From Carbon to Capability: How Corporate Green and Low-Carbon Transitions Foster New Quality Productive Forces in China

Abstract

China’s national strategies emphasize both achieving carbon peaking and neutrality (“dual carbon” objectives) and fostering high-quality economic development. This dual focus highlights the critical importance of the Green and Low-Carbon Transition (GLCT) of the economy and the development of New Quality Productive Forces (NQPF). Firms are central actors in this transformation, prompting the core research question: How does corporate engagement in GLCT contribute to the formation of NQPF? We investigate this relationship using panel data comprising 33,768 firm-year observations for A-share listed companies across diverse industries in China from 2012 to 2022. Corporate GLCT is measured via textual analysis of annual reports, while an NQPF index, incorporating both tangible and intangible dimensions, is constructed using the entropy method. Our empirical analysis relies primarily on fixed-effects regressions, supplemented by various robustness checks and alternative econometric specifications. The results demonstrate a significantly positive relationship: corporate GLCT robustly promotes the development of NQPF, with dynamic lag structures suggesting delayed productivity realization. Mechanism analysis reveals that this effect operates through three primary channels: improved access to financing, stimulated collaborative innovation and enhanced resource-allocation efficiency. Heterogeneity analysis indicates that the positive impact of GLCT on NQPF is more pronounced for state-owned enterprises (SOEs), firms operating in high-emission sectors, those in energy-efficient or environmentally friendly industries, technology-intensive sectors, non-heavily polluting industries and companies situated in China’s eastern regions. Overall, our findings suggest that corporate GLCT enhances NQPF by improving resource-utilization efficiency and fostering innovation, with these effects amplified by specific regional advantages and firm characteristics. This study offers implications for corporate strategy, highlighting how aligning GLCT initiatives with core business objectives can drive NQPF, and provides evidence relevant for policymakers aiming to optimize environmental governance and foster sustainable economic pathways.

1. Introduction

The economic consequences of greenhouse gas emissions and associated climate change present a significant global challenge [1]. Carbon emissions, in particular, are central to the complex relationship between socioeconomic development and environmental sustainability across nations [2]. As the world’s largest emerging economy, China has experienced considerable environmental degradation, largely driven by its resource-intensive growth model following economic liberalization. Amid tightening resource constraints and escalating pollution, global emphasis on environmental sustainability and low-carbon transitions has intensified. The Paris Agreement, adopted in 2016, marked a pivotal moment in global climate governance [3]. As the world’s second-largest economy and largest carbon emitter, China plays a crucial role in global mitigation efforts [4]. In 2020, China announced ambitious “dual carbon” objectives: achieving peak carbon emissions before 2030 and carbon neutrality before 2060 [5]. This commitment underscores a national strategic priority to transition towards a greener, lower-carbon economy, viewed as integral to achieving high-quality development and a comprehensive green transformation of China’s economic structure. Consequently, understanding the drivers and implications of the Green and Low-Carbon Transition (GLCT) at the firm level has become paramount.

To contextualize China’s current commitments, it is essential to consider the historical trajectory of global CO₂ emissions. Figure 1 illustrates annual CO₂ emissions by world region from 1750 to 2023, highlighting China’s dramatic rise as the largest emitter following its economic liberalization in 1978. This surge underscores the urgency of China’s “dual carbon” goals and frames its green transition as a critical response to both historical and ongoing emissions trends.

In this context, China’s policy framework has evolved to integrate environmental sustainability with economic development. Concurrent with the emphasis on GLCT, China has introduced the concept of New Quality Productive Forces (NQPF), aimed at fostering growth through strategic emerging industries, technological innovation and advanced manufacturing, thereby reshaping the drivers of productivity [7,8]. National policy discourse explicitly links these two agendas, positing that green development forms the foundation for high-quality growth and that NQPF inherently incorporates ecological sustainability [9]. The historical emissions trends, as shown in Figure 1, underscore the necessity of this integration, suggesting that achieving the “dual carbon” goals requires a fundamental reorientation of economic productivity. Achieving the “dual carbon” goals necessitates accelerating structural adjustments in industry, energy and transportation, which is expected to enhance, rather than hinder, overall productivity [10]. Within this framework, corporate adoption of GLCT practices is hypothesized to be a key driver for advancing NQPF.

From a microeconomic perspective, firms undertaking GLCT typically invest in green technologies and sustainable practices to improve energy efficiency, reduce emissions and potentially enhance operational performance, seeking synergies between environmental and economic objectives [11]. Such strategies can stimulate technological progress and optimize resource allocation. However, a strand of literature cautions that environmental regulations and transitions can impose significant costs on firms, potentially eroding competitiveness, particularly in the short term or under specific constraints related to technology, finance or market conditions [12,13]. This raises an empirical question regarding the net impact of GLCT on firm-level productivity, especially when considering the multifaceted nature of NQPF.

Existing research often proxies for GLCT using regional policy variations, such as China’s low-carbon city pilot programs [14,15]. While informative, this approach primarily captures responses to exogenous policy shocks and may not fully reflect firms’ endogenous strategic choices regarding GLCT, potentially overlooking heterogeneity and long-term dynamics. Furthermore, GLCT represents a complex, potentially latent construct not always captured by standard financial metrics. Addressing these limitations, this study employs textual analysis of corporate annual reports to construct a firm-specific measure of GLCT engagement, capturing the intensity of firms’ voluntary green initiatives. Building upon this novel measure, we investigate the following core questions: (1) Does firm-level GLCT foster the development of NQPF? (2) If so, what are the underlying mechanisms driving this relationship? (3) Does the impact of GLCT on NQPF vary systematically across firms with different ownership structures (e.g., state-owned vs. private), carbon emission intensities, geographical locations within China or industry characteristics?

To address this issue, we propose innovative firm-level measures of GLCT and NQPF, drawing upon data from annual reports of Chinese A-share listed enterprises over the period 2012–2022. Utilizing advanced textual analysis techniques, we systematically identify and quantify occurrences of GLCT-related terminology within corporate disclosures, thus generating robust indicators of corporate engagement in GLCT initiatives. Additionally, we construct a comprehensive evaluation framework for NQPF that integrates two primary dimensions: tangible assets and intangible factors. Employing these novel measures, we implement a two-way fixed effects empirical approach to investigate the relationship between GLCT activities and NQPF at the enterprise level, further examining underlying mechanisms and accounting for firm-specific heterogeneity. Our findings advance the understanding of economic consequences and internal mechanisms of GLCT initiatives, contributing significant theoretical insights and practical guidance for policy formulation.

This paper contributes to the literature in three primary ways. First, drawing on the emerging concept of NQPF, we propose and implement a novel firm-level measurement methodology, providing a foundation for future micro-level research on this topic. Second, we offer systematic empirical evidence on the relationship between corporate GLCT and NQPF, addressing a gap in understanding the microeconomic implications of large-scale green transitions. Third, by identifying specific mediating channels, our analysis provides actionable insights for managers seeking to align sustainability initiatives with productivity goals, and for policymakers designing interventions to facilitate both environmental protection and economic upgrading.

2. Literature Review and Hypothesis Development

2.1. Literature Review

2.1.1. Green and Low-Carbon Transition

The term “green” commonly signifies an “ecological bedrock,” embodying the primordial hue of life. Its association with the environment predominantly pertains to ecological sustainability, encapsulating the symbiotic relationship between humanity and nature. However, the “black development paradigm” engendered by the three industrial revolutions has inflicted profound degradation upon green ecosystems [16], spurring the emergence of novel frameworks such as “green development,” “green economy,” “green technology,” “green consumption,” “green cities,” and “green industry”. Confronted with mounting environmental constraints, China’s pursuit of sustainable economic growth necessitates the successful transition to green development.

The concept of “low-carbon” denotes the mitigation of carbon dioxide emissions, originating in the UK government’s 2003 Energy White Paper, where it delineated a “low-carbon economy”—an economic system predicated on minimal emissions, energy efficiency and reduced pollution [17]. Attaining such an economy entails fostering low-carbon urban development, curbing energy consumption and carbon output in cities [18], and leveraging innovations in low-carbon technologies alongside the production of sustainable goods [19].

GLCT by firms hinges critically on strategic decision-making [20]. Such a transition demands a departure from the conventional “pollute first, abate later” treadmill paradigm [21], instead adopting an integrated approach to sustainability across the entire value chain—from input sourcing to product distribution. Achieving this necessitates a triad of institutional backing, corporate innovation and robust societal oversight.

Governments, as the principal agents of environmental governance, implement regulatory frameworks to monitor and sanction ecologically harmful corporate practices [22]. However, beyond command-and-control mechanisms, market-based incentives also play a pivotal role in accelerating firms’ green transition. Empirical evidence demonstrates that external incentives—such as fiscal subsidies and tax relief—substantially spur technological innovation [23] while bolstering firms’ green competitiveness [24].

These incentives further function as credible signals to capital markets and stakeholders, affirming state endorsement of sustainable practices and facilitating corporate access to financing [25]. Internally, executive compensation schemes can mitigate firms’ environmental externalities. Since profit-maximizing behavior is a primary contributor to ecological degradation, managerial incentives align decision-making with sustainability goals [26]. For example, performance-linked remuneration improves the transparency and quality of environmental disclosures [27,28].

