In this section, we introduce our novel framework to elucidate the concept of socio-technical lock-in in urban consumption (see Table 1). Specifically, a lens of socio-technical systems is applied in order to understand the diverse drivers of urban consumption, due to its suitability for capturing the multiple and complexly interwoven sub-systems that make up cities. The framework is based on insights from literature describing the basic components of socio-technical systems (Geels, 2004; Seto et al., 2016; Markolf et al., 2018; Trencher et al., 2020). It consists of three broad categories: (1) physical systems of infrastructure, technology, and biogeographic conditions; (2) non-physical systems, comprising formal institutions, non-formal institutions, economics, and knowledge and competencies; and (3) human systems, consisting of demographic features. To understand how each system may lock-in unsustainable consumption behaviors, we visualize the framework in two dimensions: horizontal and vertical.

To conceive the horizontal dimension, we examined literature on socio-technical systems and systemic causes of lock-in to extract explanations of the core urban systems that influence consumption behavior (Unruh, 2000; Foxon, 2002; Seto et al., 2016; Markolf et al., 2018; Trencher et al., 2020). Previous lock-in literature has focused on interactions and self-reinforcing relationships between institutional and technological factors, often in large-scale socio-technical systems such as energy or transport (Unruh, 2000; Foxon, 2002). Recently, Seto et al. (2016) and Ürge-Vorsatz et al. (2018) extended the scope of relevant lock-in causes by integrating behavioral factors. Their framework, however, does not explicitly discuss broader contextual factors such as geographic or economic conditions. Markolf et al. (2018) later address this deficiency, incorporating ecological context in addition to social and technological dimensions. A similar conception of socio-technical lock-in is carried out by Trencher et al. (2020). Their framework, which also discusses energy systems, integrates the influence of exogenous context, rooting this in conceptions of a “landscape” from sustainability transitions literature (Geels, 2004; Geels and Schot, 2007; Edmondson et al., 2019), alongside other material, non-material, and human factors. Thus, although these recent studies consider the importance of broader contextual factors such as economic forces, natural conditions, and culture, their discussion on socio-technical lock-in is overwhelmingly placed in the context of understanding the self-perpetuating nature of fossil-fuel-based technologies or infrastructures. However, the field of sustainable consumption requires putting human agents as the focal point, situating these within the structural conditions and barriers posed by urban environments, technologies and social factors (Bengtsson et al., 2018).

To conceive the vertical dimension, we explicitly reflect the influence of both structure and agency on urban consumption behaviors. By depicting the two interlinked and interacting geographical scales of macro and micro, our framework is congruent with descriptions of multi-scalar dimensions in socio-technical systems (Giddens, 1984; Vergragt et al., 2016; Elbasha and Wright, 2017) and lock-in literature (Seto et al., 2016). The macro scale reflects the collectively provided or shared conditions that may be climatic, economic, technological, infrastructural, institutional and cultural in nature. These conditions are typically long-lasting and difficult for individuals to change, so they exert a powerful force on urban dwellers' consumption behavior (Jackson and Papathanasopoulou, 2008). In contrast, the micro scale depicts smaller-scale conditions, assets, and institutions—typically owned, accessed, or shaped by individual agents. Relevant actors at this scale include individuals, households, social networks and specific organizations, such as workplaces (Oliveira et al., 2020). At the micro scale, these actors often have more power than at the macro scale to decide or control consumption behaviors in relation to their own material possessions, environment, and non-material conditions (Jørgensen, 2012; Tao et al., 2021).

In proposing this framework comprised of two scales, macro and micro, we do not explicitly include a third scale—the so-called meso scale—that is sometimes evoked in consumption literature (Kirchberg, 2007; Maréchal, 2010; Poças Ribeiro et al., 2019). These studies explicitly describe the meso scale as situated between the macro and the micro, arising from interactions between the structural and individual conditions that shape urban environments. Yet conceiving the meso scale can be problematic, since literature concedes that its boundaries with other scales are blurry (Vergragt et al., 2016; Pekkanen, 2020). Thus, to avoid conceptual ambiguities and simply but clearly distinguish between structural and individual conditions, we incorporate in our framework only two scales: macro and micro.

