2.2.1 Resource base acquisition mechanism.

The resource base of an enterprise’s digital transformation usually comes from the enterprise’s internal resource stock and external resource support. According to the resource-based view, enterprises have limited resources of their own, and it is difficult to meet the needs of enterprises to carry out energy-consuming and high-risk organisational change activities. Therefore, when promoting digital transformation, enterprises should pay more attention to the process of acquiring and accumulating internal resources and external resources. A stable source of funding is a core component of an enterprise’s resource base, as well as an important prerequisite for ensuring the long-term development of its activities. Due to the existence of information asymmetry in the market and various types of agency problems [54], firms face different degrees of financing constraints. When enterprises face strong financing constraints [55], it is difficult to ensure a stable source of funds, which greatly limits their ability to cope with high costs and risks [56, 57], which will further weaken the incentives for enterprises to actively participate in digital transformation. In recent years, as society as a whole has paid more attention to the realisation of environmental and economic benefits for enterprises, the “green” signal emitted by enterprises has become an important factor in their ability to obtain internal and external funding [58]. Research by Chen and Shang (2023) [59] pointed out that one of the key criteria for the government to assess whether an enterprise has sustainable development values and whether it provides a series of reasonable guarantees, such as financial subsidies and tax incentives, to reduce financial distress is the amount it pays for the ‘green’ field. Lu and Li (2023) [50] also mentioned that government departments and financial institutions are more inclined to provide policy support to enterprises that carry out green innovation. In addition, from the perspective of market demand, green innovation can enhance product differentiation advantages by tailoring product development to consumer needs [53, 60]. This advantage is not only conducive to developing new markets and increasing market share, but also meets consumers’ pursuit and awareness of environmentally friendly products and brings more environmental premiums [61]. At the same time, “green investment” in the capital market also increasingly favours companies that are more likely to produce differentiated green products [62, 63]. To summarise the existing research, green innovation is an important embodiment of enterprises actively fulfilling their social responsibility, and to a certain extent, it can play the role of “endorsement”, which not only creates a good reputation and wins the favour of various stakeholders [64]; at the same time, due to the consumer’s choice of preference, it is helpful to urge enterprises to increase the supply of green products to improve internal operating capital income [58], thus reducing the threshold of enterprise financing and easing enterprise financing constraints, and ensuring strong capital reserves for enterprise digital transformation. In summary, Hypothesis 2 is proposed.

1. Hypothesis 2: If other conditions remain unchanged, enterprises actively carry out green innovation, which can effectively alleviate the constraints of enterprise financing, thus providing stable financial resource base support for enterprises’ digital transformation, and promoting enterprises’ digital transformation.

2.2.2 Resource base optimization mechanism.

Enterprises can obtain stable financial support is to lay the premise of its activities to carry out the resource base. At the same time, the ability to obtain and accumulate funds to be reasonably and effectively allocated is an important criterion to test whether the resource base ensures the needs of the enterprise’s activities and value creation [65, 66]. Many studies have shown that the ability of enterprises to manage operating capital has a profound impact on enterprise performance [67–69]. Green innovation, as an innovative activity, is embedded in the production process of enterprises and is committed to the disruptive transformation or redesign of the original factors of production, and the impact of any small change on the existing system cannot be underestimated [70]. Although in the short term, green innovation has a “double externality” [71], with obvious characteristics of high investment, high risk and low return; however, with the optimisation of production and business processes in the long term, green innovation can significantly reduce compliance costs [72], operating costs [73] and generate additional revenue for the enterprise [74], increasing the accumulation of capital resource base. From the perspective of process innovation, Lu and Li(2023) [50] illustrate the positive role of enterprise green innovation in eliminating outdated and energy-consuming processes, which can use fewer factor resources and time cost inputs to produce the same quality and quantity of products, reduce production costs [18, 51]. Porter et al. (1995) [75] argue from a product innovation perspective that the use of waste and renewable resources in green innovation has a positive impact on cost reduction. Simultaneously, green product innovation through continuous development of new market areas can increase a company’s market share and product sales revenue [76, 77], thereby further increasing the disposable strength of the enterprise’s operating capital. Furthermore, the implementation of green innovation within the patent field has been influential in providing enterprises with sustainable competitive benefits and encouraging their growth [51] through the acquisition of patent transfer advantages [78, 79]. From an innovation economics perspective, green innovation can lower production and operational costs, boost production and market share, and yield earnings from technology licensing [78–80]. This is a reflection of the optimisation of the efficiency of factor allocation in the enterprise, which is conducive to further economic efficiency [51]. Operating capital management is related to every daily aspect of a enterprise, such as supply, production, sales, enhancing competitiveness, and transformational development [81]. The amount of operating capital is the foundation of operating capital management. Filbeck and Krueger’s (2005) study [82] suggests that determining the optimal balance for funds closely related to operating capital is a primary goal of operating capital management, which is crucial for the survival and development of an enterprise and has an important impact on operating capital management decisions. Green innovation not only leads to an increase in resources, but also offers a distinct advantage in optimizing resource allocation and promoting sustainable usage. Digital transformation is challenging and gradual, demanding a high level of resource investment in terms of initial access to support and the efficient allocation and sustainable utilization of resources [11]. Therefore, it can be deduced from the above analysis that green innovation helps enterprises accumulate sales revenue, patent transfers, and a financial foundation resulting from cost reductions,thereby enhancing the enterprise’s ability to manage its operating capital, achieve a reasonable and effective reallocation of resources, subsequently increase the likelihood that the enterprise undertake high-input activities such as digital transformation. This leads to hypothesis 3.

