Digital financial inclusion aims to bring financial services to a wider range of people and businesses at a much lower cost, using big data and cloud computing to capture and share information. At the same time, the Chinese government aims to build a green and sustainable economy. Therefore, this study analyzes the impact of digital financial inclusion on green economic efficiency and identifies the moderating role of regional competition based on the empirical analysis of data from 265 prefecture-level cities in China from 2010 to 2017. Our results indicate that (1) digital financial inclusion promotes the green economy, which has a significant positive spillover effect, (2) regional competition is beneficial for green economy development, but the interaction of digital financial inclusion and regional competition is detrimental, and (3) digital financial services make the largest positive contribution to the green economy, and digital payment services have the largest negative effect.
1. Introduction
China has gradually realized that the intensive consumption of natural resources and energy is unsustainable and is now seeking green and sustainable development. Specifically, green economic efficiency (GEE) includes labor, capital, and energy as inputs and uses pollution as an undesired output and production as an intended output. GEE comprehensively considers the costs and benefits of economic growth, resource conservation, and the environment (Qian and Liu, 2013) and is a good measurement for the quality of economic growth. Two important ways to achieve green economic efficiency are government environmental policies and financial instruments.
Digital financial inclusion emerged in 2004 with the application of Alipay serving as a new financial instrument to provide online credit, insurance, and payment services. Because China has not yet built an inclusive credit system, the low-income bracket and small businesses have found it difficult to obtain credit from banks. Digital financial inclusion uses big data, artificial intelligence, and cloud computing technology to obtain, analyze, and share credit information of low-income customers and small businesses, effectively reducing transaction costs, improving the effectiveness of risk control, and expanding the scope of financial service supply. Generally, digital financial inclusion provides more people and businesses with cheaper and better services.
In China, local governments are always in regional competition for capital. As a result, these governments have come to control and treat local financial institutions as assets to attract FDI and skilled labor. Such competition boosts productivity (Liu and Tsai, 2018) and innovation (Liu et al., 2021) and optimizes the economic structure (Li et al., 2021). However, competition also means that many of the benefits of new growth can hollow out local areas as banks and investors leave. Local governments thus lose part of their control over the financial system, which is a heavy loss, especially for developing regions. This may in turn undermine ongoing green growth, not to mention the inclusion of lower income clients and small businesses. Our research questions are as follows: How does digital financial inclusion influence green economy growth? Does this effect change in the context of regional competition? This article tries to answer these questions, aiming to provide policy implications for a greener economy.
We explored and validated the impact of digital financial inclusion on the green economic efficiency based on the analysis of data from 265 Chinese prefecture-level cities from 2010 to 2017. Specifically, we paid attention to several digital financial instrument and analyzed their heterogeneous impact on green economic efficiency from a demand side.
This study makes the following contributions to the existing studies. First, our article explored the opposite effects of several digital financial instruments on green economic efficiency and tried to explain the heterogenicity with usage depth. Among those studies considering digital financial inclusion, most have examined its effects on economic growth (Mohan, 2006; Kim, 2016) and green innovations (Liu et al., 2021), and the effect of digital financial inclusion is analyzed from the supply side. Our article discussed the users’ response to easier and cheaper access to digital financial services. We also figured out major determinates that significantly influence the interaction between digital financial instruments and regional competition.
Second, our article concludes that at this moment, the interaction of digital financial inclusion and regional competition is harmful to green economic efficiency and offers an explanation for the deviation from Liu et al. (2021), which advocated for moderate regional competition while the right amount of regional competition is ambiguous. We point out, in order to make full use of digital financial inclusion, digital credit instruments should be promoted.
2. Background, literature, and hypotheses
2.1. Background
With the growing demand for cheaper and better financial services and the wide usage of big data and cloud computing, online payments and P2P credit services, such as Alipay and Jingdong Finance, emerged. These new platforms signaled that digital financial inclusion is gradually replacing traditional financial instruments and institutions. In addition, the central government introduced a series of policies to encourage digital financial inclusion. Alipay, the most popular digital financial service provider, works in the following way. Alipay rates users by automatically collecting their information to build a credit profile when they make payments with the platform, and the level of financing and insurance rates are provided accordingly. People have easy and quick access to these financial services at relatively low cost anywhere as long as they have a phone.
Aiming at a sustainable society, green development is a new way to boost the economy while staying within the ecological and resource carrying capacity. The government, firms, and citizens collaborate to realize green development. Governments propose policies that favor energy-saving technology innovations and applications. Firms that are seriously polluting are shut down, and new technologies are applied to increase resource productivity. Citizens are educated to recycle their clothing and household appliances. Recently, the imbalance of economic development under resource constraints has been widely discussed and measured with green total factor productivity (Jensen et al., 2012; Chen and Golley, 2014; Li and Wu, 2017) and green economic efficiency (Li, 2019; Ren et al., 2019). A high level of green economic efficiency indicates a good balance between the economy and the local environment and a greater output with fewer inputs and less pollution (Ye et al., 2021).
