Digital economy is an economic concept that views digital knowledge as a key factor of production and looks at modern information and communication networks as the main carrier of digital knowledge (Purnomo et al., 2022). The digital economy plays a crucial role in mitigating market imperfections, improving economic efficiency and optimizing the industrial structure. The existing definitions of the concept of digital economy are summarized by a number of researchers (Williams, 2021; Zhang et al., 2021). The core characteristics of the digital economy are systematized in (Borremans et al., 2018; Ding et al., 2021). The digital economy consists of three main components: digital infrastructure, digital technologies in economic sectors, and e-commerce. Digital infrastructure ensures the connectivity of economic agents; digital technologies and solutions transform virtually all aspects of production and consumption; and e-commerce includes the exchange of economic resources using platforms and reduces transaction costs (Akberdina and Barybina, 2022). It is worth noting that the development level of the digital economy directly correlates with the level of the material sphere. The digital economy is a superstructure over the material sector of the economy and allows increasing the efficiency of any interaction. Hence, if digital technologies are introduced in the context of the insufficient development of material production, the cumulative economic effect of digitalization will not be of decisive importance.

The digital economy in a country covers information technology, software, mobile communications, and data transmission. There is quite a lot of studies on various aspects of the digital economy in Russia (Akberdina, 2018; Basaev, 2019; Ziyadullaev et al., 2019; Belokurova et al., 2020; Gureev et al., 2020; Vlasov, 2020; Rudyk et al., 2022). The researchers note that the digital economy in Russia is developing at a high pace, transforming industries and markets and penetrating into education and intellectual activity. At that, the digital inequality of the regions and the low share of their own digital technologies serve as development constraints.

Sustainable development, green economy, circular economy and ESG-concept are the components of a worldview advocating that a just economy should be built in accordance with both social and environmental dimensions, since the economy and the environment have a tremendous mutual influence on each other (Söderholm, 2020; D’amato & Korhonen, 2021). These concepts share a common thesis that a low-carbon, resource-efficient and socially inclusive economy should improve human wellbeing and social justice, while significantly reducing environmental threats and resource scarcity (Bouchoucha, 2021; Xie et al., 2023). The bibliometric analysis indicates that there is an upward trend in the number of research in the field of green economy, circular economy and sustainable development; however, there are country-specific differences in terminology (Ali et al., 2021). At that, all researchers tend to believe that the efficient use of resource, the circular economy, innovation, social integration, ecosystem protection, etc., contribute to the coordinated development of the economy, society and the environment and the achievement of sustainable development (Ozkan et al., 2023).

Sustainability and the green economy in Russia are also deeply investigated (Bobylev and Solovyeva, 2017; Zhironkin et al., 2017; Popkova et al., 2018; Karieva et al., 2020; Tulupov et al., 2020; Lavrikova et al., 2021; Kuznetsova et al., 2022; Tagaeva et al., 2022). The researchers highlight that Russia is rich in natural resources, which has historically formed an evolution model based on commodity exports. To shift to a new paradigm of economic development, the concept of sustainability with a balanced set of economic, social and environmental components should be included in the strategic documents underlying the country’s long-term development.

Researchers typically sharing similar views within each subject area, however, express serious disagreements on the impact the digital economy has on sustainable development (Adeshola et al., 2023). On the one hand, extensive studies have shown that the digital economy and the green economy develop in sync and positively influence each other (Wu et al., 2018; Kostoska and Kocarev, 2019; Vinuesa et al., 2020). Some works analyze the overall impact of the digital economy on the total productivity of green factors of production. Researchers emphasize that information technology can increase labor productivity and promote economic growth, which are in a positive correlation with the total productivity of green factors of production (Niebel, 2018; Nguyen et al., 2020; Wang et al., 2021; Wang et al., 2022c). A number of publications put the emphasis on the relationship between digitalization and energy consumption and conclude that digital technologies cause a decrease in energy intensity (Mughal et al., 2022; Sun, 2022). For example, it was found that with a 1% increase in the digital economy index, the number of developments in the new energy domain increases by an average of 0.2% (Wang et al., 2022a). Additionally, the digital economy not only creates conditions for clean energy to develop in countries with high carbon emissions (Wang et al., 2022b), but also helps to optimize the energy structure, increase energy efficiency (Li et al., 2021; Nikitaeva and Dolgova, 2022; Pierli et al., 2022; Xue et al., 2022; Akberdina et al., 2023) and reduce energy consumption.

