1. Introduction
The urgency of accelerating low-carbon transitions is underscored by the Paris Agreement and Sustainable Development Goals, yet progress remains uneven across space (Roberts et al., Reference Roberts, Geels, Lockwood, Newell, Schmitz, Turnheim and Jordan2018; Sovacool et al., Reference Sovacool, Geels, Andersen, Grubb, Jordan, Kern, Kivimaa, Lockwood, Markard, Meadowcroft, Meckling, Moore, Raven, Rogge, Rosenbloom, Schmidt, Schot, Sharp, Stephenson and Yang2025; Yang et al., Reference Yang, Andersen and Sovacool2026). While technological advancements and declining costs of green technologies have enabled rapid uptake in some regions, systemic barriers, such as entrenched industrial legacies, policy fragmentation, and fossil fuel dependence, continue to stall diffusion elsewhere. These disparities raise a central question: How do socio-technical feedback dynamics enable or constrain low-carbon transitions?
Drawing on sustainability transitions literature (Geels, Reference Geels2002; Markard et al., Reference Markard, Raven and Truffer2012; Rip & Schot, Reference Rip and Schot2002) and socio-technical feedback studies (Alkemade & de Coninck, Reference Alkemade and de Coninck2021; Edmondson et al., Reference Edmondson, Kern and Rogge2019; Geels & Ayoub, Reference Geels and Ayoub2025; Kern & Rogge, Reference Kern and Rogge2018; Pierson, Reference Pierson1993; Rosenbloom et al., Reference Rosenbloom, Meadowcroft and Cashore2019), we examine why three U.S. states – California, Texas, and Michigan – exhibit distinct trajectories in their electric vehicle (EV) transitions, and how their industrial and climate policies interact with regional feedback loops to shape their transition paths.
Given that transportation accounts for nearly 29% of the nation's total greenhouse gas emissions (U.S. EPA, 2024), understanding the EV transitions is both analytically and policy relevant for the U.S. low-carbon transitions. Historically, U.S. EV-related policies have been fragmented. At the federal level, regulatory instruments such as the Corporate Average Fuel Economy standards mandated progressively cleaner vehicles and raised the cost of maintaining fossil-fuel-based production portfolios, nudging automakers toward electrification. At the subnational level, in 2024, California and 14 partner states adopted zero-emission vehicle (ZEV) standards that require an increasing share of new vehicle sales to be zero-emission, together covering nearly 36% of U.S. light-duty vehicle sales (NESCAUM, 2025). Since 2021, the Inflation Reduction Act has aligned EV policy with broader industrial and trade strategies, driving major automaker investments and spurring states to compete for federal grants supporting manufacturing and charging infrastructure.
However, under the second Trump administration, the federal EV-policy ecosystem is undergoing a significant reversal: regulatory mandates are being scaled back, infrastructure programs are being paused or reviewed, and state authorities are being curtailed. In this context, state governments have become the critical arenas for sustaining and advancing the transition momentum. Yet, U.S. states vary widely in their industrial structures, governance capacities, and policy orientations, particularly regarding the extent to which state governments support industrial policy intervention and climate-oriented transition strategies. Our comparative analysis shows that their distinct starting points shape different socio-technical feedback dynamics and EV transition trajectories. Understanding such variation is essential for designing context-sensitive policy interventions capable of maintaining transition momentum despite shifting national priorities.
We propose a novel framework that explains how policy interventions, industrial dynamics, and societal responses interact through reinforcing and balancing feedback loops to either accelerate or slow down transition processes. Reinforcing feedback loops, also known as positive loops, are self-amplifying processes in which change in one variable generates further change in the same direction; balancing feedback loops counteract and oppose change, seeking balance and equilibrium by self-correction (Sterman, Reference Sterman2000). Specifically, we identify five reinforcing feedback loops: industrial transformation, technological change, infrastructure investment, behavior change, and environmental legitimacy, and two balancing feedback loops rooted in legacy automotive and fossil-fuel economy lock-ins. We further argue that the operation, magnitude, and interaction of these feedback loops are conditioned by state-level contextual moderators, particularly governance capacity and policy orientation, that structure how feedback dynamics emerge and interact in different settings.
Our analysis extends existing debates on acceleration of sustainability transitions (Roberts et al., Reference Roberts, Geels, Lockwood, Newell, Schmitz, Turnheim and Jordan2018; Sovacool et al., Reference Sovacool, Geels, Andersen, Grubb, Jordan, Kern, Kivimaa, Lockwood, Markard, Meadowcroft, Meckling, Moore, Raven, Rogge, Rosenbloom, Schmidt, Schot, Sharp, Stephenson and Yang2025; Yang & Schot, Reference Yang and Schot2025) by showing how U.S. state-level EV transitions are shaped by interacting reinforcing and balancing feedback loops and institutional conditions. We make three key contributions. First, we develop a subnational feedback loop framework that extends recent positive socio-technical tipping point scholarship (Ayoub & Geels, Reference Ayoub and Geels2024; Geels & Ayoub, Reference Geels and Ayoub2023) to the state level where EV transitions occur. The framework incorporates both reinforcing and balancing loops as co-evolving features that shape EV transitions. Second, we integrate governance capacity and policy orientation as contextual moderators that explain why different states present different feedback loops that produce divergent transition outcomes – and show that these moderators can themselves be shaped by the studied socio-technical feedback dynamics. Third, we provide comparative empirical insights from California, Texas, and Michigan, representing a regulatory leader, a market-oriented follower, and a structurally constrained laggard, extending the empirical base beyond well-studied national contexts. In doing so, the paper advances a place-based understanding of low-carbon acceleration, demonstrating that effective policy design needs to align with the existing socio-technical feedback dynamics and institutional conditions of specific regions.
2. Socio-technical feedback and transition acceleration paths
2.1. Electrification of mobility as an industrial and socio-technical transformation
The electrification of mobility represents a profound industrial transformation, encompassing the restructuring of production systems, supply chains, and technological competencies. The shift from internal combustion engine (ICE) vehicles to EVs requires new manufacturing capabilities in batteries, power electronics, and software, while reconfiguring established automotive and fossil-fuel industries (Altenburg, Reference Altenburg2014). This transformation has implications for regional economies, employment structures, and technical skills, thereby positioning electrification as a driver of economic restructuring and technological upgrading (Altenburg et al., Reference Altenburg, Corrocher and Malerba2022; Meckling & Nahm, Reference Meckling and Nahm2019).
Meanwhile, mobility electrification should be understood as a socio-technical transition: a systemic reconfiguration of technologies, infrastructures, user practices, institutions, and cultural meanings (Geels, Reference Geels2002, Reference Geels2012; Geels & Schot, Reference Geels and Schot2007). The mobility system is embedded in complex social, institutional, and infrastructural arrangements that co-evolve over time. Electrification thus requires not only technological substitution but also coordinated change across multiple system elements: vehicle technologies, charging and grid infrastructures, consumer practices, policy regimes, and wider societal cultures. From this perspective, feedback loops arise from industrial and technological dynamics, and interactions between policy, user acceptance, and infrastructure readiness (Alkemade & de Coninck, Reference Alkemade and de Coninck2021).