2.2. Digital financial inclusion and green economic efficiency
There is some debate about the relationship between financial conditions and green economic efficiency. Some studies have claimed that financial development promotes green economic growth (Lei et al., 2021; Penasco et al., 2021) because more innovative enterprises obtain loans and invest in green innovations (Mohan, 2006; Kim, 2016). However, other research has claimed that financial development has not had a significant impact on green economic efficiency because the coverage of financial services has not expanded and few innovative firms have emerged (Babajide et al., 2015; Dubois and Allacker, 2015).
Digital financial inclusion emerged as a new way to provide consumers and manufacturers with online insurance, financing, and payment services. It affects green economic efficiency in the following way: On the supply side, digital financial inclusion reduces information asymmetry, and small innovative enterprises find it easier to obtain loans and thereby invest more in green innovation (Mohan, 2006; Kim, 2016), which in turn consumes less energy and produces less pollution. On the demand side, consumers can obtain cheaper financing and investment services with higher returns via digital financial inclusion. Consumption, especially in environmentally friendly forms, such as tourism and entertainment, will increase. In this case, green economic efficiency can increase.
What is more, according to the first law of geography (Tobler, 1970), regions are interrelated and interdependent. A regional change spills over to surrounding regions, as does green economic efficiency (Ren et al., 2019; Liu et al., 2021), in the following way. When a region’s green economic efficiency is high, it enjoys advances in technology (Ye et al., 2021), a large market, and substantial amounts of capital (Yu et al., 2021). New technology and green innovation spread and are applied rapidly in neighboring regions (Zeng et al., 2020); a large market resulting from this larger geographic distribution produces increased demand, and substantial capital flows then enter surrounding areas (Wu et al., 2021). We conclude that the green economic development of one region will boost that of neighboring areas. Therefore, this article proposes the following hypothesis:
Hypothesis 1 (H1): Digital financial inclusion promotes green economic efficiency and green economic efficiency has a positive spatial spillover effect.
2.3. The moderating effect of regional competition
Generally, regional competition may impact green economic efficiency in a simple way. Local governments in China have to realize goals in both economic and environmental growth (Zhang et al., 2020). Higher productivity and less pollution are undoubtedly positive factors encouraging stronger regional competitiveness. In this sense, regional competition results in greater green economic efficiency (Mundaca et al., 2016). However, things become complicated when digital financial inclusion is involved, in that the economic goal may be reachable in the short run while the environmental goal may not be.
To lead in regional competition, government investment is always a good way to boost employment and the economy (Penasco et al., 2021). Chinese local governments are self-supported by tax revenues, and there is always a fiscal gap for such investment. A typical solution is to interfere in or even control the local financial system to meet the financial needs of local governments by providing implicit guarantees for enterprises or controlling resources, including fuel, minerals, and land (Liu et al., 2021). If the economic goal rather than the environmental goal is prioritized, then digital financial inclusion will not be welcomed by governments because it squeezes out local banks through better services and lower prices. Such competition weakens digital financial inclusion’s impact on green economic efficiency. Therefore, this article proposes the second hypothesis:
Hypothesis 2 (H2): Regional competition is beneficial for green economic development and weakens the positive effect of digital financial inclusion on green economic efficiency.
3. Methods and data
3.1. Measurement of green economic efficiency
The super slack-based measurement (SBM) model, which is a modified version of the SBM model (Tone, 2002), is primarily employed to measure productivity with respect to unexpected outputs, such as environmental pollution. So, this model is applicable to the research questions in this article.
Green economic efficiency is a comprehensive indicator used to measure the effectiveness of a country or region’s allocation of resources in economic activities. With reference to Liu and Gong (2018), each prefecture-level city is regarded as an independent production decision-making unit, with inputs represented by and expected outputs by b, as well as undesired outputs by c, . We denote by u the input and output of the jth city at time t, where . We then construct a production possibility set for green economic efficiency:
Based on Tone (2002), the super SBM model we constructed is as follows:
where the slack variables for input, desired output, and undesired output are represented by , , and , respectively, while the weight vector is denoted by λj. The target function, denoted by ρ*, reflects the efficiency of the economy, with a higher value indicating greater efficiency.