On the other hand, a fairly large part of works is devoted to the inverse relationship between the digital economy and environmental pollution. For instance, researchers demonstrate that there are certain contradictions between smart digital cities and sustainable development goals (Martin et al., 2018), note that digitalization is not yet proved to be essential for reducing energy consumption and greenhouse gas emissions (Jin et al., 2018), and assume that digital equipment causes a lot of damage to the environment during production, maintenance and disposal (Kuntsman and Rattle, 2019). The main argument for the inverse relationship between the digital economy and reducing the burden on the environment is the fact that the use of digital technologies (big data, in particular) increases energy consumption (Van Heddeghem et al., 2014; Zhou et al., 2018). The researchers claim that the share of digital infrastructure in the national energy consumption can reach up to 10%–15%.

Sufficient funding is a vital prerequisite for a significant improvement in the state of the environment. Strictly speaking, sustainable development should be carried out amid the simultaneous progress in financial instruments. To handle this problem, various financing models are implemented (Cui et al., 2021; Sinha et al., 2021). Environmental protection funding was initially the state’s responsibility; however, in recent years, this function has been transferred to public-private partnerships leaving the state in charge of financing the relevant infrastructure (Ho and Park, 2019). In addition to PPP, the state actively encourages private investors to invest in environmental protection by providing tax incentives, grants and subsidies. Traditionally, there are two types of private investors—institutional and individual (Zhou et al., 2020; Akomea-Frimpong et al., 2022). Institutional investors are commercial banks, insurance companies, pension and public funds. Private capital is provided by interested companies. Recently, the market for green loans (Su et al., 2022) and green bonds (Tolliver et al., 2020) has been formed in the institutional segment. The evolution of the digital economy has led to the emergence of a new type of investor, i.e., crowdfunding platforms to finance environmental expenditures (Böckel et al., 2021).

In various studies, the term “environmental finance” is used as a synonym for such concepts as ”green finance” (Muganyi et al., 2021; Meo and Zhao, 2022), ”ecological finance” (Kihombo et al., 2021; Lee et al., 2022), ”sustainable finance” (Develay and Giamporcaro, 2023) or ”clean technology finance” (Madaleno et al., 2022). Originally, the term referred to the environmental economics paradigm and environmental investment. However, with the development of direct and derivative financing instruments, the growing impact of environmental problems and the tightening of environmental regulations (Cao et al., 2021; Feng et al., 2022) the scope of the term’s application has gradually expanded. Hence, the concept of environmental finance will be evolving adding new research aspects over time.

Publications on environmental financing in the Russian Federation cover the full range of issues identified above, focusing on the development of a green financial market and green risks (Ziyadin et al., 2019; Tulupov et al., 2020; Tyuleneva & Moldazhanov, 2020; Altunina and Alieva, 2021).

The existing literature on the digital economy primarily deals with measuring its level and effects. Currently, there is no single measurement method for selecting and evaluating indicators of the digital economy. Researchers mainly evaluate the digital economy and related indicators in terms of their specific tasks. The details of these methods are beyond the scope of the given study, but it is sufficient to refer to review articles (Bukht and Heeks, 2017). We are going to consider one of the indicators of the digital economy, namely, digital capital. This phenomenon is less popular among researchers if compared to the digital economy, and there are significant differences in studies with respect to the approaches used.

The first approach addresses digital capital from the perspective of an individual and in close connection with social and cultural capital (Resnick, 2004; Seale, 2012). These studies lie in the field of sociology and explore the extent to which people are involved in the use of digital technologies. Digital capital is interpreted as an individual’s digital technology ecosystem that determines how a user interacts with digital technologies. This characterizes the conditions for effective interaction between an individual and digital technologies, which he/she needs for their wellbeing in a digital society. The ability to purchase digital gadgets and software is a subset of an individual’s economic capital, and the material exchange takes place in areas where ICTs are used. Digital capital manifests itself in cultural capital in the form of digital skills, knowledge and competencies (Park and Park, 2017; Vartanova and Gladkova, 2020).

