul case showed that a fundamental barrier was the lack of understanding of all actors regarding how to use urban transport projects to realize effective and balanced development across Seoul.Such a lack of understanding can create a situation in which only a limited range of actors (e.g., transport experts) participate in developing an overall route plan.Long-term intended social outcomes, such as facilitating balanced development by providing better access to peripheral areas and creating local hubs at nodes, would not be considered during an operational process.

Social outcomes and integration at the micro level
Our analysis indicates that at the micro level, resource constraints and institutional ambiguity regarding procedures of integrated planning (i.e., absence of pay-off rules and choice rules) create barriers to achieving intended outcomes at the local level.It shows that operationalization of an integrated approach requires not only a networked governance of actors (Legacy et al., 2012), but also integration of plans, procedures and resources within the metropolitan government.The case study revealed that despite the keen interest of the metropolitan government, wider station area planning was often limited due to delays in the MUTP planning process (e.g., caused by absence of an integrated planning system) and consequent cost overruns (Res16, 17).
Our research clearly points out that in spatial planning, the absence of strong rules that demand that actors utilize an integrated approach to development projects around nodes negatively influences the achievement of positive outcomes at the neighborhood level.The lack of clear guidelines regarding acceptable "carrot and stick" measures for land development create a situation in which private developers are not encouraged to integrate their development activities into the surrounding area or to be concerned about the negative social consequences of their activities.As seen in the case of Seoul, many development projects around stations resulted in poorly planned pedestrian paths and station areas that were poorly-integrated in the local environment, creating negative impacts on local quality of life and accessibility (Go, 2008;C. Lee et al., 2015;Res11, 12, 20, 29).Such negative impacts were particularly notable at stations in peripheral areas, due to the resource constraints of local governments in those areas (W. Lee et al., 2011).
To achieve long-term benefits from MUTPs for the whole population, attention needs to be given not just to the institutional issues of integrated planning (e.g., Spijkerboer et al., 2017;Van Geet et al., 2019), but also to the specific socioeconomic contexts at the local level.The varied contexts of localities (e.g., geographical advantages, financial capacity of local authorities) influence local capacity to operationalize an integrated approach, resulting in spatially and socially differential outcomes (Lee et al., 2020).The socioeconomic condition of local districts in Seoul influenced the extent to which developers and district governments jointly established intended outcomes.This implies that in comparison to peripheral districts, districts with better conditions (e.g., the commercial center of a city) experience higher levels of improvement in the local environments around stations and overall social outcomes.A consequence of this situation is differential outcomes among different social groups across a city, which is a problem facing many megacities across the globe (Martens, 2012).Overall, operationalization of an integrated approach at the local level requires institutional clarity on responsibilities of actors and enhancement of the capacity of all actors to take those responsibilities.

Social outcomes and the multi-level planning process
Our research reveals that the lack of a multi-level planning process for MUTPs creates barriers to achieving outcomes that meet the various needs of neighborhoods across a city.High institutional ambiguity regarding the responsibilities of the national and metropolitan governments to operationalize macro-scale development policy (i.e., positions rules) affects the distribution of positive outcomes across localities.In the case of the 2PSD, the limited role of the metropolitan government in establishing direction for station area development resulted in a uniform approach to land development around nodes being applied by developers, irrespective of the socioeconomic conditions of each neighborhood (Res11,20,22).Consequently, the station area development was not always appropriate and often failed to address the varying priorities of the different localities (Res11,13,21,22).Our analysis supports the view that the integration of transport development and spatial development requires a bi-directional vertical form of integration (Rode, 2019;Stead & Meijers, 2009).
Our analysis suggests that the limited roles and lack of involvement of local authorities and communities in strategic decision-making has potentially led to the limited ability of planners to ensure that enhanced local accessibility and an improved living environment are actually achieved.Earlier participation of local actors in the planning process would enable planning for accessibility and quality of life at the local level, even though this is a challenging issue (Silva et al., 2017).In Seoul, the strategic stage only involved a limited range of actors, had a limited exchange of ideas, and an absence of information flow across multiple levels (Res9,11,14,20,31).During the operational stage, many requests from local communities and planners were not accepted, because acting on those requests would have delayed the process or they were too costly (Res7,15,16).Options to enhance the local environmental quality around station areas, or to improve local-level accessibility, were rarely considered during the operational stage, creating a limited level of positive change at the neighborhood level (Chang et al., 2007;Go, 2008).