The following sections describe each urban system captured by the analytical framework in more detail. We provide examples from a broad range of literature on urban consumption, climate governance, and energy transitions to describe how each can influence consumption behaviors and lock-in unsustainable living patterns. Two points deserve prior emphasis. First, we discuss these systems and scales separately for the sake of conceptual clarity. However, because of their interconnectedness and interdependence, we expect a certain overlap or conceptual blurriness when looking at the messy empirical world of living cities. Second, we do not assume that a particular consumption behavior would experience interactions and influences equally from all systems and scales. Rather, each system's relevance and influence would vary with particular consumption acts by micro-scale actors.

In studying urban consumption behavior, we identified that structural lock-in can wed agents to current and unsustainable practices as well as prevent agents from adopting sustainable alternatives. This draws on three fundamental principles from previous literature. First, lock-in arises from the persistent and multidirectional interactions that occur in multiple and interconnected urban systems, at differing scales (Seto et al., 2016). Second, feedback loops between these systems may emerge—by design or consequence. This occurs when one system influences another, which then reciprocally influences the first, reinforcing the collective impact (Wesseling and Van der Vooren, 2017; Markolf et al., 2018). Third, one factor might be more important than another, and thus may become a precondition for others or exert cascading effects on others.

Thus, behavioral lock-in, like technological lock-in, is created by persistent interactions and feedbacks occurring across the various physical, non-physical, and human systems that make up cities and urban areas. These systems are collectively provided or shaped at the macro scale or individually owned or shaped at the micro scale. The ability of individual agents to break free from these hampering conditions can be determined by the extent of their dependence on the collectively provided structural conditions at the macro scale, and by their ability to overcome various restraining factors originating from the micro scale.

To systematically capture these dynamic forces affecting urban consumption behavior, this study follows the distinction between structure and agency in structuration theory (Whittington, 2015). As shown in Figure 1, we apply two perspectives to visualize the origin and direction of these shapers of consumption behavior: (1) top-down, describing forces originating from the macro scale (i.e., from structural conditions), and (2) bottom-up, describing forces originating at the micro scale (i.e., from conditions under the control of individual agents and organizations).

In conceiving the top-down perspective, we make two proposals about the nature of consumption drivers and lock-in forces that originate from the macro scale. The first is that various systems at the macro scale will interact and self-reinforce, magnifying the cumulative influence of collectively provided structural conditions in urban areas (Bettencourt, 2013). Road transport, if viewed as a socio-technical system comprising several interlinked, self-reinforcing components (Geels, 2004), is an illustrative example. Here, a nexus of co-evolving, interlocked, and self-perpetuating factors at the macro scale may arise from policies that support passenger vehicle use through the provision of infrastructure like roads and bridges, the rise of a middle-income class, increasing production or availability of passenger vehicles in the market, and cultural identification with the car as a symbol of desire or a transportation norm (Cohen, 2012; Seto et al., 2016; Sovacool, 2017; Canitez, 2019).

The second proposal is that individual conditions at the micro scale are directly influenced and perpetrated by the collectively provided structural conditions at the macro scale (Janssen and Jager, 1999). For instance, increasing GDP at the macro scale often manifests as higher incomes for individuals and households, providing more opportunities for the consumption of resources, goods, and services (Spangenberg, 2014; Zaharia et al., 2019). This relationship has also been noted when the urban population size directly influences the density of the workplace or dwelling sizes due to less land availability (Aranda-Mena et al., 2018). Also commonly discussed, shared social norms or values can pose constraints or pressures on an agent's beliefs and preferences (Foxon, 2011; Biswas and Roy, 2015; Pekkanen, 2020). These can influence the practices of individuals, households, and organizations in diverse living activities such as bathing habits (Sovacool and Griffiths, 2020), dietary preferences (Einhorn, 2020), and choice of transport mode (Belgiawan et al., 2017).

The bottom-up perspective is inspired by the idea of dynamic agency (Elbasha and Wright, 2017). This encompasses two views. Conditions under the control or influence of individual agents may interact and self-reinforce to perpetuate unsustainable consumption behavior at the micro scale. Individual agents, organizations, or communities may also exercise their agency to contribute to continuing or reproducing larger structural factors originating from the macro scale (Phipps, 2001; Whittington, 2015). In addition, though it is beyond the scope of this paper, certain individual agents may succeed in transforming structural conditions (Corman, 2008).

