Promoting or inhibiting? The impact of supporting policy for resource-exhausted cities on corporate green technology innovation: evidence from China

Abstract

Promoting green technology innovation (GTI) is a critical pathway to concurrently address economic development and environmental conservation. Thus, could the supporting policies implemented in resource-exhausted cities foster local firms’ GTI? To investigate this question, we collect the data of Chinese listed manufacturing companies from 2003 to 2018, take the “supporting policy for resource-exhausted cities (SPREC)” as a quasi-natural experiment, and apply the DID methodology with multiple time periods to examine how the SPREC effects firms’ GTI. Our findings reveal a significant inhibitory effect of SPREC on firms’ GTI, and this effect is achieved through mechanisms that involve the alleviation of financing constraints, an increase in the proportion of fixed assets, the escalation of rent-seeking costs, and the suppression of R&D expenditures. In the heterogeneity analysis, we discover that the SPREC has a more obvious inhibitive role in GTI of companies that are state-owned, located in the mid-western area, or under a lower degree of marketization. However, it shows a considerable promotion impact on the GTI of companies that are under a higher level of marketization. This study enhances our comprehension of SPREC’s impacts and offers empirical evidence advocating for the advancement of green and sustainable development in resource-exhausted cities.

1 Introduction

In this world, there are many famous resource-based cities or regions, such as Ruhr Industrial Base in Germany, Pittsburgh in the USA, Lorraine in France, etc. These resource-based cities are crucial to the swift economic growth of their nations due to huge profits and taxes from processing raw materials like coal mines and woods (Li et al., 2013). However, with the gradual increase in extraction and excessive consumption of resources, some resource-based cities inevitably step into the stage of resource exhaustion. A city is characterized as a resource-exhausted city when the accumulated resource extraction amounts to more than 70% of the entire extractable amount (Xiao et al., 2021; Zhang et al., 2018).

From the Millennium Development Goals in 2000 to the Sustainable Development Goals in 2015, the United Nations has consistently emphasized the safeguarding of non-renewable resources and the promotion of environmental sustainability (Biermann et al., 2017; Ilcan & Phillips, 2010; Sachs, 2012). An increasing number of nations and organizations have recognized the significance of sustainable development and actively assumed responsibility. Noteworthy instances encompass international initiatives such as the Mining Mineral and Sustainable Development program (Azapagic, 2004; Fitzpatrick et al., 2011), Sustainable Minerals Roundtable (SMR) gatherings in the USA (Gomes et al., 2015; Shields et al., 2003), Canada’s Minerals and Metals projects (Bond & Quinlan, 2018), and the European Industrial Minerals Association program in Europe (Regueiro & Alonso-Jimenez, 2021; Tato Diogo, 2020). Concurrently, certain resource-dependent cities, supported by diverse policies, have undertaken self-rescue efforts and successfully transitioned toward sustainability. Illustrative examples include Lorraine, France’s industrial transformation strategy, the adoption of the “Smoke Control Act” in Pittsburgh, USA, the implementation of the “Future Movement for Mining and Metallurgy” in Germany’s Ruhr region, and Japan’s “Environmental Technology Industrialization Project” in Kitakyushu (He et al., 2017). These case studies lead us to the conclusion that both government intervention and market forces serve as pivotal drivers in the transformation of resource-dependent cities (Ruan et al., 2020). Nevertheless, it is undeniable that numerous resource-exhausted cities continue to grapple with significant challenges, particularly in emerging economies and other developing nations (Dou et al., 2023; Li et al., 2013).

These challenges can be broadly categorized into two main areas: on the one hand, because of the homogeneous business structure and poor production technology, these cities’ economic development is severely limited by the exhaustion of resources, which subsequently gives rise to pressing issues, including unemployment, poverty, and crime, etc. On the other hand, due to the over-exploited and abused resources and crude corporate production patterns, the ecological environments have been destroyed which hinders the green development of cities (Dong et al., 2007; Dou et al., 2023; Yang et al., 2021; Yu et al., 2022). Thus, it is urgent for resource-exhausted cities to find a path to attend simultaneously to both economic growth and environmental protection. So how to achieve this path? Actually, numerous research literature has shown that green technology innovation (GTI) is the crucial remedy to unlocking the conundrum of harmonious coexistence between economic growth and environmental conservation (Shen et al., 2021), which is achieved by improving energy efficiency, adjusting energy structure, and allocating resources rationally (Jin et al., 2022; Xu & Cui, 2020). Therefore, it is essential for resource-exhausted cities to improve GTI.

Innovation efforts focused on advancing green technologies, which include preservation of energy and lowering of emissions, cleaner production, and the adoption of renewable energy sources, are referred to as “green technology innovation (GTI)” (Wang et al., 2019a). In recent years, the data demonstrate that there is still persistent problem with insufficient GTI, despite the policies relating to it have been promulgated constantly. For instance, in China, less than 10% of the total patent applications are for green patents, which is a comparatively low percentage. To explore the reasons, the GTI is marked by large expenditures, elevated risk, long cycle, and also has strong externalities. Therefore, market forces alone cannot address the issue of inadequate GTI, and government forces are required to intervene, guide and stimulate. So are there any special government policies for resource-exhausted cities? Could these government policy interventions promote their GTI?

Actually, to save resource-exhausted cities that are gradually going to the “valley of death,” many national governments have already extended their “helping hands.” Taking China as an illustration, in 2007, the State Council promulgated “supporting policy for resource-exhausted cities (SPREC),” which initially proposed the establishment of transfer payments for resource-exhausted cities.¹ Subsequently, the lists of resource-exhausted cities were released in three batches in 2008, 2009, and 2011 separately, including 69 cities in total.² What needs to be explained here is that although the earliest list of resource—depleted cities was released in 2008, government documents and news reports indicate that the central government commenced transfer payments to these cities as early as 2007 (Ai et al., 2023; Lu et al., 2022; Wu et al., 2023). Up to 2018, the cumulative amount of fiscal transfer to support these cities has reached about 160 billion yuan.³ Figure 1 illustrates the evolutionary progression of SPREC, delineated along a timeline. So, for GTI in resource-exhausted cities, is the support policy “a timely help” or just like “closing the stable door after the horse has bolted”?