2.1.2. New Quality Productive Forces

The notion of “New Quality Productive Forces” emerged in Chinese policy discourse only in 2023. This concept encapsulates a dual emphasis on both novelty (“new”) and substantive transformation (“quality”). With respect to novelty, Liu contends that NQPF transcend mere incremental upgrades [29], instead embodying a tripartite renewal of productive factors, production modalities and production relations [30]. Its manifestations span five key domains: innovation, efficiency, openness, frontier industries and sustainability [31].

The more pivotal dimension, however, lies in its qualitative attributes—characterized by radical, “zero-to-one” breakthroughs that signify a paradigmatic shift in productive capabilities [32]. Yan elaborates that new quality productive forces entail a systemic leap over conventional productive forces [33], redefining industrial structures and their underlying components through fundamental transformations.

Scholars have further dissected this concept through the classical triad of productive forces. Guo and Tang posit that new quality productive forces epitomize advanced productivity, marked by substantial improvements in labor quality, capital goods, production inputs and their synergistic optimization [34]. Pu and Xiang delineate its distinct elements: high-skilled labor, next-generation capital goods (“new-medium” means of labor) and advanced-factor inputs (“new-material” objects of labor) [35]. Empirical analyses, such as Xiao, underscore its technological underpinnings—evidenced by cognitive-intensive labor, AI-augmented tools and digitized production factors—which collectively drive its transformative potential [36].

Building on prior research, the measurement framework for NQPF is structured along two key dimensions. First, it captures the core manifestations of “newness” and “quality” through indicators of informatization, connectivity, digitalization, intelligence and automation. Second, it aligns with the classical triad of productivity determinants—labor inputs, objects of labor and means of production. In related literature, total factor productivity (TFP) is also employed as a proxy for this conceptual construct.

2.1.3. Review

The extant literature has primarily examined how external factors shape corporate decision-making in the green and low-carbon transition and how firms adapt to exogenous environmental shifts. However, it has largely overlooked the catalytic role of internal incentives in facilitating this transition, as well as the synergistic resonance-diffusion effects generated by the interplay between internal incentives and external regulatory frameworks. For firms undergoing transition, while external drivers—such as environmental regulations and policy incentives—are undeniably critical, internal factors like organizational culture, resource-allocation mechanisms and corporate social responsibility consciousness exert equally significant influence on transition-related decisions. Furthermore, prior studies have predominantly focused on the short-term performance implications of green and low-carbon transition, with scant attention paid to long-term determinants, including but not limited to new quality productive forces. A notable limitation of this body of work is its reliance on endogenous theoretical frameworks, which often fail to incorporate the mediating role of external channels—such as technological innovation capacity, competitive market positioning and government subsidies—in shaping firms’ sustained developmental trajectories.

In light of the prevailing methodological heterogeneity in quantifying NQPF, we propose a synthesized framework that systematically incorporates the dual conceptual pillars of novelty (“new”) and qualitative advancement (“quality”) into the classical tripartite structure of productive forces—labor, capital and intermediate inputs. By doing so, we establish a more robust and theoretically grounded indicator system for empirical assessment.

Building on the preceding analysis, this study underscores the imperative to examine—through the lens of China’s distinctive institutional and economic landscape—how firms undergoing green and low-carbon transitions shape the evolution of new quality productive forces. A granular analysis of the mechanisms underpinning this relationship would advance the literature on sustainable productivity growth while informing both firm-level strategies and macro-level policymaking in pursuit of carbon peaking and neutrality (“dual carbon”) objectives.

2.2. Hypothesis Developme

Theoretical frameworks and empirical evidence yield ambiguous predictions regarding the impact of corporate green transitions on productivity, highlighting both potential gains and significant costs. This section synthesizes relevant theories and evidence to frame how environmental regulations—particularly in the context of China’s carbon-neutrality objectives—influence the relationship between corporate GLCT and the development of NQPF, as illustrated in Figure 2. We then derive testable hypotheses concerning the net impact and mediating mechanisms of GLCT.

China’s carbon peak and neutrality targets constitute a significant environmental regulatory intervention with implications for firm behavior and performance. Such regulations introduce a well-documented trade-off: firms incur compliance costs that can distort production decisions, potentially altering factor inputs, reducing plant-level productivity and adversely affecting competitiveness [12,37,38]. Firms respond to these pressures through emission reductions and clean energy adoption—actions central to GLCT. This transition often requires substantial investments in pollution control, low-carbon technologies and process upgrades, potentially crowding out other productive investments and depressing short-run productivity growth [39,40].

Conversely, the Porter Hypothesis suggests that well-designed environmental regulations can stimulate innovation that may offset compliance costs and enhance competitiveness [41]. Regulatory pressure can incentivize firms to increase Research and Development (R&D) investment, leading to both environmental improvements and productivity gains [42], and encourage operational optimizations that improve resource efficiency [43]. Empirical studies offer support for this view; for instance, evidence suggests sustained productivity growth following environmental regulation in Japan’s manufacturing sector [44], and increased R&D investment stimulated by flexible regulations in the EU [45]. Nevertheless, it is important to acknowledge that the implementation of GLCT can entail substantial short-term trade-offs alongside potential longer-term gains. Specifically, firms often face significant financial burdens due to the large upfront investments required for retrofitting facilities, acquiring low-carbon technologies, and adapting processes. There is also the risk of technology lock-in, whereby early commitment to specific green technologies may constrain firms’ flexibility to adopt superior solutions in the future. Additionally, adjustment frictions—including workforce reallocation, retraining requirements and transitional disruptions to established supply chains—can impose temporary costs and operational inefficiencies. These considerations suggest that the net productivity impact of GLCT may not be immediate but could unfold gradually over time, highlighting the need to empirically assess both short-term and lagged effects within a unified framework. Integrating these perspectives leads to our first hypothesis regarding the net effect:

Hypothesis 1 (H1). 

Corporate GLCT has a net positive effect on NQPF.

Beyond the potential net effect, we propose that corporate GLCT influences NQPF primarily through three channels, representing both responses to regulatory pressures and strategic opportunities for productivity enhancement:

(i) Financing Optimization. The capital-intensive nature of GLCT—requiring investments in retrofitting, new technologies and process innovation—can expose firms to significant financing constraints due to large upfront costs, long payback periods and return uncertainty [46,47]. Policy interventions within China’s evolving green finance ecosystem aim to mitigate these frictions [48]. Key instruments include:

Green Bonds: Facilitated access, potentially through inclusion in central bank collateral frameworks, can enhance liquidity and lower the cost of capital for decarbonization [49].

Green Credit: Preferential lending conditions incentivize the adoption of cleaner technologies and processes [50].

Green Funds: Targeted funds support transition efforts, particularly in emission-intensive sectors [51].

Carbon Markets: Emissions trading schemes can generate revenue streams that improve the financial viability of abatement investments [52].

This institutional framework aims not only to alleviate financing constraints but also to signal policy commitment, potentially crowding in private investment towards productivity-enhancing green innovation. By reducing the risk-adjusted cost of capital for sustainable projects, these mechanisms can accelerate the adoption and diffusion of advanced technologies integral to NQPF [53].

(ii) Collaborative Innovation. The dual challenge of reducing emissions while enhancing productivity often necessitates technological advancements beyond the capabilities of individual firms, particularly given existing innovation system characteristics [54]. GLCT can thus compel firms to engage more deeply in open and collaborative innovation models:

Network Expansion: Firms may strengthen ties with universities, suppliers and even competitors to access complementary knowledge and capabilities [55].

Resource Pooling: Consortia and partnerships can emerge to address shared technological hurdles, especially in areas like clean production processes [56].

Knowledge Spillovers: Enhanced collaboration within innovation ecosystems can accelerate the diffusion of best practices and new technologies [57].

Such collaborative approaches can enhance the innovation production function by helping to internalize positive externalities [58] that might otherwise lead to underinvestment in R&D. The recombination of diverse knowledge and capabilities through networks can foster breakthrough innovations and drive shifts in the technological frontier, contributing fundamentally to NQPF [59].

(iii) Resource-Allocation Efficiency. Enhanced allocative efficiency is another key channel through which GLCT may foster NQPF. Extending the Porter Hypothesis [41], environmental regulations can induce firms to re-evaluate and optimize the allocation of resources, particularly labor and capital:

Labor Allocation: While potentially exerting complex effects on labor productivity [60], the transition towards GLCT incentivizes firms to invest in human capital aligned with green technologies and sustainable management practices, potentially shifting labor towards higher value-added activities [61].

Capital Allocation: Environmental regulations can discourage investment in inefficient, pollution-intensive activities and potentially redirect capital towards higher-productivity firms or projects [62]. By highlighting inefficiencies (e.g., in energy use), regulations, viewed through a bounded rationality lens, can prompt firms to identify and pursue improvements [62]. Firms respond by optimizing energy consumption, investing in energy-saving equipment and adopting cleaner processes [63].

Firms anticipating stricter environmental constraints or seeking competitive advantages may proactively optimize resource allocation. This systematic improvement in the efficiency of labor and capital deployment enhances overall productivity and underpins the development of NQPF.

Based on these mechanisms, we propose:

Hypothesis 2 (H2). 

Corporate GLCT enhances NQPF through mechanisms including improved financing conditions, stimulated collaborative innovation and increased resource-allocation efficiency.