3.1.1. Spatial weight matrix
The spatial weight matrix includes 4 forms: spatial adjacent weight matrix, geographic distance weight matrix, economic distance weight matrix, and nested spatial weight matrix. These forms quantify the level of interdependence between specific economic or geographic attribute values across different cross-sectional spatial units. The spatial diffusion of economic efficiency is influenced significantly by both geographical distance and economic differences (Peri, 2005; Thompson and Fox-Kean, 2005; Singh and Marx, 2013; Li, 2014). To this end, drawing on the methodology used by Lin et al. (2005), we adopt a nested spatial weight matrix, which has the following form:
where
where wd refers to a matrix that weights geographic locations, dij refers to the distance between two cities calculated according to latitude and longitude, and we refers to a matrix that weights economic distances. The local economy Y is measured by gross domestic product (GDP) per capita, and the smaller the economic gap between 2 regions, the greater its weighting.
3.2. Empirical strategy
Since digital financial inclusion affects not only the local economy but also the strategies of neighboring regions, we used the spatial Durbin model. This model accounts for the spatial correlation between the dependent variables, as well as the relationships between the explanatory variables and the error term:
where ρ reflects the extent of mutual influence of the endogenous variable gee in neighboring areas, β and θ are the vectors with k × 1 dimension that represent the estimated parameters, and w is a nonnegative spatial weight matrix with a dimension of N × N, and the spatial effect, time effect, and error term are represented by σi, ut, and εit, respectively.
We added an interaction term to Equation 4 reflecting the interaction effect of digital financial inclusion and regional competition, denoted as dfi×rc, and obtained the following equation:
To test the application of the spatial Durbin model, we used the Lagrange multiplier (LM) test, Wald statistic test, and likelihood ratio (LR) test.
Furthermore, in order to classify the heterogeneous effects of digital financial inclusion on green economic efficiency, the following model is constructed based on Model 5:
where dfi_x are the insurance index (dfi_ins), credit index (dfi_cre), and payment index (dfi_pay), respectively.
3.3. Variables and data
3.3.1. Measurement of green economic efficiency and data
Green economic efficiency, which comprehensively considers the cost of economic growth, resource conservation, and environmental impact based on traditional input–output economic efficiency (Qian and Liu, 2013), is a good proxy for the quality of economic growth. As shown in Table 1, its inputs include labor and capital, and its output includes both desired and undesired outputs. Real GDP is the desired output, for which 2002 is used as the base period and the consumer price index of each region is deflated; unexpected output is industrial wastewater, SO2 emission, and dust (Kaneko and Managi, 2004).
Measurement of the green economic efficiency
. | Input . | Desired Output . | Undesired Output . | |||
---|---|---|---|---|---|---|
Indicator | Labor | Capital | GDP | Industrial Waste Water | Industrial SO2 Emission | Industrial Dust |
Unit | 10,000 | 100 million | 100 million | 10,000 tons | 10,000 tons | 10,000 tons |
Source | China Statistical Yearbook | China Statistical Yearbook | China Statistical Yearbook | China Statistical Yearbook | China Statistical Yearbook | China Statistical Yearbook |
. | Input . | Desired Output . | Undesired Output . | |||
---|---|---|---|---|---|---|
Indicator | Labor | Capital | GDP | Industrial Waste Water | Industrial SO2 Emission | Industrial Dust |
Unit | 10,000 | 100 million | 100 million | 10,000 tons | 10,000 tons | 10,000 tons |
Source | China Statistical Yearbook | China Statistical Yearbook | China Statistical Yearbook | China Statistical Yearbook | China Statistical Yearbook | China Statistical Yearbook |
In this article, we use panel data from 265 cities at the prefecture level in China from 2010 to 2017. The statistical description is shown in Table 2. The data are taken from the Peking University Digital Financial Inclusion Index, the China Statistical Yearbook, the China City Statistical Yearbook, and the Report on City and Industrial Innovation in China (2017). Missing data were completed with linear interpolation.