Fig. 1

Timeline of SPREC implementation

Upon reviewing the existing literature on SPREC, we observe a lack of consensus regarding its effects: Some scholars argue that the SPREC promotes the transformation of manufacturing (Yu & Ma, 2022) and high-quality urban development (Zhang et al., 2022). Conversely, others argue that it inhibits regional innovation (Lu et al., 2022) and slows down urban economic growth (Zhang et al., 2019). Thus, the usefulness of the SPREC is still a question to be verified. In addition, we discover that the majority of the existing study on the SPREC’s effects is concentrated on the city level, but there is little literature to study its impact on enterprises, which is a more microscopic perspective. As enterprises are the core strength of a city’s development and the hope of countless families, we believe that it is vital to investigate the consequences of the SPREC from the perspective of enterprises.

Therefore, we chose the data of Chinese listed manufacturing companies from 2003 to 2018 as our sample, take SPREC as a quasi-natural experiment, apply the DID with multiple time periods to explore how the SPREC effects firms’ GTI.

Why do we choose China as our research context? There are several main reasons: First, China is a resourceful country with 118 resource-based cities, constituting approximately 20% of the nation’s total cities. These cities have contributed significantly to Chinese economic development (Wang & Guo, 2012). Nowadays, with more than half of the resource-based cities exhausted, China urgently needs to find a way to help resource-exhausted carry out green transformation and sustainable growth (Lu et al., 2022; Yang et al., 2021). Second, as a major emerging economy in the world, China has a greater responsibility and challenge to balance environmental protection and green development while pursuing high-quality economic growth. Third, our study, which is conducted in China, could provide empirical evidence for the green and sustainable growth of resource-depleted cities. This experience could become an important part of “Good Governance” and be extended to more developing countries, and even make a pivotal role in the transformation of resource-based cities globally.

The main contributions are as follows: Firstly, our research augments the current literature concerning the micro-level impacts of the SPREC. Prior studies have predominantly centered on the effects of the SPREC at the city or regional levels. By investigating the impact of the supporting policy on firms’ GTI, we provide valuable insights from a microscopic perspective. In addition, SPREC is one of the instruments of government intervention, and there has been a constant academic debate around the “visible hand” of government, so our paper also enriches research on government intervention. Secondly, in line with the government’s efforts to achieve carbon neutrality, our paper considers the importance of environmental preservation and sustainable growth in the context of resource-exhausted cities, which face limitations in pursuing green development. Thirdly, we employ a structural equation model and Bootstrap test in our mechanism analysis to examine multiple mediating factors, thereby offering a more profound and detailed understanding of the mechanisms through which the SPREC influences GTI. Lastly, by employing the DID method with multiple time periods, we address potential endogeneity issues, account for variable interactions, and better align our study design with the real-world implementation of the policy pilot, thus enhancing the robustness of our results.

Following is the structure of the remaining parts: The literature review is presented in Sect. 2. Sect. 3, the theoretical analysis and hypotheses are covered. The research design is discussed in Sect. 4, followed by Sect. 5 analysis of the empirical results, Sect. 6 further analysis, and Sect. 7 conclusion and policy recommendations.

2 Literature review

2.1 SPREC

While traditional economic theory generally considers good natural resource endowment as the basis for the start of industrialization and the driver of growth in economy (Shao et al., 2020), Auty (1993) first proposes the famous paradox of the “resource curse,” which means that abundant resources of nature are a limitation to regional economic growth, and this argument has attracted widespread attention and debates in academic circles. Subsequently, numerous studies have revealed that the impacts of “resource curse” are prevalent in several countries worldwide (Papyrakis & Gerlagh, 2004; Sachs & Warner, 1995; Shao & Yang, 2010). There have been many different perspectives to explain this paradox, summarized as the Dutch disease effect (Corden & Neary, 1982; Shao et al., 2020), the crowding-out effects of human capital and technology innovation (Papyrakis & Gerlagh, 2003; Sachs & Warner, 2001), the institutional weakness effect (Leite & Weidmann, 1999). Based on previous studies, Shao and Yang (2010) argue that while abundant natural resources can bolster regional economic development, overdependence on these resources is the culprit of the “resource curse” effect.

Up to now, research on SPREC has predominantly focused on its effects, resulting in varied evaluations due to differing research perspectives. For example, several studies highlight the positive impact of SPREC on high-quality development (Yu & Ma, 2022; Zhang et al., 2022), Yu and Ma (2022) posit that the SPREC facilitates the upgrading of the manufacturing industry by fine-tuning production and relieving the tax burden on high-tech enterprises. In their investigation, Yu et al. (2022) contend that SPREC encourages urban areas to diminish their reliance on finite resources, consequently enhancing local energy efficiency. Furthermore, Sun et al. (2020) observe that SPREC, through strategies such as bolstering marketization, expanding openness, and fostering alternative industries, contributes to increased local employment rates and per capita GDP. Moreover, both Li and Wang (2022) and Ai et al. (2023) substantiate that SPREC significantly diminishes urban carbon emissions by propelling technological advancement. However, certain studies cast doubt on the feasibility of SPREC. Zhang et al. (2019) identify a potential “incentive trap” associated with SPREC, which could decelerate urban economic growth. Lu et al. (2022) find that SPREC hinders regional innovation, attributing this inhibitory effect to three primary factors: the increase in zombie enterprises, shifts in government subsidy preferences, and distortions in the land market. Additionally, Sun et al. (2020) contend that SPREC does not stimulate the development of the service industry.

In fact, the SPREC can be considered as one of the instruments of government intervention. Government intervention, which is regarded as a remedy to market failure, is one of the key predictors of innovation (Wang, 2018). There are two primary contrasting viewpoints surrounding the impacts of government intervention on corporate technological innovation: according to the first point of viewpoint, government intervention could encourage firm s’ innovation (Guellec & Potterie, 2003; Howell, 2017; Lee & Cin, 2010; Wang, 2018). Scholars who hold this view think that the government’s “visible hand” can intervene in the economy to compensate for the failings of the market’s “invisible hand” (Rao, 2016), mitigate the risk associated with companies’ R&D expenditures (Guo et al., 2014), and compensate for the loss of revenue due to the externalities of technological innovation (Kang & Park, 2012). Through such interventions, the government could effectively promote corporate innovation. According to the second viewpoint, government intervention actually crowds out enterprises’ innovation (Busom, 2000; Li et al., 2020; Marino et al., 2016). Scholars holding this view argue that government intervention may trigger a “moral hazard” that makes firms use subsidies for activities other than innovation (Wu et al., 2022). In addition, government funding of R&D may increase the need for R&D resources in a whole market, which may raise the cost of R&D and thus discourages firms from innovating (Marino et al., 2016).