3. Research Design

3.1. Variable Construction

3.1.1. Dependent Variable

This study constructs a firm-level evaluation index for New Quality Productive Forces (NQPF) by adapting the advanced conceptual framework proposed by Han [64]. This framework is adopted for its theoretical superiority in capturing the essence of NQPF. According to Han [64], NQPF is defined not as a simple upgrade of inputs, but as a systemic leap driven by the interplay of two distinct types of elements: “substantive factors” and “penetrative factors”. Substantive factors are the tangible foundations of production—new laborers, new means of labor and new objects of labor—that are themselves undergoing a qualitative upgrade. The framework’s key innovation lies in its explicit identification of “penetrative factors”—new technology, production organization and data elements—which are non-physical enablers. While prior studies have often focused only on the substantive factors, they have largely overlooked the indispensable role of these penetrative factors. In fact, these enablers do not participate in production directly but rather permeate the substantive factors to enhance efficiency, optimize processes and ultimately drive the overall productivity leap. Translating this dual-dimensional framework to the firm level provides crucial micro-foundations for a macroeconomic concept. The specific indicators are selected based on established literature and the availability of firm-level data for Chinese A-share listed companies.

The substantive factors comprise (i) new laborers, (ii) new means of labor and (iii) new objects of labor. For new laborers, we measure the share of R&D personnel, their compensation share, the proportion of highly educated employees and whether management possesses a digital background. These metrics collectively capture the scale, quality and composition of a firm’s human capital, reflecting its investment in talent. New means of labor is proxied by firm-level robot adoption rates, fixed asset intensity, geographic proximity to high-speed rail networks and access to Fifth-Generation mobile network (5G) infrastructure—indicators that gauge modernization in production tools and industrial foundations. New objects of labor focuses on ecological sustainability and innovation potential, operationalized through R&D intensity and intangible asset ratios.

The penetrative factors encompass the core drivers that empower and transform the substantive factors: (i) technological innovation, (ii) production organization and (iii) data utilization. Technological innovation, the core driver of NQPF, is measured by R&D expenditures and patent applications, reflecting a firm’s innovative capacity. Production organization evaluates advancements in smart manufacturing, green production and industrial integration. The level of smart manufacturing is quantified by a composite ’Intelligent Transformation Index,’ which assesses both technology-related financial investments and the application of intelligent technologies based on textual analysis. The detailed construction method and indicators are provided in Appendix A.1. A firm’s progress in green production is specifically proxied by its “Green Innovation” output, which is quantitatively measured as the logarithm of the number of green patent applications filed by the firm in a given year, identified by matching patents against the WIPO’s IPC Green Inventory. Data utilization assesses a firm’s digital transformation level. This is captured by a composite “Digitization Index,” which is constructed via textual analysis of corporate annual reports based on a structured five-dimension keyword dictionary. The full construction details and keyword lists are available in Appendix A.2.

Table 1. Evaluation index system for enterprise-level new quality productive forces.

Dimension	Component	Sub-Indicator	Metric	Sign
Tangible Factors	Innovative Labor	Innovative Labor Quantity	R&D Personnel Share	+
		Innovative Labor Compensation	R&D Payroll Intensity	+
		Educational Attainment	Share of College-Educated Workers	+
		Digital Expertise in Management	Management’s Digital Background Dummy	+
	Advanced Capital	Automation Intensity	Robot Adoption Rate	+
			Fixed Asset Intensity	+
		Infrastructure Modernization	High-Speed Rail(HSR) Access Dummy	+
			5G Infrastructure Dummy	+
	Novel Inputs	Environmental Stewardship	Environmental, Social and Governance(ESG) Environmental Score	+
		Innovation Commitment	Direct R&D Expenditure Ratio	+
			Intangible Asset Intensity	+
Intangible Factors	Technology	R&D Effort	R&D Intensity	+
		Innovation Yield	Patent Stock (log)	+
	Production System	Smart Manufacturing	Artificial Intelligence(AI) Adoption Index (log)	+
		Green Transition	Green Innovation (log)	+
		Industrial Synergy	Information and Communications Technology(ICT)-Industry Integration Dummy	+
	Digital Assets	Digital Transformation	Digitization Index (log)	+

Notes: This table presents the comprehensive evaluation framework for measuring enterprise-level New Quality Productive Forces (NQPF). The index system consists of two main dimensions: Tangible Factors and Intangible Factors. Tangible Factors include innovative labor, advanced capital and novel inputs. Intangible Factors encompass technology, production systems and digital assets. All indicators are normalized and aggregated using principal component analysis or factor analysis methods. The “+” sign indicates a positive relationship with the NQPF index.

details the complete indicator system.

To ensure objective weighting, we apply the entropy method. The procedure involves: (1) standardizing raw indicators to neutralize scale effects, (2) computing information entropy and redundancy for each metric, (3) deriving weights from entropy redundancy and (4) aggregating standardized values by their weights to obtain a composite NQPF score [65].

To address the potential limitations of the entropy method and ascertain the scientific robustness of our findings, we have implemented a comprehensive testing strategy, which is detailed in the subsequent robustness check section. This strategy ensures our results are not an artifact of a single measurement choice by employing four key approaches. This strategy will confirm that our findings are robust by employing four key approaches: (1) adopting an alternative NQPF measurement framework from established literature (e.g., Song et al., 2024 [66]); (2) using Principal Component Analysis (PCA) as an alternative weighting scheme to reconstruct the index; (3) employing Factor Analysis to validate the internal construct of our primary index; and (4) using the traditional Total Factor Productivity (TFP) as an alternative dependent variable. The consistency of results across these rigorous tests will affirm that our measured NQPF index is robust.

3.1.2. Independent Variable

The green and low-carbon transition ($GLCT$) represents a strategic adaptation undertaken by firms in response to their operational and regulatory environment. To objectively and systematically quantify this transition, we construct a firm-level GLCT index through a multi-step process involving textual analysis of corporate annual reports.

The process began with the construction of a comprehensive, custom-built dictionary of terms. To ensure relevance and coverage, we identified keywords from influential national policy directives—including the Government Work Report [67], the “14th Five-Year Plan” for Industrial Green Development [68] and the State Council’s Action Plan for Energy Conservation and Carbon Reduction [69] (2024–2025)—alongside extant literature. This initial pool of candidate terms was then carefully refined by merging synonyms and removing ambiguous words, resulting in a final lexicon of 82 specific keywords. These keywords span five domains: (i) advocacy and public commitments, (ii) strategic orientation, (iii) technological innovation, (iv) emissions abatement and (v) monitoring and compliance. Representative examples of these keywords, in both Chinese and English, are provided in Appendix A.3. This lexicon was subsequently compiled into a custom dictionary for use with Jieba (Python 3.12). library for Chinese text segmentation, ensuring that multi-character technical terms (e.g., “carbon neutrality”) were accurately identified as single units.

With the dictionary in place, we acquired the full texts of all corporate annual reports for our sample firms from the CNINFO database. Each entire annual report was then processed using the Jieba tokenizer, which was pre-loaded with our custom dictionary. After segmenting the text, we calculated the frequency of each of the 82 keywords. The raw keyword count for each firm-year is the simple sum of these frequencies. To create a comparable measure of disclosure intensity that controls for document length, this raw count was normalized to represent the frequency per one thousand words. The final GLCT index used in our regression analysis is therefore calculated as follows:

$$GLC{T}_{i,t}=\left({\displaystyle \frac{\mathrm{Keyword}{\mathrm{Count}}_{i,t}}{\mathrm{Total}\mathrm{Word}{\mathrm{Count}}_{i,t}}}\right)\times 1000$$

(1)

A higher value of the $GLC{T}_{i,t}$ index indicates a greater number of green and low-carbon transition keywords mentioned per one thousand words in the corporate annual report. The application of this methodology results in indices for the five constituent domains of GLCT.

Table 2. Descriptive statistics of GLCT sub-dimensions.

Category	Variable	Mean	Median	Std. Dev.	Min	Max	N
GLCT Sub-Dimensions	Advocacy & Commitments	0.45	0.46	0.19	0	1	33,768
	Strategic Orientation	0.28	0.24	0.20	0	1	33,768
	Technological Innovation	0.03	0.00	0.07	0	0.91	33,768
	Emissions Abatement	0.21	0.19	0.14	0	1	33,768
	Monitoring & Compliance	0.02	0.00	0.04	0	1	33,768

presents the descriptive statistics for these sub-dimensions, offering insights into the nature of corporate green disclosures.

3.1.3. Control Variables

Following prior studies [70,71], we include a comprehensive set of control variables spanning three key dimensions: (1) Firm characteristics: firm age (Age) and size (Size); (2) Financial metrics: leverage (Lev), capital intensity (Capital), liquidity ratio (Liquid), return on assets (ROA), price-to-book ratio (PB) and Tobin’s Q (TobinQ); (3) Governance variables: audit opinion (AO), ownership concentration (measured by the largest shareholder’s stake, First), divergence between control and cash-flow rights (Seperation), share of independent directors (BI) and sales growth (SGrow).

3.2. Empirical Strategy

3.2.1. Baseline Model and Endogeneity Mitigation

To empirically assess whether green and low-carbon transition enhances firms’ new-quality productivity, our primary approach utilizes a two-way fixed-effects panel data model. This model is specified as:

$${\mathrm{NQPF}}_{it}={\beta }_0+{\beta }_1{\mathrm{GLCT}}_{it}+{\gamma }^{\prime }{X}_{it}+{\mu }_i+{\lambda }_t+{ϵ}_{it}$$

(2)

where for firm i at time t, ${\mathrm{NQPF}}_{it}$ is our dependent variable, and ${\mathrm{GLCT}}_{it}$ is the key explanatory variable. $X_{it}$ is a vector of control variables, while ${\mu }_i$ and ${\lambda }_t$ represent firm and year fixed effects, respectively, to control for unobserved time-invariant firm characteristics and common macroeconomic shocks.