Integrating insights from industrial transformation and socio-technical transition literatures enables a more comprehensive understanding of how socio-technical feedback loops shape transition trajectories. The industrial transformation perspective highlights processes of investment, production, and policy that drive economic restructuring (Andreoni & Chang, Reference Andreoni and Chang2019; Breetz et al., Reference Breetz, Mildenberger and Stokes2018), whereas the socio-technical transition framework emphasizes interactions between niche innovations, regime structures, and landscape pressures that condition the direction and pace of system change (Geels, Reference Geels2012). This dual lens captures how policy interventions and system dynamics jointly determine the dynamics of mobility electrification across states.
2.2. Feedback loops in the electrification of mobility
Although recent work has linked policy feedback to socio-technical transitions (Allen et al., Reference Allen, Moallemi, Raven and Sharp2026; Edmondson et al., Reference Edmondson, Kern and Rogge2019; Fesenfeld et al., Reference Fesenfeld, Schmid, Finger, Mathys and Schmidt2022; Geels & Ayoub, Reference Geels and Ayoub2025), limited research has examined how such socio-technical feedbacks interact differently within subnational jurisdictions where much policy implementation and industrial transformation occur. Policy feedback scholarship primarily examines how policies reshape political coalitions, institutional capacities, governance arrangements, and future policy trajectories, largely at the national level (Béland et al., Reference Béland, Rocco and Waddan2019; Daugbjerg & Kay, Reference Daugbjerg and Kay2020; Pierson, Reference Pierson1993). In parallel, socio-technical transition studies have emphasized the co-evolution of technologies, industries, infrastructures, user practices, and markets, often treating policy as an enabling or constraining context rather than as co-evolving and being shaped by feedback effects (Ayoub & Geels, Reference Ayoub and Geels2024; Geels & Ayoub, Reference Geels and Ayoub2023, Reference Geels and Ayoub2025). The mechanisms through which political, industrial, technological, and infrastructural dynamics recursively reinforce or constrain one another during transition processes remain insufficiently specified, particularly across regional and state-level contexts. Existing approaches also describe these interactions schematically, paying limited attention to how institutional variation across states shapes their operation and outcomes.
To address the gap, this paper develops a framework that conceptualizes the electrification of mobility as a complex adaptive socio-technical system shaped by interacting reinforcing and balancing feedback loops (Figure 1). In this paper, reinforcing feedback loops refer to self-amplifying dynamics through which initial transition developments generate cumulative effects that further accelerate system change. For example, EV adoption may stimulate infrastructure investment, technological learning, and industrial scaling, which subsequently further increase EV adoption. By contrast, balancing feedback loops refer to stabilizing dynamics that counteract or slow system transformation by reinforcing incumbent institutional, industrial, or infrastructural arrangements. Reinforcing loops amplify transition momentum once established, whereas balancing loops stabilize incumbent structures and slow change.
Socio-technical feedback loops and contextual moderators shaping EV transition dynamics.
Figure 1 Long description
EV transitions at the center, surrounded by various factors. Reinforcing feedback loops include Infrastructure Investment, Environmental Legitimacy, Behavior Change, Technological Change and Industrial Transformation. Infrastructure Investment is described as the expansion of charging infrastructure, reducing range anxiety and enabling broader EV adoption. Environmental Legitimacy notes that growing EV adoption improves air quality and strengthens support for environmental policy. Behavior Change involves changes in social norms, values, culture and routines. Technological Change highlights improved EV performance through learning by doing, economies of scale and knowledge spillovers. Industrial Transformation mentions manufacturing and service sector investments that accelerate technological learning and strengthen socio-political support. Balancing feedback loops include Oil Industry Resistance and Legacy Auto Lock-in. Oil Industry Resistance is characterized by pushback from fossil fuel industry interests. Legacy Auto Lock-in involves sunk costs, embedded skills, supply chains and institutional commitments that weaken EV production shifts by automakers. The diagram is enclosed in an oval with Policy Orientation on the left and Governance Capacity on the right. Reinforcing feedback loops are indicated by solid lines, while balancing feedback loops are indicated by dashed lines.
In our framework, we propose five core reinforcing loops: industrial transformation, technological change, behavior change, infrastructure investment, and environmental legitimacy. When activated simultaneously, these loops interact cumulatively, creating system-wide acceleration. At the same time, we propose two balancing loops, legacy automotive lock-in and fossil-fuel resistance, that constrain transition speed. Drawing on inspirations from policy feedback scholarship, our framework further examines how policy orientations and governance capacities shape the formation and evolution of these socio-technical feedback loops, while also being shaped over time by the socio-technical transition dynamics they produce.
Industrial transformation describes how state industrial policies and manufacturing incentives reduce investment risks and stimulate production capacity (Kemp, Reference Kemp2005; Mazzucato, Reference Mazzucato2013) and generate tangible economic gains that strengthen political support for continued policy commitment (Mazzucato, Reference Mazzucato2013; Rodrik, Reference Rodrik2014). Technological change captures cumulative innovation processes, whereby regulatory pressure and R&D support stimulate learning-by-doing and technological improvements (Hsieh et al., Reference Hsieh, Pan, Chiang and Green2019; Nemet, Reference Nemet2006; Nykvist & Nilsson, Reference Nykvist and Nilsson2015; Wilson, Reference Wilson2012), leading to higher performance (Barwick et al., Reference Barwick, Kwon, Li and Zahur2025; Goetzel & Hasanuzzaman, Reference Goetzel and Hasanuzzaman2022), lower costs, and wider diffusion (Arthur, Reference Arthur1989; David, Reference David1985; Forsythe et al., Reference Forsythe, Gillingham, Michalek and Whitefoot2023; Nemet, Reference Nemet2009). Behavior change reflects the role of social learning and changing norms as growing visibility and familiarity with EVs’ strengthen public acceptance and reinforce market demand (Axsen et al., Reference Axsen, Orlebar and Skippon2013; Liao et al., Reference Liao, Molin and Van Wee2017; Rezvani et al., Reference Rezvani, Jansson and Bodin2015; Sierzchula et al., Reference Sierzchula, Bakker, Maat and Van Wee2014). Infrastructure investment highlights the mutually reinforcing relationship among charging infrastructure, consumer confidence, market expansion (Gallagher & Muehlegger, Reference Gallagher and Muehlegger2011; White et al., Reference White, Carrel, Shi and Sintov2022), and complementary investment in electricity networks (Markard et al., Reference Markard, Geels and Raven2020; Yang & Schot, Reference Yang and Schot2025). Finally, environmental legitimacy captures how increasing public concern over air quality and climate risks reinforces political support for sustained policy intervention and continued EV deployment (Bayham et al., Reference Bayham, Burkhardt, Coffman, Hayashida and La Croix2022; Su & Diao, Reference Su and Diao2025).