2.2 Green technology innovation

Another relevant literature to this paper focuses on the determinants of corporate GTI. Previous research has rigorously investigated both internal and external factors that influence corporate GTI (Takalo & Tooranloo, 2021). Concerning internal factors, entrepreneurs with environmental consciousness (Rui & Lu, 2021), academically qualified executives (He et al., 2021), green organizational culture (Gürlek & Tuna, 2018), and the practice of corporate social responsibility (Yuan & Cao, 2022) all have the potential to stimulate GTI within a company. Conversely, the presence of quality management (Li et al., 2018), ineffective corporate governance (Amore & Bennedsen, 2016) and an excessive concentration of power (Gao et al., 2022) can hinder a company’s GTI. Turning to external factors, elements such as environmental regulations (ER) (Borsatto & Bazani, 2021), digital financial solutions (Liu et al., 2022), financing options (Xiang et al., 2022), pressures from stakeholders (Rui & Lu, 2021), and collaborative ventures between organizations (Zhu et al., 2017) all possess the capacity to exert an influence on a company’s GTI. Among these factors, the influence of ER on GTI remains a subject of debate, with three primary viewpoints. Firstly, the renowned “Porter Hypothesis” (Porter & van der Linde, 1995) postulates that ER may compel companies to engage in green innovation to enhance competitiveness. This, in turn, helps offset the additional costs incurred due to ER, leading to an innovation compensation effect. Subsequent studies by Rubashkina et al. (2015), Song et al. (2019), and Zhong and Peng (2022) have provided support for this hypothesis. Secondly, neoclassical economic theory suggests that ER might escalate production costs and diminish financial flexibility, thereby inhibiting GTI (Du et al., 2021; Lanoie et al., 2008; Li et al., 2021). Thirdly, some scholars propose a nonlinear, inverted U-shaped relationship between ER and GTI (Wang et al., 2019c; Pan et al., 2021; Song et al., 2020b).

2.3 Gap in literature

By combining the literature, we can see that although the relevant research has been fruitful, there are still the following shortcomings: First, whether it is the SPREC or more generally government intervention, their impact on firms’ technology innovation is controversial and no definite conclusions have been reached. In addition, although there are many studies on how government intervention affects firms’ technology innovation, a scarcity of research exists in a sample of especially resource-exhausted cities. Compared with other cities, resource-exhausted cities have their particularity, which is that the local resources have been exhausted, economic development and environmental conditions are both worrying, so at this time whether the SPEC is “a timely help” or just like “closing the stable door after the horse has bolted” is worthy of our in-depth discussion. Next, we find that the majority of the existing research on the effects of the SPREC is concentrated at the city level, but there is little literature to study its impact on enterprises, which is a more microscopic perspective. Finally, the majority of current research on resource-depleted cities ultimately strives to support the urban economies, but less attention is paid to their green development. However, the resource-depleted cities rely more heavily on local resources and environments for growth compared to other cities, but years of exploitation and abuse have seriously harmed the local environment, so our study aims to address the equilibrium between economic quality and sustainable development in these cities.

3 Theoretical analysis and hypothesis

Through literature reading and theoretical derivation, we argue that the impact of the SPREC on GTI is unclear, with two possibilities of promoting or inhibiting effects.

On the one hand, we think that the SPREC’s impact might be positive.

Firstly, externalities, either positive or negative, occur when the production or consumption of one person or entity affect the existence and well-being of another (Dahlman, 1979). GTI exemplifies a classic case of “positive externality” linked to knowledge spillovers. Specifically, when certain companies implement GTI and introduce it to the market, a portion or the entirety of the innovative knowledge becomes public, leading to the replication, imitation, or even advancement of new technologies. In such scenarios, other participants in the market may benefit from this wave of GTI, while the companies that invested in it might not fully realize the fruits of their labor. Despite the overall societal benefits, where spillover advantages accrue to the public rather than the companies implementing GTI, these innovators may grow disheartened as they bear the entire innovation cost, possibly reducing their investments in GTI (Acemoglu et al., 2012). To address the issue of “market failure,” government subsidies and interventions can compensate for the income losses incurred by green technological innovators due to positive externalities (Rao, 2016). The central fiscal transfer payments provided by SPREC can, to some extent, directly alleviate the financial constraints faced by resource-exhausted cities, consequently increasing local government expenditures on technology and research and development subsidies for enterprises (Yu & Ma, 2022). Additionally, during the implementation of SPREC, resource-exhausted cities stimulate the growth of high-end manufacturing and eco-friendly enterprises through tax incentives, reducing the tax burden on such businesses (Song et al., 2020a). Therefore, it is evident that SPREC offers direct government subsidies and tax incentives to enterprises, boosting their research and development investments and effectively counteracting the positive externalities associated with technological innovation, thereby encouraging businesses to engage in GTI (Mamuneas & Ishaq Nadiri, 1996).

Secondly, GTI exhibits distinct characteristics such as substantial investment requirements, heightened risks, and protracted research and development cycles. These factors underline the imperative need for companies to secure long-term and stable financial support (Huang et al., 2019). Nevertheless, owing to the imperfections within China’s capital market, companies frequently grapple with agency issues and information asymmetry. This predicament results in escalated external financing expenses for companies and the looming threat of financing constraints (Lin & Ma, 2022; Liu et al., 2021). Research conducted by Yu et al. (2021) underscores that financing constraints can lead to a shortage of funds allocated to research and development activities, impeding their capacity to engage in GTI, and thereby influencing their operational capabilities and developmental trajectories. Leveraging signaling theory, SPREC, functioning as a conspicuous public signal within the market, inherently carries the government’s endorsement. It possesses the inherent capability to transmit affirmative signals to external investors and creditors, thereby reinforcing their confidence in making investments in these companies (Jiang et al., 2023; Tong et al., 2022). The dissemination of such signals concurrently diminishes information asymmetry between companies and external investors (Kong et al., 2022). Consequently, companies benefiting from SPREC support can gain increased access to financial resources from shareholders and banks (Xiao et al., 2023). This, in turn, indirectly mitigates their financing constraints and amplifies their endeavors in the realm of GTI.