While the fixed-effects model helps address omitted variable bias, the relationship between GLCT and NQPF may still be influenced by reverse causality and dynamic effects. To mitigate these concerns more effectively and to capture the possibility that the impact of GLCT unfolds over time, we estimate the following distributed lag model:

$${\mathrm{NQPF}}_{it}={\beta }_0+{\beta }_1{\mathrm{GLCT}}_{it}+{\beta }_2{\mathrm{GLCT}}_{i,t-1}+{\beta }_3{\mathrm{GLCT}}_{i,t-2}+{\gamma }^{\prime }{X}_{it}+{\mu }_i+{\lambda }_t+{ϵ}_{it}$$

(3)

A significant and positive ${\beta }_2$ and/or ${\beta }_3$ in this specification provide stronger evidence of a temporal and potentially causal relationship between earlier green initiatives and subsequent improvements in new quality productive forces, consistent with the results reported in the table in Section 4.1.

3.2.2. Causal Inference via Difference-in-Differences (DID)

To further ensure the credibility of our findings, we employ a comprehensive suite of robustness checks, which are detailed in Section 4. This policy was implemented in a staggered manner across three distinct batches of cities (the specific city list and timing are provided in Appendix A.4). The model compares the change in NQPF for firms located in the pilot cities (the treatment group) after the policy implementation, against the change for firms in non-pilot cities (the control group) over the same period. The staggered DID model is specified as:

$${\mathrm{NQPF}}_{it}={\delta }_0+{\delta }_1({\mathrm{Treat}}_i\times {\mathrm{Post}}_t)+{\delta }^{\prime }{X}_{it}+{\mu }_i+{\lambda }_t+{ϵ}_{it}$$

(4)

where ${\mathrm{Treat}}_i$ is a dummy variable equal to 1 if firm i is located in a city that becomes a low-carbon pilot, and 0 otherwise. ${\mathrm{Post}}_t$ is a dummy variable equal to 1 for the years after the policy is implemented in that specific city. The coefficient of interest, $\delta$, captures the average treatment effect of the policy on the treated firms. A positive and significant ${\delta }_1$ would provide strong, policy-based evidence that exogenously driven green transition pressures foster the development of NQPF.

3.3. Data and Sample Construction

Our empirical analysis employs a sample of A-share firms listed on the Shanghai and Shenzhen stock exchanges between 2012 and 2022. We choose 2012 as the starting point, on the one hand, to ensure the data quality and consistency of key variables. On the other hand, this timing allows us to bypass the lingering effects of the global financial crisis and to focus on a new era characterized by significantly strengthened environmental governance policies in China, thus making our findings more targeted and relevant. Financial statement data are collected from CNINFO, while additional firm-level variables are sourced from the CSMAR and CNRDS databases. To ensure sample integrity, we implement the following screening procedures: (1) Excluding financial-sector firms due to their distinct regulatory framework; (2) Dropping observations associated with special treatment status (ST/*ST) or non-active listings (suspensions/delistings); (3) Removing entries with incomplete data on core variables; (4) Imputing limited missing values via linear interpolation where applicable. The final balanced panel comprises 33,768 firm-year observations. Summary statistics for key variables are reported in

Table 3. Descriptive statistics of variables.

Category	Variable	Symbol	Mean	Med.	Std.	Min	Max	N
Dependent Var.	New Quality Productive Forces	$NQPF$	7.31	5.49	4.52	0.04	20.28	33,768
Core Indep. Var.	Green & Low-Carbon Transit. Index	$GLCT$	0.07	0.04	0.09	0	1.20	33,768
Control Vars.	Firm Age	$Age$	2.88	2.94	0.34	1.61	3.50	33,768
	Firm Size (log assets)	$Size$	22.17	21.98	1.32	18.71	26.07	33,768
	Leverage	$Lev$	0.42	0.40	0.21	0.06	0.94	33,768
	Capital Intensity	$Capital$	2.57	1.94	2.24	0.40	15.09	33,768
	Liquidity Ratio	$Liquid$	0.58	0.59	0.20	0.10	0.97	33,768
	ROA	$Roa$	0.04	0.04	0.07	−0.32	0.27	33,768
	Price-to-Book	$PB$	3.92	2.73	4.30	−2.26	34.02	33,768
	Tobin’s Q	$TobinQ$	2.07	1.62	1.40	−2.09	9.06	33,768
	Clean Audit Opinion	$AO$	0.97	1.00	0.18	0	1.00	33,768
	Largest Shareholder (%)	$First$	34.29	31.95	14.98	9.09	84.11	33,768
	Control-Cash Divergence	$Seperation$	4.53	0.00	7.25	0	29.73	33,768
	Board Independence (%)	$BI$	0.38	0.36	0.05	0.22	0.57	33,768
	Sales Growth	$SGrow$	0.16	0.10	0.43	−0.62	2.78	33,768

Notes: This table presents descriptive statistics for all variables used in the analysis. The sample consists of 33,768 firm-year observations from Chinese listed companies. All continuous variables are winsorized at the 1% and 99% levels. $NQPF$ represents the New Quality Productive Forces index. $GLCT$ is the Green and Low-Carbon Transition Index. Control variables include firm characteristics, financial performance and governance measures.

.

4. Empirical Results

4.1. Baseline Regression Results

This study investigates the effect of corporate green and low-carbon transition ($GLCT$) on new quality productive forces ($NQPF$) by estimating Equation (1).

Table 4. Baseline regression: the effect of GLCT on NQPF.

	Dependent Variable: New Quality Productive Forces (NQPF)
	(1)	(2)	(3)	(4)
	Core Explanatory Variable: Contemporary GLCT		Core Explanatory Variable: Lagged GLCT
GLCT	3.118 ***	2.476 ***	1.447 **	1.552 **
	(0.598)	(0.597)	(0.600)	(0.634)
L.GLCT			1.441 ***	0.703
			(0.609)	(0.532)
L2.GLCT				1.096 *
				(0.640)
age		−0.132	−0.310	−0.085
		(0.457)	(0.524)	(0.625)
size		0.467 ***	0.438 ***	0.444 ***
		(0.062)	(0.065)	(0.068)
lev		−0.277	−0.376	−0.407
		(0.227)	(0.239)	(0.253)
capital		−0.018	−0.026 *	−0.022
		(0.015)	(0.016)	(0.017)
liquid		−1.039 ***	−0.962 ***	−0.952 ***
		(0.211)	(0.224)	(0.241)
roa		−0.014	0.003	0.004
		(0.009)	(0.010)	(0.011)
TobinQ		−1.217 ***	−1.022 ***	−1.025 ***
		(0.325)	(0.330)	(0.334)
PB		0.070 ***	0.004	0.009
		(0.024)	(0.026)	(0.028)
AO		0.151	0.123	0.086
		(0.115)	(0.115)	(0.114)
top1		−0.001	0.002	0.007
		(0.004)	(0.004)	(0.004)
Seperation		−0.003	−0.003	−0.004
		(0.006)	(0.006)	(0.007)
BI1		−0.923	−0.872	−0.780
		(0.587)	(0.601)	(0.631)
SGrow		−0.058	−0.084 **	−0.085 **
		(0.038)	(0.039)	(0.040)
Constant	7.085 ***	−1.879	−0.436	−1.149
	(0.043)	(1.821)	(2.022)	(2.315)
Firm Fixed Effects	YES	YES	YES	YES
Time Fixed Effects	YES	YES	YES	YES
Observations	33,768	33,768	29,258	24,955
R2	0.758	0.761	0.771	0.781

Notes: This table presents the baseline regression results. The dependent variable is NQPF (New Quality Productive Forces). All models are estimated using a two-way fixed-effects specification. Standard errors are in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1. Columns (1) and (2) examine the effect of the contemporary GLCT variable. Columns (3) and (4) examine the effect of lagged GLCT variables to mitigate endogeneity and test for dynamic effects.

reports the regression results. Column (1) includes only year and firm fixed effects to account for macroeconomic shocks and time-invariant unobserved heterogeneity at the firm level. The coefficient on $GLCT$ is positive and statistically significant at the 1% level. In Column (2), we augment the specification with firm-level controls to address potential omitted-variable bias. The estimated effect of $GLCT$ remains statistically significant at the 1% level.

To address potential reverse causality and capture the dynamic relationship between GLCT and NQPF, Columns (3) and (4) incorporate lagged measures of GLCT. In Column (3), the one-period lag (L.GLCT) yields a coefficient of 1.441, statistically significant at the 5% level. This indicates that the positive effect of GLCT on NQPF persists into the subsequent period, suggesting a sustained rather than fleeting impact. The persistence aligns with the expectation that green transitions, involving structural changes in corporate practices, require time to translate into productivity gains. Column (4) extends this analysis by including both one-period (L.GLCT) and two-period lags (L2.GLCT). Here, the two-period lag coefficient is 1.906, significant at the 10% level, while the one-period lag coefficient drops to 0.703 and becomes statistically insignificant. This shift suggests that the benefits of GLCT may not fully manifest until two periods after implementation, potentially reflecting a delayed realization of innovation or efficiency gains. The insignificance of the one-period lag in this specification could stem from collinearity between the lagged terms, as successive lags may capture overlapping variation. These findings underscore the importance of a dynamic perspective in evaluating green transitions, highlighting that their contribution to NQPF unfolds gradually, consistent with theoretical models of adjustment costs and learning-by-doing in sustainability initiatives.