1. Hypothesis 3: Under the same other conditions, enterprises actively carry out green innovation, which is conducive to the accumulation of capital resources by optimizing production operations, thus affecting the management of operating capital and further influencing the digital transformation of enterprises.

3.2.1 Dependent variable.

Enterprise digital transformation (DT). According to Qi et al. (2020) and Zhang et al. (2021) [83, 84], the notes to the financial reports of listed companies disclose the proportion of the digital transformation-related portion of the year-end intangible asset breakdown items to the total intangible assets. This serves as a measure of the digitalisation level of enterprises. Specifically, when the detailed item of intangible assets includes keywords related to Digital transformation technology such as “software”, “network”, “client”, “management system”, “intelligent platform”, as well as related patents, the detailed item is defined as “digital technology intangible assets”, and then several digital technology intangible assets of the same company in the same year are summed up to calculate their proportion in the intangible assets of the current year, The final result is determined as the proxy variable of the degree of digital transformation of the enterprise and expressed by DT. Effective use of digital technology was additionally chosen as a proxy measure for evaluating the degree of digital transformation within the enterprise, for the sake of conducting robustness testing.

3.2.2 Independent variable.

Green innovation (GI). As stated previously, implementing green innovation is capable of enhancing production and business procedures and endorsing establishments by signaling their eco-friendliness, ultimately leading to a constructive outcome of both cost reduction and efficiency elevation. Such a dual effect of cost-cutting and productivity enhancement can be realised. Therefore, it is imperative to guarantee the provision of a sufficient amount of green innovations, as well as to concentrate on enhancing innovative quality to accomplish sustainable development. With this in mind, green innovation quantity and quality are considered as proxy variables for green innovation. The volume of green patent applications is a more accurate representation of the green innovation achievement of enterprises [85, 86]. Specifically, the quantity of green innovation is evaluated by summing up the number of green invention patents and utility model patents [86–88], while the quality of green innovation is measured by considering the number of green invention patent applications exhibiting superior technological content [86]. These metrics are identified as GIN and GIQ, correspondingly.

3.2.3 Mechanism variable.

Mechanism variable (MV). This paper selects two indicators, financing constraints and operating capacity, to represent the financial status and management capability of enterprises, respectively. The mechanism variable, referred to as MV, is used to represent these indicators.