Thirdly, currently, sustainability is a popular framework. The growing acceptance of the concept of “sustainability” can be attributed to the escalating conflict between economic development and environmental degradation. The rapid economic progress has elevated material living standards for all individuals. However, environmental deterioration, including air, water pollution, soil contamination and the following extreme weather, lead to a significant destroy of life and life expectancy (Polasky et al., 2019). Consequently, humanity has initiated a critical reevaluation of the limitations of “Anthropocentrism” and has come to regard sustainable development as the optimal choice for the future (Kopnina, 2014). Nevertheless, this presents a formidable challenge. Achieving this goal necessitates comprehensive changes in the behaviors of market participants. Consumers are encouraged to embrace “green” consumption practices (Testa et al., 2021); while, businesses are urged to adopt “green” production methods (Xie et al., 2019). It is noteworthy that a study conducted by Opentext discovered that nearly 88% of surveyed consumers expressed a preference for purchasing products from companies with ethical sourcing practices over others, based on a survey of 27,000 respondents worldwide.⁴ Thus, environmentally friendly production is not merely an issue of production ethics but also a proactive response to market demands. Therefore, in light of external pressures, the implementation of SPREC will lead to enhanced governmental oversight of enterprises, with external stakeholders demonstrating increased interest in corporate development. These dynamics collectively encourage companies to overcome inertia, cultivate a positive social image, and bolster their motivation to engage in GTI (Henriques & Sadorsky, 1996; Li & Xiao, 2020).

Finally, Florida (2000) demonstrates that lifestyle elements, including amenities and environmental quality, hold equal significance to economic determinants in both attracting talent and fostering the growth of high-tech regional economies. SPREC underscores the urgency to upgrade educational infrastructure and healthcare facilities, and intensify environmental conservation endeavors in resource-depleted cities (Song et al., 2020a). These initiatives foster a more conducive living and working environment, facilitate the residents’ daily lives, subsequently aiding in the cultivation and attraction of highly skilled professionals. This reservoir of talent serves as the cornerstone for fostering the advancement of GTI within enterprises (Sun et al., 2020).

Therefore, we suggest hypothesis H1a.

H1a.

The SPREC could promote corporate GTI.

On the other hand, we believe that the SPREC may inhibit corporate GTI in resource-exhausted cities. The possible reasons are mainly the following.

Firstly, although in the previous section, we discussed how SPREC can directly or indirectly alleviate a company’s financial constraints, some researchers contend that this alleviation has no positive impact on GTI (Liu et al., 2019; Marino et al., 2016; Xiaobao & Wu, 2022). In the absence of effective supervision within companies, the risk of “moral hazard” emerges, potentially leading to a diversion from innovation-related uses of subsidies (Mao & Xu, 2018). For instance, after receiving government subsidies and bank loans, companies might choose to invest in large-scale equipment rather than engage in research and development activities. Consequently, the increase in the proportion of fixed assets is at the cost of R&D and human capital investment, thereby impeding the GTI of the company.

Secondly, rent-seeking is the practice of individuals, groups, or entities pursuing special privileges to acquire economic rents or benefits without generating new wealth or enhancing economic productivity (Krueger, 1974). In the context of China, local governments wield substantial influence over crucial factors, including financial resources and land allocation, along with extensive control over fiscal spending (Mao & Xu, 2018). Consequently, to garner support from SPREC, certain enterprises may actively cultivate rent-seeking relationships with local government officials. Nevertheless, these enterprises inevitably incur rent-seeking costs, which fall within the category of non-productive expenses (Yu et al., 2010). As posited by Shleifer and Vishny (1994), rent-seeking activities may be more enticing than productive endeavors, provided the associated costs do not exceed those of production. Once such rent-seeking connections are established, the government often prioritizes affiliated enterprises when allocating financial resources and land assets. This allocation approach weakens the importance of enterprise performance and prospects, detrimentally affecting not only the interests of other enterprises but also eroding market fairness (Aidt, 2016). The rent-seeking behavior of firms could impede the GTI, consequently hindering long-term development. Specifically, substantial rent-seeking costs may crowd out a company’s research and development expenditures, impeding innovation in green technologies (Du et al., 2020). Furthermore, enterprises may easily secure substantial government support or subsidies through rent-seeking, and the excessive profits obtained through “unearned gains” may diminish their motivation to invest in high-risk, high-investment GTI (Yao et al., 2023).

Finally, in resource-exhausted cities, resources depletion makes the local economy on the verge of collapse. Productive elements including labor, capital, and technology have already flowed out (Yu & Ma, 2022). In this circumstance, implementing the SPREC seems like “closing the stable door after the horse has bolted.” Once coupled with poor government governance and weak local institutional environments, high subsidies therefore exacerbate corruption and resource misallocation (Song et al., 2020a). Corporate GTI is hard to carry on.

Therefore, we suggest hypothesis H1b.

H1b.

The SPREC could inhibit corporate GTI.

4 Research design

4.1 Sample selection and data sources

We select Chinese manufacturing enterprises with A-share listings from 2003 to 2018 as our study sample. The following three components listed below are the primary data sources: ① financial data of companies. That data in this part are processed as follows after being taken from the CSMAR database. First, in accordance with the categorization criteria of the “Industry Classification Guidelines for Listed Companies (2012 Edition),” we only keep firms in the manufacturing industry, excluding those in the primary and tertiary industries. Then, we exclude the companies with incomplete financial data or special treatment. Lastly, winsorization is performed on all continuous variables. ② Corporate green innovation data. In accordance with the classification criteria of the “International Patent Classification Green Inventory (IPC Green Inventory)” released by the WIPO, we manually gather and compile the green patent data of firms in the search column of the China National Intellectual Property Administration. By matching, sorting, and computing the raw data, we obtain micro panel data for a total of 17,240 sample observations from 1956 observed firms during 16 consecutive years.

