Policy–technology–finance synergy and sustainable industrial low-carbon transition

Abstract

Digital transformation is widely recognized as a critical driver of sustainable industrial development and the low-carbon transition. However, empirical evidence suggests that policy-driven digitalization does not automatically translate into improved environmental performance. This study examines whether and under what conditions the interaction among digital transformation, industrial policy, technological investment, and financial support can effectively promote industrial low-carbon transition, with particular attention to the role of policy–enterprise fit. Using a firm-level panel dataset of manufacturing enterprises during a period of rapid digital transformation (over 12,000 firm-year observations), we construct an integrated analytical framework linking digitalization, policy intervention, technological input, and financial support. Fixed-effects models and interaction analyses are employed to identify both direct effects and conditional synergies. The results show that digital transformation, technological investment, and financial support significantly reduce carbon intensity, whereas policy intervention alone exhibits weak and unstable effects. Importantly, a strong policy–technology–finance synergy emerges only when policy instruments are well-aligned with firm-level capabilities. Under high policy–enterprise fit, the joint interaction effect is substantially amplified and statistically significant, while under low fit, policy intervention weakens the emission-reduction effects of digital transformation, indicating a crowding-out mechanism. Further mechanism analysis suggests that information distortion, incentive misalignment, and compliance cost crowding-out are key channels through which policy mismatch undermines low-carbon outcomes. These findings highlight the conditional nature of policy effectiveness and provide evidence-based insights for designing more precise, adaptive, and sustainability-oriented industrial policies to support low-carbon transition in digitally transforming economies.

1 Introduction

Digital transformation has become a central force shaping industrial upgrading and low-carbon transition. With the diffusion of artificial intelligence, industrial internet platforms, smart sensors, and data-driven production systems, digital technologies are increasingly expected to improve production efficiency, optimize factor allocation, and support energy-saving and emission-reduction practices in manufacturing firms (Lyu et al., 2023; Wang and Li, 2022). Yet the environmental consequences of digitalization are not uniform. For firms, digital transformation is not only a technological process but also an organizational and institutional one, because its low-carbon payoff depends on policy incentives, innovation capability, and access to finance. The key question is therefore not merely whether digitalization promotes low-carbon transition, but under what conditions digitalization can be converted into sustained environmental improvement (Zhao et al., 2026; Bianchini et al., 2023).

Existing studies have largely examined the separate effects of digitalization, policy support, or green finance, but have paid limited attention to their joint synergy in shaping firm-level low-carbon transition (Wei and Liu, 2025). More importantly, the literature often assumes that policy support exerts homogeneous effects across firms, while the role of policy–enterprise fit remains underexplored (Liu and Song, 2023; Liu et al., 2024; Kyriakopoulos, 2024). Recent green-development studies reinforce the need for a more integrated perspective. For example, dynamic CGE evidence on the rollout of China's national carbon emissions trading scheme shows that policy sequencing and institutional design materially affect decarbonization outcomes. Research on green common institutional ownership further suggests that external governance arrangements can influence firms' energy-saving technological innovation. In addition, evidence on firm-perceived economic policy uncertainty and corporate greenwashing indicates that poorly aligned policy environments may induce symbolic rather than substantive green behavior. Taken together, these studies suggest that low-carbon transition is shaped not only by technological change itself, but also by the broader policy, governance, and financial environment surrounding firm transformation (Xue et al., 2022; He and Chen, 2025).

Despite these advances, important gaps remain in literature. First, most existing studies examine the independent effect of digitalization or policy support, but pay less attention to low-carbon transition from an integrated policy–technology–finance synergy framework. Second, prior research often implicitly assumes that policy support is broadly effective for firms, while overlooking the heterogeneity of policy effects across enterprises and the context dependence of policy implementation (Xue, 2025; Yang et al., 2024). Third, although the idea of policy–enterprise fit has been discussed in policy implementation and organizational studies, it has largely remained at a theoretical level, and there is still a lack of operationalized firm-level quantitative evidence on how such fit shapes the environmental outcomes of digital transformation. As a result, the conditional nature of policy effectiveness in the process of low-carbon transition remains insufficiently understood (Andrews et al., 2025; Waheed, 2025).

Using panel data of manufacturing firms in Guangzhou from 2015 to 2023, this study investigates whether and how digitalization-driven low-carbon transition depends on the joint interaction of policy support, technological investment, and financial development. We develop a firm-level indicator of policy–enterprise fit to capture the degree of alignment between policy design and enterprise characteristics, and then employ fixed-effects models, interaction analyses, heterogeneity tests, and endogeneity checks to identify the conditional synergy mechanism (Palma and Scotti, 2025; Zhao et al., 2026).

This study makes three contributions corresponding to the three gaps identified above. First, it constructs a policy–technology–finance synergy framework to explain industrial low-carbon transition as the outcome of multiple interacting forces rather than a single-factor process. Second, it introduces and quantifies policy–enterprise fit at the firm level, thereby extending the discussion of policy implementation from a conceptual argument to an empirically testable mechanism. Third, it reveals the conditional nature of policy effectiveness by showing that policy support strengthens low-carbon transition only when it is compatible with firm-level capabilities, whereas under low fit it may generate crowding-out effects that weaken the environmental benefits of digitalization. The remainder of this paper is structured as follows. Section 2 develops the theoretical framework and research hypotheses. Section 3 describes the data, variables, and empirical strategy. Section 4 presents empirical results. Section 5 discusses the findings and their implications. Section 6 concludes the paper.

2 Literature review and hypothesis development

2.1 Digitalization and industrial low-carbon transition

Digitalization has increasingly become a central driver of industrial transformation and environmental upgrading. In manufacturing sectors, digital technologies such as artificial intelligence, industrial internet platforms, smart sensors, and data-driven production systems can reshape production organization, improve operational efficiency, and support low-carbon transition. Existing studies suggest that digitalization may contribute to industrial low-carbon transition through at least three interrelated mechanisms (Wade and Vochozka, 2021). Digitalization improves resource allocation efficiency. By enhancing data availability, real-time information processing, and interdepartmental coordination, digital systems reduce information asymmetry and transaction costs, enabling firms to allocate labor, capital, and energy resources more efficiently. This improvement in internal and inter-firm coordination helps reduce redundant input use and carbon-intensive inefficiencies.

Digitalization helps optimize production processes and energy-use management. Smart manufacturing systems, predictive maintenance, automated scheduling, and digital monitoring tools allow firms to track production status and energy consumption in real time, identify waste points, and improve process control. These functions can reduce unnecessary energy use, lower production losses, and improve carbon efficiency (Blake et al., 2021). Digitalization promotes green innovation and knowledge diffusion. Digital platforms facilitate information sharing, cross-organizational learning, and the rapid diffusion of cleaner technologies and best practices. At the firm level, digital capability also strengthens innovation capacity by improving the identification of technological opportunities and reducing the uncertainty associated with experimentation and upgrading. The environmental benefits of digitalization are not automatic. A growing body of literature also points to several restrictive or countervailing effects. Digital infrastructure itself consumes substantial electricity and may generate additional carbon emissions, especially in the short term. Moreover, digital transformation often requires large upfront investment in software, equipment, training, and organizational restructuring, which may increase firms' cost burdens before efficiency gains are realized. In addition, when firms lack sufficient absorptive capacity, managerial competence, or organizational readiness, the returns to digitalization may be unstable or even disappointing. Under such conditions, digitalization may fail to produce substantive environmental improvement and may instead intensify operational complexity or energy demand. Digitalization should not be viewed as an automatically effective pathway to low-carbon transition. Its impact depends on whether firms can translate digital inputs into real organizational and technological upgrading. Nonetheless, under suitable conditions, digitalization is generally expected to facilitate low-carbon transition.