The inclusion of firm-level controls in Columns (2) through (4) aims to mitigate omitted-variable bias. In empirical research, control variables play a critical role in isolating the net effect of the independent variable on the dependent variable. By incorporating potential confounding factors, they enhance the credibility of causal inferences. In this study, we selected eight control variables—such as firm size, liquidity and Tobin’s Q—based on established theoretical frameworks and prior literature. The analysis results show that some control variables have a robust and significant impact on new-quality productivity. Among these, the coefficient for company size (size) is consistently significantly positive at the 1% level, indicating that larger companies have higher levels of new-quality productivity, which aligns with the theoretical expectations of economies of scale and resource endowments. Notably, certain financial and market indicators exhibit significant negative relationships. The coefficients for company liquidity (liquid), sales growth rate (SGrow) and the Tobin Q ratio (TobinQ), which measures market growth expectations, are all significantly negative across all models. This may suggest that companies facing weaker financial constraints or in a phase of rapid sales expansion may have relatively insufficient motivation or resource allocation for long-term, fundamental innovation (i.e., new-quality productivity development); simultaneously, the intrinsic value of new-quality productivity may not be fully recognized by capital markets in a timely manner. However, certain control variables, including firm age and leverage, did not demonstrate statistical significance within the model. This lack of significance is likely attributable to the inclusion of fixed effects, which effectively control for unobserved heterogeneity and absorb the independent effects of these variables. Retaining these statistically insignificant control variables is justified, as it mitigates the risk of omitted variable bias, thereby preserving the robustness and explanatory power of the model.

The analysis reveals that GLCT positively influences NQPF, with lagged effects indicating a delayed but strengthening impact over two periods. Among controls, size and liquid consistently shape NQPF, while others appear insignificant, likely due to the model’s structure or sample characteristics. These insights refine our understanding of green transitions’ economic implications, offering a rigorous foundation for policy and future research.

4.2. Robustness Checks

The baseline estimates suggest a positive effect of green and low-carbon transition on firms’ new quality productive forces. To address potential confounders and bolster the robustness of our findings, we implement a battery of robustness checks: (1) alternative core explanatory variables; (2) alternative dependent variables; (3) sample scope adjustments; (4) alternative standard error clustering approaches; (5) propensity score matching (PSM) and entropy balancing for sample reweighting; (6) alternative empirical specifications; and (7) instrumental variable two-stage least squares (2SLS) estimation. Comprehensive results are reported in

Table 5. Robustness tests I: alternative variable specifications.

VARIABLES	(1)	(2)	(3)	(4)	(5)
	Explanatory Variable	Dependent Variable: New Quality Productive Forces (NQPF)
	Robustness	Robustness
	Alt. Explan. Var.	Alt. Measure	PCA	Factor Analysis	TFP
GLCT	0.220 ***	1.144 ***	0.470 *	0.210 ***	0.305 ***
	(0.043)	(0.414)	(0.084)	(0.049)	(0.106)
Controls	YES	YES	YES	YES	YES
Firm Fixed Effects	YES	YES	YES	YES	YES
Year Fixed Effects	YES	YES	YES	YES	YES
Observations	33,768	30,370	33,768	33,768	30,431
R2	0.761	0.825	0.550	0.774	0.905

Notes: Standard errors are in parentheses. *** p < 0.01, * p < 0.1. Column (1) replaces the explanatory variable with an alternative measure. Column (2) replaces the dependent variable with an NQPF index constructed using the method of [66]. Column (3) uses an NQPF index constructed from Principal Component Analysis (PCA). Column (4) uses an NQPF index constructed from Factor Analysis. Column (5) replaces the dependent variable with Total Factor Productivity (TFP) estimated via the Olley-Pakes method.

,

Table 6. Robustness tests II: endogeneity and methodological checks.

VARIABLES	(1)	(2)	(3)	(4)	(5)
	Sample	Alt. Class.	PSM	Entropy	DID
	Screening	Method		Balancing
GLCT	2.824 ***	2.340 ***
	(0.698)	(0.588)
GLCT_Matched			0.227 ***	0.191 ***
			(0.069)	(0.056)
DID					0.367 ***
					(0.130)
Controls	YES	YES	YES	YES	YES
Firm Fixed Effects	YES	YES	YES	YES	YES
Year Fixed Effects	YES	YES	YES	YES	YES
Observations	33,768	33,764	23,352	33,768	33,768
R2	0.760	0.764	0.788	0.767	0.760

Notes: Standard errors are in parentheses. *** p < 0.01. The dependent variable is New Quality Productive Forces (NQPF). Column (1) applies data screening methods (winsorizing). Column (2) uses alternative classification methods for standard errors. Columns (3) and (4) employ propensity score matching (PSM) and entropy-balancing techniques, respectively. Column (5) uses a difference-in-differences (DID) estimation based on the low-carbon city pilot policy.

and

Table 7. Instrumental variable regression results.

VARIABLES	(1)	(2)
	GLCT	NQPF
GLCT		0.42 ***
		(0.013)
City_GLCT	0.88 **
	(0.45)
Control Variables	YES	YES
Firm Fixed Effects	YES	YES
Year Fixed Effects	YES	YES
Observations	26,862	26,862

Notes: Standard errors in parentheses. *** p < 0.01, ** p < 0.05. Column (1) presents the first-stage regression with City_GLCT as the instrumental variable. Column (2) shows the second-stage results with the instrumented GLCT as the independent variable and NQPF as the dependent variable.

.

4.2.1. Alternative Measure of the Explanatory Variable (GLCT)

Following Ding [72], we reconstruct our explanatory variable as the natural logarithm of one plus the count of green and low-carbon-related keywords in listed firms’ annual reports, which proxies for the intensity of corporate green transition. As reported in Column (1) of Table 5, the coefficient remains statistically significant at the 1% level, confirming that firms’ green transformation exerts a positive effect on new quality productive forces. This result further corroborates the robustness of Hypothesis 1.

4.2.2. Alternative Measure of the Dependent Variable (NQPF)

To ensure the credibility of our main findings, we address potential concerns regarding our primary NQPF index from multiple angles. The primary concerns include: (1) the potential subjectivity in the initial selection of indicators; (2) the sensitivity of the results to the entropy weighting method; and (3) the overall construct validity of our novel index. We therefore conduct a comprehensive series of robustness checks by reconstructing our dependent variable using four distinct and well-established methods.

a. Alternative Conceptual Framework. We first reconstruct the NQPF index based on the “two-factor theory of productivity” as proposed by Song et al. [66]. This framework uses a different set of 11 indicators based on a theory of “labor force” and “means of production,” which are also weighted using the entropy method. This test directly addresses the concern of indicator selection sensitivity. The result in Column (2) of Table 5 shows that the coefficient on GLCT remains positive and highly significant (1.144), suggesting our findings hold even when the NQPF index is constructed under a different theoretical framework and with a different set of indicators.

b. Alternative Weighting and Structuring Methods. To ensure our results are not artifacts of the entropy weighting technique, we reconstruct our NQPF index using two distinct multivariate statistical methods.

Principal Component Analysis (PCA): This method addresses the sensitivity to the weighting scheme. We apply PCA to our original set of indicators and extract the first eight principal components, which collectively explain 73.3% of the total variance. A new NQPF index is then constructed as a weighted average of these components, with weights determined by their proportion of explained variance. As shown in Column (3), the regression coefficient using this PCA-based index is 0.47 and remains statistically significant, confirming that our findings are robust to a different objective weighting scheme.

Factor Analysis: This method serves a dual purpose: it validates the internal structure of our indicators and creates another alternative index. The data’s suitability is confirmed by a Kaiser-Meyer-Olkin (KMO) measure of 0.729 and a highly significant Bartlett’s test of sphericity (p < 0.001). The analysis extracted six common factors (eigenvalues > 1) explaining 62.11% of the total variance. A composite index is then calculated from the weighted factor scores. As reported in Column (4), the coefficient on this factor-based index is 0.21 and highly significant, further corroborating the robustness of our results and the construct validity of our primary measure.

c. Alternative Productivity Proxy (Total Factor Productivity). Finally, to ensure our findings are not merely capturing nuances of our novel NQPF construct, we test whether GLCT has a similar effect on a traditional, widely-accepted measure of firm productivity: Total Factor Productivity (TFP). We estimate TFP using the semi-parametric method developed by Olley and Pakes [73]. This approach addresses the endogeneity problems inherent in production function estimation by using the firm’s investment decision as a proxy for unobserved productivity shocks. The result of regressing this OP-estimated TFP on our variable of interest shows that the coefficient on GLCT is 0.305 and remains highly significant (p-value = 0.004). This demonstrates that the positive impact of GLCT extends beyond our comprehensive NQPF measure to a classic, rigorously estimated measure of productivity, which strongly reinforces our core conclusion.

4.2.3. Data Trimming Procedure

To address potential bias from outliers, all continuous variables are winsorized at the 1st and 99th percentiles. As reported in Column (1) of Table 6, the coefficient on green low-carbon transition ($GLCT$) retains its economic and statistical significance (1% level) post-trimming, corroborating the robustness of the baseline estimates.