Two postulations emerge from here. In the first, multiple systems at the micro scale can interact, reinforcing each other's influence to perpetuate the conditions that drive unsustainable consumption. For example, higher levels of disposable income or access to credit can incite the purchase of larger households or business premises, resulting in greater demand for energy, furniture, and household appliances due to the increased flooring area (Li et al., 2018; Sakah et al., 2019).

In the second postulation, individual factors at the micro scale can exert an upward influence on collective conditions at the macro scale, even stimulating change. Some have argued that agents have little control over structural forces (Jackson and Papathanasopoulou, 2008). But examples do exist where individual behavior has succeeded in influencing structural conditions at the macro scale (Spangenberg, 2014; Pekkanen, 2020). For example, the intellectual resources of individuals agents or organizations, when widely shared, can enrich the collective knowledge base of society (Kanger and Schot, 2016). And, though it is an atypical example, the individual beliefs and travel behavior of Greta Thunberg have circulated globally, arguably influencing collective norms and knowledge on climate change in many countries (Martiskainen et al., 2020).

Our attempt to conceptually integrate concepts of socio-technical lock-in with discussions of unsustainable consumption behavior draws inspiration from others. Notably, Kim and Kim (2020) highlighted that lock-in mechanisms can explain the loyalty of some consumers to certain consumption behaviors. Meanwhile Jackson and Papathanasopoulou (2008) explain that lock-in to unsustainable consumption patterns is exacerbated by the general inability of consumers to control the broader structural conditions shaping their surrounding urban environment.

Yet by adopting theories of lock-in and socio-technical systems in the domain of urban consumption, we acknowledge several challenges that arise. For example, though human agents can play a crucial role in triggering or inhibiting change in socio-technical systems (Borrás and Edler, 2014), the sustainability transitions literature on socio-technical systems and lock-in has been criticized for overemphasizing the influence of structural conditions and paying inadequate attention to individuals and agency (Duygan et al., 2021). However, by applying insights about socio-technical systems and lock-in while heeding the limitations of scholarship on urban consumption and sustainability transitions, an opportunity arises for a novel research perspective with a more comprehensive view of urban consumption. Importantly, this perspective can account for structural and individual factors, as well as interactions and lock-in that arise from the socio-technical systems that comprise cities.

This case examines mobility behavior in Bangkok City (henceforth referred to as Bangkok). Surrounded by five vicinities, Bangkok lies at the heart of the broader Bangkok Metropolitan Region. The whole region comprises a diversity of density and land uses, ranging from dense downtown areas, semi-urban and industrial areas, over a vast area of 7,762 km². National level policies are also examined when they affect urban development in Bangkok.

Triggered by its accelerating development and population growth, Bangkok's traffic situation is worsening, provoking the environmental and social consequences of a steep surge in privately owned passenger vehicles. The city's population increased from 6.4 million in 2000 to 10.7 million in 2021, a rate of 2.6% per year. It currently accommodates more than 13% of Thailand's total population and is home to some 6 million on-road passenger vehicles. The ratio of 0.6 vehicles per resident is twice as high as Tokyo's and three times higher than Beijing's.³

This explosion in the use of private passenger vehicles (see Figure 2) has occurred in a series of waves, one from 2011 to 2013 and another from 2018 to 2019. Despite the availability of diverse transport modes, such as buses and rapid rail transit, privately owned passenger vehicles have become the dominant mode of travel in Bangkok, making up around of 40% of journeys, followed by motorcycles with nearly 24% (OTP, 2018). The rapidly expanding fleet of motor vehicles has worsened the city's traffic conditions and air quality and has raised the urban temperature by contributing to the heat island effect (Attavanich, 2017; Khamchiangta and Dhakal, 2019). Motor vehicles thus damage public health, human welfare, and the economy (Uttamang et al., 2017). Indeed, the impacts of air pollution on health and life satisfaction have been estimated to cost around $6.17 billion annually during the first wave of car purchasing (Attavanich, 2017). Bangkok is often also alerted with days of harmful air pollution, its annual average for PM2.5 levels being more than four times the World Health Organization's target (IQAir, 2019). The dominant sources coming from traffic activities, PM2.5, PM10, O3, and SO3, are observed to be heavily contributing to non-accidental mortality, cardiopulmonary diseases, and lung cancer in the city (Chuersuwan et al., 2008; Guo et al., 2014; Fold et al., 2020).