Counteracting these reinforcing mechanisms are two balancing feedback loops. Legacy auto lock-in arises from incumbent firms’ sunk investments, established supply chains, and organizational capabilities associated with ICE technologies (Bohnsack et al., Reference Bohnsack, Kolk, Pinkse and Bidmon2020; Klitkou et al., Reference Klitkou, Bolwig, Hansen and Wessberg2015), reducing the pace of industrial restructuring (Wesseling et al., Reference Wesseling, Farla and Hekkert2015a, Reference Wesseling, Niesten, Faber and Hekkert2015b). Oil industry resistance reflects the political, institutional, and economic influence of fossil-fuel interests (Bohnsack et al., Reference Bohnsack, Kolk, Pinkse and Bidmon2020; Culhane et al., Reference Culhane, Hall and Roberts2021), which can delay or weaken policies supporting transport electrification (Newell et al., Reference Newell, Geels and Sovacool2022).
Together, these interacting reinforcing and balancing loops shape both the pace and direction of EV transitions. Table 1 summarizes the characteristics, mechanisms, and expected system-level effects of each feedback loop.
Overview of reinforcing and balancing feedback loops in EV transitions
Table 1 Long description
Columns list Feedback Loop, Type, Changing Mechanism, Expected Effect, and Literature for seven feedback loops related to electric vehicle transitions. Reinforcing loops are industrial transformation, technological change, behavior change, infrastructure investment, and environmental legitimacy. Industrial transformation links industrial policy to reduced investment risk, expanded manufacturing capacity, economies of scale, learning by doing, employment creation, localized supply chains, and stronger political support, with expected effects of expanded capacity, lower production costs, stronger competitiveness, and stronger political commitment. Technological change links research and development support, regulation, and emissions standards to innovation, learning by doing, knowledge spillovers, improved electric vehicle performance, and cost reductions through cumulative development and experience curves, with expected effects of faster innovation, improved performance and competitiveness, and increased diffusion. Behavior change links financial incentives with social norms, environmental values, consumer experience, and visibility of electric vehicles, with expected effects of reinforced consumer demand and faster market diffusion. Infrastructure investment links charging expansion to convenience, consumer confidence, further public and private investment, and reinforcement from growing electric vehicle fleets to grid modernization, distributed energy resources deployment, and vehicle to grid technology development, with expected effects of improved system attractiveness, stronger energy mobility integration, and support for large scale deployment. Environmental legitimacy links concern over air pollution and climate change to support for regulatory intervention and policy continuity, with expected effects of maintained long term political commitment and political legitimacy. Balancing loops are legacy auto lock in and oil industry resistance. Legacy auto lock in lists sunk investments, embedded skills, supplier networks, production capabilities, and labor institutions tied to internal combustion engine technologies, with expected effects of slower industrial transformation and electric vehicle diffusion, while oil industry resistance lists political influence, lobbying, and institutional power that preserve existing infrastructures and policy arrangements, with expected effects of delayed or weakened electrification policies and reinforced fossil fuel dependence.
Our framework extends existing analyses of positive tipping points in low-carbon transitions. While previous studies have traced national-level feedbacks between techno-economic improvements (Lenton et al., Reference Lenton, Benson, Smith, Ewer, Lanel, Petykowski, Powell, Abrams, Blomsma and Sharpe2022; Nijsse et al., Reference Nijsse, Sharpe, Sahastrabuddhe and Lenton2024) and the reorientation of actors and interest groups (Ayoub & Geels, Reference Ayoub and Geels2024; Geels & Ayoub, Reference Geels and Ayoub2023; Rosenbloom et al., Reference Rosenbloom, Meadowcroft and Cashore2019), this paper foregrounds subnational political economy by examining how state-level governance capacities, incumbent-state relationships, and coalitional politics condition whether feedback loops gain traction, stabilize, or stall. In addition, our framework incorporates balancing and reinforcing feedback loops as co-evolving dynamics that jointly shape processes of acceleration and deceleration, rather than treating acceleration primarily as a phenomenon that emerges only after positive tipping points have been crossed.
2.3. Contextual moderators: governance capacity and policy orientation
The feedback loops discussed above primarily capture the internal dynamics of socio-technical systems once transition processes are underway. However, such within-system processes alone do not sufficiently account for institutional variations across states (Lockwood, Reference Lockwood2015, Reference Lockwood2022). We argue that the operation of these feedback dynamics is embedded within broader institutional and political configurations, including prevailing sources of political legitimacy, dominant problem framings, and governance capacities that shape both the initiation and sustaining of a transition (Gallagher, Reference Gallagher2013). Existing research suggests that EV adoption in many U.S. states is strongly conditioned by environmental ideology and partisan politics, with conservative states exhibiting weaker policy commitment despite rising EV uptake (Hayashida et al., Reference Hayashida, La Croix and Coffman2021). Importantly, the policy orientation and governance capacity are not external to the system but co-evolve with the socio-technical feedback dynamics overtime. Accordingly, as illustrated in Figure 1, we conceptualize state-level governance capacity and policy orientation as mutually constitutive with the strength and direction of the identified reinforcing and balancing feedback loops.
Governance capacity refers to a state’s institutional ability to coordinate across agencies, mobilize fiscal resources, and sustain long-term regulatory commitments (Domorenok et al., Reference Domorenok, Graziano and Polverari2021; Van Popering-Verkerk et al., Reference Van Popering-verkerk, Molenveld, Duijn, Van leeuwen and Van Buuren2022). Policy orientation captures whether a state relies on state-led intervention or market-based approaches to steer transitions (Wesseling, Reference Wesseling2016). These two factors condition the degree to which policies align with industrial strategies and infrastructure development (Meckling & Nahm, Reference Meckling and Nahm2018). States with strong regulatory and fiscal capacities, such as California, can design and sequence interventions that reinforce positive feedback and mitigate balancing constraints (Meckling et al., Reference Meckling, Sterner and Wagner2017). By contrast, states with institutional preferences for market-led investment (e.g., Texas) or those facing limited fiscal capacity and partisan competition (e.g., Michigan) tend to rely more heavily on federal incentives, resulting in weaker coordination across industrial, infrastructural, and environmental domains.
Over time, governance capacity and policy orientation may evolve depending on policy outcomes and institutional learning. Successful policy outcomes, such as improved air quality, green job creation, and technological leadership, may reinforce public trust and environmental consciousness and generate broader political support for continued policy intervention, thereby accelerating the transition (Allan et al., Reference Allan, Lewis and Oatley2021). In this sense, governance capacity and policy orientation are not fully exogenous conditions but co-evolve with socio-technical feedback dynamics. Where incumbent industries hold concentrated economic and political weight, their interests may become embedded in governance arrangements over time, making state capacity partly endogenous to the lock-in dynamics that the transition must overcome. This paper therefore treats these factors as contextual moderators while examining how they are reshaped through ongoing socio-technical change. This framework allows for cross-state comparison by identifying how the presence, strength, and interaction of feedback loops explain variations in EV transition speed and scale. Below, we will apply this framework to California, Texas, and Michigan to assess how their policy choices interact with state-specific system dynamics.