4.2 Definition of variables

4.2.1 Dependent variables

The GTI is represented by the amount of green patent applications (GP) of the firm. There are the following two main causes: The first benefit of green patents is their quantifiability, which may most intuitively reflect the output of firm’s efforts in GTI. Furthermore, compared with R&D inputs, patents have a clear technology categorization, allowing patent data to be subdivided based on numerous technological properties to indicate the various values of innovative initiatives. Second, to analyze the influence of the pilot policy on GTI activities in a more timely manner, we utilize patent application data instead of patent grant data due to the lengthy patent application process. Referring to Xu and Cui (2020), we use the logarithmic calculation of the amount of GP to evaluate the degree of enterprises’ GTI.⁵ In addition, to further depict the different innovativeness and meanings of green patents, we also separate green patent applications (GP) into two dimensions, which are green invention patent applications (GIP) and green utility patent applications (GUP) as sub-studies to test our hypothesis.

4.2.2 Independent variable

The implementation of the SPREC (${\mathrm{Treat}}_{c}\times {\mathrm{Post}}_{c,t})$ is the independent variable and is represented by the interaction of the policy dummy ${(\mathrm{Treat}}_{c})$ and the time dummy ${(\mathrm{Post}}_{c,t})$.

4.2.3 Control variables

Considering the potential impacts of other corporate elements on GTI, some control variables are also set.

• 1. The age of a firm ($\mathrm{age}$). A firm’s age often indicates how mature the firm is. According to Jin et al. (2022), companies that are more mature also tend to be more innovative. The logarithm of the length of time since a business was founded is applied to calculate its age.

• 2. Size ($\mathrm{size}$). According to Bu et al. (2020), size shows a significant impact on corporate innovation. Larger firms usually invest more consistently in R&D for sustainable growth. The logarithm of the total capital serves as a gauge for companies’ size.

• 3. Leverage ($\mathrm{lev}$). Numerous literature has shown that leverage affects corporate innovation. Some scholars believe that excessive firm leverage will inhibit innovation (Barclay & Smith, 1995; Rajan & Zingales, 1995), while others argue that moderate firm leverage could improve firms’ innovation (Margaritis & Psillaki, 2010; Meyer, 1998).

• 4. Return on sales ($\mathrm{ros})$, the ratio of profit to operating revenue, is a reflection of an enterprise’s profitability. More profitability can provide more endogenous capital for R&D, but there are also some companies that are less motivated to innovate due to excessive profitability (Xu & Cui, 2020).

• 5. Variables related to the market value ($\mathrm{grow}$), cash flow level ($\mathrm{cfo}$), and governance structure ($\mathrm{top}10$) of the firm. Drawing on (Yang et al., 2023), the impact of factors such as a firm’s market value ($\mathrm{grow}$), cash flow level ($\mathrm{cfo}$), and governance structure ($\mathrm{top}10$), which could have impacts on GTI, are also controlled. The market value is calculated by the market value to asset ratio, and the level of cash flow is evaluated by the net cash flow from operations to assets ratio.

Table 1 indicates the definition, and Table 2 indicates the descriptive statistics.

Table 1 Definition of variables

Table 2 Descriptive statistics

4.3 Empirical model

To ascertain whether the SPREC promotes or inhibits GTI, we design the DID model with multiple time periods by using the policy as a quasi-natural experiment, which has been extensively employed in previous policy studies (Lu et al., 2022; Xu & Cui, 2020; Yu & Ma, 2022).

First, according to the criterion of “whether they are listed as resource-exhausted cities,” we divide the sample of cities from 2003 to 2018 into an experimental group (cities with implemented policies) and a control group (cities without implemented policies) and set policy dummy variables. Then, we construct time dummy variables with the approval year (the year 2007, 2009, and 2011) of the three batches of resource-exhausted cities.⁶ At last, we construct the following DID model with multiple time periods.

$${\text{G}}{{\text{P}}_{ict}} = {\alpha_0} + {\alpha_1}{\text{Trea}}{{\text{t}}_c} \times {\text{Pos}}{{\text{t}}_{c,t}} + {X_{ict}}^{\prime}\theta + {p_i} + {q_t} + {s_{pt}} + {\varepsilon_{ict}}$$

(1)

where the firm, city, province, and year are denoted by i, c, p and t, in turn. ${\text{GP}}_{ict}$ indicates how many green patents have been applied by a firm i in city c in year t. ${\mathrm{Treat}}_{c}$ is a policy dummy variable that indicates if a city belongs to the resource-depleted cities. If the city c is present in any of the three batches of resource-exhausted city lists, ${\mathrm{Treat}}_{c}$ takes the value of 1, and 0 otherwise. ${\mathrm{Post}}_{c,t}$ is denoted as a time dummy variable. The cities in the experimental group are not subject to the policy shock until the year after they are listed. Thus, if the city has already been listed in year t, the value of ${\mathrm{Post}}_{c,t}$ is 1, otherwise it is 0. As for the cross multiplication term ${\mathrm{Treat}}_{c}\times {\mathrm{Post}}_{c,t}$, when it is 1, it indicates the implementation of the policy in city c in year t, otherwise it is not implemented. A group of control variables are referred to as ${X}_{ict}$. ${p}_{i}$ is an acronym for individual-fixed effects, ${q}_{t}$ stands for time-fixed effects, ${s}_{pt}$ stands for joint fixed effects of provinces over time. The random error term is indicated by ${\varepsilon }_{ict}$. In model (1), only the amount of GP by companies is listed, while GIP and GUP are analyzed separately in the sub-studies below.

5 Analysis of empirical results

5.1 Baseline regression results

The regression findings of the SPREC’s impact on the amount of corporate green patent applications are shown in Table 3. Column 1 displays the findings without the inclusion of control variables and fixed effects, Columns 2 and 3 show the findings with the inclusion of fixed effects but not control variables, and the last two columns show the findings with the inclusion of them both. The results in the five columns demonstrate that the regression coefficients of the independent variable ${\mathrm{Treat}}_{c}\times {\mathrm{Post}}_{c,t}$ are all significantly negative, proving that the implementation of the SPREC significantly inhibits the corporate GTI in resource-depleted cities. These findings support hypothesis H1b.

Table 3 Baseline regression results

5.2 Sub-index regression results

To confirm the outcomes of the baseline regression, we subdivide green patent applications (GP) into two dimensions for further testing, and Table 4 presents the outcomes. No matter whether the dependent variable is assumed to be GIP or GUP, the regression coefficients of the ${\mathrm{Treat}}_{c}\times {\mathrm{Post}}_{c,t}$are significantly negative. The results again support the hypothesis H1b that the SPREC significantly restricts the corporate GTI in resource-exhausted cities.