A multi-level integrated approach to enhancing social outcomes
Our analysis provides valuable insights into the opportunities to enhance the social outcomes from MUTPs that would come from a multi-level integrated approach.In order to facilitate accessibility and improved quality of life for all across a city from MUTP development (Hall & Pain, 2006), there should be institutional harmonization at the metropolitan level that enables the integration of transport route planning with city-level spatial development planning.In order to facilitate positive changes in local neighborhoods (Rydin, 2010), what matters is greater institutional clarity and a strong regulatory framework for local-level planning that clearly outlines the permitted and prohibited actions for an integration of transport and spatial development.Furthermore, to achieve a fair distribution of desirable outcomes across localities (Jones & Lucas, 2012), multi-level governance-e.g., strategic decision-making at the micro level, and operational follow-up at the macro level-needs to be in place.
In applying the IAD Framework, our research advances recent and current discussions (Bertolini, 2012;Heeres et al., 2016;Lee, 2020) about an integrated approach to MUTPs for socially sustainable urban development.To realize the social outcomes from projects, we suggest that institutional harmonization between transport and spatial planning should require different strategies for decision-making stage than for the operational stage (Heeres et al., 2016;Van Geet et al., 2019).The strategic decision-making stage needs to involve spatial planners in co-producing a city-level plan for MUTPs with transport planners, while the operational stage should focus on engaging transport planners in quality control of spatial outcomes experienced in local neighborhoods.In order to arrive at new integrated knowledge, the strategic planning stage needs joint action that goes beyond simply piecing together sectoral perspectives (Stead & Meijers, 2009), while the operational stage requires institutionalization of integrated planning instruments, such as accessibility measures (Silva et al., 2017).
The institutional analysis also clearly suggests that operationalization of an integrated approach requires strengthening different rules at each level.For example, there should be rules at the macro level that enforce establishment of an effective interface that allow flows of ideas between actors and the co-production of a project vision and project plans (i.e., aggregation and information rules).At the local level, there should be rules that enable integration of different systems, and rules that specify clear incentives and allocate greater responsibility for operationalizing an integrated approach (i.e., choice and payoff rules) to a wider range of actors.We argue that networked forms of governance (Curtis & James, 2004;Legacy et al., 2012) alone would be unlikely to deliver operationalization of an integrated approach.It is the careful combination of structural changes and the networked coordination of actors that will lead to the integration of spatial development and transport development, and ultimately to improved social outcomes from MUTPs.
Finally, in line with Cole and McGinnis (2017), our research strongly points out that more attention needs to be given, not only to the institutional rules, but also to the varied contextual factors that structure MUTP planning processes.In particular, the varied capacities of actors (e.g., in terms of knowledge and resources) to produce and implement integrated policy and the wider socioeconomic contexts of projects (e.g., stage of urban development) influence the extent to which the rules are applied, thus affecting the fulfillment of intended social outcomes.This situation indicates that ensuring social outcomes requires establishing context-specific approaches to integration that consider the different opportunities and challenges facing localities and cities.

Conclusion
Our research identified the institutional barriers to and opportunities for developing and operationalizing an integrated approach to transport development and spatial development.The goal of this research was to enhance the social outcomes from mega urban transport projects.Using Ostrom's Institutional Analysis and Development Framework (Ostrom, 2005), we considered the rules-in-use in the context of an MUTP and how these rules shape the way actors select actions, interact, and how the intended social outcomes, such as accessibility to opportunities for all, and a fair distribution of benefits across a city, are realized.
Our institutional analysis of the Seoul Second Phase Subway Development Project provided insights into the ongoing difficulties and opportunities for an integrated approach to transport development and spatial development, and how these might affect the social outcomes from MUTPs in rapidly-urbanizing metropolitan cities.It showed that the planning process in Seoul-which was characterized by imbalances between transport and spatial development, and incongruence between the macro and micro levels-negatively influenced the resultant social outcomes.
Our research indicated that the intended social outcomes from MUTPs cannot be achieved without institutional congruence and institutional clarity between transport planning and spatial development planning at multiple levels.At the macro level, there should be institutional congruence, especially in scope.This requires institutional arrangements that jointly establish and implement a macro-scale integrated plan, which has the aim of facilitating balanced development across a city.At the micro level, it is important to establish institutional clarity regarding inter-sectoral procedures for the operationalization of an integrated approach, as well as regarding the responsibilities of all actors to ensure quality of development around nodes.
We also illustrate the importance of rules that facilitate a multi-level planning process that ensures a fair distribution of benefits across varied localities.Our research suggests that institutional arrangements that allow participation of local actors and information sharing across territorial levels during the strategic stage are crucial to ensure the delivery of enhanced accessibility and quality of life in local neighborhoods.Moreover, metropolitan governments need to consider context-specific approaches to integration and provide local institutions with clear directions about how to operationalize land-use and transport integration in localities with differing needs and priorities.
Finally, achieving intended social outcomes requires not only institutional support at the macro level, but also paying attention to the technical and financial capacity of local actors to operationalize an integrated policy.What matters is to carefully consider the differential capacities and resources of the local actors (institutions) across a city.There needs to be strategic guidelines for converting an integrated plan into desired results at the local level, and especially to support those local districts that lack financial capacity or appropriate strategic vision.