As we now demonstrate, the lock-in of unsustainable mobility behavior in Bangkok is the result of long-term interactions among multiple urban systems, at both macro and micro scales, which promote the purchase and use of privately owned passenger vehicles. The lock-in mechanism is summarized in Figure 3.

We begin with vehicle purchases, both new and used. We focus first on two clusters of lock-in mechanisms, shown as [1] and [2] in Figure 3. The first cluster shows how informal institutions, formal institutions, economic forces, and technology availability at the macro scale intertwine to exert a downward impact on informal institutions at the micro scale. The second cluster starts with informal institutions at the micro scale, showing how these exert an upward influence on economic forces at the macro scale. To illustrate the persistence of the lock-in situation over time, while also addressing the changing relevance of various factors, we focus explicitly on different time periods when describing each cluster. When referring to the different systems of the lock-in framework, we use codes in the text and in Figure 3. For example, [ECO-MA] in upper case refers to economic forces at the macro scale, while [eco-mi] in lower case refers to economic forces at the micro scale.

The description of cluster [1] is strongly based on the first wave of car buying, between 2011 and 2013. Beginning with economic forces at the macro scale [ECO1-MA], rising GDP has boosted households' income and expenditure levels [eco2-mi].

In 2010, with the economy rebounding, Thailand's GDP per capita regained positive growth of 7.2% compared to 2009 (Pasuk and Pornthep, 2012). Bangkok's average monthly income per household continued to be ranked the highest in Thailand, hitting a peak of 49,000 THB (Thai baht) in 2011–2013 (NSO, 2019). The driving effect of this rising income on car ownership is stressed in multiple studies (Pongthanaisawan and Sorapipatana, 2010). For example, Taylor and Yue (2011) predicted that car ownership surpasses motorcycle ownership when household income exceeds around 50,000 THB per month. Thus, the increasing income triggered by economic growth gave many Bangkok residents the opportunity to purchase their first passenger vehicle.

Thailand has been by far the largest automobile producer [ECO3-MA] among ASEAN countries for the last 20 years (Figure 4). The high availability of automobiles for the domestic market [TECH1-MA] provides Thais with wide opportunities for purchasing. Indeed, Thailand has been called the “Detroit of Asia” (Wad, 2009; Wonglimpiyarat, 2016) owing to its success at attracting foreign automobile brands, mainly Japanese, to build factories on the outskirts of Bangkok (Warr and Kohpaiboon, 2017). Thus, compared to the populations of other ASEAN countries, Thai people benefit from low purchase costs for new vehicles, since locally produced vehicles are exempted from import tariffs.

Informal institutions at the macro scale [INFORIN1-MA] also contributed to increased purchasing of private automobiles. National government policy propagated a political ideology around the automotive industry's important position in ensuring economic growth, stressing its creation of employment (Thailand Board of Investment, 2017) and its contribution of 12% to average annual GDP since 2012 (Pollio, 2012). Policies favorable to the automobile market were implemented as a result, aimed at shortening the gap between customers' income and purchase prices. For example, since the early development of the automotive industry in the 1990s, low import duties [FORIN1-MA] have persistently provided attractive market conditions, by ensuring affordable price choices of new vehicles [ECO4-MA] (Wu and Pojani, 2016). The used-vehicle market was also targeted by policies at the macro scale: the rate of annual registration taxes [FORIN2-MA] was lowered to make automobiles more affordable for lower-income households (Techakanont and Leelahanon, 2015).