Mechanism variable 1: Financing constraints (FC). Usually, financing constraints can be measured using multiple indicators. However, Kaplan and Zingales (1997) [55] were the first to combine operating cash flow, dividend payout ratio, gearing ratio, cash holdings, and Tobin’s Q to construct a comprehensive financing constraint indicator, referred to as the KZ index. This index has been widely utilised. Drawing on Kaplan and Zingales (1997) and Wei et al. (2021) [55, 89], the specific approach is: Each year of the full sample is classified according to operating net cash flow/previous period’s total assets (CF), cash dividends/previous period’s total assets (Div), cash holdings/previous period’s total assets (C), gearing ratio (Lev) and Tobin’s Q (TobinQ). Kz1 is assigned a value of 1 if CF is below the median and 0 otherwise. Kz2 is assigned a value of 1 if Div falls below the median and 0 otherwise. Similarly, kz3 is assigned a value of 1 if C is below the median and 0 otherwise. Kz4 is assigned a value of 1 if Lev is above the median and 0 otherwise, and kz5 is assigned a value of 1 if Q is above the median and 0 otherwise. Calculate the KZ index by adding kz1 to kz2 to kz3 to kz4 to kz5. Employ Ordered Logistic Regression (OLR) to regress the KZ index on CF, Div, C, Lev, and Q, used as independent variables. Estimate the regression coefficients of each variable. Using the above regression model’s estimation results, we can calculate the KZ index of each listed company’s degree of financing constraints. A higher KZ index indicates a higher degree of financing constraints and lower efficiency of financing.

Mechanism variable 2: Operating capacity (OC). Operating capacity can accurately indicate an organisation’s ability to create value through managing operating capital and allocating resources. Currently, commonly used indicators for measuring an enterprise’s operating capacity include the total asset turnover ratio, the current asset turnover ratio, and other individual metrics. Additionally, certain research methods are used to calculate comprehensive indicators for the enterprise asset turnover ratio. It is essential to use these indicators to measure an enterprise’s operating capacity. Comprehensive indicators rely more on human factors compared to a single indicator, which can be heavily influenced by subjectivity. Therefore, this paper utilizes the total asset turnover ratio (the ratio of operating income to average total assets) as a proxy variable for measuring operating capacity, following Cheng et al. (2021) [66].

3.2.4 Control variable.

In order to enhance the precision of the investigation, this paper also incorporates a range of control variables into the model. These consist of firm size (Size, the natural logarithm of total assets at the end of the year), leverage, commonly referred to as the gearing ratio (Lev, total liabilities at the end of the year/total assets at the end of the year), return on total assets (ROA), and growth (Growth, current year’s operating income/previous year’s operating income-1). No changes needed as the text adheres to the principles of objective and concise academic writing. Shareholder ownership (Top1, percentage of total shares held by largest shareholder), Tobin’s Q (TobinQ), firm age (FirmAge, the current year minus the year of the company’s establishment plus one, transformed to natural logarithm), and capital intensity (CAPINT, total assets divided by operating income).

4.3.1 Robustness tests for endogeneity problems.

To reduce the influence of endogeneity on the results, this study extends the time frame for examining the effects of green innovation on enterprises’ digital transformation, as shown in Table 3. As an approach, model (1) utilises lagged core explanatory variables (GIN and GIQ) by one period. The findings presented in Table 3 demonstrate that green innovation continues to positively impact enterprises’ digital transformation over the extended observation period. The regression coefficients are positive and successfully pass the 10% and 5% statistical significance tests.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Robustness test results for extended observation windows.

https://doi.org/10.1371/journal.pone.0301266.t003

4.3.2 Two-way causal robustness test.

Previous theoretical analyses and benchmark regression results have demonstrated that increasing the level of green innovation will drive the progress of enterprise digital transformation and boost its performance. Nevertheless, the investigations by Li et al. (2022) and Luo et al. (2023) [93, 94] indicated that digital transformation also notably improves green innovation performance. To tackle the issue of two-way causality in the study, we adopt the industry averages of the independent variables GIN and GIQ as the instrumental variables for the amount (GIN) and quality (GIQ) of green innovation, respectively, following the standard practice of most scholars in this area. On the one hand, the green innovation level of the industry in which an enterprise is situated may influence its green innovation behavior, ultimately impacting the enterprise’s green innovation performance and fulfilling the instrumental variable correlation hypothesis. On the contrary, the industry’s average level of green innovation principally impacts companies’ choices and behaviours concerning green innovation but does not directly influence their digital transformation process. This finding fulfils the exogeneity hypothesis of the instrumental variable. Table 4 outlines the findings of the second stage of instrumental variable regression analyses. In both models (1) and (2), the Kleibergen-Paap rk LM successfully passes the significance test at the 1% statistical level, thereby rejecting the initial hypothesis of inadequate instrumental variable identification. Similarly, both the Cragg-Donald Wald F-statistics exceed the critical value of 16.38 for the Stock-Yogo F-test for weak instrumental variables at the 10% significance level, which nullifies the original assumption of weak instrumental variable. Therefore, it indicates that the instrumental variables chosen in this paper are justified. Moreover, the regression coefficients of GIN and GIQ are both significantly positive, which reaffirms the fundamental conclusion that environmentally-friendly innovation enhances enterprise digital transformation performance.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Robustness test results for two-way causation problem.