4.2.4. Adjustments to Clustering Methodology

Since listed firms within the same Chinese city often share substantial economic and operational linkages, we cluster standard errors at the city level to account for potential spatial correlation. The regression estimates reveal that the coefficient remains statistically significant at the 1% level, confirming the robustness of our findings to alternative clustering specifications.

4.2.5. Propensity Score and Entropy-Balancing Methods

To mitigate concerns of omitted variable bias arising from the possibility that firm-specific developmental factors jointly determine both new quality productive forces and green transition decisions, we implement propensity score matching (PSM) and entropy balancing. Firms are stratified into low- and high-transition groups based on whether their green transition intensity falls below or above the sample mean. Post-matching diagnostics indicate that standardized biases for all covariates are below 5%, with most exhibiting no statistically significant differences across treatment and control groups. The corresponding regression outputs are reported in Table 6.

4.2.6. Alternative Identification Strategy

Since environmental regulation policies amplify the pressure on firms to undergo green and low-carbon transformation, we employ a difference-in-differences (DID) design to estimate the relative shifts in competitiveness attributable to firms’ transition following the policy intervention. Building on Yang [74], we leverage the staggered rollout of China’s three-phase low-carbon city pilot program as a quasi-experimental variation and estimate a staggered DID model. Column (5) of Table 6 reports a coefficient of 0.367, significant at the 1% level, corroborating the robustness of our baseline findings.

4.2.7. Instrumental Variable Estimation

To further mitigate endogeneity, we implement an instrumental variables ($City\_GLCT$) strategy, using city-level regulatory intensity for green transition ($City\_GLCT$) as our instrument. The exclusion restriction is justified on two grounds: First, local governments’ regulatory stringency directly shapes firms’ decisions regarding green transition, as stricter environmental policies compel firms to invest in green innovation and adopt cleaner production methods. Second, city-level green transition policies are designed based on aggregate environmental and economic targets rather than firm-specific new quality productive forces, ensuring exogeneity.

The 2SLS results, presented in Columns (1)–(2) of Table 7, validate the instrument’s relevance: The first-stage coefficient is 0.88 (significant at 5%), and the second-stage estimate for is 0.42 (significant at 1%), confirming that green transition robustly enhances new quality productive forces. The F-statistic surpasses the Stock-Yogo weak IV test threshold (20% maximal size), reinforcing the instrument’s strength. These results further substantiate the credibility of our baseline estimates.

4.3. Mechanism Analysis

To investigate whether financing optimization, collaborative innovation and improved resource allocation serve as plausible mechanisms through which firms’ green and low-carbon transition enhances new quality productive forces, we estimate the following empirical specification:

$$NQP{F}_{i,t}={\gamma }_0+{\gamma }_1{M}_{i,t}\times GLC{T}_{i,t}+{\gamma }^{\prime }{X}_{it}+{\mu }_i+{\lambda }_t+{ϵ}_{it}+{\epsilon }_{i,t}$$

(5)

where $M_{i,t}$ denotes the mediating variables—proxies for financing efficiency, innovation synergy and resource-allocation efficiency—and all other variables align with the baseline regression.

4.3.1. Financing-Optimization Effect

Enhancing new quality productive forces necessitates substantial technological innovation and sustained capital investment [75]. However, firms frequently encounter financing constraints when engaging in innovative R&D. Green and low-carbon transformation mitigates these constraints by improving access to preferential financing and policy support, thereby optimizing capital allocation and securing essential funding for productivity advancement. This study investigates how green transformation facilitates new quality productive forces through financing optimization along three key mechanisms.

Financing constraints. We measure financing constraints using the WW index [76] The index is constructed based on the formula:

$$\begin{array}{cc}\hfill WW=& -0.091\times C{F}_{it}-0.062\times DIVPO{S}_{it}+0.021\times LT{D}_{it}\hfill \\ \hfill & -0.044\times LNT{A}_{it}+0.102\times IS{G}_{it}-0.035\times S{G}_{it}\hfill \end{array}$$

(6)

Here, $C{F}_{it}$ represents the cash flow to total assets ratio; $DIVPO{S}_{it}$ is a dummy variable for dividend payments (1 if dividends are paid, 0 otherwise); $LT{D}_{it}$ is the long-term debt to total assets ratio; $LNT{A}_{it}$ is the natural logarithm of total assets; $IS{G}_{it}$ is industry sales growth; and $S{G}_{it}$ is the firm’s own sales growth. A higher (less negative) value of the WW index indicates a greater degree of financing constraints.

We then test its moderating role in the relationship between green transformation and productivity. Column (1) of

Table 8. Mechanism tests.

	(1)	(2)	(3)	(4)	(5)	(6)
	Dependent Variable: New Quality Productive Forces ($\mathit{NQPF}$)
$GLCT\times WW$	−0.086 ***
	(0.027)
$WW$	0.219 ***
	(0.069)
$GLCT\times Taxp$		5.116 *
		(2.783)
$Taxp$		0.210
		(0.259)
$GLCT\times Subsidy$			50.09 **
			(21.23)
$Subsidy$			−4.60 *
			(2.41)
$GLCT\times ConInv$				0.970 ***
				(0.283)
$ConInv$				0.155 ***
				(0.042)
$GLCT\times EI$					−2.691 ***
					(1.209)
$EI$					−0.090
					(0.088)
$GLCT\times Ineff$						−7.499 ***
						(3.542)
$Ineff$						0.154
						(0.434)
$GLCT$	1.526 ***	2.415 ***	1.96 ***	1.339 ***	3.561 ***	2.881 ***
	(0.644)	(0.813)	(0.54)	(0.615)	(0.736)	(0.735)
Control Variables	YES	YES	YES	YES	YES	YES
Firm Fixed Effects	YES	YES	YES	YES	YES	YES
Year Fixed Effects	YES	YES	YES	YES	YES	YES
Observations	28,186	33,566	33,566	33,238	33,352	25,605
R2	0.774	0.479	0.750	0.763	0.763	0.768

Notes: Standard errors in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1. Column (1) examines the financing constraint mechanism using theWWindex. Column (2) tests the tax incentive mechanism. Column (3) investigates the subsidy effect. Column (4) explores the collaborative innovation mechanism. Columns (5) and (6) test the resource-allocation efficiency mechanisms.

reveals a significantly negative interaction term alongside a positive coefficient for green transformation, implying that financing constraints attenuate—while their alleviation amplifies—the productivity benefits of green transformation. This aligns with the intuition that eased constraints free up capital for innovation.

Tax incentives. Following Li [77], we proxy tax incentives (Taxp) by the ratio of tax refunds to total tax payments. Column (2) demonstrates a positive interaction term, signifying that tax incentives strengthen the productivity gains from green transformation. This mechanism operates through cost reduction and improved financing conditions, which incentivize R&D expenditure.

Environmental subsidies. Adopting Yu [78]’s approach, we manually compile subsidies tied to environmental keywords (e.g., “green,” “sustainability”) from financial reports and adjust for firm size. Column (3) shows a positive interaction, underscoring subsidies’ role in easing financial pressures and signaling policy commitment, thus accelerating productivity growth.

4.3.2. The Moderating Role of Collaborative Innovation

This study examines the role of collaborative innovation by employing the logarithmic count of jointly filed patents as a proxy for inter-firm knowledge co-creation. Table 8 reveals a statistically significant positive coefficient (1% level) for the interaction term, suggesting that collaborative innovation amplifies the positive impact of green and low-carbon transition on new-quality productivity. Mechanistically, the transition fosters greater inter-organizational cooperation in innovation, which in turn accelerates the diffusion and commercialization of advanced technologies, driving productivity gains.

4.3.3. The Optimization Effect of Resource Allocation

Building upon the analytical framework of Chen [79], we decompose firm-level resource-allocation efficiency into two distinct dimensions: labor-allocation efficiency and capital-allocation efficiency. We then conceptualize the simultaneous improvement in both dimensions as the resource-allocation-optimization effect. This study investigates whether such optimization, driven by green and low-carbon transition, fosters the development and enhancement of new quality productive forces.

Labor-Allocation Efficiency. To measure labor-allocation efficiency, we estimate the level of “excess employees” by drawing on the methodology of Zeng and Chen (2006). The logic is that a firm’s normal employment level is determined by its fundamental characteristics, and any deviation from this level can be interpreted as a form of labor misallocation.

$${\displaystyle \frac{{\mathrm{Employees}}_{it}}{{\mathrm{Assets}}_{it}}}={\beta }_0+{\beta }_1{\mathrm{Size}}_{i,t-1}+{\beta }_2{\mathrm{Capital}}_{i,t-1}+{\beta }_3{\mathrm{Growth}}_{i,t-1}+{\mu }_i+{\lambda }_t+{\epsilon }_{it}$$

(7)

where Employees is the number of employees, Assets is total assets, Size is the natural logarithm of total assets, Capital is the ratio of fixed assets to total assets, and Growth is the sales growth rate.

Second, we calculate the excess employees ($E{l}_{i}t$). We use the estimated coefficients from the equation above to predict the normal employment ratio. We then multiply this predicted ratio by the firm’s total assets to obtain the predicted number of normal employees. Finally, “excess employees” is the actual number of employees minus this predicted normal number. A higher value of this proxy (El) signifies greater over-staffing and thus lower labor-allocation efficiency. Column (5) of Table 8 reveals a statistically negative coefficient for the interaction term, implying that green transition mitigates labor misallocation, thereby elevating new quality productive forces. Mechanistically, firms undergoing green restructuring streamline their workforce deployment, reducing redundancies and reallocating labor toward innovation-intensive activities. This reallocation enhances marginal productivity, aligning with endogenous growth theory.