Stage
Payoff rules assign costs and benefits to actors in light of the outcomes Sources 1-2 Pa1 SMG, as a constructer and an operator of the urban rail, was responsible for financing funds by using its own fund and (or) profits from operation, and (or) profits from station area development, and (or) issuing urban rail bond, and (or) receiving financial support from the central government.
D2 (11), I7, I8, I9, I15, I16, I17 1-2 Pa2 The central government could provide financial supports or loan to the city government for developing urban rail system if urban transport infrastructure passed more than two cities.

Figure 2 .
Figure 2. Application of the IAD Framework to the planning process for a mega urban transport project (Source: authors).

Figure 3 .
Figure 3. Subway maps for Seoul.Insert a: First Phase Subway Development.Insert b: Second Phase Subway Development (Source: adapted from Seoul Metro, 2008).

Figure 4 .
Figure 4. Multiple action situations of second phase subway development in Seoul (Source: authors).

Figure 5 .
Figure 5.An evaluation of the planning process for the subway development in Seoul (Source: authors).

D7
prepared a land-use plan, details of urban design, and traffic management plan, in line with a specific guideline for designated areas development D7, D9, D10, D11, I20-24, I27-32 3 Ch18 Developers gained incentives if they donated private lands to SMG for public amenities and pedestrian paths D8, D10, D11, D12, I19, I20, I23, I24, I25 3 Ch19 SMG reviewed a development proposal developed by private developers with reference to the Guideline in the District Plan (e.g., floor ratio, building height, and building coverage ratio) D9, D10, I20, I30 3 Ch20 District governments and SMG discussed the details of a development plan with developers and encouraged improving a proposal by integrating local environments D10, D13, I19, 120, I23, I27 D2(14), I2, I3, I7, I8, I9, I15, I16, I17 2 Pa3 SMG (SMDD), as a developer of urban infrastructure projects, was financially responsible for land acquisition for subway development and compensation for any disruption on private lands.D7, I15, I16, I17 2 Pa4 The owners of private lands (buildings) paid to SMDD for constructing connection gates between stations and properties if they desired to have such connecting gates.D1, I9, I15, I16, I17 2-3 Pa5 Private developers donated their lands to SMG/district governments if they want to get permission to increase density of land development.have the effective use of land by facilitating station-oriented development across, facilitate balanced spatial development across cities D1, D15, I3, I7, I8 1 Sc2 SMG would reduce traffic congestion by increasing ridership, and expand transport services into wider areas D1, D7 I1,3, I7, I8 1 Sc3 Metropolitan areas should be developed in a balanced way D7, D10, D11, D17, I1, I2, I3, I5 1-2 Sc4 SMG would have transport networks, which were cost-effective and did not disturb private lands D1, I9, I11, I13, I15, I16, I17 2 Sc5 Subway development would facilitate the high level of station area development to trigger effective land development D9, D10, D13, I20, I30 3 Sc6 Land development in District Plan Area would induce effective land use by the increasing development density of the area D8, D10, D11, D12, I19, I20, I23, I24, I25 3 Sc7 Station areas would be redeveloped by integrating surrounding areas and by facilitating mixed use land development D14,D19, I27, I28, I29 .

Table A3 .
Choice rules.Ch13 SMDD assigned transport experts to conduct Transport Impact Assessment, in order to examine impacts on local mobility and to develop mitigation measures.SMDD considered applying suggested mitigation measures and improving a plan

Table A4 .
Information rules.In3 Detailed plans were shared and consulted with local communities and district governments during local district consultation after SMDD and experts finalized a draft plan.

Table A5 .
Aggregation rules.Stage Aggregation rules determine "who is to decide" which action or set of activities is to be undertaken Sources 1 Ag1 The President and the Office of Prime Minister made a final decision on subway development proposal and an initial plan D1, I9, I15, I16 1-2 Ag2 SMG and transport experts working in the policy institutes at national and metropolitan levels decided a route plan D1, I1, I3, I4, I5 2 Ag3 SMG with support of other respective parties such as expert committees and other divisions in SMG decided a final route plan of subway development