https://doi.org/10.1371/journal.pone.0301266.t004

4.3.3 Robustness test on sample selection bias.

To minimize the impact of sample selection bias on causal inference results, this paper employs propensity score matching (PSM) for robustness testing [95]. To ensure the reliability of the PSM results, this paper defines samples with green innovation quantity (GIN), and green innovation quality (GIQ) above the median as the experimental group, and samples below the median as the control group, respectively. After the sample grouping, the PSM radius matching method was used for processing, and a matching quality test was conducted on the overall sample. Finally, the causal relationship between the explanatory variables (GIN and GIQ) and the dependent variable (DT) was inferred for the matched samples, as shown in Table 5. The coefficients of GIN and GIQ in columns (1) and (2) of Table 5 are significantly positive at the 1% and 5% statistical levels, indicating that the sample selection issue did not lead to significant estimation errors, and the baseline regression results remain reliable.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Robustness test results using the PSM method.

https://doi.org/10.1371/journal.pone.0301266.t005

4.3.4 Other robustness tests.

A. Methods for replacing variables. In order to further validate the precision of the model, we conduct robustness tests using the replacement of dependent and independent variables as our selected method. Digital transformation, following Wu et al.’s (2021) [96] methodology, involves searching, matching, and counting the frequency of feature words related to artificial intelligence, blockchain, cloud computing, and big data technologies in the annual report. This process classifies and summarizes the word frequencies of the key technological directions, ultimately yielding a metric (DT1) that reflects the effective implementation of digital technology and serves as an alternative measure of the enterprise’s digital transformation. As for the quantity of green innovation and the quality of green innovation, the replacement variables lnGIN and lnGIQ were obtained by following the approach of most studies to add 1 to the variables to take the logarithm for the model robustness test. Columns 1 and 2 of Table 6 show the regression results with the replacement of the dependent variables. The regression coefficients of GIN and GIQ on DT1 are each significantly positive at the 1% statistical level. Columns 3 and 4 of Table 6 show the regression results with the independent variables replaced. lnGIN and lnGIQ regression coefficients of 0.004 and 0.005 on DT pass the significance test at the 5% and 1% levels respectively. The results in Table 6 show that the accuracy of the model is validated in a statistically significant way whether the dependent or independent variables are replaced.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 6. Robustness test results of substituted dependent variable and independent variable.

https://doi.org/10.1371/journal.pone.0301266.t006

B. Methods for adding missing variables. Restricted by objective conditions such as data availability, it is often straightforward to overlook many additional factors influencing the dependent variable in empirical studies. Thus, to minimise bias caused by omitted variables, this study introduces enterprise market competitiveness and degree of marketization as control variables in the main effects regression. Columns 1 and 2 of Table 7 show the regression outcomes integrating the missing variables. The regression coefficients of GIN and GIQ on DT are positively significant at the 5% and 1% levels, respectively. Therefore, the conclusion that green innovation improves digital transformation’s effectiveness aligns with prior research.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 7. Robustness test results by adding missing variables and changing sample size.

https://doi.org/10.1371/journal.pone.0301266.t007

C. Approaches to changing sample size. Although the concept of digital transformation has long been established, it was not until 2016 that the digital economy began to gain national strategic significance. China rapidly advanced the construction of information technology in various areas, continuously consolidating its digital foundations and promoting the concept of digitisation. The government, led by policies, emphasised the importance of digital technology and began to accelerate the development and application of new digital technologies, as well as promoting their integration with the real economy. Since then, enterprises have embraced the current trend and joined the movement towards digital transformation, causing it to flourish and develop. As a result, the research sample period has been revised to cover 2016–2020, and the model is being retested for its robustness. The results align with the baseline regression findings as the regression coefficients of GIN and GIQ on DT in columns 3 and 4 of Table 7 pass the significance test at the 1% statistical level.