Other informal institutions have also propelled vehicle purchases. After experiencing severe consequences from the Asian economic crisis of 1997 (Kittiprapas, 2000) and the recession during post-Thaksin's regime (2001–2006), the new government of Yingluck Shinawatra in 2011 continued the populist development strategy from their predecessor. This notably included establishing preferential policies to support the outputs of the domestic vehicle production sector (Chambers, 2013; Pollio and Rubini, 2021). In the second half of 2011, the government's first-time car-buyer tax rebates scheme [FORIN3-MA] was introduced to stimulate automobile purchasing over motorcycles, which were historically Thais' preferred choice for urban mobility. The scheme offered a tax rebate up to 100,000 THB (Muthitacharoen et al., 2019), equivalent to the average yearly income of low-income households. At the same time, banks' lending policies [ECO5-MA] acted as a supporting economic force, providing more opportunities for the rising urban middle class to obtain personal credit [eco6-mi]. The additional car purchases were even boosted in this period by competition across the banking industry, as many introduced subprime loans (Muthitacharoen et al., 2019). With loans authorized for many low-income applicants, and the above-mentioned subsidy accepted as a down payment, this interconnected web of economic forces and formal institutions at the macro scale shrank the gap between income and affordability. It thus shifted the purchase preferences of many first-time buyers [inforin2-mi] from motorcycles to automobiles. As a consequence of these driving forces originating from the macro scale, national sales reached 1.44 million in 2012, up 80% from 2011. This momentum was sustained, with 1.33 million purchases in 2013 (Krungsri Research, 2020). Demonstrating the shift, in Bangkok the number of registered automobiles exceeded, for the first time ever, the number of registered motorcycles in 2012 and 2013 (Figure 5).

Cluster [2] is linked to the second wave of car buying, from 2018 to 2019. Although car sales dropped dramatically after the first wave, the accumulated car ownership at the micro scale [tech2-mi] triggered a series of economic repercussions at the macro scale. The vast number of cars purchased during the first car buyer scheme (2011–2013) [FORIN3-MA] led to demand for replacement [inforin3-mi] as households became accustomed to the convenience of private automobile use. This occurred after around five years of ownership, largely during 2018 and 2019, as owners finished loan repayments and opted for replacement after considering the increasing repair costs for vehicles exiting the warranty period [eco7-mi] (Krungsri Research, 2020).

Macro-scale conditions reacted to this new market opportunity. Car dealers adopted aggressive marketing activities [ECO8-MA], such as cash-back deals for owners trading in vehicles early. Vehicle upgrading then reinforced many Bangkok residents' preference for private automobiles [inforin2-mi]. This created a stock of affordable used vehicles, lowering the financial barrier for lower-income households to acquire their first automobile. At the same time, Thailand's per capita GDP [ECO1-MA] recovered after a slight drop due to the influence of political uncertainty on the economy (BOT, 2014), reaching ~10% growth in 2017 (KNOEMA, 2021). The stimulated economic activity at the macro scale raised household incomes further [eco2-mi] (NSO, 2019).

Though the second wave was less pronounced than the first, dependence on private automobiles in Bangkok deepened once again, as a result of these interactions across multiple sub-systems, pushing annual car sales to pass 1 million in the following 2 years in 2018 and 2019 (Figure 2).

We now examine how various urban systems have collectively fostered and locked-in urban mobility behavior that depends on private car use. We focus on two clusters of intertwined factors, marked as [3] and [4] in Figure 3. Cluster [3] illustrates an upward trajectory: how technology availability, informal institutions, and knowledge at the micro scale collectively influence biogeographic conditions at the macro scale. Cluster [4] illustrates a downward trajectory: how formal institutions along with knowledge and competencies at the macro scale collectively influence informal institutions at the micro scale. In contrast with the previous section, where we focused on two temporal events of vehicle purchasing, here we draw on different time periods and pieces of evidence to show how a temporally accumulated array of factors is perpetuating dependence on automobile use.

In cluster [3], mobility behavior is closely correlated with the availability of privately owned automobiles in households [tech2-mi]. Such ownership leads to more driving, and decreases the likelihood of owners walking, cycling, or using public transport (Pongprasert and Kubota, 2017; Witchayaphong et al., 2020). This causal trend is also stimulated by economic forces at the micro scale because, as more households enjoy higher incomes [eco2-mi], they become less sensitive to travel costs (Witchayaphong et al., 2020).