• H1: Digitalization promotes industrial low-carbon transition.

2.2 Policy support, technological investment, and financial support in low-carbon transition

Industrial low-carbon transition is not driven by digitalization alone. Because green transformation often involves high uncertainty, long payback periods, and significant externalities, firms usually require broader institutional and financial support to sustain transformation efforts. In this context, policy support, technological investment, and financial support represent three important complementary forces (Hmamed et al., 2025).

Policy support can positively contribute to low-carbon transition by correcting market failures and reducing the uncertainty associated with environmental and digital upgrading. Government subsidies, tax incentives, environmental regulations, industrial guidance programs, and digitalization-related support policies can provide firms with incentives to adopt cleaner production modes and invest in low-carbon technologies. Policy support may also serve as a signaling mechanism, guiding firms toward environmentally preferred activities and strengthening expectations about future regulatory and market trends. At the same time, however, policy effects are not always stable. If policy instruments are too broad, too rigid, or poorly targeted, firms may engage in symbolic compliance rather than substantive transformation (Ding et al., 2025). Technological investment and financial support constitute two essential foundations of industrial low-carbon transition. On the one hand, investment in R&D, cleaner production processes, energy-saving equipment, and production innovation can improve resource utilization, reduce carbon intensity, and enhance firms' capacity to respond to environmental and market pressures, especially in manufacturing sectors where carbon reduction relies heavily on technological upgrading and process redesign. On the other hand, financial support facilitates low-carbon transition by easing financing constraints and lowering transformation costs, thereby enabling firms to sustain investments in digital systems, innovation activities, and energy-saving technologies despite uncertain or delayed returns. Therefore, policy, technology, and finance should be understood not as isolated factors, but as complementary support for industrial low-carbon transition, whose respective baseline contributions need to be established before examining their joint interaction.

Based on these arguments, we propose the following hypotheses:

• H2a: Policy support positively contributes to industrial low-carbon transition.

• H2b: Technological investment positively contributes to industrial low-carbon transition.

• H2c: Financial support positively contributes to industrial low-carbon transition.

2.3 Policy–technology–finance synergy

While policy support, technological investment, and financial support may each affect low-carbon transition independently, their combined effect may be substantially greater when they operate as a coordinated system. Industrial transformation is a multidimensional process in which firms must simultaneously respond to regulatory signals, upgrade technological capabilities, and secure sufficient financial resources. In such a setting, complementarity among policy, technology, and finance becomes especially important (Yang et al., 2025).

Policy support can shape the direction of enterprise upgrading by setting incentives and reducing uncertainty. Technological investment provides the substantive capability needed to implement cleaner production and innovation. Financial support supplies the resources required to sustain these adjustments and absorb transition risks. When these three elements are aligned, they can form a mutually reinforcing mechanism: policy support encourages firms to undertake innovation, technological investment improves the productivity of financial and policy inputs, and financial support ensures that firms can carry out policy-guided upgrading without severe capital constraints (Wang and Gong, 2025). This synergy perspective moves beyond a single-factor explanation of low-carbon transition by emphasizing that firms often fail to achieve substantial green transformation not because one factor is entirely missing, but because policy, technology, and finance are insufficiently coordinated. Policy without technological capability may lead only to superficial adaptation, technology without financial support may remain underfunded, and finance without clear policy guidance may not be directed effectively toward low-carbon activities. Therefore, the effectiveness of each factor depends partly on the presence and quality of the others, and this study accordingly argues that policy support, technological investment, and financial support can generate a joint synergistic effect on industrial low-carbon transition. H3: Policy support, technological investment, and financial support generate a joint synergistic effect on industrial low-carbon transition.

2.4 Policy–enterprise fit and conditional effectiveness

Although policy–technology–finance synergy may promote industrial low-carbon transition, such synergy is unlikely to emerge uniformly across firms. Policy implementation theory and organizational research both suggest that policy effectiveness depends not only on the formal strength of policy instruments, but also on the degree to which those instruments are compatible with firm-level characteristics. This study conceptualizes this compatibility as policy–enterprise fit (Waheed, 2025).

Policy–enterprise fit refers to the degree of alignment between policy objectives, implementation requirements, and enterprise capabilities. Firms differ substantially in technological readiness, organizational capacity, financial resilience, managerial competence, and industrial characteristics. When policy design is well-matched to these firm-level conditions, enterprises are more likely to understand policy goals, utilize policy tools effectively, and convert external support into substantive low-carbon outcomes. Under high fit, policy support can strengthen the interaction among technology and finance, thereby amplifying synergistic effects (Bernard et al., 2022).

When policy–enterprise fit is low, policy intervention may become ineffective or even counterproductive, as excessive compliance burdens, mismatched standards, distorted incentives, and administrative pressures can divert firms' resources away from substantive innovation and real carbon-reduction activities. In such cases, especially when policy demands exceed firms' digital and organizational capacities, policy support may weaken rather than enhance the environmental benefits of digitalization. Therefore, policy–enterprise fit is a critical boundary condition that shapes both the effectiveness of policy–technology–finance synergy and whether policy intervention ultimately strengthens or undermines industrial low-carbon transition.

• H4a: Policy–enterprise fit positively moderates the synergistic effect of policy, technology, and finance on industrial low-carbon transition.

• H4b: Under low policy–enterprise fit, policy intervention weakens the low-carbon benefits of digitalization.

2.5 Conceptual framework

Based on the above literature review and theoretical analysis, this study develops a conceptual framework in which industrial low-carbon transition is shaped by both the direct effect of digitalization and the complementary roles of policy support, technological investment, and financial support. Digitalization is expected to promote low-carbon transition by improving resource allocation efficiency, optimizing production and energy-use processes, and stimulating green innovation and knowledge diffusion. However, because digitalization may also involve energy costs, investment burdens, and organizational constraints, its low-carbon effect is not automatically guaranteed (Bernard et al., 2022). To explain this conditionality, the study further introduces a policy–technology–finance synergy framework. Policy support, technological investment, and financial support are expected to provide complementary foundations for low-carbon transition and to generate stronger effects when jointly aligned. At the same time, the realization of such synergy depends critically on policy–enterprise fit. When policy instruments are compatible with firm capabilities, synergistic effects are likely to be strengthened. When fit is weak, policy intervention may crowd out firms' substantive transformation efforts and reduce the effectiveness of digitalization. Accordingly, the hypotheses of this study are summarized as follows:

• H1: Digitalization promotes industrial low-carbon transition.

• H2a: Policy support positively contributes to industrial low-carbon transition.

• H2b: Technological investment positively contributes to industrial low-carbon transition.

• H2c: Financial support positively contributes to industrial low-carbon transition.

• H3: Policy support, technological investment, and financial support generate a joint synergistic effect on industrial low-carbon transition.

• H4a: Policy–enterprise fit positively moderates the synergistic effect of policy, technology, and finance on industrial low-carbon transition.

• H4b: Under low policy–enterprise fit, policy intervention weakens the low-carbon benefits of digitalization.

This framework responds directly to the reviewer's concern that digitalization should not be assumed to be uniformly effective and that the hypothesis structure should better reflect hierarchical theoretical logic. It also provides a clearer bridge between the literature review, the empirical model, and the later interpretation of synergy and boundary effects.

As illustrated in Figure 1, this study proposes a conditional synergy framework.

Figure 1

3 Research methodology and data sources

3.1 Data sources and sample construction

This study uses a firm-level panel dataset of manufacturing enterprises in Guangzhou covering 2015–2023. Guangzhou is selected for three reasons. First, it is one of China's major manufacturing hubs and an early adopter of digital-transformation and green-upgrading policies. Second, the city has implemented multiple policy programs related to industrial digitalization, green manufacturing, and intelligent upgrading, which provides an appropriate institutional setting for examining policy–technology interactions. Third, comparatively rich firm-level information is available for this region (Palma and Scotti, 2025; Zhao et al., 2026). Firm financial and production data are obtained from the China Industrial Enterprise Database and supplemented with annual reports. Information on government subsidies and policy support is compiled from official fiscal disclosure records and local government databases. Energy-consumption data are used to calculate firm-level carbon emissions using standard IPCC emission coefficients (Wang et al., 2022).