3. Methodology and research design
This study employs a comparative case study approach to explore how distinct feedback loops shape the pace and direction of EV transition at the subnational level in the United States in the past two decades (George & Bennett, Reference George and Bennett2004). We focus on the state level as the unit of analysis. The U.S. states and local governments have the leeway to design and implement policies in areas such as transportation, infrastructure, health, environment, and energy to promote sustainable development (Bedsworth & Hanak, Reference Bedsworth and Hanak2013). In line with a theoretical sampling strategy (Eisenhardt, Reference Eisenhardt1989; Eisenhardt & Graebner, Reference Eisenhardt and Graebner2007), we select three U.S. states – California, Texas, and Michigan – to capture how contrasting feedback loops, moderated by their contextual factors, shape their distinct EV transitions. The three states vary in EV registrations, annual EV market share, and their share of the total vehicle stock, which reflect differences in their EV transition pace and in socio-economic contexts, such as industrial bases, automotive manufacturing employment, and political conditions (Table 2).
Comparative EV transitions of California, Texas, and Michigan
Table 2 Long description
Three states, California, Texas, and Michigan, are compared across EV transition pace, economic profile, political context, EV adoption measures for 2023, and 2023 motor vehicle manufacturing employment in thousands. EV transition pace is listed as rapid speed for California, moderate speed for Texas, and lagging speed for Michigan. Economic profile is leading digital economy for California, oil economy for Texas, and strong auto-industry for Michigan; political context is Democratic majority for California, Republican majority for Texas, and politically competitive for Michigan. EV registration in 2023 is 0.23 million in Texas and 0.05 million in Michigan. EV annual market share in 2023 is 6.14% in Texas and 3.96% in Michigan. EV share of the total vehicle stock in 2023 is 3.4% in California, 0.9% in Texas, and 0.6% in Michigan; EV per 10,000 people in 2023 is 322 in California, 75 in Texas, and 50 in Michigan. Motor vehicle manufacturing employment in 2023 is 32.5 in California, 16 in Texas, and 48.9 in Michigan, with units shown as in thousands. The EV annual market share row includes a 2023 footnote marker.
Source: U.S. Department of Energy; Environmental Defense Fund, 2025; U.S. Bureau of Labor Statistics; Alliance for Automotive Innovation, 2023.
Notes: Summarized by authors.
a The EV market share data reflects the first three quarters of 2023.
California represents a case of a leading EV transition, underpinned by strong regulatory capacity, long-standing environmental policy commitments, and a dynamic industrial innovation system. Texas, by contrast, exhibits a moderate pace of transition, with a strong oil-and-gas-based economy, a diversified industrial base, evolving political priorities, and expanding but uneven EV adoption. Michigan represents a case of a relatively slow EV transition, characterized by a strong legacy automotive industry and more incremental policy changes. The cases also reflect distinct partisan environments: California’s stable Democratic governance has supported long-standing environmental policy commitments; Texas’ predominantly Republican governance has shaped a different regulatory orientation; and Michigan's swing-state dynamics have produced more fluctuating political and policy priorities.
Data collection involved two steps: desk research and interviews with stakeholders. Policy documents, reports, news articles, and EV-related statistics were collected and analyzed to provide context and gain a broader understanding of the EV transition at the state level. The second step involved conducting semi-structured interviews with key stakeholders, including government officials, researchers, charging companies, utilities, the EV industry, and analysts from think tanks and NGOs, at both the federal and state levels. A total of 31 interviews were conducted between July and October 2025, each lasting an average of 45 minutes, and all were conducted in English (Table 3). Interview consent was obtained verbally or in writing, with anonymity and confidentiality ensured.
Interviewees’ distributions across states
Table 3 Long description
Counts of interviewees are listed by Region or State and Stakeholder Sector. Michigan has Industry 1, Academia 3, Government 4, and NGO 2. Texas has Academia 4, NGO 5, and Utilities 1. California has Academia 2, Government 2, and Industry 4. Federal has NGO 3. The Total row reports 31 interviewees.
We adopted an abductive research approach, iteratively moving between an initial theoretical feedback-loop framework and empirical patterns emerging from the data. The conceptual framework provided initial guidance but was refined as interview insights revealed new feedback dynamics and contextual moderators. All interview transcripts were coded in NVivo. We first applied deductive codes derived from the theoretical framework and then added inductive codes to capture unanticipated themes. This enabled within-case analysis and cross-case comparison across California, Texas, and Michigan, ensuring that our final analytical framework reflected both theory and emergent empirical evidence. The synthesis of the comparative analysis results is presented in Table 4.
Comparison of three states’ results
Table 4 Long description
Three columns compare California, Texas, and Michigan across reinforcing feedback loops, balancing feedback loops, and one contextual moderator. Reinforcing loops list five areas: Industrial transformation, R1 is marked xx for California and Texas and x for Michigan; Technological change, R2 is xx for California and x for Texas and Michigan; Behavior change, R3 is xx for California and x for Texas with Michigan blank; Infrastructure investment, R4 is xx for California and x for Texas with Michigan blank; Environmental legitimacy, R5 is xx for California, blank for Texas, and x for Michigan. Balancing loops list two areas: Legacy auto industry lock in, B1 is blank for California and Texas and xx for Michigan; Oil industry resistance, B2 is blank for California, xx for Texas, and x for Michigan. The contextual moderator row lists, for California, Strong governance capacity and direct guidance; for Texas, Market driven with no clear policy direction; and for Michigan, Partisan divisions and policy discontinuity. Marks use xx, x, or blank to indicate the strength or absence of evidence for each loop.
Note: XX = strongly observable feedback loop; X = weak or emerging feedback loop; Blank = no clear evidence of the loop.
4. Comparing EV transitions: California, Michigan, and Texas
4.1. California: a regulatory leader
California’s EV transition represents one of the most mature examples of mutually reinforcing feedback loops among environmental legitimacy, industrial transformation, technological change, behavioral change, and infrastructure investment. The state’s governance capacities evolved from air-quality regulation (1970s–1980s), to technology-push mandates and market formation policies (1990s–2010s), and eventually toward a comprehensive, coordinated multiagency governance architecture (2020s) that integrates climate, industrial, infrastructure, and innovation policies.
The origins of this transition lie in California’s response to severe air pollution. The establishment of the California Air Resources Board (CARB) in 1967 and increasingly stringent vehicle emission standards created both the regulatory capacity and political legitimacy for later EV policies. Public concern over smog fostered durable political coalitions among environmental organizations, public health agencies, and regulators, embedding environmental leadership within the state’s policy orientation while strengthening the administrative capacity to pursue industrial and climate objectives (Collantes & Sperling, Reference Collantes and Sperling2008; Slifko & Rigby, Reference Slifko and Rigby1995; Sperling & Eggert, Reference Sperling and Eggert2014) (R5).
As one interviewee observed: ‘The shift really goes back to the history of the California Air Resources Board, which became a robust well-staffed agency with significant technical, policy and legal expertise because of the air pollution crises of the 1960s and 1970s in California’ (Interview with California EV and Energy Policy Analyst, 2025).
The shared framing of EVs as solutions to both environmental pollution and economic stagnation, especially during the post–Cold War decline of the aerospace and defense industries in Southern California, generated a powerful political coalition that bridged environmental, industrial, and labor constituencies (Slifko & Rigby, Reference Slifko and Rigby1995). This alignment enabled local advocacy groups and state agencies to jointly construct the EV transition as simultaneously an environmental imperative and a strategy for industrial restructuring and economic development (R1, R5).