Moreover, for high-income Thais, the persistence of informal institutions at the micro scale, such as personal preference for car usage [inforins4-mi], can be explained not only by the relative levels of convenience, comfort, and safety—just as importantly, vehicle ownership gives them a desirable symbol of economic success and social status (Dissanayake and Morikawa, 2010; Pongthanaisawan and Sorapipatana, 2010; Matsuyuki et al., 2020). The correlation between private automobile ownership and social status has been described as a way for Bangkok's wealthy to save face and avoid the embarrassment of using public transport. These individual perceptions are influenced by widely circulated norms and societal values at the macro scale [INFORIN5-MA], with Thai society rooted in traditional class divisions separating the wealthy from the working class [DEM1-MA].

The literature also describes a self-reinforcing loop of exhaust fumes from the growing fleet of on-road vehicles and higher vehicle usage rates. Among all road transportation, passenger cars are the major contributor to worsening air quality in Bangkok [BIO1-MA] (Chuersuwan et al., 2008; Pongthanaisawan and Sorapipatana, 2010; Cheewaphongphan et al., 2017). Interestingly, the deterioration of air quality is exacerbated by low levels of environmental awareness [knocom1-mi], because many of Bangkok's citizens accept environmental problems such as air pollution as part of urban living. Furthermore, the extreme climate conditions [BIO2-MA] posed by high temperatures are worsened by the combined influence of climate change and the heat island effect. Bangkok's average surface temperatures have increased by more than 6°C from 1991 to 2016 (Khamchiangta and Dhakal, 2020); the use of automobiles, though it is not the major cause, has contributed to this anthropogenic heat flux (Khamchiangta and Dhakal, 2019). Alongside concerns about residents' physical and mental health on account of heat stress (Arifwidodo and Chandrasiri, 2020), the high relative humidity all year in Thailand (Taweekun and Tantiwichien, 2013) causes perspiration and body odor when walking to public transport stations. The air-conditioned comfort of private automobiles thus offers an attractive mobility option for many Bangkok residents [inforin4-mi].

In cluster [4], we highlight a further web of interactions and self-reinforcing loops that collectively contribute to locking-in reliance on private automobiles. In particular, we draw attention to the aggregated impacts of formal and informal institutions, infrastructure development, and the deficiency of knowledge and competences around integrated, sustainable transport planning. The first set of relations occurs between formal institutions and the expansion of road infrastructure. Thailand's road policies [FORIN4-MA] aim to meet the increasing demand for road space that has ensued from the surge in private vehicle purchases and the increase of freight traffic serving the automotive industry (Charoentrakulpeeti et al., 2006).

Specifically, the construction of vehicle-production factories on the outskirts of Bangkok has necessitated the investment of road infrastructure [INFRA1-MA] to attract foreign development investment and enable supply-chain development (Lecler, 2002; Chalermpong and Ratanawaraha, 2015). Japanese investments in particular have powered this trend. Indeed, 41% of foreign direct investment flows to Thailand for automobile and other industries between 2010 and 2017 have come from Japanese companies and government agencies (OECD, 2021). In parallel, after the first and second waves of car buying, successive infrastructure-development policies have sought to increase the convenience of car ownership by easing traffic congestion (Peungnumsai et al., 2020; Witchayaphong et al., 2020). This has resulted in a disproportionate allocation of public spending toward increasing Bangkok's road networks over other transport infrastructure, such as rail. Government data shows that road development has accounted for about 80% of the total annual investment budget for infrastructure development since 2009 (Ministry of Transport, 2020).

In particular, from 2016 to 2020, national highway development was the only category of transport infrastructure to receive increasing investment from government budgets and loans, rising around 10% compared to 2011–2015 (Ongkittikul, 2014). This directed public spending was underpinned by the government's goal to add 6,612 km of roadways across the country as part of its Intercity Motorway Development Master Plan. Formal commitments to expanding Thailand's road network continue. Setting a 20-year development plan for 2017–2036, the government is currently working to add more than 390,000 km of roads. These public investment policies have widened the gap between the development of road infrastructure and of alternative transport modes (OECD, 2020). And the resulting superior road network has increased people's preferences for using cars [inforin4-mi] (Wu and Pojani, 2016).