The sample construction follows standard data-cleaning procedures. First, firms with missing information on key financial, production, or energy-related variables are excluded to ensure data completeness (Adams et al., 2019; Boda and Úradníček, 2020). Second, to reduce the influence of extreme values, all continuous variables are winsorized at the 1st and 99th percentiles. Following standard practice in corporate finance and panel-data studies, all continuous variables were winsorized at the 1st and 99th percentiles to mitigate the influence of outliers (Boda and Úradníček, 2020). Third, the analysis is restricted to manufacturing firms with continuous observations for at least three consecutive years, which allows for reliable panel estimation and dynamic analysis. After applying these screening criteria, the final dataset consists of an unbalanced panel of approximately 1,500 firms, yielding more than 12,000 firm-year observations over the sample period.

3.2 Variable definitions

3.2.1 Dependent variable: low-carbon transition (LCT)

Following the mainstream literature on firm-level low-carbon transition, this study measures the dependent variable using carbon intensity, defined as firm-level carbon emissions per unit of industrial output, as shown in Equation 1:

Where CO_{2, it} represents firm-level carbon emissions calculated from energy consumption, and Output_it denotes total industrial output. A lower value of LCT indicates better low-carbon performance. As robustness checks, alternative measures such as energy intensity and logarithmic transformations are also used. Conceptual rationality. Carbon intensity is one of the most widely used outcome variables in studies of low-carbon transition because it captures both the environmental side of production (emissions) and the efficiency side of economic activity (output generated per unit of carbon released). Compared with total emissions, carbon intensity better captures whether firms are achieving decarbonization alongside industrial output expansion. Literature basis. Existing research commonly adopts closely related indicators such as carbon intensity, carbon emissions per unit of output, carbon performance, and industrial carbon efficiency when evaluating the environmental consequences of digitalization and industrial upgrading (Ding et al., 2022; Bianchini et al., 2023). These measures are especially suitable for identifying whether technological and organizational upgrading improves emissions efficiency rather than merely reflecting firm scale. Applicability to this study. Because the research objects are manufacturing firms, a unit-output-based indicator is more appropriate than total emissions for assessing green transformation efficiency. Total emissions are strongly influenced by firm size and production expansion, whereas carbon intensity more directly reflects whether digitalization, technological investment, and financial support jointly improve cleaner production efficiency. In the empirical analysis, a lower value of LCT indicates better low-carbon transition performance. As robustness checks, alternative dependent variables, including energy intensity and the logarithm of carbon emissions, are also employed.

3.2.2 Core explanatory variables

Digitalization: Digitalization is measured as the ratio of digital-related investment to total assets. Digital-related investment includes expenditure on software, intelligent equipment, information systems, and other digital infrastructure-related capital expenditure disclosed in firm financial statements and annual reports. Compared with text-frequency indicators or simple digitalization dummies, this input-based measure more directly reflects firms' substantive resource commitment to digital transformation and is therefore better suited to capturing the actual intensity of enterprise digital upgrading (Xie et al., 2023; Dai et al., 2025). Recent methodological reviews also note that investment-based indicators are particularly valuable when the research focus is on firm-level transformation capacity rather than rhetorical disclosure alone (Xie et al., 2023). This measure reflects substantive digital transformation rather than symbolic digital labeling and is conceptually consistent with the later discussion of symbolic compliance under policy mismatch.

Policy Support: Policy support is measured by the logarithm of government subsidies related to digitalization, technological upgrading, or environmental protection. This variable reflects the intensity of policy intervention received by each firm (Li et al., 2024). Technological Investment: Technological input is measured by R&D expenditure as a proportion of operating revenue, capturing firms' innovation effort in production and process upgrading. Financial Support: Financial development is proxied by a composite index constructed from firms' access to credit, interest expenses, and leverage conditions, reflecting financing availability and cost (Hossain and Sultana, 2024).

3.2.3 Policy–enterprise fit (PERF)

To strengthen the theoretical and empirical basis of policy–enterprise fit, this study draws on policy implementation theory, absorptive-capacity theory, and the literature on institutional compatibility and firm heterogeneity (Waheed, 2025; Lyu et al., 2018). The core idea is that policy effectiveness depends not only on policy intensity, but also on the compatibility between policy requirements and firm characteristics. Accordingly, PERF is constructed as a composite indicator rather than a directly observable variable. PERF includes three dimensions. Industry Match indicates whether a firm operates in sectors prioritized by local digitalization, green manufacturing, or low-carbon policies. Scale Compatibility reflects whether firm size aligns with the size range targeted by the policy program, indicating its capacity to absorb policy resources. Digital Readiness measures whether a firm's prior digital investment intensity is close to or above the industry benchmark, indicating whether it has the basic technical conditions for effective policy implementation. Policy–enterprise fit is calculated as shown in Equation 2:

Each dimension is first standardized to a 0–1 scale using min-max normalization, after which the three standardized sub-indices are aggregated using equal weights. Equal weighting is adopted because the three dimensions correspond to complementary but theoretically coequal aspects of policy compatibility: policy targeting, organizational capacity, and technical preparedness. A higher value of PERF indicates stronger alignment between policy requirements and firm capabilities. To address the concern that the index may be sensitive to subjective construction, the robustness analysis additionally reconstructs PERF using alternative weighting methods and verifies whether the main results remain unchanged.

3.3 Econometric model specification

To examine the relationship between digitalization, policy intervention, and low-carbon transition, the following fixed-effects model is estimated.

Model 1 is the baseline model used to test Hypotheses H1 and H2a–H2c, as specified in Equation 3.

where:

LCT_it denotes the low-carbon transition level of firm i in year t; DIG_it, POL_it, TECH_it, and FIN_it represent digitalization, policy support, technological investment, and financial support, respectively; X_it is a vector of control variables; μ_i and λ_t denote firm and year fixed effects; and ε_it is the error term. Model 1 estimates the direct effects of digitalization, policy support, technological investment, and financial support on low-carbon transition.

Model 2 is the interaction model used to test the pairwise complementarity implied by Hypothesis H3, as specified in Equation 4.

Model 2 introduces the three pairwise interaction terms to examine whether policy, technology, and finance exhibit preliminary complementary effects before considering full-system synergy.

Model 3 is the triple-interaction model used to test Hypothesis H3, as specified in Equation 5.

Model 3 focuses on whether policy support, technological investment, and financial support generate a joint synergy effect on low-carbon transition beyond the sum of their separate contributions.

Model 4 is the moderated triple-interaction model used to test Hypotheses H4a and H4b, as specified in Equation 6.

Model 4 introduces policy-enterprise fit as the key boundary condition. The four-way interaction term captures whether PERF amplifies the policy-technology-finance synergy, while the interaction under low-fit conditions is used to identify the crowding-out mechanism implied by H4b. In addition to the full-sample estimation, the empirical analysis also reports grouped regressions for high-PERF and low-PERF firms as a robustness and interpretation aid. In all models, i denote firms and t denotes years. LCT_it is the dependent variable measuring firm-level carbon intensity. DIG_it, POL_it, TECH_it, and FIN_it, and PERF_it denote digitalization, policy support, technological investment, financial support, and policy-enterprise fit, respectively. X_it is a vector of control variables, including firm size, firm age, ownership type, export status, and industry competition. μ_i and λ_t denote firm and year fixed effects, respectively, and ε_it is the idiosyncratic error term.