The second approach examines digital capital in the context of companies’ intangible assets (Crouzet and Eberly, 2019; Ayyagari et al., 2020; McGrattan, 2020; Tambe et al., 2020; Wu et al., 2020). Firms invest in both manufacturing and digital equipment to enhance their production capacity. ICT equipment (servers, routers, online shopping platforms and basic Internet software) acts as a tangible part of digital capital. In order to benefit from new technologies, digital-focused companies not only require investments in digital technologies but also in intangible assets. These intangible assets include staff training, new decision-making structures, and new business models to generate profits from digital activities (Eisfeldt and Papanikolaou, 2013; Bughin and Manyika, 2018). These investments often result in higher overall costs compared to the costs of digital technologies alone. These intangible assets make up the intangible part of digital capital. Similar to other forms of capital, digital capital can depreciate over time and needs to be replenished through additional investments. However, unlike tangible assets, the value of the intangible part of digital capital is closely tied to a specific company and is influenced by external economic conditions. As a result, the value of intangible assets tends to fluctuate more strongly than the value of tangible ICT assets, which are more easily exchangeable and have active secondary markets. As digital capital becomes an increasingly crucial component of a company’s overall capital reserves, differences in digital capital among firms can explain variations in the performance of new digital-focused companies compared to older firms. These differences in digital capital can be attributed to the accumulated reserves and variations in the marginal costs of investing in digital capital. In summary, the presence and management of digital capital play a significant role in determining the success and performance of digital-focused firms in comparison to traditional firms (Tambe et al., 2020).

The third approach to investigating digital capital lies in the field of the regional economy and characterizes the extent to which digital capital of a country or region is formed. The existing studies in this domain are not numerous. A number of publications on the assessment of the country’s digital capital as a combination of digital technologies and digital competencies explore its relationship with socio-economic and demographic characteristics such as income, age, education level, and place of residence, etc. (Ragnedda, 2018; Ragnedda et al., 2020). The techUK trade association holds a regular study of the Local Digital Capital Index (LDC Index) in the UK regions (LDCI, 2021; LDCI, , 2022). This index incorporates eight components, these are digital skills, digital technologies, data ecosystems, digital infrastructure, finance and investment, research and innovation, trade support, and cooperation. The LDC Index evaluates the impact that digital technologies can exert on the region, demonstrates its strengths and sets the direction for further development. The Index can be applied when formulating public policy to address a range of issues faced by the region and the entire country. The LDC Index also provides data to regional innovation ecosystems, including industry, government, universities and the public.

Despite the fact that the mutual impact of digitalization and sustainable development is being studied in depth, the issues of the relationship of financial flows underlying these processes have not been investigated. Our research should fill this gap, initiate such research, and substantiate the directions for clarifying public policy.

According to the above theoretical analysis and study design, to test the impact of the digital economy on environmental financing, we will use the least squares method (LSM) with clustering of annual standard deviation and individual fixed effects. Figure 2 presents the set of the tested models.

Hypothesis H1

is tested using model M1: E F i t = α 0 + α 1 D F i t + α 2 C t + ε i t (1) where E F i t denotes environmental financing of industry i in time period t; D F i t is digital financing of industry i in time period t; C t is a vector of control variables in time period t; ε t denotes random term; α 1 and α 2 are the coefficients to be estimated.

To test hypothesis H2, models M1.1 and M1.2 are used, respectively: E I i t = β 0 + β 1 D F i t + β 2 C t + μ i t (2) C E P C i t = γ 0 + γ 1 D F i t + γ 2 C t + δ i t (3) where E I i t denotes environmental investment of industry i in time period t; C E P C i t is current environmental protection costs of industry i in time period t; D F i t is digital financing of industry i in time period t; C t denotes a vector of control variables in time period t; μ t and δ t are random terms; β 1 , β 2 , γ 1 and γ 2 are the coefficients to be estimated.