Table 4 Sub-index regression results

5.3 Parallel trend test

The potential prerequisite underlying the DID for use in policy assessment is that the experimental group and the control group should exhibit a parallel trend, meaning that in the absence of policy intervention, both the experimental and control group should exhibit similar GTI trends. Therefore, drawing inspiration from Beck et al. (2010), we perform a parallel trend test using event analysis and put up the following model:

$${\text{G}}{{\text{P}}_{ict}} = \gamma + \mathop \sum \limits_{k = - 8}^{k = 11} {\beta_k} \times {D_{c,\;\;t + k}} + {X_{it}}^\prime \eta + {p_i} + {q_t} + {s_{pt}} + {\varepsilon_{ict}}$$

(2)

In model (2), $D_{c, t+k}$ is denoted as a set of dummy variables. When $k\ge 0$ if city c is a resource-exhausted city and is in the kth year after the implementation of the SPREC, then ${D}_{c, t+k}$adopts the value of 1, otherwise 0. When $k<0$, if city c is a resource-exhausted city and is in the − kth year before the implementation of the SPREC, then ${D}_{c, t+k}$takes the value of 1, otherwise 0. It is necessary to explain that since the SPREC is implemented in three batches of 2007, 2009, and 2011 in different cities, we need to generate a relative time value k by subtracting the policy’s implementation year from each year within the sample period. Thus, the sample period of our study covers 8 years before ($k=-8$) and 11 years after ($k=11$) the implementation of the policy.

Considering such a long time span before or after the implementation of the SPREC, we refer to Wang and Ge (2022), and truncate the samples that are too early or too late for the implementation of the policy. The specific approach is to aggregate data earlier than 5 years before the policy implementation to period $t-5$, and to aggregate data later than 4 years after the policy implementation to period $t+4$.

According to the model (2), the coefficient ${\beta }_{k}$ reflects that if there are any discrepancy between two groups at the kth year before or after the policy’s initiation. If none of the values of ${\beta }_{k}$ is significant during $k<0$, it is evidence that two groups meet the parallel trend assumption. By contrary, if the values of ${\beta }_{k}$ are partially significant during $k<0$, it is evidence that the common trend assumption is not met and the two groups have a significant disparity prior to the adoption of the policy. With the GP as the dependent variable, Fig. 1 depicts the parallel trend test.

Figure 2 demonstrates that, prior to the implementation of the SPREC, significant difference does not exist in the degree of corporate GTI between the two group of cities, indicating that the research sample passes the parallel trend test.

Fig. 2

Parallel trend test

5.4 Placebo test

Other undetectable time-varying urban features might interfere with the findings, which is another obstacle with the DID. Different cities have their characteristics in terms of governance level, resource environment and cultural deposits, etc. Despite the preceding section’s inclusion of the fixed effects of provinces over time, there are still some undetectable time-varying urban features level that may not be controlled. For this, we run a placebo test with reference to Liu and Lu (2015) and Zhao et al. (2020).

After the random sampling of 1000 times, Fig. 3 demonstrates the significance of the estimated coefficients and their distributions. The findings demonstrate that the distribution of the “pseudo-policy dummy variable” is primarily centered around zero, and the corresponding p values are greater than 0.1, which indicates that the study passes the placebo test and other unobserved factors do not have a significant impact on it.

Figure 3

Placebo test

5.5 Other robustness tests

5.5.1 Replacing core variables

It may take some time for the SPREC’s influence on corporate GTI, which means a time lag effect may exist. Therefore, we treat the dependent variables with one and two lags, respectively, before conducting regression analysis, and the findings displayed in Table 5 agree with those of the baseline regression.

Table 5 Robustness tests: replacing core variables

5.5.2 Reducing selection bias

To lessen the sample selection bias, the PSM-DID method is adopted for robustness tests. A stricter caliper radius (0.0001) is chosen due to the adequacy of the sample and in an effort to eliminate the effect of “selection bias” as much as possible. The first three columns in Table 6 list the DID estimations after matching. The findings indicate that the SPREC continues to inhibit GTI at the 1% level of significance.

Table 6 Robustness tests: reducing selection bias

Furthermore, municipalities directly under the central government, as the frontier of China’s reform and opening-up and regional economic centers, have significantly higher economic, locational, and political superiority than other regions. Therefore, we re-run the regression test after eliminating the four municipalities directly under the central government from the sample, and the outcomes in the last three columns of Table 6 concur with those of the baseline regression.

5.5.3 Excluding the interferences of other policies

During the same time when the SPREC is implemented, enterprises in some energy-exhausted cities might also be affected by other policies. It has been proven that the policy for low-carbon cities (Chen et al., 2022), which started to be implemented gradually in 2010, and the fiscal policy for energy saving and emission reduction (Jin et al., 2022), which started to be implemented gradually in 2011, both significantly advance the corporate GTI. Thereby, to exclude these interferences, we run the regression analysis after controlling for the two policies mentioned above. The findings, as displayed in Table 7, are in accord with those of our baseline regression.

Table 7 Robustness tests: excluding the interferences of other policies

6 Further analysis

6.1 Mechanism

As an instrument of government intervention, the SPREC is designed to assist resource-exhausted cities in emerging from the “woods” and promoting sustainable development. However, the above-mentioned study shows that the SPREC inhibits corporate GTI, which is a key way to balancing economic growth with environmental protection and achieving sustainable development. To explore the specific reasons for the failure of the policy to meet expectations, we further examine the mediating mechanisms by which the SPREC inhibits firms’ GTI on the foundation of baseline regressions.

In the previous section of the research hypothesis, we analyze the possible paths by which the SPREC inhibits corporate GTI. On the one hand, the SPREC not only directly alleviates the financial pressure of enterprises, but also serves as a public signal of the market with the natural nature of government endorsement, which could indirectly alleviate the financing constraints of the supported enterprises. However, after receiving support funds and bank loans, firms may use the funds for other activities than innovation, such as raising the proportion of fixed assets, which inhibits corporate GTI. Therefore, there are two possible paths: “Implementing the SPREC—increasing the proportion of fixed assets—inhibiting GTI” and “Implementing supportive policies—alleviating financing constraints—increasing the proportion of fixed assets—inhibiting GTI.”