This environmental legitimacy underpinned California’s pioneering ZEV Mandate in 1990. By requiring automakers to supply increasing shares of ZEVs and introducing tradable ZEV credits, the mandate combined technology-push regulation and market-pull incentives, creating a protected niche for EV innovation (Collantes & Sperling, Reference Collantes and Sperling2008; Kemp, Reference Kemp2005) (R2). Although incumbent automakers initially resisted the policy, sustained regulatory commitment reduced market uncertainty, improved technological performance, and strengthened industrial capability building, gradually transforming EV development from a matter of regulatory compliance into a source of long-term competitiveness (Collantes & Sperling, Reference Collantes and Sperling2008; Shnayerson, Reference Shnayerson1996) . Simultaneously, growing public acceptance of EVs, reinforced by environmental concerns and changing social norms, strengthened behavioral feedbacks that supported market expansion, which in turn further strengthened industry actors’ confidence (R1, R2, R5).
Industrial transformation accelerated during the early 2010s following Tesla’s acquisition of the former NUMMI Fremont factory, which demonstrated the commercial viability of large-scale EV manufacturing and anchored California's emerging EV industrial cluster (R1). At the same time, governance evolved beyond emissions regulation toward coordinated industrial development. Executive Orders such as B-16-12 (2012), 2013 ZEV Action Plan, and subsequent initiatives aligned CARB, the California Energy Commission (CEC), the California Public Utilities Commission (CPUC), and the Governor’s Office of Business and Economic Development (GO-Biz) around complementary objectives, including manufacturing, charging infrastructure, workforce development, and market expansion. Together, these policies created reinforcing interactions between regulatory certainty, industrial investment, infrastructure deployment, and technological innovation, reducing investment risk while stimulating localized manufacturing and private capital mobilization (R1, R2, R4, R5). By 2025, California hosted more than 54 ZEV and related component manufacturers, leading the United States in EV production and employment (California Energy Commission, 2025).
California also illustrates how balancing feedbacks can weaken over time. Although fossil fuel and automotive incumbents initially opposed the ZEV mandate and later challenged broader climate policies, including attempts to suspend key regulatory instruments such as AB 32 through Proposition 23, which was ultimately rejected by voters (Collantes & Sperling, Reference Collantes and Sperling2008). At the federal level, periodic efforts to revoke California's Clean Air Act waiver introduced regulatory uncertainty, temporarily dampening investment confidence in the EV sector (Lemphers et al., Reference Lemphers, Bernstein, Hoffmann and Wolfe2022). However, these interventions increasingly failed to reverse the trajectory of California's EV transition. Strong public support, institutional resilience, and sustained political commitment prevented incumbent resistance from overturning the reinforcing dynamics generated by industrial development, infrastructure investment, and environmental legitimacy.
Overall, California demonstrates how environmental crises generated enduring legitimacy for ambitious state intervention, how technology-push and demand-pull mechanisms evolved into industrial transformation, and how governance capacities expanded through iterative institutional learning and cross-agency coordination. The co-evolution of policy, industry, infrastructure, and social acceptance enabled California to sustain one of the world’s most advanced EV transitions.
4.2. Texas: a market-oriented follower
In Texas, the EV transition has become increasingly self-reinforcing through market dynamics that expand adoption beyond affluent early adopters. The decision to purchase EVs is shifting from being an aspirational or luxury choice to a practical and financially sensible one, as purchase costs of EVs remain a key determinant for adopters (Adepetu & Keshav, Reference Adepetu and Keshav2017; Rahman et al., Reference Rahman, Suryawan, Suhardono, Nguyen and Lee2025; Tseng et al., Reference Tseng, Wu and Liu2013). The growing menu of EV options, which range from high-end Tesla and BMW models with luxury-oriented technological features to more affordable options such as the Chevy Bolt, Nissan Leaf, and various Kia models, broadens access across income groups. This diversification normalizes EV ownership and generates peer visibility and social proof, reinforcing consumer confidence and illustrating the technological and behavioral change loops (R2; R3). For residents in Texas, the decision calculus is increasingly economic: with long driving distances and high monthly gasoline bills, the electric charging option effectively lowers the cost of driving. New adoptions incentivize further investment in charging infrastructure and dealership offerings, creating a positively reinforcing loop between consumer demand, affordability, and market response (R4). That is, the EV transition in Texas can be illustrated as a self-reinforcing market ecosystem in which consumer preferences and cost advantages drive the transition.
In addition, Texas has fostered another reinforcing loop of industrial development and economic strategy (R1). State and local incentives, such as manufacturing capacity, land availability, and economic incentives, have attracted major EV manufacturers and battery producers to establish large-scale facilities, such as Tesla’s Gigafactory in Austin, Texas (Hopkins & Lazonick, Reference Hopkins and Lazonick2024). Tesla received $64.5 million in state and local subsidies for the construction of this factory (Hopkins & Lazonick, Reference Hopkins and Lazonick2024). Furthermore, Texas has no individual state income tax, and it is set to receive $407.8 million between the 2022 and 2026 fiscal years from the National Electric Vehicle Infrastructure (NEVI) program to build out the EV charging network (Kraus et al., Reference Kraus, Kuzio, Askariyeh, Shukla, Naranjo, Harbin, Grossmann, Ramani and Ettelman2025). Critically, these investments generate visible employment and supply chain development that build a robust manufacturing ecosystem and local economic buy-in, a defining feature of the industrial transformation reinforcing dynamics. As the EV production base expands, political and financial interests within the state become increasingly tied to its success, reinforcing a positive feedback loop in which economic incentives, infrastructure growth, and manufacturing commitments sustain Texas’ role in the EV supply chain in the United States.
Yet, the EV transition in Texas is not without resistance. Legacy industrial capabilities tied to the ICE economy, including sunk capital investments, specialized labor pools, supplier networks, and petroleum-linked infrastructure, create path dependencies that reinforce the existing transportation regime. Unlike the traditional Michigan auto-belt dynamic, where lock-in emerges through vertically integrated automobile manufacturing and legacy automakers, the Texas case is rooted more deeply in the state’s broader hydrocarbon political economy, including oil extraction, refining, petrochemicals, and fuel distribution networks (Yang, Reference Yang2025). Specifically, the politically influential oil and gas sector in Texas introduces both friction and subtle reinforcement into the EV transition process (B2). Fossil fuel industry interests have expressed opposition to EVs, in part because EV adoption reduces gasoline demand. The declining gas tax revenues have prompted policymakers to introduce EV registration fees to help offset highway funding shortfalls. For example, Texas imposes a $200 annual registration fee on EVs to compensate for lost motor fuel tax revenue, which incurs additional costs for EV owners and dampens demand.