Capital-Allocation Efficiency. Adopting the methodology of Richardson [80], we estimate firm-specific investment inefficiency as the residual from a model of optimal investment. Excess investment (Ineff) captures allocative distortions, with larger values indicating severe inefficiency. Column (6) demonstrates that the interaction is negative at the 5% significance level, underscoring the role of green transition in curbing capital misallocation. The underlying mechanism involves disciplined capital expenditure: firms prioritize high-return projects (e.g., clean technology R&D) over unproductive investments, thus bolstering new quality productive forces through improved dynamic efficiency.

4.4. Heterogeneity Analysis

4.4.1. Ownership-Based Heterogeneity

Firms under distinct ownership structures differ substantially in governance mechanisms, strategic priorities and resource allocation [81], potentially inducing heterogeneous responses to green and low-carbon (GLC) transition in terms of new-quality productivity. We estimate Equation (1) separately for state-owned enterprises (SOEs), private firms and foreign-invested enterprises (FIEs).

Table 9. Heterogeneity analysis by firm characteristics.

VARIABLES	Firm Ownership		Carbon Emission Intensity
	(1)	(2)	(3)	(4)
	SOEs	Non-SOEs	High-Emission	Low-Emission
GLCT	1.754 **	2.382 ***	4.526 ***	1.045
	(0.841)	(0.832)	(1.095)	(0.675)
Controls	YES	YES	YES	YES
Firm FE	YES	YES	YES	YES
Year FE	YES	YES	YES	YES
Observations	11,378	22,333	11,542	21,769
R2	0.790	0.755	0.797	0.747

Notes: This table reports heterogeneity results based on firm characteristics. The dependent variable is NQPF. Columns (1)–(2) split the sample by ownership type. Columns (3)–(4) split the sample by whether the firm is in a high- or low-emission industry. Standard errors are in parentheses. *** p < 0.01, ** p < 0.05.

summarizes the results.

Firms under distinct ownership structures differ substantially in governance mechanisms, strategic priorities and resource allocation [81], potentially inducing heterogeneous responses to the green and low-carbon transition (GLCT) in terms of new-quality productivity. We estimate Equation (1) separately for state-owned enterprises (SOEs) and non-state-owned enterprises (non-SOEs). Table 9 summarizes the results.

For SOEs, the coefficient on GLCT ($GLCT$) is 1.754 and statistically significant at the 5% level. Private firms exhibit a larger coefficient ($2.508$), also significant at 5%. In contrast, FIEs show an insignificant coefficient ($-1.913; p>0.1$).

For SOEs, the coefficient on GLCT is 1.754 and statistically significant at the 5% level. Non-SOEs exhibit a larger coefficient (2.382), significant at the 1% level.

These results imply that GLCT robustly enhances new-quality productivity in both SOEs and non-SOEs, with a stronger effect observed in the latter. The impact is also more pronounced in high-emission firms. This heterogeneity likely reflects institutional and strategic differences: Non-SOEs, due to their market agility and adaptive capabilities, are better positioned to leverage GLCT-driven opportunities and translate them into productivity gains. SOEs, although benefiting from stronger policy alignment and resource access, may face structural inefficiencies in innovation and market responsiveness.

4.4.2. Heterogeneity in Firm-Level Carbon Emission Intensity

To examine potential differential effects of green and low-carbon transition across firms, we stratify the sample into high- and low-carbon-intensity subgroups based on the median carbon intensity per unit size (defined as total emissions scaled by firm size). Regression estimates in Panel A of Table 9 (columns 4–5) reveal a statistically significant coefficient of 4.526 ($p<0.01$) for the transition variable ($GLCT$) in the high-intensity subgroup, compared to an insignificant 1.045 ($p>0.1$) in the low-intensity subgroup. This asymmetry implies that the productivity gains from decarbonization are concentrated among high-emission firms.

The mechanism likely stems from regulatory stringency: High-intensity firms face binding emission constraints, compelling them to innovate via green R&D, which amplifies productivity spillovers. Conversely, low-intensity firms operate closer to the efficiency frontier, where incremental abatement entails higher marginal costs and weaker regulatory impetus, attenuating the transition’s productivity effects.

4.4.3. Sectoral Heterogeneity

The impact of Green and Low-Carbon Transition (GLCT) on New Quality Productive Forces (NQPF) is unlikely to be uniform across all industries, as their strategic positions, regulatory environments and core production structures differ significantly. To provide a comprehensive picture, we investigate this heterogeneity from three parallel perspectives: (1) a firm’s alignment with national green development strategy, (2) its exposure to environmental regulation and (3) its core factor intensity.

Heterogeneity based on Strategic Policy Alignment.

As a preliminary test, we first examine whether the effect of GLCT is more pronounced in industries that are officially designated as central to China’s green development strategy. We classify our sample into Energy-Saving, Environmental Protection and Clean Industries and “Non-Clean Industries” based on the official Strategic Emerging Industry Classification. The results, presented in

Table 10. Heterogeneity analysis by industry characteristics.

VARIABLES	Strategic Alignment		Factor Intensity			Regulatory Pressure
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
	Clean Industries	Non-Clean Industries	Tech-Intensive	Labor-Intensive	Capital-Intensive	Heavily Polluting	Non-Heavily Polluting
GLCT	2.388 ***	2.258	2.923 ***	2.102 *	2.464 **	2.608 **	2.631 ***
	(0.666)	(1.745)	(1.105)	(1.205)	(1.008)	(1.016)	(0.807)
Controls	YES	YES	YES	YES	YES	YES	YES
Firm FE	YES	YES	YES	YES	YES	YES	YES
Year FE	YES	YES	YES	YES	YES	YES	YES
Observations	26,280	7469	15,732	11,548	6128	9518	24,237
R2	0.759	0.753	0.758	0.758	0.739	0.738	0.769

Notes: This table reports heterogeneity results based on industry characteristics. The dependent variable is NQPF. Columns (1)–(2) test based on strategic alignment (Clean vs. Non-Clean). Columns (3)–(5) test based on factor intensity. Columns (6)–(7) test based on environmental regulatory pressure (Heavily Polluting vs. Non-Heavily Polluting). Standard errors are in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1.

, Columns (1) and (2), support this hypothesis. For firms in the “Energy-Saving, Environmental Protection and Clean Industries” group, the coefficient on the GLCT variable is 2.388 and highly significant at the 1% level. In contrast, the coefficient for the “Non-Clean Industries” group is statistically insignificant. This suggests that the productivity-enhancing effects of the green transition are particularly concentrated in sectors at the forefront of the green economy. The potential reason for this is that these firms likely benefit from a dual advantage: their core business models are inherently aligned with GLCT goals, which allows them to more effectively leverage stronger policy support (e.g., green finance and subsidies) to transform green initiatives into tangible productivity gains.

Heterogeneity based on Factor Intensity.

Finally, to investigate whether the impact of GLCT varies with industries’ production factor structures, we classify firms into three categories: technology-intensive, labor-intensive and capital-intensive industries. The regression results for these subsamples are presented in Table 10, Columns (3), (4) and (5). The findings reveal significant heterogeneity. The positive effect of GLCT on NQPF is most pronounced in “technology-intensive industries,” with a coefficient on GLCT of 2.923, significant at the 1% level. For “capital-intensive industries,” the coefficient is also positive and significant at 2.464 (p < 0.05). In contrast, the effect is weaker and only marginally significant for “labor-intensive industries” (coefficient = 2.102, p < 0.1). The potential reasons for this distinct pattern lie in the different ways these industries interact with green transition policies. For technology-intensive firms, the effect is most pronounced because they can readily leverage GLCT mandates as an opportunity to innovate and develop the cutting-edge green technologies that are central to NQPF. In capital-intensive firms, the significant effect is likely driven by the high costs associated with their existing capital stock; stricter regulations compel them to make substantial investments in cleaner and more efficient equipment, which in turn leads to direct productivity upgrades. In contrast, the weakest effect is observed in labor-intensive industries, as their core production and cost structures are less sensitive to the technology- and energy-focused incentives of GLCT. Consequently, their path to compliance may involve less innovation and more direct, non-productive costs.

Heterogeneity based on Environmental Regulatory Pressure.

Next, to examine how the effect of GLCT varies with the level of environmental regulatory pressure, we classify firms into “heavily polluting” and “non-heavily polluting” groups, based on the official classification of 16 heavily polluting sectors by Chinese environmental authorities. As shown in Table 10, Columns (6) and (7), the effect of GLCT on NQPF is statistically significant in both groups: for heavily polluting industries, the coefficient is 2.608 (p < 0.05), while for non-heavily polluting industries, the coefficient is slightly higher at 2.631 (p < 0.01). The comparable magnitude and significance across both subsamples suggest that GLCT drives productivity improvements regardless of regulatory pressure level. For heavily polluting sectors, stringent environmental regulations may catalyze process upgrades and green innovation—supporting a regulation-induced innovation mechanism in line with the Porter Hypothesis. Conversely, for non-heavily polluting industries, the positive effect likely stems from voluntary strategic behavior, where firms proactively invest in green initiatives to meet investor expectations, capture ESG premiums or anticipate future regulation. These results underscore that both regulatory compliance and market-driven incentives can serve as effective channels for green transition to enhance firm-level productivity.