On the other hand, the SPREC may trigger enterprises to actively establish rent-seeking relationships with local governments, in this process, not only the high cost of rent-seeking might cause enterprises to spend less on R&D expenditures, but also enterprises might lose the motivation to conduct R&D due to the excessive profits gained after rent-seeking, both of which will inhibit corporate GTI. Therefore, there are two possible paths: “Implementing the SPREC—increasing rent-seeking costs—inhibiting GTI” and “Implementing the SPREC—increasing rent-seeking costs—crowding out R&D expenditures—inhibiting GTI.”

According to Fee et al. (2009), the FC index⁷ is utilized to gauge the severity of corporate financing constraint. The larger the FC, the severer the financing constraint, indicating the greater the difficulty or the higher the relative cost of financing for the firm. The R&D expenditure is expressed as the logarithm of the enterprise’s R&D expenditure. As for rent-seeking costs, referring to Dechow et al. (1995), we use excess administration expenses (EAE) to measure them. Specifically, we incorporate various factors affecting administration expenses into model (3) and calculate the expected management expenses of firm i in year t. Then, the gap between the actual management expenses and the expected management expenses is the EAE.

$$\begin{aligned} {\text{Managemen}}{{\text{t}}_{it}} & = {\theta_0} + {\theta_1}{\text{Lnsal}}{{\text{e}}_{it}} + {\theta_2}{\text{Le}}{{\text{v}}_{it}} + {\theta_3}{\text{Growt}}{{\text{h}}_{it}} + {\theta_4}{\text{Board\;siz}}{{\text{e}}_{it}} \\ & \quad + {\theta_5}{\text{Staf}}{f_{it}} + {\theta_6}{\text{Audito}}{{\text{r}}_{it}} + {\theta_7}{\text{List\;ag}}{{\text{e}}_{it}} + {\theta_8}{\text{Price\;leve}}{{\text{l}}_{it}} \\ & \quad + {\theta_9}{\text{Captain\;intensit}}{{\text{y}}_{it}} + {\theta_{10}}H{1_{it}} + {\theta_{11}}{\text{Margi}}{{\text{n}}_{it}} + {\sigma_{it}} \\ \end{aligned}$$

(3)

Management is the ratio of corporate management expenses to operating income for the purpose of normalizing corporate management expenses. Other variables in order represent the logarithm of sales revenue in the previous period, leverage, sales revenue growth rate, board size, whether the audit is from the Big4 accounting firms, years of listing, price index (measured by the average salary of active employees), capital intensity, equity concentration (Herfindahl–Hirschman index of the top five shareholders), and gross margin.

In the section of mediating mechanism tests, first, we use Stata to construct a structural equation model (SEM), the coefficients and significance of each path are shown in Fig. 4. Then, we further test the chain mediation models by using Bootstrap analysis (repeating sampling 5000 times and calculating 95% confidence intervals), the outcomes are displayed in Table 8.

Fig. 4

Results of the SEM test for chain mediation

Table 8 Results of the Bootstrap analysis for chain mediation

6.1.1 Analysis of the path “the SPREC—the proportion of fixed assets—GTI”

According to Fig. 4, we find that the coefficient of the path “the SPREC—the proportion of fixed assets” is 0.056 (p < 0.05), the coefficient of the path “the proportion of fixed assets—GTI” is − 0.275 (p < 0.001), indicating that the SPREC makes the proportion of fixed assets of enterprises increase significantly, and consequently has an effect that crowds out GTI.

Next, the outcomes of Bootstrap in the second row of Table 8 show that the β value of the path “the SPREC—the proportion of fixed assets—GTI” is − 0.015 and the 95% confidence interval is [− 0.023, − 0.009], which excludes 0. The outcomes indicate that the mediating effect of the path exists obviously, which is consistent with that of the structural equation model.

6.1.2 Analysis of the path “the SPREC—financing constraints—the proportion of fixed assets—GTI”

According to Fig. 4, we find that the coefficients of the path “the SPREC—financing constraints—the proportion of fixed assets—GTI” are − 0.01 (p < 0.001), − 0.075 (p < 0.001) and 0.275 (p < 0.001) in turn, indicating that the SPREC alleviates the financing constraint of enterprises, brings them more abundant funds and makes the proportion of their fixed assets rise significantly, and consequently has an effect that crowds out GTI.

Next, the outcomes of Bootstrap in the third row of Table 8 show that the β value of the path “the SPREC—financing constraints—the proportion of fixed assets—GTI” is − 0.002 and the 95% confidence interval is [− 0.003, − 0.001], which excludes 0. The outcomes indicate that the mediating effect of this path exists obviously, which is consistent with that of the structural equation model.

6.1.3 Analysis of the path “the SPREC—rent-seeking costs—GTI”

According to Fig. 4, we find that the coefficients of the path “the SPREC—rent-seeking costs—GTI” are 0.013(p < 0.05) and − 0.803 (p < 0.001) in turn, indicating that the SPREC triggers rent-seeking behavior of enterprises and increases their rent-seeking costs, and thus inhibits their GTI.

Next, the outcomes of Bootstrap in the fourth row of Table 8 show that the β value of the path “the SPREC—rent-seeking costs—GTI” is − 0.01 and the 95% confidence interval is [− 0.016, − 0.005], which excludes 0. The outcomes indicate that the mediating effect of this path exists obviously, which is consistent with that of the structural equation model.

6.1.4 Analysis of the path “the SPREC—rent-seeking costs—R&D expenditures—GTI”

According to Fig. 4, we find that the coefficients of the path “the SPREC—rent-seeking costs—R&D expenditures—GTI” are 0.013(p < 0.05), − 8.902 (p < 0.001) and − 0.016 (p < 0.001) in turn, indicating that the SPREC triggers rent-seeking behavior of enterprises, and the rent-seeking cost would crowd out their R&D expenditure, and thus might inhibit their GTI.

Next, the outcomes of Bootstrap in the fifth row of Table 8 show that the β value of the path “the SPREC—rent-seeking costs—R&D expenditures—GTI” is − 0.002 and the 95% confidence interval is [− 0.281, − 0.185], which excludes 0. The outcomes indicate that the mediating effect of this path exists obviously, which is consistent with that of the structural equation model.