Hypothesis H3

is tested using model M1.1.1: E I i t = ρ 0 + ρ 1 D I i t + ρ 2 C D C i t + ρ 3 D C C i t + ρ 4 C t + τ i t (4) where E I i t denotes environmental investment of industry i in time period t; D I i t denotes digital investment of industry i in time period t; C D C i t is current digital costs of industry i in time period t; D C C i t is digital competence costs of industry i in time period t; C t denotes a vector of control variables in time period t; τ t is random term; ρ 1 , ρ 2 , ρ 3 and ρ 4 are the coefficients to be estimated.

For testing hypothesis H4, model M2 is designed: E F i t = χ 0 + χ 1 D C t + χ 2 C t + ω i t (5) where E F i t denotes environmental financing of industry i in time period t; D F i t is digital capital in time period t; C t is a vector of control variables in time period t; ω t denotes random term; χ 1 and χ 2 are the coefficients to be estimated.

As indicated above, we offer our own index methodology to evaluate digital capital: D C t = D P t ∙ D B t ∙ D G t ∙ D I t 4 (6) where D C t is digital capital index in time period t; D P t is digital population index in time period t; D B t is digital business index in time period t; D G t denotes digital government index in time period t; D I t is digital interaction index in time period t.

Similar approach to integrating sub-indices into a composite index produced a positive outcome in the case of the digital space index (Akberdina et al., 2022).

The first sub-index D P t describes digital competencies of the population that are assessed through the following indicators: the number of mobile broadband Internet access subscribers per 100 population, the share of households with broadband Internet access, the share of the population that are active Internet users, the number of graduates in the program Computer Science and Computer Technology per 10,000 population, and the share of people employed in the ICT sector in the total number of the employed.

The second sub-index D B t is related to the level of companies’ digital transformation. The following indicators are applied to assess it: the share of organizations that provided additional training for employees in the field of ICT, the volume of investments in ICT, the share of organizations using Internet access at a speed of at least 2 Mbps, the share of organizations having special software to manage the procurement of goods (works, services), the share of organizations having special software to manage the sales of goods (works, services), the share of organizations using ERP systems, the share of organizations using CRM systems, and the share of organizations using electronic document management systems.

The third sub-index D G t shows the extent to which digital technologies have penetrated the sphere of public administration and is assessed using the following indicators: the share of public authorities and local governments using the Internet at a speed of more than 256 Kbps, the share of public authorities and local self-governments with a data transfer rate of at least 2 Mbps, the share of public authorities and local self-governments using electronic digital signature means, and the share of public authorities and local self-governments utilizing electronic document management systems.

The fourth sub-index D I t reflects the extent to which the interactions between the population, companies and the state are digitized. The following indicators are applied to assess it: the share of public authorities and local governments using automatic data exchange, the share of public authorities and local governments providing access to databases, the share of orders for state and municipal needs placed via electronic trading platforms, the share of the e-document management system in the interaction between public authorities, the share of organizations using the Internet to receive certain types of state and municipal services, the share of organizations using electronic data interchange between internal and external information systems, the share of organizations placing orders for goods (works, services) on the Internet, the share of organizations receiving orders for goods (works, services) via the Internet, the share of organizations using digital platforms, the share of the population using the Internet to get state and municipal services, and the share of the population using the Internet to order goods and (or) services.

The indicators of the sub-indices were normalized. The maximum value for each indicator over a period of time was equated to one, and the values for the remaining years were normalized in relation to it.

Finally, hypothesis H5 is tested using models M1.2 and M2.2: E I i t = ν 0 + ν 1 D C t + ν 2 C t + φ i t (7) C E P C i t = η 0 + η 1 D C t + η 2 C t + ψ i t (8) where E I i t is environmental investment of industry i in time period t; C E P C i t denotes current environmental protection costs of industry i in time period t; D C t is digital capital in time period t; C t is a vector of control variables in time period t; φ t and ψ t are random terms; ν 1 , ν 2 , η 1 and η 2 are the coefficients to be estimated.