4.4.4. Regional Heterogeneity

Given the significant disparities in economic development, industrial structure and resource endowments across China, we expect the impact of GLCT on NQPF to exhibit regional heterogeneity (

Table 11. Heterogeneity analysis by region.

VARIABLES	(1)	(2)	(3)	(4)
	Eastern	Central	Western	Northeastern
GLCT	2.500 **	3.088 *	2.500 **	4.223
	(1.189)	(1.746)	(1.189)	(3.091)
Controls	YES	YES	YES	YES
Firm FE	YES	YES	YES	YES
Year FE	YES	YES	YES	YES
Observations	4338	4627	4338	1327
R2	0.754	0.763	0.754	0.753

Notes: This table reports heterogeneity results based on the firm’s registered region. The dependent variable is NQPF. The regional classification follows the official definition from China’s State Council. Standard errors are in parentheses. ** p < 0.05, * p < 0.1.

). Following the official economic-geographical classification from China’s State Council, we partition our sample into four major regions: Eastern, Central, Western and Northeastern China. The specific provinces included in each region are detailed in Appendix A.5. We then estimate our baseline model for each of these four subsamples. The estimated coefficient for GreLc is 2.500 (significant at the 5% level) in Eastern China, 3.088 (significant at the 10% level) in Central China, 2.500 (significant at the 5% level) in Western China and 4.223 (statistically insignificant) in Northeastern China. These findings reveal significant cross-regional variation in the productivity-enhancing effects of green transition.

The results indicate that firms across most regions benefit from green and low-carbon transition, albeit to varying degrees. In eastern China, where economic development and industrial sophistication are more advanced, the productivity gains from are particularly pronounced, though potential diminishing marginal returns may temper the effect. Western China, despite its comparatively weaker economic infrastructure, exhibits substantial productivity improvements, likely attributable to greater developmental headroom and technological spillovers from the more industrialized east. Meanwhile, central China—a region undergoing rapid growth, industrial relocation and policy-driven support—experiences a positive yet less statistically robust effect, possibly due to intra-regional heterogeneity in development trajectories. The insignificant result for Northeastern China, despite having the largest coefficient in magnitude, may suggest that while the potential for GLCT-driven growth is high, it is currently hindered by deep-rooted structural barriers inherent to its traditional heavy-industry base.

5. Discussion

5.1. Summary of Key Findings

In the context of China’s ambitious carbon peak and neutrality objectives, understanding the economic consequences of corporate GLCT is paramount. This study contributes by theoretically outlining and empirically investigating the mechanisms through which firm-level GLCT influences NQPF. We develop novel metrics for both constructs and utilize a comprehensive panel dataset of Chinese A-share listed firms (2012–2022) to estimate the causal impact of GLCT on NQPF.

Our empirical analysis yields three main findings:

(1) Corporate GLCT exerts a statistically significant and economically meaningful positive effect on NQPF. This result is robust across various specifications, including instrumental variable approaches designed to address potential endogeneity concerns. Specifically, our instrumental variable estimations confirm that policy-driven GLCT initiatives significantly improve firms’ productivity by encouraging proactive innovation and enhancing operational efficiency, validating the causal relationship posited by the Porter Hypothesis.

(2) The positive impact of GLCT on NQPF appears to operate through at least three mediating channels: improved firm financing conditions, enhanced collaborative innovation activities and increased resource-allocation efficiency. Empirical tests explicitly demonstrate that firms undertaking GLCT benefit significantly from reduced financing constraints, likely owing to enhanced policy credibility and market confidence. Furthermore, our analysis highlights that collaborative innovation, stimulated by GLCT, substantially accelerates technological breakthroughs and knowledge diffusion, providing direct empirical evidence that the collaborative innovation mechanism is a key pathway for productivity enhancement. Lastly, improved resource-allocation efficiency emerges clearly in our findings, underscoring how GLCT directs capital and labor toward higher-value, sustainability-oriented productive activities.

(3) The effect of GLCT on NQPF exhibits significant heterogeneity. The positive impact is notably stronger for state-owned enterprises (SOEs), firms in high-emission sectors, firms operating within designated energy-saving and environmental protection industries, technology-intensive and capital-intensive firms compared to labor-intensive firms and non-heavily polluting industries, and firms located in the more economically developed eastern regions of China. These heterogeneous effects suggest that policy interventions and market conditions significantly modulate the effectiveness of GLCT, with SOEs and high-emission firms benefiting disproportionately due to their greater exposure to regulatory frameworks and policy support. Additionally, firms in eastern regions likely capitalize on advanced technological infrastructure and superior market conditions to better leverage GLCT into productivity gains, highlighting the need for tailored policy strategies that consider regional disparities.

5.2. Theoretical and Practical Implications

Our results broadly corroborate prior evidence that green transition policies generate substantial productivity enhancements [62]. Notably, the mitigating role of financing constraints aligns with findings in Yu [46], which demonstrate that green financing accessibility augments the innovation spillovers of environmental regulations. However, our identification of more persistent lagged effects contrasts with the analysis in Chen [48], which predominantly captured short-run impacts—implying that the economic returns to Green Low-Carbon Transition (GLCT) may materialize over extended horizons. Additionally, our sectoral decomposition reveals heterogeneous treatment effects, with technology- and capital-intensive industries exhibiting disproportionately larger gains. This finding extends the clean/non-clean industry dichotomy established in Ding [72] by elucidating how factor intensity mediates the productivity dividends of sustainability transitions. Collectively, these comparative analyses delineate this study’s marginal contributions while underscoring the necessity of incorporating temporal and structural contingencies in evaluating corporate sustainability’s economic efficacy.

These findings offer several considerations for policy aimed at simultaneously fostering sustainable development and enhancing economic productivity:

1. Strengthening Green Finance Mechanisms: Our results highlight the importance of the financing channel. This suggests value in developing a multi-tiered green finance system employing a diversified toolkit (e.g., green bonds, targeted credit policies, investment funds, potentially fiscal incentives) to alleviate financing constraints associated with GLCT. This recommendation is empirically reinforced by our mechanism tests, which demonstrate that easing financing constraints significantly amplifies the productivity benefits derived from GLCT. Particular attention might be warranted for high-carbon emitting sectors and potentially small and medium-sized enterprises (SMEs), where tailored financial instruments could facilitate their transition towards lower-carbon operations and NQPF development.

2. Fostering a Collaborative Innovation Ecosystem: The identified role of collaborative innovation suggests benefits from policies that promote knowledge sharing and joint R&D related to green technologies. This could involve facilitating platforms for collaboration between firms, universities and research institutions, potentially across regions and industries. The empirical confirmation of collaborative innovation as a key mediator indicates that strengthening intellectual property rights (IPR) protection related to green innovations and ensuring effective mechanisms for technology diffusion would further incentivize the development and adoption of productivity-enhancing sustainable technologies.

3. Enhancing Factor Market Efficiency: The finding that GLCT improves resource-allocation efficiency points towards the importance of complementary policies that enhance factor market flexibility. Our detailed empirical analysis reveals clear productivity gains associated with reductions in labor redundancy and excess capital investment, underscoring the necessity of policies aimed at developing human capital skills relevant to green industries and facilitating more efficient resource allocation.

4. Designing Differentiated and Targeted Transition Policies: The significant heterogeneity in the impact of GLCT underscores the potential need for tailored policy interventions. Differentiated approaches based on firm characteristics (ownership, sector, emission intensity) and location may be more effective. Our empirical findings explicitly highlight that targeted support is particularly effective for SOEs and firms in strategic sectors or regions facing stricter environmental regulations. Policymakers should thus leverage differentiated regulatory and supportive measures to maximize the productivity dividends from GLCT across diverse economic contexts.

5.3. Limitations and Future Research

Despite its strengths, our study is not without limitations. The reliance on Chinese firm-level data may constrain the generalizability of our findings to other institutional or economic contexts. Moreover, while this study focuses primarily on the net productivity effects of GLCT, it does not explicitly disentangle potential short-term adjustment costs, such as increased financial burdens, technology lock-in risks or operational frictions during the transition process. Although our inclusion of lagged effects provides some insight into dynamic patterns, future research could more precisely quantify the timing and magnitude of these short-term trade-offs relative to longer-term gains. Additionally, while our text-based GLCT index is innovative, it may not fully capture the multifaceted nature of green transition, potentially underrepresenting unobservable efforts or outcomes. Future research could address these gaps by exploring alternative GLCT measures, extending the analysis to other countries or examining the long-term dynamics of green transition effects using extended time horizons. Additionally, future research could examine whether foreign-invested enterprises respond differently to GLCT, for example due to parent-firm strategies or differential exposure to domestic regulation. Investigating the role of firm-level capabilities or specific policy instruments in mediating the GLCT-NQPF relationship could further enrich this agenda. Finally, due to limited sample size and data constraints, our analysis does not fully explore the specific mechanisms shaping the response of foreign-invested enterprises to GLCT. Future research could address this by incorporating richer firm-level and cross-border variables.

In conclusion, this study advances the growing literature on the economic consequences of environmental policies by demonstrating that green and low-carbon transition serves as a catalyst for enhancing new quality productive forces in firms. As global economies confront the twin imperatives of environmental sustainability and economic competitiveness, our findings offer actionable insights for policymakers and firm managers alike, highlighting the transformative potential of aligning green strategies with productivity goals.