3.4 Identification strategy and robustness checks

To ensure the robustness of the results, several additional tests are conducted: Alternative variable measurements: Using energy intensity and logarithmic carbon emissions as alternative dependent variables. Lagged explanatory variables: Mitigating potential reverse causality. Subsample analyses: Splitting firms by size, ownership, and industry pollution intensity. Endogeneity checks: Employing lagged policy variables as instruments. Robust standard errors: Clustered at the firm level. These tests confirm that the main conclusions are not driven by model specification or sample selection (Barros et al., 2020).

4 Identification strategy and robustness checks

4.1 Descriptive statistics and baseline regression results

Table 1 reports the descriptive statistics of the main variables. During the sample period, the mean value of corporate carbon intensity (LCT, measured as carbon emissions per unit of output value) is 2.15, with a standard deviation of 1.08, indicating substantial heterogeneity in low-carbon transition performance across firms. The mean values of digitalization, policy support, technological investment, and financial support are 0.45, 3.20, 0.78, and 0.62, respectively, suggesting that although the sample firms have undertaken digital and innovation-related upgrading to some extent, the intensity of transformation remains uneven. These descriptive results also provide preliminary support for the empirical strategy because sufficient cross-firm variation exists in both the dependent and explanatory variables. Several descriptive implications deserve emphasis. The coexistence of relatively moderate digitalization and substantial dispersion in LCT suggests that firm-level low-carbon transition is unlikely to be explained by scale alone and is more plausibly related to differences in organizational upgrading and resource allocation. The average value of policy support is not low, yet the later regression results show only a weak standalone policy effect. This pattern is consistent with the argument that policy resources may be widely distributed but not necessarily well targeted or effectively absorbed by firms. Third, the mean value of policy–enterprise fit is 0.51, which is only moderately above the midpoint, indicating that the alignment between policy requirements and firm capabilities remains imperfect and still leaves considerable room for improvement in policy design and implementation.

Table 2 presents the baseline regression results from the fixed-effects model. Models (1) to (4) sequentially introduce the core explanatory variables. The coefficient on digitalization remains significantly negative at the 1% level throughout, with the full-model estimate equal to −0.402, supporting Hypothesis H1. Because the dependent variable is carbon intensity, this result indicates that greater digital investment is associated with lower emissions per unit of output. The finding is consistent with the view that digital transformation can improve monitoring, process coordination, and production efficiency, thereby contributing to firm-level decarbonization. The coefficients on technological investment and financial support are also significantly negative in the baseline model, which supports Hypotheses H2b and H2c rather than the synergy hypothesis. These results suggest that innovation input and financing conditions constitute two direct enabling conditions for low-carbon transition: technological investment improves firms' capability to undertake cleaner production and process upgrading, while financial support relaxes liquidity constraints and allows firms to sustain costly transformation. By contrast, the coefficient on policy support (POL) is negative but only weakly significant at the 10% level in Model (4), providing only limited support for Hypothesis H2a. Economically, this pattern implies that policy support alone is not a sufficient driver of low-carbon transition. Instead, its effectiveness appears to depend on whether policy intervention can be combined with firms' technological capability, financial capacity, and implementation readiness. Taken together, the baseline results show that the direct effects of digitalization, technological investment, and financial support are clearer and more stable than the direct effect of policy support. This distinction is important because it prepares the ground for the later interaction analysis: policy is not expected to operate primarily as an isolated driver, but rather as a conditional institutional force whose effectiveness depends on complementarity and fitness.

4.2 Testing for policy-technology-finance synergy effects

A more nuanced result emerges when the pairwise interactions are examined in Model (2). The interaction terms POL × TECH and POL × FIN are statistically insignificant, whereas TECH × FIN is significantly negative. This pattern is theoretically informative. It indicates that technology and finance exhibit relatively direct complementarity because financing helps firms sustain R&D, cleaner equipment adoption, and process upgrading, while technological capability improves the return to financing by converting capital into substantive low-carbon innovation. By contrast, policy support does not generate strong effects when paired with only one factor. In other words, policy is not an independent engine of low-carbon transition but a conditional amplifier whose effectiveness depends on firms simultaneously possessing technological capability and financial capacity. The fact that the triple interaction is significant while the two policy-related pairwise interactions are not therefore highlights the system-level role of policy: policies matter most when they coordinate and magnify existing technological and financial foundations rather than operate as stand-alone subsidies (Wang et al., 2022; Xue and Xi, 2025).

Table 3 reports the synergy and moderating effect tests.

4.3 Analysis of the moderating role of policy-enterprise fit (PERF)

To test Hypotheses H4a and H4b, the sample is divided into high-fit and low-fit groups according to the median value of policy–enterprise fit, and the regressions are re-estimated separately. This step is central to the paper because it directly examines whether the same policy environment produces heterogeneous outcomes depending on firm-level compatibility with policy requirements.

In the high-PERF group, the coefficient on the triple interaction term POL × TECH × FIN is −0.068 and significant at the 1% level, which is substantially larger in absolute value than the corresponding estimate in the full sample. At the same time, the coefficient on digitalization is also more negative 0.523. These results indicate that when policy design is well-aligned with firm characteristics, policy support plays a lubricating and amplifying role: it helps firms translate technological investment and financial support into more effective low-carbon outcomes. Under high fit, policy does not merely add resources; it improves the coordination efficiency of existing resources, lowers implementation frictions, and strengthens the environmental return to digital and technological upgrading. This finding provides strong support for Hypothesis H4a.

In the low-PERF group, the picture changes markedly. The coefficient on POL × TECH × FIN becomes positive and statistically insignificant 0.017, indicating that the expected synergy disappears when policy requirements are poorly matched with firm conditions. More importantly, the absolute value of the coefficient on digitalization falls sharply to −0.210 and its significance weakens. This pattern suggests a crowding-out mechanism: when policy and enterprise conditions are misaligned, firms may divert managerial attention, financial resources, and organizational effort toward formal compliance, administrative adaptation, or unsuitable investment, thereby weakening the emission-reduction benefits of their own digital transformation. This result is consistent with Hypothesis H4b. More broadly, the contrast between the high- and low-fit groups implies that the same policy instrument can generate fundamentally different outcomes across firms, which confirms that policy effectiveness is conditional and heterogeneous rather than uniform.

4.4 Robustness checks and endogeneity treatment

A series of robustness and endogeneity tests are conducted to verify that the main findings are not driven by variable construction, reverse causality, or spurious group effects. Rather than serving as a mechanical checklist, each test addresses a distinct empirical concern and therefore strengthens a different part of the identification logic.

First, the dependent variable is replaced with alternative indicators, including energy intensity and the logarithm of carbon emissions. The results remain substantively unchanged, which suggests that the main conclusions do not depend on a single operationalization of low-carbon transition and are not an artifact of the carbon-intensity measure alone.

Second, all core explanatory variables are lagged by one period to mitigate potential reverse causality. This test addresses the possibility that firms with better low-carbon performance may subsequently receive more policy attention or invest more heavily in digitalization, technology, and finance. The persistence of the main results after lagging the regressors provides stronger support for the proposed causal direction.

Third, instrumental-variable estimations (2SLS) are employed to address potential endogeneity in digitalization (DIG) and policy support (POL). Specifically, the average digitalization level of other firms in the same industry and the average policy intensity for non-manufacturing firms in the same city are used as instruments. The Kleibergen–Paap rk LM test and the Cragg–Donald Wald F statistic rejects the weak-instrument concern, and the 2SLS estimates remain consistent with the baseline findings. This result suggests that the observed effects are unlikely to be driven primarily by omitted variables or simultaneous determination (Chang and Kang, 2019).