Meanwhile, energy producers are simultaneously diversifying into renewables: Texas leads the United States in wind electricity generation in 2024, contributing 28% of national output (U.S. EIA, 2025). Because Texas’ electric grid remains dependent on natural gas, these incumbents paradoxically benefit from higher electricity demand due to EV growth. This creates a complex dynamic within Texas’ EV ecosystem, as one interviewee noted: ‘while solar and wind increase the renewable share of the electricity mix, natural gas continues to coexist alongside these sources in powering the grid and by extension, electric vehicles’ (interview with Texas energy analyst, 2025).
Another point of challenge emerges within car dealerships. Interviewees repeatedly highlight inadequate training and limited knowledge of EVs among salespeople, hindering EV promotion and sales. Salespeople who lack specific knowledge of EV performance, charging, or maintenance reflect a lack of institutional investments in EV education, potentially creating friction in the positively reinforcing feedback loop. Lastly, the reinforcing environmental legitimacy dynamic appears considerably weaker in the case of Texas compared with the other two cases. Although several metropolitan areas in Texas continue to experience air-quality improvement due to EV adoption, environmental concern has not consistently translated into robust social awareness or sustained support for environmental or climate-oriented policies.
4.3. Michigan: a structurally constrained laggard
Michigan’s EV transition has been shaped by its legacy automotive sector and by governance constraints partly rooted in that industrial inheritance. The auto industry contributed around 27% of gross state product and over 20% of private sector employment (MichAuto, 2024). This concentration has generated not only economic path dependence but also a political economy in which the investment decisions, employment base, and lobbying power of the Big Three limit the state’s institutional capacity to pursue an independent transition strategy.
As a state energy official observed:
The state itself is trying to find policies that accommodate the OEMs (Original Equipment Manufacturers) for their business practices versus leading with mandates such as ‘you must electrify all your sales by this date.’ … It really is, because of the economic impact of auto OEMs, [a matter of] saying ‘we're kind of taking your lead — what do you need, and we'll make sure that there's an environment here. (Interview, 2025)
Michigan’s governance weaknesses are substantially shaped by the Legacy Auto Lock-in Loop (B1). This dynamic is mostly reflected in the state's policy instability, which constrains its ability to establish a more effective Industrial Transformation Loop (R1). Governor Granholm (2003–2010) promoted a green economy strategy to reduce the heavy reliance on ICE automotive manufacturing, but these diversification initiatives were largely scaled back under Governor Snyder (2011–2018), who favored market-oriented approaches more aligned with the preferences of incumbent automakers and their supplier networks. Governor Whitmer (2019–present) revived state-led industrial policy for the EV and battery industries through the Strategic Outreach and Attraction Reserve (SOAR) Fund. Yet the removal of SOAR’s appropriation from the FY2026 budget underscores the fragility of the political constituency for proactive industrial transformation.
The reinforcing feedback mechanism required for R1 has not materialized in Michigan, as successive administrations have reset the policy landscape before industrial policies could generate sufficient employment and supply-chain gains to build durable coalitional support. In addition, the Big Three’s partial EV reorientation has been driven primarily by regulatory signals from California and the EU rather than by Michigan’s own policies.
Michigan possesses globally significant engineering and automotive R&D capacity. However, these capabilities were oriented toward ICE optimization for decades, and their redeployment toward EV-specific technical challenges has proceeded only incrementally. In Michigan, the limited diffusion of EV-related knowledge spillovers has been driven by corporate strategies that historically treated EVs as compliance products (Wesseling et al., Reference Wesseling, Niesten, Faber and Hekkert2015b). As one automotive industry researcher recounted in the early 2010s: ‘everyone else did their arithmetic, and says, no way you can make money on a hybrid, let alone on anything electric’ (Interview, 2025). The Big Three’s product planning long remained guided by ICE-dominated demand forecasts (Edwards et al., Reference Edwards, Hill and Szakaly2007; Siegel et al., Reference Siegel, Burrows and LaCivita1978), limiting the incentive to invest in rapid EV-specific technological learning.
Interdepartmental coordination is also difficult without clear legislation and dedicated funding streams. For example, the MI Healthy Climate Plan (2022), which sets an economy-wide carbon-neutrality goal for 2050, is a nonbinding roadmap. Because it establishes no regulatory mandates, its implementation depends largely on follow-up legislation, budget allocations, and agency actions. Part of what sustains this institutional fragility is the absence of a strong Environmental Legitimacy Loop (R5). Unlike California, where EV policy draws legitimacy from the transportation-air-quality nexus, Michigan’s environmental politics have centered on water quality and industrial contamination.
A similar stagnation is evident in charging infrastructure. Public direct current fast-charging deployment remains heavily concentrated in southeast Michigan, with most of the rest of the state served by a thin Level 2 network that limits long-distance EV travel (Anderson Economic Group, 2024). Federal NEVI funding, alongside the state’s Charge Up Michigan grant program, has begun to support charging deployment. However, these programs have functioned more as catch-up responses to demand than as market-shaping investment (Michigan, 2024). Where charging density remains limited, adoption signals to utilities and private capital remain weak, preventing the political coalition for further infrastructure appropriations from consolidating and leaving the Infrastructure Investment Loop stalled (R4).
Michigan’s EV ownership is concentrated in dense, higher-income, Democratic-leaning counties, while ICE familiarity and ‘buy local’ loyalty continue to shape consumer expectations across most of the rest of the state. These geographic and partisan cleavages also carry institutional consequences: in 2025, Michigan lawmakers passed a road-funding agreement that made Michigan EV owners’ vehicle registration fees the highest in the nation. Thus, the Behavior Change Loop (R3) is not only underdeveloped but also being actively undermined through the legislative process.
Alongside these stalled reinforcing dynamics, active resistance from fossil-fuel incumbents has become more visible. Marathon Petroleum’s Detroit refinery – the only petroleum refinery in the state – received an EGLE (Michigan Department of Environment, Great Lakes, and Energy, the state regulatory agency responsible for environmental protection) permit in September 2024, lifting prior production capacity limits, signaling continued institutional accommodation of fossil-fuel infrastructure. A particularly salient manifestation of the oil industry resistance feedback loop (B2) occurred in January 2026, when Michigan Attorney General Dana Nessel initiated a federal antitrust lawsuit against BP, Chevron, ExxonMobil, Shell, and the American Petroleum Institute, alleging coordinated efforts to suppress competition from EVs and renewable energy within Michigan’s markets.
5. Explaining the cross-state differences
5.1. Feedback loops and EV transition dynamics
As illustrated in the cases, the divergent paces of the EV transition in California, Michigan, and Texas are manifested through distinct configurations of five reinforcing (R1–R5) and two balancing (B1 and B2) feedback loops, moderated by their different institutional contexts. Figure 2 visualizes how these dynamics interact to accelerate or restrain EV transitions across states, capturing these processes in a coherent framework.
Dynamics of reinforcing and balancing loops in electric vehicle transitions.