6.2 Heterogeneity

According to the analysis above, the SPREC exists a significant inhibitory effect on the GTI of manufacturing enterprises based on the entire sample. The different degrees of regional economic growth and marketization in China both have significant differences, and the different ownership nature also has many effects on the development of enterprises. So, does the above conclusion still hold for the samples with diverse ownership natures, in diverse regions, and under diverse degrees of marketization? Addressing this issue will be essential to give us a deeper understanding of the study. Table 9 displays the specific outcomes.

Table 9 Heterogeneity analysis

In accordance with the type of ownership, the businesses are separated into state-owned enterprises (SOEs) and non-state-owned enterprises (non-SOEs) for group regressions, and the first two columns of Table 9 present the outcomes. It is clear that the SPREC exists a significant inhibitory effect on GTI in SOEs, while it does not exist in non-SOEs. A possible explanation is that during China’s rapid economic growth, there are still inevitable problems of government bureaucracy and formalism, breeding rent-seeking by enterprises, and corruption by officials, which are more common in resource-based cities. At the same time, Chinese SOEs have natural government resources that facilitate the establishment of government-business linkages. Therefore, when confronted with the SPREC released by the government, SOEs are more inclined to pursue rent-seeking activities than technological innovation to obtain excess returns than non-SOEs (Dong & Zhang, 2021).

According to the location, we group the companies according to being in the eastern region and the mid-western region and then regress them separately, and the outcomes are presented in Columns 3 and 4. We discover that the SPREC causes a significant inhibitory effect on GTI of firms located in the mid-western regions, while facilitating those in the eastern regions, with a T-value of 1.33, which is already relatively close to the 10% significance level. To explore the possible reasons for this, compared to Chinese backward mid-western region, the eastern region is relatively developed, with a more advanced state of economic development, a better financial system, a more deeply rooted concept of sustainable development, more standardized political governance and more open and transparent information about government affairs. Therefore, when resource-depleted cities in the eastern region obtain government support, the use of support funds and access to financial loans would be more standardized and efficient, and it would be more difficult to implement rent-seeking behaviors that hinder the GTI. Thus, the SPREC could encourage corporate GTI in resource-exhausted cities in the eastern region, while inhibit those in the mid-western regions.

As for the degree of marketization, we utilize the regional index of marketization measured by Wang et al., (2019b). Depending on the median of marketization, the study samples are split into two types: those with high and those with low levels of marketization, and the SPREC’s effects on GTI in manufacturing companies under different degrees of marketization are further examined, and the last two columns present the outcomes. We can see that the SPREC remains a disincentive to GTI of companies under a low degree of marketization, while under a high degree of marketization, the SPREC unexpectedly promotes corporate GTI significantly, which support the initial intention of the central government in formulating the policy. This result enlightens us that the SPREC is not useless and we should focus more on the construction and governance of the institutional environment while implementing it.

7 Conclusions and policy recommendations

7.1 Conclusions

We investigate the SPREC’s impact on corporate GTI. The following are the research conclusions.

Firstly, there is a significant inhibitory effect of the SPREC on corporate GTI, and the finding is still valid after robustness tests including tests of the dependent variable sub-dimensions, placebo tests, use of dynamic panel models, reduction selection bias, and exclusion of interfering policies.

Secondly, there are multiple paths for the inhibitory effect. On the one hand, the SPREC not only directly alleviates firms’ financial pressure, but also indirectly alleviates firms’ financing constraints as a public signal from the market. However, firms use the support funds and loans they receive for other activities than innovation, such as increasing the proportion of fixed assets. On the other hand, supportive policies may trigger enterprises to establish rent-seeking relationships with local governments, which would not only generate high rent-seeking costs to crowd out enterprises’ R&D expenditures, but also make enterprises lose their R&D incentives due to the excessive profits brought by rent-seeking. All of these above would inhibit enterprises’ GTI.

Third, the inhibitory effect shows sample heterogeneity depending on enterprises’ ownership nature, location characteristics, and degrees of marketization. The SPREC significantly inhibits GTI among SOEs, enterprises located in the mid-western region, and enterprises under a low degree of marketization. In addition, we unexpectedly find the SPREC is not useless and presents a significant promotion effect on GTI of enterprises under a high degree of marketization.

7.2 Policy recommendations

The initial intention of the SPREC is to advance the transformation, upgrading, and sustainable development of resource-exhausted cities, yet our study shows that it inhibits GTI of manufacturing firms in general and shows variability under different heterogeneity. It can be seen that that there is still a long way to go in the transformation and green growth of resource-exhausted cities. The following policy recommendations may be available to us.

• 1. The government cannot be static in the formulation and implementation of supporting policies, but should establish a dynamic adjustment mechanism and a strict evaluation mechanism. First of all, when local governments screen and identify support targets, they should conduct the qualification review and result announcement in an open and fair manner. Secondly, the supported enterprises should disclose the information and progress on time and accept public supervision, and the government should increase the punishments on enterprises that violate the rules, abuse, and speculate, so as to stop enterprises from “rent-seeking.” Finally, the government should enhance its oversight, regularly assess the performance of supported enterprises, timely track and audit the use of funds, and adjust the decision of whether to continue support based on the results dynamically and promptly. Additionally, the borrowing system of financial institutions should be improved, and financial institutions should be urged to earmark the special loans obtained by enterprises.

• 2. Government’s “visible hand” and market’s “invisible hand” should be coordinated adequately with each other. The government should not only formulate supporting policies according to local circumstances, but also promote the level of local marketization. Our findings indicate that the SPREC’s effects on corporate GTI varies significantly under different conditions. Financial transfer payments alone are not sufficient to promote green and sustainable growth of enterprises. It is also vital to help improve the economy, environment, and people’s livelihood in the backward mid-western regions, eliminate discrimination and injustice against non-state enterprises and fully protect their rights and interests, improve the marketization of cities, reduce government control over key factors and resources and let the market decide the allocation of resources. All of these above would not only regulate the behavior of enterprises and reduce opportunities for rent-seeking, but also encourage enterprises to pursue technological innovation and development actively.

• 3. Strengthen anti-corruption efforts at all levels of government and establish a healthy and transparent relationship between government and business. In recent years, the Chinese government’s determination to fight corruption has been unprecedentedly high, but there is still a long road to go. The government should improve governance capacity, increase anti-corruption efforts, and establish a proper and transparent relationship with enterprises, especially with state-owned enterprises, to avoid damage to the market due to rent-seeking and bribery, and create a fair and positive economic environment.