Based on the research purpose and data availability, we have developed dependent variables, main independent variables, control variables, and intermediate variables. Their specific values, calculation methods and data sources are presented in Table 2.

The study used data from the Federal State Statistics Service of the Russian Federation (Rosstat) on environmental protection expenditures for 2012–2022 by industry, including investment and current costs. In the research, the data are given by types of industry-specific economic activity—in aggregate (sections B, C, D, E according to the OKVED-2 classifier [OKVED-2 is the Russian National Classifier of Types of Economic Activity]) and in detail (industry sectors—decimal codes according to the OKVED-2 classifier).

To perform regression modeling of the relationship between environmental financing and digital financing, the data for 2015–2021 were taken, since the statistics on digitalization by type of industry expenses has been collected only since the approval of the state program Digital Economy of the Russian Federation. To carry out the regression assessment of the relationship between environmental financing and digital capital, data for 2012–2022 were used. The indicators for calculating the digital capital index and sub-indices are presented in the Consolidated Monitoring of the Development of the Information Society in the Russian Federation, provided by the Federal State Statistics Service of the Russian Federation.

The first group of the research models dealt with the direct relationship between digital financing and environmental financing. Table 3 presents empirical results based on panel data for 31 Russian industries for 7 years (2015–2021). As we can see, the main model M1 (Eq. 1) gives quite good results: the regression coefficient of the impact of digital financing ( D F i t ) on environmental financing ( E F i t ) is positive and passed the test for significance at the 1% level.

We suppose that in this case the effect of digital technologies penetrating into environmental protection technologies is triggered. Digital technologies such as the Internet, the IoT, artificial intelligence, big data, digital twins, etc., are widely used in environmental protection, which changes resource consumption patterns, pollution reduction, and higher energy efficiency. These technologies are instrumental in analyzing big data obtained through environmental monitoring, automating management decision-making in real time, and predictive forecasting of potentially hazardous natural phenomena and objects. Owing to the use of GISs, Earth space sensing data and unmanned observation, it possible to control landfills, identify flora and fauna objects, etc. The effect of the digital economy is also evident when solving any engineering and environmental problem. Software for design and automation of technological preparation of production is of great importance for cleaner production to progress. Increasingly scrupulous attention is paid to the latest achievements in artificial intelligence and neural networks applied to produce optimal technological solutions, i.e., to optimize resource consumption, reduce emissions of harmful substances, and cut down energy consumption.

On the other hand, there is a substitution effect: outdated production technologies are replaced by new digital solutions, which ultimately leads to a significant decrease in environmental pollution and resource savings in industry. This, in turn, reduces the need for environmental facilities construction funding and lowers current environmental protection costs. For example, industrial robots replace human labor for automated production, intelligent design improves the efficiency of allocation of production factors and productivity. Digital technologies also contribute to reducing the volumes of raw materials required. With electronic sensors of various sizes, virtually any change in the production system’s operating state can be monitored. This allows not only tracking CO₂ emissions, but also controlling the level of emissions related to the company’s entire value chain. The effects of penetration and substitution are manifested in different growth rates of digital financing and environmental financing. As evidenced by the case of Russia, digital financing is increasing annually at a faster pace than environmental financing. This led to the fact that over 7 years the share of digital financing in GDP increased 1.8 times, while the rise in the share of environmental financing in GDP was only 1.3 times (Figure 3). This absolutely does not mean that the digital economy in Russia is prioritized over sustainable development; this is merely a manifestation of the abovementioned effects. Thus, we can conclude that hypothesis H1 has been confirmed.

Models M1.1 and M1.2 (Eqs 2, 3) were developed to test hypotheses about the impact of digital financing on the elements of environmental financing, namely, environmental investment and current environmental protection costs. Table 3 demonstrates the situation that we had predicted. There is a sustainable positive relationship between digital financing ( D F i t ) and environmental investment ( E I i t ). At the same time, when evaluating the relationship between digital financing ( D F i t ) and current environmental protection costs ( C E P C i t ), we can see that the regression coefficient is a positive number of 0.290, but it fails the test for significance indicating that there is no relationship between the indicators.