Figure 2 Long description
Various feedback loops affecting electric vehicle (EV) transitions. Key components include EV Demand, Grid Capacity, Reliable Electricity Supply, Charging Infrastructure and EV Registration Fees. The loops are labeled R1 to R5 and B1 to B2, representing different feedback mechanisms. R1 is the Industrial transformation loop, involving Economic & Employment Gains from EV, Manufacturing Capacity, Manufacturing Incentives and Industrial Policy. R2 is the Technological change feedback loop, involving EV Performance, EV Profitability and R&D Investments. R3 is the Behavior change feedback loop, involving Visibility, Climate Policy, Social Awareness, Air Quality Improvement and EV Adoption. R4 is the Infrastructure investment feedback loop, involving Grid Infrastructure Investments, Utilities Profits, Electricity Demand and NEVI. R5 is the Environmental legitimacy feedback loop, involving Social Awareness, Air Quality Improvement and EV Adoption. B1 is the Incumbent protection balancing loop, involving ICE Resistance, ICE Market Share, ICE Profitability, ICE Vehicle Demand and ICE Vehicle Adoption. B2 is the Oil industry resistance loop, involving Lobbying & Resistance from Oil Industry, Gasoline Demand and ICE Vehicle Adoption. Arrows indicate positive or negative influences between components, showing how these dynamics interact to accelerate or restrain EV transitions.
As shown in Table 4, California demonstrates strong and mutually reinforcing positive loops across the domains of industrial transformation, technological change, infrastructural investment, behavior change, and environmental legitimacy, generating a cumulative virtuous cycle. By contrast, Texas exhibits more moderate reinforcing dynamics that are consistently offset by powerful balancing forces rooted in the state’s hydrocarbon-dependent economy. Michigan shows relatively weak reinforcing loops, which are constrained by a powerful balancing loop arising from the legacy automotive sector and its entrenched institutional path dependencies.
Across cases, the Industrial Transformation Loop (R1) shows vast divergence. California’s coordinated regulatory and investment instruments produce a robust virtuous cycle linking industrial capacity, policy commitment, and market expansion. For example, CARB’s regulatory mandate coordinates with GO-Biz’s investment facilitation and the California Energy Commission’s financing programs, illustrating how industrial transformation can reinforce both economic and climate objectives. Texas displays partial industrial momentum, but it remains uneven due to structural dependence on fossil-fuel sectors. Michigan’s reinforcing potential is repeatedly interrupted by partisan divisions and lock-in to the legacy automotive industry, preventing a stable industrial transformation loop from forming.
The Technological Change (R2) and Behavior Change (R3) loops also show stronger self-reinforcement in California. The combination of ZEV mandates and Silicon Valley’s entrepreneurial ecosystem helped initiate a technological-change feedback loop that enabled the emergence of numerous EV startups backed by venture capital. California’s ZEV mandate exemplifies this dynamic: continuous tightening of fleet requirements catalyzed ‘market-pull’ effects by creating predictable demand for ZEVs, while the regulatory targets and credits incentivized automakers, and finally accelerated technological upgrading in both domestic and foreign automakers, creating a virtuous cycle of innovation and deployment. By contrast, in Texas and Michigan, weaker demand signals and more ambiguous policy trajectories inhibit cumulative technological and behavioral feedbacks. In Michigan, the technological change loop remains externally dependent – the Big Three’s EV pivot responds to external regulatory signals rather than to a self-reinforcing internal innovation dynamic. Behavioral feedbacks are similarly stalled as EV adoption concentrates in only a few urban counties, leaving a constituency too thin to achieve consensus on the EV transition.
The Infrastructure Investment Loop (R4), linking EV penetration, grid capacity, and charging network expansion, follows similar patterns. California’s long-standing regulatory mandates and sustained public investment introduce a reinforcing cycle of reliability, deployment, and adoption. The growing electricity demand incentivizes utilities and governments to invest in grid modernization, which enhances consumer confidence and supports further EV growth. Higher grid capacity also encourages charging infrastructure development, and this reflects how an expanding charging network, supported by NEVI programs and other state incentives, reduces range anxiety and logistical barriers, thereby increasing EV demand as well. The infrastructure loop connects key stakeholders, including automakers, utilities, the oil and gas industry, government, and consumers. In Texas, rising electricity demand spurred grid expansion and EV-charging deployment, which partially offsets resistance from oil and gas interests, as natural-gas producers benefit from increased electricity generation. Michigan, however, with limited fiscal capacity and slower EV uptake, exhibits the weakest infrastructure reinforcement.
Finally, the Environmental Legitimacy Loop (R5) is strong in California, where air-quality concerns and long-standing public support for environmental measures provide a durable legitimacy base for successive policy interventions, stabilizing long-term investment expectations and justifying stronger regulatory trajectories. By contrast, Texas’ strong fossil-fuel orientation and comparatively weak regulatory pressure reduce the state’s capacity for sustained climate policy intervention. While in Michigan it faces a different challenge: the lack of comprehensive climate legislation and frequent partisan turnover have constrained its ability to maintain consistent and durable policy support.
Counteracting these reinforcing dynamics, two balancing loops exert significant influence. First, the Legacy Auto Lock-in Loop (B1) is most pronounced in Michigan, where the legacy automakers’ economic weight shapes its local industrial structure, governance capacity, and policy orientation. The state's institutional accommodation to OEM preferences means policy reacts to corporate strategies rather than directing industrial transformation. California overcame analogous automaker resistance in the 1990s through regulatory persistence backed by the air-quality mandate. The Oil Industry Resistance Loop (B2) operates most strongly in Texas, where fossil-fuel interests mobilize political and regulatory pressure that dilutes EV incentives and slows transition momentum.
An important implication of our framework is that the identified feedbacks should not be understood as unfolding through a fixed or linear temporal sequence. While some feedback mechanisms may emerge earlier than others in particular regional contexts (such as environmental legitimacy occurring earlier in California), socio-technical transitions are inherently nonlinear, recursive, and co-evolutionary processes. Reinforcing and balancing dynamics frequently interact simultaneously, overlap temporally, and mutually shape one another across different dimensions of the transition system. As such, our framework does not propose a deterministic stage-based model of EV industrial development but instead seeks to capture the evolving and path-dependent character of regional transition dynamics. This perspective is particularly important in explaining spatial divergence across U.S. states.
5.2. Regional contexts and institutional legacy
The above-identified feedback loops are fundamentally shaped by their states’ differing governance capacities and policy orientations. California’s approach, characterized by strong state capacity and a clear strategic direction that supports the EV transitions without picking specific technological winners, consistently amplifies its reinforcing feedback loops, leading to the most rapid EV acceleration. Conversely, Texas’ market-driven strategy, with no explicit direction, fails to generate sufficiently strong policy signals to overcome its powerful balancing loops, resulting in a moderate and market-contingent pace. The state’s deregulated electricity market further embodies this market-oriented governance model, in which state authorities exercise limited intervention in electricity generation or infrastructure investment decisions. Michigan’s limited fiscal capacity and policy pendulum create regulatory uncertainty that sustains the balancing loop of the legacy auto industry while failing to incentivize the reinforcing loops of industrial transformation and technological change necessary for the EV transition.
These comparative insights highlight the importance of context-specific strategies for accelerating low-carbon transitions. Regional political economies generate distinct constellations of feedback loops, which in turn shape transition trajectories. California’s stable regulatory environment strengthens reinforcing policy feedback, whereas Michigan and Texas exhibit fragmented, reactive approaches that attenuate cumulative momentum and rely more heavily on federal-level incentives.
