ARMA analysis of the green innovation technology of core enterprises under the ecosystem – Time series data

The characteristics of green innovation define that the green innovation ecosystem is unusual, and the construction of the system can positively affect the research on the evolution of green innovation in enterprises. As current researches fail to make a systematic elaboration on the generation mechanism of the green innovation ecosystem, this paper, by taking into consideration the characteristics of the innovation ecosystem and green innovation, defines the green innovation ecosystem as reducing environmental pollution, improving green innovation capabilities, and achieving sustainable development, through cooperation on funds and technologies, forming an orderly yet complexed system in which the core innovation entities, auxiliary innovation entities and the innovation environment coexist in an open, sharing, competitive, cooperative and higher yielding environment. Therefore, a green innovation ecosystem is constructed in this paper to analyse its core subjects, auxiliary subjects and their roles, and external supporting environment. The evolution process and driving forces of core enterprises’ green innovation from the perspective of the green innovation ecosystem is discussed to set a referential example for core enterprises that seek sustainable development.

The green innovation ecosystem is composed of green innovation core entities, green innovation auxiliary entities, the supporting environment for green innovation and system boundaries, as shown in Figure 1. The first three layers are interconnected and exchange elements with extremely high efficiency.

Fig. 1

The core entities include core enterprises, upstream and downstream enterprises in the green supply chain, universities, scientific research institutes, alternative enterprises, complementary enterprises and users, which lay the foundation of the green innovation ecosystem. Through the green supply chain, the knowledge chain, and the capital chain, they are connected with every entity in the whole process of green research and development, technology application and product production. The upstream and downstream enterprises of the green supply chain provide the core enterprises with raw materials and export green products, ensuring the efficiency of green product production. Universities and research institutes are the major players in knowledge innovation, where they help to transmit knowledge resources to the system and promote the research and development of green technology. Alternative enterprises and complementary enterprises are powerful competitors and helpers of core enterprises, which stimulate them to carry out green innovations. Users are the demanders of green products who can provide core companies with feedback on actual market demand.

Auxiliary entities include governments, financial institutions and intermediary agencies. Governments play a vital role in supporting the operation of core enterprises through imposing environmental regulations, handing out green innovation subsidies, and encouraging green procurement. Financial institutions provide financial support to core enterprises, which has a significant positive effect on the development, production, sales and promotion of green products. Intermediary agencies offer information consultation, which helps improve the efficiency of the operation of the entire system, and coordinate innovation resources and connect innovation entities.

Environmental support includes the economic level, knowledge environment, institutional guarantee, modern service and market development. Together, they form the environmental elements underpinning the green innovation ecosystem, which helps stimulate the willingness of core enterprises to embrace green innovation, standardise the green innovation process of core enterprises, and sustain healthy and orderly development of the ecosystem.

All circles in the green innovation ecosystem are interconnected, exchanging resources and information with each other, and jointly improving the overall efficiency. There is a relationship of competition, cooperation and symbiosis among them, which drives the formation of a green innovation ecosystem. The improvement of core entities’ green innovation and their collaboration with multiple subjects to broaden the limits of cooperation have also played a key role in shaping a green innovation system. Green innovation ecosystem is characterised by openness, ecology, systematism, symbiosis and evolution. It is precisely because of these characteristics that the system can realise the close interaction of innovation subjects and the full flow of innovation elements.

According to its responsibility annual report, the core entities in BYD's green innovation ecosystem include the government, industry/standards associations, financial institutions/media, non-governmental organisations and communities, and research institutions/academics. As shown in Figure 2, the focus of each entity on green innovation is different. BYD as the builder of the green innovation ecosystem forms a dynamic, complex and coupled relationship with other players [19]. Collaborative innovation and close cooperation among them have jointly improved the quality of BYD's green innovation.