We argue that environmental investment and current environmental protection costs are of a different nature and determined by different factors. This is clearly illustrated in Figure 4.

The dynamic graphs of environmental investment and current environmental protection costs are configured in a completely manner. Current environmental protection costs are related to production volumes and resource consumption. Digital technologies have little effect on the costs associated with previous technological solutions. We believe that environmental investment is determined by the willingness of companies to invest and the availability of sufficient funding. Any investment in technology, therefore, will have unidirectional dynamics and a close statistical relationship. These arguments, in our view, support hypothesis H2.

Model M1.1.1 (Eq. 4) is a variation of model M1.1 and supposed to reveal the relationship of environmental investment ( E I i t ) with digital financing components, such as digital investment ( D I i t ), current digital costs ( D C C i t ) and digital competence costs ( C D C i t ). All the independent variables exert a positive effect on environmental investment, but only digital investment is of high significance, which confirms hypothesis H3. Digital financing is unevenly distributed across different industries. Our study has shown that 75% of the funds allocated for digitalization were distributed between 10 industries (in OKVED, industry covers more than 30 types of activities). Among the sectors leading in investment in industry digitalization are the energy industry, production of petroleum products, gas and oil production, metallurgy and production of metal products, electronics and machine tool manufacturing. At that, it is these industries that generate the main investment inflow in new environmental technologies and set environmental goals.

However, we prove that there is a relationship between environmental investment and GDP, but it is rather weak. Moreover, this relationship is significant only in the regression without a time lag, while a 1-year lag notably worsens the regression values (Figure 5).

In 2021, we analyzed the structure of digital financing in Russia. Of the total funds allocated for digital transformation in industry, 72% were directed to internal expenses, such as the purchase of information and communication equipment, software, staff training, etc. The remaining 28% of the budget was allocated to external expenses, such as digital equipment rental, software, technical support, and database access. The internal costs of digital transformation were involved in the acquisition of digital machinery and equipment, and 40.2% of these were associated with the purchase of computers and office equipment. However, the share of digital production equipment purchases remained small. Software accounted for 22% of the total domestic digitalization budget in 2021. Employee training comprised just 3% of all internal digital transformation spending. In industry, there are practically no costs incurred in the formation of digital context. Thus, the internal funds for digital transformation exceed the external ones, and most of them are aimed at purchasing hardware and software. Employee training and creating digital context in industry remain less significant expenses.

In the second part of our study of the relationship between the digital economy and environmental financing, we applied the digital capital index, for which we had previously proposed a definition and an assessment method. The digital capital index characterizes the environment, where technological segments of traditional industries develop. To assess the impact of the digital capital index ( D C t ) on environmental financing ( E F i t ), we substantiated model M2 (Eqs 5, 6) and related models M2.1 and M2.2 (Eqs 7, 8), i.e., the impact of the index on environmental investment ( E I i t ) and current environmental protection costs ( C E P C i t ). The results of modelling are presented in Table 3. As can be seen, the regression coefficients and their significance are worse than models including digital financing; however, there is an overall positive relationship. Models M2 and M2.1 showed a more significant statistical relationship, which can be due to the fact that current environmental protection costs are in principle insensitive to anything other than production volumes.

The digital capital index has an indirect impact due to the cumulative synergistic effect of digitalization of the population, companies and the state. As follows from Figure 6, in 2022 the digital capital index in Russia was 0.835 out of the maximum possible value of 1. If the digital population index is close enough to the maximum value, and the digital business index and the digital government index are 18%–19% behind the maximum value, then the digital interaction index remains at a relatively low level of 0.744. With the growing importance of factors affecting the digital interaction index (e.g., the share of companies receiving orders via the Internet, the share of companies using digital platforms, the share of the population using the Internet to order goods and services, etc.), its contribution to the digital capital index will increase, so will the importance of the environmental financing index. These arguments support hypotheses H4 and H5.