Fourth, subsample regressions excluding state-owned enterprises and restricting the sample to high-tech firms yield qualitatively similar conclusions. This indicates that the main relationships are not confined to a specific ownership structure or industrial segment, therefore have broader relevance within the manufacturing sample. A placebo test is performed by randomly generating a pseudo policy–enterprise fit indicator and repeating the estimation 500 times. These patterns reinforce the paper's core claim that low-carbon transition is conditional and heterogeneous rather than uniform across enterprises (Neumark and Young, 2021; Sraer and Thesmar, 2018).

4.5 Mechanism test: evidence from mediating effects

A series of robust and endogeneity tests are conducted to verify that the main findings are not driven by variable construction, reverse causality, or spurious subgroup differences. First, the dependent variable is replaced with alternative indicators, including energy intensity and the logarithm of carbon emissions. The core coefficients remain directionally and statistically consistent, which suggests that the conclusions do not depend on a single operationalization of low-carbon transition.

Second, all core explanatory variables are lagged by one period to mitigate potential reverse causality, namely the possibility that firms with better low-carbon performance subsequently receive more policy attention or invest more heavily in digitalization, technology, and finance. The persistence of the main results after lagging the regressors provides stronger support for the proposed causal direction. Third, instrumental-variable estimations (2SLS) are employed to address potential endogeneity in digitalization (DIG) and policy support (POL). Specifically, the average digitalization level of other firms in the same industry and the average policy intensity for non-manufacturing firms in the same city are used as instruments. These instruments are relevant because they are correlated with firm-level DIG and POL through common technological and policy environments, but they are plausibly exogenous to the focal firm's contemporaneous carbon intensity after firm and year fixed effects are included. The Kleibergen–Paap rk LM statistic rejects under-identification, and the first-stage statistics indicate that weak-instrument concerns are not driving the estimates. The 2SLS coefficients remain consistent with the baseline findings, which strengthens the credibility of the identification strategy (Barros et al., 2020; Chang and Kang, 2019).

Fourth, several additional specification checks are implemented. Subsample regressions excluding state-owned enterprises and restricting the sample to high-tech firms yield qualitatively similar conclusions, indicating that the main relationships are not confined to a specific ownership structure or industrial segment. A placebo test is also performed by randomly generating a pseudo policy–enterprise fit indicator and repeating the estimation 500 times. The resulting coefficients are centered around zero, indicating that the moderating effect of PERF is unlikely to be a statistical artifact of arbitrary grouping. Taking together, these tests suggest that the main conclusions are robust to alternative measurement, timing, specification, and identification choices.

4.6 Mechanism test: evidence from mediating effects

To further uncover why low policy–enterprise fit weakens low-carbon transition, this study examines three mechanisms: information distortion, incentive misalignment, and compliance cost crowding-out. For each mechanism, the analysis proceeds from theoretical expectation to proxy construction and then to empirical evidence, thereby linking the moderation results to a more concrete behavioral explanation.

The first mechanism is information distortion. The theoretical expectation is that when policy requirements are poorly matched with firm capabilities, firms may misreport, selectively disclose, or strategically reframe environmental information to appear compliant without undertaking substantive transformation. This channel is proxied by the degree of abnormality in corporate environmental information reporting. The results show that in the low-PERF group, stronger policy intensity significantly increases this abnormality measure, which in turn raises carbon intensity. This finding suggests that policy mismatch can distort information transmission and weaken the substantive effectiveness of environmental governance.

The second mechanism is incentive misalignment. When policy evaluation emphasizes visible formal indicators rather than actual emission-reduction outcomes, firms may engage in symbolic green behavior instead of substantive innovation. This channel is captured by the proportion of strategic patents within green patents. The empirical results indicate that in the low-PERF group, stronger policy support increases strategic green innovation behavior, and this behavior is associated with higher carbon intensity. The third mechanism is compliance cost crowding-out. The theoretical expectation is that excessive procedural and administrative burdens may absorb resources that would otherwise be used for genuine low-carbon upgrading. This channel is proxied by the management expense ratio. The results show that in the low-PERF group, policy intensity significantly raises management expenses, which in turn hinder carbon reduction. By contrast, these mediating effects are not significant in the high-PERF group. Taking together, the mechanism tests complement the main findings by showing not only that policy mismatch weakens synergy, but also how it does so: through distorted information, symbolic compliance, and the diversion of resources away from substantive green transformation (Ding et al., 2022).

5 Discussion

This study advances the discussion on industrial low-carbon transition by moving beyond a simple “digitalization is good for decarbonization” narrative. The empirical evidence shows that the effects of digitalization, policy intervention, technological investment, and financial support are structurally unequal: digitalization, technology, and finance exhibit relatively stable direct emission-reduction effects, whereas policy support is mainly effective as a conditional coordinating force. More importantly, the results reveal that policy–technology–finance synergy is significantly strengthened only when policy instruments are well-aligned with firm-level capabilities. This pattern allows the discussion to engage more directly with the broader literature on digital green transition, policy implementation, and firm heterogeneity, and it clarifies why policy-driven transformation often produces mixed outcomes across firms and contexts (Zhang et al., 2025; Chang, 2025).

5.1 Why digitalization does not automatically lead to low-carbon outcomes

The baseline results confirm that digitalization is associated with lower carbon intensity, which is consistent with studies arguing that digital technologies can improve information processing, real-time energy monitoring, production scheduling, and green innovation efficiency (Chen et al., 2024; Xie et al., 2023). In this sense, digitalization works as an enabling infrastructure for cleaner production: it improves firms' visibility over resource use, reduces coordination frictions, and helps managers identify where energy waste and process losses occur. This interpretation is also consistent with the significant direct effect of digitalization observed in this study, even after controlling technology, finance, and firm fixed effects.

On the one hand, digital technologies improve energy efficiency and enable cleaner production. On the other hand, digital transformation requires substantial capital investment, organizational restructuring, and energy-intensive infrastructure, which may offset environmental gains in the short term. Without appropriate institutional support and alignment, digitalization may even increase energy demand and operational complexity.

However, the findings also suggest that digitalization is not automatically effective. This is an important nuance because some policy and managerial narratives implicitly equate digital upgrading with immediate environmental gains. Digital transformation requires substantial upfront expenditure on software, equipment, data systems, staff training, and organizational redesign, and these investments may raise costs and energy demand before efficiency gains are realized. Prior research has likewise noted that the green effects of the digital transition can be uneven, delayed, or dependent on complementary capacities (Bianchini et al., 2023; Chang, 2025). Our results refine this literature by showing that the environmental payoff of digitalization depends on whether firms also possess the technological capability, financial room, and implementation readiness required to convert digital inputs into operational change.

5.2 The conditional nature of policy effectiveness

A second major finding is that policy support alone has a weak and unstable standalone effect. This result is partly consistent with policy-implementation research, which emphasizes that policy outcomes depend on target firms' absorptive capacity and implementation context rather than on policy intensity alone (Neumark and Young, 2021; Englmaier et al., 2018). It also departs from studies reporting a more directly positive role of environmental or digital policies in promoting enterprise transformation (Zhang et al., 2025; Zhu et al., 2023). A plausible explanation is that policy is fundamentally an external incentive, not an internal capability. When firms lack technological readiness, financial resilience, or managerial capacity, subsidies and formal support may be translated into symbolic compliance, low-quality adoption, or short-term response rather than substantive carbon reduction. The recent literature on policy uncertainty and corporate greenwashing further supports this interpretation by showing that policy environments can induce performative environmental behavior when incentives are misaligned (Dai et al., 2025).