Fig. 2

Established in 1995, BYD targeted its business on batteries and became the second-largest rechargeable battery manufacturer in the world. In 2008, BYD produced the first dual-mode electric model F3 DM. In 2013, by acquiring 77% of Qinchuan Automobile's shares, BYD officially entered the automobile industry and the new energy automobile industry, embarking on a new journey of manufacturing home-grown brands. In March 2021, Chuanfu Wang, chairman of the company, proposed to focus on the digital transformation of automobiles and carried out in-depth cooperation with partners in intelligent networked systems, supply chain informatisation, and intelligent manufacturing. BYD's green innovation has undergone many changes. Table 1 shows some of the landmark events in the process of its green innovation. Both the new energy vehicle business and the ‘smart city’ project have a key impact on the evolution of its innovation.

Since July 2005, BYD has gradually established a green supply chain, and it has released a series of files such as “Requirements on BYD's suppliers”, “Development, Assessment and Management Procedures for Productive Material Suppliers” and “Audit and Management Rules for BYD's Suppliers”, which has made strict requirements and clear operating guidelines for suppliers’ environmental materials management. For the raw material end of the supply chain, BYD has always committed to green procurement and has established a procurement system, characterised by green suppliers and green raw materials, with the procurement for its headquarters as the guide and departments, branches, and factories as the main force, which sets a standard for the green management in the production process, ensuring that every outsourcing component can meet the standard of environmental protection. In the procurement and management of the green supply chain, BYD strengthened the orderly supply of greens in all links of the supply chain, under the environmental policy and development strategy, and built a cooperation bridge for mutual benefit, win-win and common development with all suppliers. In this way, the best cost resources and efficiency have been secured, and its own competitiveness has been properly sustained and improved.

For the downstream enterprises of the core enterprises in the green innovation ecosystem, BYD actively establishes a coordinated relationship with them. In April 2021, the first batch of BYD's new pure electric tractor Q3 purchased by Budweiser, known as the ‘King of Beer,’ was put into operation in Wuhan, Tangshan and other factories, striving to achieve the company's green logistics carbon emission reduction by 2025. The goal is to continue to promote low-carbon logistics development with green technology.

As it quickens its development, BYD also faces threats from competitors. In 2017, SAIC Roewe launched the Internet car RX5, which achieved sales of 100,000 units in three months. In the same year, BAIC New Energy Automobile completed sales of 103,200 units. In 2018, more car manufacturers launched their vehicles, such as Xiaopeng Motors, Weilai, Weimar and other car companies. The unceasing launch of green vehicles by other car manufacturers constantly erodes BYD's leading advantages. However, BYD is also continuously increasing R&D investment, striving to achieve an absolute advantage in green innovation.

BYD has always adhered to three green dreams: electric vehicles, energy storage power stations, and solar power stations. They provide solutions to exhaust gas pollution in cities, ‘electricity storage,’ a worldwide hot potato, and the shift of human energy structure respectively. The ultimate goal is to achieve sustainable development.

As it pursues its green dream, BYD has been committed to research and development, during which, the company established the Central Research Institute, Electronic Technology Research Institute, Verdi Battery Research Institute, Power Research Institute, Automotive Engineering Research Institute, Product Planning and New Automobile Technology Research Institute, Ford Power Research Institute, Ford Science and Technology Research Institute, Commercial Vehicle Research Institute, Light Rail Transit Research Institute, and Semiconductor Research Institute. There are 11 research institutes in total. Besides, it created a talent pool of more than 20,000 highly sophisticated technical personnel who conducted R&D and made innovations in new materials, automobiles, new energy, rail transit and other fields, giving the industry a considerable leg-up. In recent years, BYD has continued to increase R&D investment, developed several green technologies and many green products, which plays a key role in promoting the process of green innovation.

To sum up, the evolution of green enterprise innovation is a process, jointly driven by the core entities in the green innovation ecosystem, formed when core enterprise supply chains and potential competitors are connected together. The internal transformation and upgrading of core enterprises, technological changes and the coordination of the main structure also fall under this process.