Institutional legacies further structure these dynamics. In Michigan, deeply embedded ICE manufacturing systems create technological and institutional lock-in (Unruh, Reference Unruh2000). High sunk costs, employment dependencies, and political alignment with incumbent automakers anchor the state in incremental change. This lock-in limits policy reinforcement and slows structural reallocation of labor and capital toward electrification.
In contrast to Michigan’s manufacturing-based lock-in, Texas exhibits a fossil–industrial path dependency rooted in oil and gas extraction, refining, and electricity generation. The state’s fiscal structure and political institutions remain closely aligned with the interests of the fossil-fuel sector, shaping policy priorities and regulatory practices. Institutions such as the Public Utility Commission of Texas have historically prioritized energy reliability, low-cost electricity, and market competition over decarbonization or electrification mandates. These institutional arrangements perpetuate a ‘carbon lock-in’ dynamic (Seto et al., Reference Seto, Davis, Mitchell, Stokes, Unruh and Ürge-vorsatz2016), dampening positive feedback between policy, infrastructure, and market development, even as the state increasingly attracts EV manufacturing and battery investment.
6. Conclusion
This study demonstrates that socio-technical feedback loops provide a powerful analytical lens for understanding the uneven dynamics of low-carbon mobility transitions. Accelerating decarbonization requires not only strengthening reinforcing feedback loops that generate and sustain transitional momentum, but also weakening balancing loops that stabilize incumbent socio-technical regimes. However, as the three cases demonstrate, the emergence, interaction, and durability of these feedback dynamics are strongly shaped by regional governance capacities and policy orientations. Consequently, transition strategies cannot rely on standardized policy instruments alone. Effective policy interventions must be adapted to regional political–economic contexts and institutional legacies.
Our findings further suggest that transition policies should not only target immediate technological deployment or market expansion but also foster the longer-term institutional, infrastructural, and industrial conditions that enable reinforcing socio-technical dynamics to emerge and persist. In this regard, the framework developed in this paper provides a place-sensitive approach for analyzing how industrial structures, governance arrangements, infrastructural development, and societal legitimacy interact to shape differentiated transition trajectories across regions.
Our paper makes three key contributions to the literature on the acceleration of sustainability transitions and socio-technical feedback loops. First, we develop a subnational feedback loop framework that extends recent positive tipping point scholarship to the state level, where most industrial and climate policy implementation occurs. In doing so, we conceptualize how balancing loops co-evolve alongside reinforcing dynamics that jointly shape transition dynamics. Second, we explain why different states produce divergent socio-technical feedback dynamics by integrating governance capacity and policy orientation as contextual moderators. We argue that these contextual moderators can shape the strength and direction of both reinforcing and balancing feedback dynamics, while demonstrating that these moderators may themselves be shaped by socio-technical feedbacks. Third, we provide comparative empirical evidence from three U.S. states with distinct institutional legacies, thereby extending the empirical base for transition acceleration research beyond well-studied national contexts.
Overall, our central argument is that effective transition policies cannot be designed in isolation from regional contexts, but need to be targeted and tailored to the existing system dynamics and feedback loops to sustain long-term decarbonization. Amid recent federal policy rollbacks threatening EV progress, our findings underscore that transition momentum will depend on state-level governance capacities and interventions that generate local feedback dynamics. Recognizing and strengthening these regional reinforcing feedback loops is thus essential for maintaining long-term acceleration toward low-carbon mobility in the United States.
There are several avenues for future research. While we argue that our proposed framework has broader analytical applicability, its transferability to other geographical and institutional contexts requires further empirical validation. In particular, future research could examine how different political economies shape the operation of the seven feedback loops identified in this study. As our three cases already suggest, variation in governance capacities and policy orientations is likely to condition their dynamics. More comparative work is therefore needed to unpack the mechanisms through which such variation operates, including its effects on coalition formation, regulatory stability, and investment commitment.
Second, while our framework offers heuristic value for unpacking EV transition dynamics over time, its current application remains relatively static, capturing feedback loop configurations at a particular historical moment. Although our empirical analysis traces the historical development of EV policy and industrial dynamics to explain how these feedback loops emerge and evolve, it does not systematically reconstruct specific policy and investment events. Future research could adopt a more fine-grained, longitudinal, and event-based design that explicitly tracks critical historical developments, including policy shifts and investment decisions, to better capture how feedback loops are constituted, reinforced, or disrupted over time. Such an approach would enable a more explicit analysis of causality and temporal dynamics in socio-technical transitions. It would also allow for the investigation of interactions among the seven feedback loops, which are not addressed in our study due to our focus on their individual manifestation and their overall impact in shaping EV transitions across the three cases.
Third, future studies could develop a more scale-sensitive framework of socio-technical feedback loops, with important implications for place-based transformative innovation policy. Our analysis highlights that the spatial scale at which feedback loops operate may itself be evolving, particularly in an era of heightened geopolitical uncertainty and a shifting global economic order. While earlier phases of EV-related technological diffusion were strongly enabled by globalization and transnational knowledge flows, emerging geopolitical tensions, trade fragmentation, and industrial policy competition suggest a potential re-localization of certain dynamics. Some feedback mechanisms may increasingly operate at national or subnational scales, while others remain globally embedded. Future research could therefore systematically examine how global and local logics interact in shaping socio-technical feedback loops under conditions of rising protectionism and reconfigured global supply and value chains. This would enhance understanding of how multi-scalar dynamics shape feedback configurations, generate differentiated transition pathways, and enable more context-sensitive policy interventions.
Acknowledgements
We thank all interviewees for generously sharing their time, experiences, and insights, which were invaluable to this research. We gratefully acknowledge Professor Kelly Sims Gallagher for her guidance and insightful advice throughout this study. We also thank Shruti Katiyar for her excellent research assistance, especially her support in arranging and coordinating the participant interviews. We are grateful to the editor and the two anonymous reviewers for their constructive comments and suggestions that helped improve the quality of this manuscript. Any remaining errors are our own. ChatGPT (OpenAI) was used to assist with grammar checking and language editing of the manuscript. All intellectual content, analysis, and conclusions are the sole responsibility of the authors.
Author contributions
All authors contributed substantially and collaboratively to the conceptual development, methodological design, empirical investigation, and interpretation of findings. All authors participated in refining the theoretical framework, analyzing the empirical material, and revising the manuscript throughout the research process.
Data availability statement
The interview data are not publicly available due to confidentiality and ethical considerations concerning the research participants.
Funding statement
This research was supported by the Energy Foundation (G-2006-31570), Rockefeller Brothers Fund (22-601 and 25-420), and William and Flora Hewlett Foundation (2022-00792-GRA and 2026-07038-GRA). The first author also gratefully acknowledges the Climate Policy Lab (CPL) at The Fletcher School of Law and Diplomacy, Tufts University, for the visiting fellowship opportunity and for providing a stimulating research environment and opportunities for academic exchange that facilitated the development of this research and collaboration among the authors.
Competing interests
The authors declare no conflicts of interest.
Open access