5.3 Policy–enterprise fit as a key boundary condition

The high-fit vs. low-fit comparison provides the clearest evidence for the paper's core argument. Under high policy–enterprise fit, the triple interaction among policy support, technological investment, and financial support becomes substantially stronger and highly significant, indicating that policy, technology, and finance form a closed-loop support system when firms are sufficiently prepared. In this configuration, policy provides direction, legitimacy, and institutional coordination; technological investment provides the substantive capability required for cleaner production and innovation; and financial support relaxes capital constraints and allows firms to sustain long-horizon upgrading. The resulting synergy is consistent with research stressing that green and digital transformation is most effective when institutional guidance, innovation capability, and financial resources reinforce one another rather than operate in isolation (Wang et al., 2022; Xue and Xi, 2025).

By contrast, the same synergy does not appear under low fit. This contrast is theoretically important because it shows that policy effectiveness is not additive but configurational. The policy signal matters most when firms can interpret it, respond to it, and integrate it with existing technological and financial resources. That is why the three-way interaction is significant even though the two policy-related pairwise interactions are not: policy does not function as an autonomous production input, but as a coordinating and amplifying mechanism that becomes effective only when the other two pillars are simultaneously present. This finding extends prior work on heterogeneous policy effects by providing micro-level evidence that “readiness” is the condition under which external support is converted into real low-carbon performance (Zhang et al., 2025; Neumark and Young, 2021).

5.4 Mechanisms: why policy mismatch undermines low-carbon transition

The mechanism analysis further explains why crowding-out emerges under low policy–enterprise fit. When firms face policy requirements that exceed their readiness, additional compliance obligations can redirect managerial attention and scarce resources away from substantive green upgrading toward administrative coping. Information distortion becomes more likely because firms have incentives to frame compliance symbolically; incentive misalignment intensifies when evaluation systems reward visible inputs rather than emission outcomes; and compliance costs rise as reporting, coordination, and procedural burdens expand. These channels are consistent with the broader literature on environmental disclosure distortion, regulatory burden, and crowding-out effects (Li et al., 2026; De Benedetto, 2018). Therefore, the crowding-out observed in the low-fit group should not be understood simply as policy failure in the abstract, but as the result of information asymmetry, misaligned incentives, and rising transaction costs that compress the space for effective investment. This interpretation also explains why the coefficient on digitalization becomes markedly weaker under low fit: digital spending remains present, but part of its potential environmental return is absorbed by compliance-oriented rather than transformation-oriented activities.

5.5 Implications for theory and policy

From a theoretical perspective, the main contribution of this study is to recast the digital green transition as a conditional governance process. Existing research has often emphasized the direct environmental effects of digitalization or the average effect of industrial policy (Zhang et al., 2025; Chang, 2025). The present findings suggest that a more accurate interpretation is interactional: environmental outcomes depend on whether digital capability, technological upgrading, finance, and policy design are mutually compatible. This helps bridge the literature on digital transformation, environmental economics, and policy implementation by showing that policy intensity alone is a poor predictor of success unless it is embedded in a fit-sensitive support system. The policy implications should therefore be derived directly from the empirical results rather than stated in general terms. Four implications follow from the evidence.

First, digitalization has a significant overall effect but is not automatically green, governments should avoid treating digital upgrading as equivalent to decarbonization. Policy should promote the deep integration of digital technologies with concrete energy-saving and emission-reduction scenarios, such as energy management, process optimization, predictive maintenance, and cleaner production systems. This implication follows directly from the finding that digitalization reduces carbon intensity, but only more effectively when complementary capacities are present. Second, because the standalone effect of policy support is weak, policy design should shift from input-oriented logic to an outcome-oriented logic. Rather than emphasizing equipment installation, project enrollment, or platform access as core performance indicators, policymakers should prioritize measurable improvements in carbon intensity, energy efficiency, and verified green innovation outcomes. This recommendation follows from the result that policy alone does not deliver stable emission-reduction effects and is therefore more useful as a quality-enhancing governance instrument than as a simple spending tool (Saba, 2025).

Third, the triple synergy becomes significant only under high policy–enterprise fit, policy tools should be more differentiated and targeted. Support intensity, implementation standards, and financing arrangements should be calibrated to firms' scale, sector, digital readiness, and innovation capability. A layered policy architecture can increase the probability that policy support, technological input, and finance reinforce one another rather than remain fragmented. This implication is directly grounded in the high-PERF results, which show that readiness is the condition under which three-way coordination becomes effective (Shao and Xi, 2025).

Fourth, because low fit produces a crowding-out effect, governments should reduce unnecessary administrative procedures and compliance burdens that displace substantive green investment. Policies should minimize reporting duplication, discourage symbolic compliance, and build feedback mechanisms that detect mismatch early and allow timely adjustment (Flatø, 2024). In substantive terms, the results suggest that the central challenge is not whether governments should intervene, but how they can intervene without pushing firms away from real innovation and toward administrative response. Overall, the transition from policy expansion to policy precision is essential for sustainable industrial decarbonization.

6 Conclusion

This study investigates the conditional effects of digitalization, policy intervention, and financial support on industrial low-carbon transition, with a particular focus on the role of policy–enterprise fit. Using firm-level panel data from manufacturing enterprises in Guangzhou during 2015–2023, we construct an integrated analytical framework and provide empirical evidence on how policy-driven digital transformation affects carbon performance under different institutional conditions. Several key conclusions emerge (Chang, 2025).

First, digitalization, technological investment, and financial support each contribute positively to industrial low-carbon transition, but their effects are not automatic. Policy intervention alone does not guarantee emission reductions, suggesting that the effectiveness of digital transformation depends critically on complementary conditions. Second, significant synergy among policy, technology, and finance exists only when policy instruments are well-aligned with firm-level capabilities. When such alignment is absent, policy intervention may weaken or even offset the emission-reduction effects of digitalization. This finding helps explain the mixed results reported in previous studies and highlights the limitations of one-size-fits-all policy approaches (Howden-Chapman et al., 2020).

Third, policy–enterprise fit plays a decisive moderating role. Firms with higher levels of policy–enterprise fit are better able to translate policy support and digital investment into tangible environmental performance. In contrast, low fit leads to information distortion, incentive misalignment, and compliance cost crowding-out, thereby undermining low-carbon outcomes. Taken together, these findings suggest that policy effectiveness in the digital era should be understood as a conditional outcome rather than an automatic consequence of policy intensity. Digital transformation can facilitate low-carbon transition only when institutional design, firm capabilities, and market conditions are properly aligned (Zhu et al., 2023).

6.1 Policy implications

The findings of this study imply that effective low-carbon governance should be designed around empirically observed mechanisms rather than around the assumption that more policy support always generates better outcomes. Specifically, the results of this paper point to four tightly linked policy implications.

• Because digitalization is generally beneficial but not automatically so, policy should encourage the combination of digital technologies with concrete decarbonization applications rather than rewarding digital investment in the abstract.

• Because policy support alone has only a weak standalone effect, evaluation systems should move from input-based targets to outcome-based carbon-performance assessment, thereby reducing the incentive for symbolic compliance (De Benedetto, 2018).

• Because policy–technology–finance synergy is significant only when policy–enterprise fit is high, governments should adopt differentiated and precision-oriented support packages based on firms' scale, digital readiness, and sectoral characteristics.

• Because low fit generates compliance-cost crowding-out, implementation procedures should be streamlined, and dynamic feedback mechanisms should be strengthened so that policy mismatch can be corrected before it undermines substantive green innovation (Flatø, 2024).

6.2 Limitations and future research

This study has several limitations that offer directions for future research. The analysis focuses on firms in a single metropolitan area, which may limit the generalizability of the findings. Future studies could extend the analysis to other regions or conduct cross-country comparisons. Second, although multiple robustness checks are conducted, potential endogeneity issues cannot be fully eliminated (Gao et al., 2024). Future research may employ quasi-natural experiments or instrumental variable approaches to strengthen causal inference (Deng et al., 2019). The measurement of policy–enterprise fit relies on proxy indicators; future work could benefit from survey-based or qualitative data to capture this concept more precisely (Feng et al., 2024). Despite these limitations, this study provides robust evidence that digital transformation alone is insufficient to guarantee low-carbon transition (Chen et al., 2024). Effective policy design must account for firm heterogeneity and institutional compatibility (Englmaier et al., 2018). By highlighting the conditional nature of policy effectiveness, this research contributes to a more nuanced understanding of how digitalization can support sustainable industrial development (Xie et al., 2023).