As the green innovation process is continuous, the input will affect current output and future output. Therefore, there is a certain connection between current output and future output. As an output indicator of innovation activities, patent statistics are unique and open. Scholars have begun to pay attention to patent statistics and use them as a measure of innovation activities [20]. The output of green innovation activities is classified into two types: expected output and undesired output. Expected output is the key output of core entities. Therefore, this paper evaluates the expected output of the company's green innovation activities by counting the number of green patent authorisation. Using the green standard IPC classification numbers listed in the ‘Green List of International Patent Classifications’ issued by the World Intellectual Property (WIPO) and ‘BYD Company Limited’ as keywords, the author searches the number of green output authorisation on the national intellectual property official website, and then makes a classification and summary [21].

ARMA model is formed based on the autoregressive model and moving average model. It has three basic forms: autoregressive model, moving average model and mixed model. It can predict the future according to the existing data. Besides, the ARMA model is one of the most commonly used linear models in time series and is often used for the short-term analysis of time series [22]. Sun Wenqing [23] used ARMA(1,1) to study the supply chain performance of perishable products. Through analysing BYD's patent authorisation data, an ARMA model is constructed in this paper. And through empirical study, the author found out how the current patent authorisation influences the future patent authorisation. As BYD secured green patent authorisation in 2002, the time span selected in this article is from 2002 to 2019.

For sample data, the k-order autocorrelation regression equation for estimation is suggested to adopt: (1) yt=β0+β1yt−1+…+βkyt−k+εt y_{t} = \beta_{0} + \beta_{1}y_{t-1} + \ldots{} + \beta_{k}y_{t-k} + \varepsilon_{t}

Another time series model is the ‘moving average process,’ the first-order moving average process is MA(1): (2) yt=μ+εt+θεt−1 y_{t} = \mu + \varepsilon_{t} + \theta \varepsilon_{t-1}

Assuming that {ε} obeys independent and identical distribution and obeys the normal distribution N(0,σ²), if ε_t−1 is known, then: (3) yt|εt−1~N(μ+θεt−1,σε2) y_{t}|\varepsilon_{t-1} \sim N \left(\mu + \theta \varepsilon_{t-1},\sigma^{2}_{\varepsilon}\right)

In general, for the q-order moving average process, it is denoted as MA(q): (4) yt=μ+εt+θ1εt−1+θ2εt−2+…+θqεt−q y_{t} = \mu + \varepsilon_{t} + \theta_{1}\varepsilon_{t-1} + \theta_{2}\varepsilon_{t-2} + \ldots{} + \theta_{q}\varepsilon_{t-q}

To achieve a better effect of regression, the autocorrelation process and the moving average process are combined to consider, and it is ARMA(p,q): (5) yt=β0+β1εt−1+…+βpεt−p+εt+θ1εt−1+…+θqεt−q y_{t} = \beta_{0} + \beta_{1}\varepsilon_{t-1} + \ldots{} + \beta_{p}\varepsilon_{t-p} + \varepsilon_{t} + \theta_{1}\varepsilon_{t-1}+ \ldots{} + \theta_{q}\varepsilon_{t-q}

First, the software is used to perform autocorrelation and partial autocorrelation tests on the amount of green patent authorisation (zls) and its residuals. The results are shown in Tables 2 and 3.

According to Table 2, the first-order Q statistic of the variable is noticeable, indicating that the AR(1) model is appropriate. The residuals 1–5 are not significant, indicating that there is no autocorrelation between the residuals, and the ARMA(1,1) model can be used to make an evaluation. The estimation results of AR(1) and ARMA(1,1) are shown in Tables 4 and 5.

From Table 3, the regression results in ARMA(1,1) are insignificant, that is, it is more appropriate to use AR(1) for regression estimation of the time series data of green patent authorisation. The specific results are shown in Table 4. In the AR(1) regression results, the first-order coefficient of green patent grants is positive, indicating that the current patent grants positively promote the next patent grants, reflecting that most of the innovation input is transformed into the output of the current year. It has an obvious promoting effect on the future output, but the proportion of this effect is relatively low. This process also reflects the dynamic and continuity of green innovation.