Use of private automobiles is also stimulated by other macro-scale formal institutions, such as government policies on buildings, fuel taxes, and tollway fees. The Thai Building Control Regulation [FORIN4-MA] has been criticized for failing to limit how much parking space can be installed around residential and office buildings (Chalermpong and Wibowo, 2007; Pongprasert and Kubota, 2018). This has led to a marked increase of parking spaces [infra2-mi] in private-sector-led real estate development, boosting the convenience of park-and-ride behavior for car users. Government policies—e.g., subsidies and tax breaks for consumers and the oil and gas industry—also incite vehicle use by reducing the burden of fuel costs [FORIN6-MA] (ADB, 2015). Meanwhile, tollway prices are subsidized to encourage road use (Charoentrakulpeeti et al., 2006).

With regard to the underdevelopment of public transportation, this is largely explained by the prioritization of road development in formal institutions [FORIN4-MA]. This has manifested as frequent political debates on the need for the state to invest in public transport. The development of public bus and rail networks has received inconsistent and less support [FORIN7-MA] than road development in terms of public funds, private investments, and development plans (Charoentrakulpeeti et al., 2006; Hossain, 2006) despite the its overloaded capacity (Peungnumsai et al., 2020). For example, a lack of political support has delayed the development and expansion of Bangkok's bus rapid transit system (BRT). The BRT—originally planned for 12 routes covering 185 km, as proposed in 2004—was delayed several times due to change in government and complicated institutional structures and bidding procedures. It finally opened in 2010, with coverage confined to a single 15 km route with low ridership capacity (Wu and Pojani, 2016). The BRT's limited coverage and capacity [INFRA3-MA] (Munira et al., 2013; Sanko et al., 2014; Pongprasert and Kubota, 2017) pose physical barriers to Bangkok residents able to access or ride public transport. These structural conditions at the macro scale thus reinforce preferences at the micro scale for using privately owned automobiles [inforins4-mi] (Witchayaphong et al., 2020).

Development of the public transportation network is also hampered by the impact of knowledge and competencies on infrastructure development. This occurs in two sequences of lock-in causes. The first starts from the low degree of coordination between the different government departments and agencies involved in Bangkok's transportation system [KNOCOM2-MA]. Though several national and city departments oversee urban transport planning, operation, regulation, monitoring, and so on, there is no agency that leads this effort or oversees collaboration among the relevant bodies. This has led to fragmented transport planning and inter-agency conflict (Hossain, 2006; Marks, 2020); it has also ultimately prevented the formation of an authoritative multi-modal transport master plan (Wu and Pojani, 2016).

The lack of coordinated competencies and planning to guide the sustainable development of Bangkok's transportation system has also hampered the emergence of efforts to integrate the different modes of transport (BRT, Metropolitan Rapid Transit, skyrail, etc.) [INFRA4-MA] (Ongkittikul, 2014). For example, Bangkok transport planners have missed an important opportunity to link and streamline fare collection for Metropolitan Rapid Transit and skyrail, which currently use separate systems (Chalermpong and Ratanawaraha, 2015). Most of BRT, meanwhile, lacks any integration with rail stations (Wu and Pojani, 2016). Inconveniences posed by the absence of an integrated ridership experience reduce accessibility to public transport, again increasing the probability of Bangkok residents using privately owned automobiles to fulfill their mobility needs.

The second sequence of lock-in causes relates to government agencies' limited competences in the design of safe and user-friendly infrastructures to promote the use of public transport [KNOCOM3-MA]. Understaffing and a lack of qualified technical consultants in public transportation agencies has resulted in poorly designed interconnectivity infrastructure [INFRA5-MA] around stations (Wu and Pojani, 2016). This occurs with the design of walkways in stations themselves (Prasertsubpakij and Nitivattananon, 2012) as much as the footpaths, pedestrian bridges, and shading devices (e.g., roofing or trees) around stations in the urban environment. The few Bangkok residents who choose to walk or cycle around thus feel uncomfortable and unsafe [inforin5-mi] (Prasertsubpakij and Nitivattananon, 2012; Munira et al., 2013). These structural conditions discourage residents from accessing stations on foot and push them further away from public transport, perpetuating demand for motorized travel, even for trips of just a few hundred meters.