References

• 1

  AdamsJ.HayungaD.MansiS.ReebD.VerardiV. (2019). Identifying and treating outliers in finance. Financ. Manag. 48, 345–384. doi: 10.1111/fima.12269

  • CrossRef
  • Google Scholar
  • View reference in article

• 2

  AndrewsD. S.FossN. J.FainshmidtS. (2025). From multilateral organizations to transactions: reconceptualizing the governance of public goods. Strateg. Organ. 23:14761270251375682. doi: 10.1177/14761270251375682

  • CrossRef
  • Google Scholar
  • View reference in article

• 3

  BarrosL. A.BergmannD. R.CastroF. H.SilveiraA. D. M. D. (2020). Endogeneity in panel data regressions: methodological guidance for corporate finance researchers. Rev. Bras. Gest. Neg. 22, 437–461. doi: 10.7819/rbgn.v22i0.4059

  • CrossRef
  • Google Scholar
  • View reference in article

• 4

  BernardA. B.DhyneE.MagermanG.ManovaK.MoxnesA. (2022). The origins of firm heterogeneity: a production network approach. J. Political Econ. 130, 1765–1804. doi: 10.1086/719759

  • CrossRef
  • Google Scholar
  • View reference in article

• 5

  BianchiniS.DamioliG.GhisettiC. (2023). The environmental effects of the “twin” green and digital transition in European regions. Environ. Resour. Econ. 84, 877–918. doi: 10.1007/s10640-022-00741-7

  • CrossRef
  • Google Scholar
  • View reference in article

• 6

  BlakeR.MichalkovaL.BilanY. (2021). Robotic wireless sensor networks, industrial artificial intelligence, and deep learning-assisted smart process planning in sustainable cyber-physical manufacturing systems. J. Self-Gov. Manag. Econ. 9:4. doi: 10.22381/jsme9420214

  • CrossRef
  • Google Scholar
  • View reference in article

• 7

  BodaM.ÚradníčekV. (2020). Methodology of industry statistics: averages, quantiles, and responses to atypical values. EandM Econ. Manag. 23, 120–137. doi: 10.15240/tul/001/2020-3-008

  • CrossRef
  • Google Scholar
  • View reference in article

• 8

  ChangJ. (2025). The impact of digitalization on low-carbon transition in China: the moderating role of dual development targets. Environ. Dev. Sustain. 27, 25841–25866. doi: 10.1007/s10668-024-04764-6

  • CrossRef
  • Google Scholar
  • View reference in article

• 9

  ChangJ.KangY. (2019). Instrumental variable estimates of the effect of management practices on firm performance in Korean firms. J. Labor Res. 40, 106–125. doi: 10.1007/s12122-018-9273-z

  • CrossRef
  • Google Scholar
  • View reference in article

• 10

  ChenX.HuangY.GaoY. (2024). Can urban low-carbon transitions promote enterprise digital transformation?Finance Res. Lett. 59:104807. doi: 10.1016/j.frl.2023.104807

  • CrossRef
  • Google Scholar
  • View reference in article

• 11

  DaiJ.KiawJ. P. A.HiungE. Y. T. (2025). Is some uncertainty better than none? Nonlinear relationships between climate policy uncertainty and corporate green performance. Int. Rev. Econ. Finance102:104336. doi: 10.1016/j.iref.2025.104336

  • CrossRef
  • Google Scholar
  • View reference in article

• 12

  De BenedettoM. (2018). Effective law from a regulatory and administrative law perspective. Eur. J. Risk Regul. 9, 391–415. doi: 10.1017/err.2018.52

  • CrossRef
  • Google Scholar
  • View reference in article

• 13

  DengT.HuY.MaM. (2019). Regional policy and tourism: a quasi-natural experiment. Ann. Tour. Res.74, 1–16. doi: 10.1016/j.annals.2018.10.001

  • CrossRef
  • Google Scholar
  • View reference in article

• 14

  DingX.AppolloniA.ShahzadM. (2022). Environmental administrative penalty, corporate environmental disclosures and the cost of debt. J. Clean. Prod. 332:129919. doi: 10.1016/j.jclepro.2021.129919

  • CrossRef
  • Google Scholar
  • View reference in article

• 15

  DingY.BiC.SunP. (2025). Low carbon constraints, innovation driven and carbon neutral technological innovation: empirical evidence based on multiple policy combinations. Sci. Rep. 15:22912. doi: 10.1038/s41598-025-06207-2

  • CrossRef
  • Google Scholar
  • View reference in article

• 16

  EnglmaierF.FossN. J.KnudsenT.KretschmerT. (2018). Organization design and firm heterogeneity: towards an integrated research agenda for strategy. Organ. Des.40, 229–252. doi: 10.1108/S0742-332220180000040008

  • CrossRef
  • Google Scholar
  • View reference in article

• 17

  FengT.CuiM.ZhangM.LiuL. (2024). How to construct a carbon asset management system for Chinese power enterprises: a survey-based approach. Energies17:3978. doi: 10.3390/en17163978

  • CrossRef
  • Google Scholar
  • View reference in article

• 18

  FlatøH. (2024). “Chapter 22: Environmental feedback mechanisms and state responsiveness,” in Handbook on Climate Change and Environmental Governance in China, eds. X. Zang, and X. Zhang (Cheltenham: Edward Elgar Publishing), 375–394. doi: 10.4337/9781035316359.00032

  • CrossRef
  • Google Scholar
  • View reference in article

• 19

  GaoD.ChenX.ZhangL. (2024). Digital industrial policy and digital transformation in banking: evidence from a quasi-natural experiment in China. SSRN Electron. J. 4964675. doi: 10.2139/ssrn.4964675

  • CrossRef
  • Google Scholar
  • View reference in article

• 20

  HeK.ChenW. (2025). Can digital transformation improve corporate green innovation?Technol. Anal. Strateg. Manag. 37, 1509–1525. doi: 10.1080/09537325.2024.2319603

  • CrossRef
  • Google Scholar
  • View reference in article

• 21

  HmamedH.CherrafiA.ShokriA.Garza-ReyesJ. A.FosterM. (2025). Barriers to low-carbon technology adoption: insights from the automotive manufacturing sector in emerging economies. Bus. Strateg. Environ. 35, 4422–4443. doi: 10.1002/bse.70337

  • CrossRef
  • Google Scholar
  • View reference in article

• 22

  HossainM. S.SultanaM. (2024). Digitalization of corporate finance and firm performance: global evidence and analysis. J. Financ. Econ. Policy16, 501–539. doi: 10.1108/JFEP-04-2023-0109

  • CrossRef
  • Google Scholar
  • View reference in article

• 23

  Howden-ChapmanP.KeallM.WhitwellK.ChapmanR. (2020). Evaluating natural experiments to measure the co-benefits of urban policy interventions to reduce carbon emissions in New Zealand. Sci. Total Environ. 700:134408. doi: 10.1016/j.scitotenv.2019.134408

  • CrossRef
  • Google Scholar
  • View reference in article

• 24

  KyriakopoulosP. (2024). Revisiting research on firm-level innovation in rural areas: a systematic literature review and future research directions. J. Rural Stud. 111:103437. doi: 10.1016/j.jrurstud.2024.103437