5.2.1 Heterogeneity of enterprise scale.

The paper examines the impact of “green innovation to digital transformation” from a full-sample perspective. However, considering the heterogeneity among firms, the effects of green innovation on digital transformation may show asymmetry in the presence of different firm attributes. Therefore, this study further investigates the outcomes on a split-sample basis. Referring to previous studies, large enterprises refer to those with a higher enterprise size than the mean within the same year. SMEs, on the other hand, refer to those whose enterprise size is equal to or below the mean. Table 11 demonstrates that amongst the large enterprise group, the facilitating effect of green innovation on the digital transformation of enterprises passed the significance test at a 1% level. This paper contends that in comparison to SMEs, large enterprises possess greater scale and resource advantages and sufficient market power. Consequently, these enterprises tend to prioritise the optimisation and upgrading of production and business processes, and have the ability to withstand the high investment and risk of “greening” such processes. They also possess the capacity to govern and allocate their own resources in a rational manner, allowing them to excel in the conversion of environmental performance into economic performance. SMEs, conversely, contend with the fundamental issue of survival and lack the ability to withstand the ramifications of ventures with high-input and high-risk and low-yield. They also possess the capacity to govern and allocate their own resources in a rational manner, allowing them to excel in the conversion of environmental performance into economic performance. As a consequence, these enterprises are unable to effectively convert environmental performance into economic performance through green innovation.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 11. Heterogeneity of enterprise scale.

https://doi.org/10.1371/journal.pone.0301266.t011

5.2.2 Leverage ratio heterogeneity.

Previous research has shown that the level of leverage has a significant effect on innovation performance. This study classifies firms with leverage ratios above the mean as high leverage enterprises and those below as low leverage enterprises. Table 12 displays the outcomes of the grouping by leverage heterogeneity. It has been discovered that green innovation substantially boosts firms’ performance in digital transformation, regardless of their leverage status (all regression coefficients pass the 1% statistical level test and are positively significant). However, the impact of green innovation quantity on enterprises’ digital transformation is more prominent in the high leverage group (regression coefficients are significant at the 5% level). For one thing, the quality of green innovation has made a significant progress from “quantitative change” to “qualitative change”, which conveys “green” signals more clearly and optimises production and business processes. This qualitative shift is more crucial for enhancing the performance of digital transformation than the quantity of green innovations. Hence, it proves to be more effective in improving the digital transformation performance of enterprises. Secondly, a low leverage ratio implies that companies have inadequate solvency, which could easily lead to insufficient financial backing for these companies, thereby impeding their green innovation performance.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 12. Leverage ratio heterogeneity.

https://doi.org/10.1371/journal.pone.0301266.t012

5.2.3 Regional heterogeneity.

Differences in policy, economic, and environmental conditions within a firm’s location result in disparities in their resource base, leading to variations in firm performance. According to Table 13, in the Central region, there is a significant positive correlation between green innovation and digital transformation (Regression coefficients all pass the 1% statistical level test.). Enterprises in the eastern region have a greater overall capacity and abundant resource reserves required for transformation and growth. This makes it easier for businesses to access support from various resources, and creates opportunities for companies to build up a sustained resource reserve. Therefore, in the eastern region the impact of enterprises that provide resource support for the advancement of digital transformation via green innovation to improve digital transformation performance appears to be insignificant. In the western region, the economy suffers from underdevelopment and enterprises lack adequate financial, human and technological resources. This results in low motivation for innovation and challenges in enterprise development, leading to poorer performance. Compared to the eastern and western regions, the central region’s enterprises demonstrate more balanced strengths. They do not suffer the severe innovation limitations of the west and, while somewhat inferior to the eastern region, are able to acquire and create resources. By utilizing green innovation to create and accumulate internal and external resource bases, businesses in the central region can greatly promote their digital transformation. Therefore, the creation and accumulation of internal and external resource bases through environmentally friendly green innovation can greatly facilitate the process of digital transformation for enterprises. The above also highlights the apparent regional imbalances in the co-management of the environment and the economy.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 13. Regional heterogeneity.

https://doi.org/10.1371/journal.pone.0301266.t013