  • CrossRef
  • Google Scholar
  • View reference in article

• 25

  LiS.ChenL.JiangT.WangY.ShenC. (2024). Multidimensional financial development and natural resources: a path for sustainable development via natural resources and digitalization. Resour. Policy88:104400. doi: 10.1016/j.resourpol.2023.104400

  • CrossRef
  • Google Scholar
  • View reference in article

• 26

  LiX.ShenY.MengX. (2026). Credit crowding-out and risk transmission: the impact of extended local government debt maturity on corporate debt structure. Emerg. Mark. Financ. Trade62, 290–310. doi: 10.1080/1540496X.2025.2569718

  • CrossRef
  • Google Scholar
  • View reference in article

• 27

  LiuQ.LiZ.WangY. (2024). The heterogeneous impact of economic policy uncertainty on export recovery of firms: the role of the regional digital economy. Reg. Sci. Policy Pract. 16:100044. doi: 10.1016/j.rspp.2024.100044

  • CrossRef
  • Google Scholar
  • View reference in article

• 28

  LiuY.SongP. (2023). Digital transformation and green innovation of energy enterprises. Sustainability15:7703. doi: 10.3390/su15097703

  • CrossRef
  • Google Scholar
  • View reference in article

• 29

  LyuK.BianY.YuA. (2018). Environmental efficiency evaluation of industrial sector in China by incorporating learning effects. J. Clean. Prod. 172, 2464–2474. doi: 10.1016/j.jclepro.2017.11.163

  • CrossRef
  • Google Scholar
  • View reference in article

• 30

  LyuY.ZhangL.WangD. (2023). The impact of digital transformation on low-carbon development of manufacturing. Front. Environ. Sci. 11:1134882. doi: 10.3389/fenvs.2023.1134882

  • CrossRef
  • Google Scholar
  • View reference in article

• 31

  NeumarkD.YoungT. (2021). Heterogeneous effects of state enterprise zone programs in the shorter run and longer run. Econ. Dev. Q. 35, 91–107. doi: 10.1177/0891242420982341

  • CrossRef
  • Google Scholar
  • View reference in article

• 32

  PalmaG.ScottiF. (2025). Synergies in sustainability: assessing the innovation effects of digital and green investments in EU cohesion policy. Sustainability17:10446. doi: 10.3390/su172310446

  • CrossRef
  • Google Scholar
  • View reference in article

• 33

  SabaA. (2025). “Digital agriculture and the enhancement of results-based agri-environmental and climate commitments,” in Legal Implications of Results-based Agri-Environmental and Climate Commitments in the EU Common Agricultural Policy (Cham: Springer Nature Switzerland), 149–189. doi: 10.1007/978-3-031-82540-8_5

  • CrossRef
  • Google Scholar
  • View reference in article

• 34

  ShaoX.XiChenK. (2025). Science and technology financial policies and enterprise digital technology innovation. Finance Res. Lett.83:107646. doi: 10.1016/j.frl.2025.107646

  • CrossRef
  • Google Scholar
  • View reference in article

• 35

  SraerD.ThesmarD. (2018). A sufficient statistics approach for aggregating firm-level experiments. Natl. Bur. Econ. Res. Work. Pap. w24208. doi: 10.3386/w24208

  • CrossRef
  • Google Scholar
  • View reference in article

• 36

  WadeK.VochozkaM. (2021). Artificial intelligence data-driven internet of things systems, sustainable industry 4.0 wireless networks, and digitized mass production in cyber-physical smart manufacturing. J. Self Gov. Manag. Econ. 9, 48–60. doi: 10.22381/jsme9320214

  • CrossRef
  • Google Scholar
  • View reference in article

• 37

  WaheedJ. (2025). Linking absorptive capacity and project performance in environmental uncertainty: a perspective on implementation arrangements. Rev. Policy Res. 42, 252–275. doi: 10.1111/ropr.12591

  • CrossRef
  • Google Scholar
  • View reference in article

• 38

  WangB.GongS. (2025). How does digital transformation drive green technology MandA under the carbon cap and trade policy?Technol. Soc. 81:102868. doi: 10.1016/j.techsoc.2025.102868

  • CrossRef
  • Google Scholar
  • View reference in article

• 39

  WangS.LiM. (2022). Does industrial intelligence improve resource misallocation? An empirical test based on China. Environ. Sci. Pollut. Res. 29, 77973–77991. doi: 10.1007/s11356-022-21085-1

  • CrossRef
  • Google Scholar
  • View reference in article

• 40

  WangX.LiuS.TaoZ.CaoQ. (2022). The impact of industrial policy and its combinations on the innovation quality of wind power enterprises: a study from the perspective of financing modes. Renew. Energy188, 945–956. doi: 10.1016/j.renene.2022.02.057

  • CrossRef
  • Google Scholar
  • View reference in article

• 41

  WeiJ.LiuX. (2025). Do digitalization policies promote enterprise innovation? The logic of cost heterogeneity of enterprises. Humanit. Soc. Sci. Commun. 12:980. doi: 10.1057/s41599-025-05413-7

  • CrossRef
  • Google Scholar
  • View reference in article

• 42

  XieF.GuanX.ZhuJ.RuanJ.WangZ.LiuH.et al. (2023). Environmental protection goes digital: a policy perspective on promoting digitalization for sustainable development in China. Sustainability15:10673. doi: 10.3390/su151310673

  • CrossRef
  • Google Scholar
  • View reference in article

• 43

  XueL.ZhangQ.ZhangX.LiC. (2022). Can digital transformation promote green technology innovation?Sustainability14:7497. doi: 10.3390/su14127497

  • CrossRef
  • Google Scholar
  • View reference in article

• 44

  XueR. (2025). The impact of digital transformation on the new productivity of enterprises: an empirical analysis of industry heterogeneity differences. J. Educ. Humanit. Soc. Sci. 58, 78–90. doi: 10.54097/c2bpyq35

  • CrossRef
  • Google Scholar
  • View reference in article

• 45

  XueT.XiX. (2025). Empirical study on the synergistic effects of the composite system of technological finance in the technology industry, technological innovation, and technological funding. J. Knowl. Econ. 16, 8612–8639. doi: 10.1007/s13132-024-02242-y

  • CrossRef
  • Google Scholar
  • View reference in article

• 46

  YangJ.LvX.XuX.ChenX. (2024). Heterogeneous treatment effects of digital transformation on firms' innovation: based on BART method. Pac. Basin Financ. J. 85, 102372. doi: 10.1016/j.pacfin.2024.102372

  • CrossRef
  • Google Scholar
  • View reference in article

• 47

  YangY.RenH.LiuY.YangY. (2025). Integration of technology and finance, digital transformation and corporate green innovation. Finance Res. Lett. 71:106444. doi: 10.1016/j.frl.2024.106444

  • CrossRef
  • Google Scholar
  • View reference in article

• 48

  ZhangH.LiuX.QuanT. (2025). The impact and mechanism of the “specialized and new” policy on enterprise digital transformation: evidence from China. Technol. Anal. Strateg. Manag. 37, 2443–2456. doi: 10.1080/09537325.2024.2361425

  • CrossRef
  • Google Scholar
  • View reference in article

• 49

  ZhaoG.ZhangJ.ChenW. (2026). Environmental regulation policy and enterprise digital transformation: evidence from low-carbon city policies in China. Int. J. Emerg. Mark. 21, 212–243. doi: 10.1108/IJOEM-03-2024-0526

  • CrossRef
  • Google Scholar
  • View reference in article

• 50

  ZhuJ.BakerJ. S.SongZ.YueX. G.LiW. (2023). Government regulatory policies for digital transformation in small and medium-sized manufacturing enterprises: an evolutionary game analysis. Humanit. Soc. Sci. Commun. 10, 1–18. doi: 10.1057/s41599-023-02250-4

  • CrossRef
  • Google Scholar
  • View reference in article