Statistical summary
Variables . | Unit . | Mean . | SD . | Minimum . | Maximum . | Observation . |
---|---|---|---|---|---|---|
Green economic efficiency (gee) | / | 0.146 | 0.160 | 0.018 | 1.000 | 2,120 |
Use depth of digital financial inclusion (dfi) | / | 1.435 | 0.607 | 0.125 | 3.257 | 2,120 |
Insurance service (dfi_ins) | / | 2.675 | 1.297 | 0.014 | 5.882 | 2,120 |
Credit service (dfi_cre) | / | 1.086 | 0.427 | 0.000 | 1.953 | 2,120 |
Payment service (dfi_pay) | / | 1.460 | 0.725 | −0.470 | 3.760 | 2,120 |
Regional competition (rc) | 1,000 million | 0.784 | 1.428 | 0.000 | 14.005 | 2,120 |
Regional financial development (rfd) | / | 1.341 | 0.565 | 0.371 | 7.203 | 2,120 |
Population density (pd) | 100 people/km2 | 4.451 | 3.176 | 0.051 | 26.481 | 2,120 |
Industrial structure (inst) | / | 0.487 | 0.096 | 0.020 | 0.821 | 2,120 |
Innovation capacity (ic) | / | 0.119 | 0.410 | 0.000 | 8.085 | 2,120 |
Environmental governance (eg) | / | 0.150 | 0.121 | 0.000 | 1.000 | 2,120 |
Variables . | Unit . | Mean . | SD . | Minimum . | Maximum . | Observation . |
---|---|---|---|---|---|---|
Green economic efficiency (gee) | / | 0.146 | 0.160 | 0.018 | 1.000 | 2,120 |
Use depth of digital financial inclusion (dfi) | / | 1.435 | 0.607 | 0.125 | 3.257 | 2,120 |
Insurance service (dfi_ins) | / | 2.675 | 1.297 | 0.014 | 5.882 | 2,120 |
Credit service (dfi_cre) | / | 1.086 | 0.427 | 0.000 | 1.953 | 2,120 |
Payment service (dfi_pay) | / | 1.460 | 0.725 | −0.470 | 3.760 | 2,120 |
Regional competition (rc) | 1,000 million | 0.784 | 1.428 | 0.000 | 14.005 | 2,120 |
Regional financial development (rfd) | / | 1.341 | 0.565 | 0.371 | 7.203 | 2,120 |
Population density (pd) | 100 people/km2 | 4.451 | 3.176 | 0.051 | 26.481 | 2,120 |
Industrial structure (inst) | / | 0.487 | 0.096 | 0.020 | 0.821 | 2,120 |
Innovation capacity (ic) | / | 0.119 | 0.410 | 0.000 | 8.085 | 2,120 |
Environmental governance (eg) | / | 0.150 | 0.121 | 0.000 | 1.000 | 2,120 |
3.3.2. Variables
Digital financial inclusion (dfi): The direct result of promoting digital financial inclusion is the depth of use of digital financial inclusion. This article uses the index of the depth of digital financial inclusion (Guo et al., 2020) as a proxy variable. To keep the data matched, the index is divided by 100. To identify the impact of different types of digital financial inclusion on the efficiency of the green economy, we further use insurance (dif_ins), credit (dfi_cre), and payment (dfi_pay) as dependent variables.
Regional competition (rc): Researchers have constructed a comprehensive index of regional competition (Fu et al., 2016; Luo and Peng, 2016). Such indices reflect patterns of regional competition, but complex indices lead to collinearity. Regional competition will eventually turn into a competition for capital, and foreign capital is the main object of such competition (Bai and Nie, 2018) and a competitive source for development. We used foreign capital used in the focal year (100 million U.S. dollars; Li and He, 2018) as the proxy variable for regional competition.
Control variables include the following: (1) The level of regional financial development (rfd), expressed as the ratio of the balance of deposits in financial institutions at the end of the year (10,000 yuan) to the regional GDP (10,000 yuan), (2) population density (pd) was set to 100 people/km2, (3) industrial structure (inst) is expressed as the share of secondary industry value added in GDP, (4) innovation capacity (ic) is defined by the city innovation index in the China Cities and Industry Innovation Report 2017, divided by 100, and (5) environmental governance (eg) is the ratio of reduced industrial sulfur dioxide emissions (tons) in real GDP.
4. Results
4.1. Spatial correlation test
Existing studies have shown that geographical distance and economic distance are the important factors that shape regional competition. We conducted both the LM test and robust LM test to verify our spatial analysis. The null hypothesis is the nonspatial panel model, and the alternative hypotheses are the spatial autoregressive model (SAR) or the spatial error model (SEM). To test whether there is spatial correlation, an ordinary panel model is used for the LM test. If the result rejects the null hypothesis, the SAR model or the SEM should be selected. If the result accepts the null hypothesis, the ordinary panel regression model should be applied. The LM test of the spatial interaction form should be conducted based on fixed effects, and the possible coexistence of spatial effects and time effects can affect the outcome of the test. Therefore, we need to select for effects before the LM test.
Our results (Table 3) showed that there are spatial fixed effects. The results of the LM test indicate a spatial correlation in the model. We further conducted Wald statistics and LR statistics analysis to test the form of the model. Here, the null hypothesis is that SAR and SEM should be applied. If this null hypothesis is rejected, the spatial Durbin model is better.
Model spatial correlation test
. | (1) . | (2) . | ||
---|---|---|---|---|
Test . | Coefficient . | P Value . | Coefficient . | P Value . |
Lagrange multiplier (LM) test no spatial lag | 4.311 | 0.038 | 4.510 | 0.034 |
Robust LM test no spatial lag | 11.777 | 0.001 | 12.548 | 0.000 |
LM test no spatial error | 8.687 | 0.003 | 9.159 | 0.002 |
Robust LM test no spatial error | 16.152 | 0.000 | 17.197 | 0.000 |
. | (1) . | (2) . | ||
---|---|---|---|---|
Test . | Coefficient . | P Value . | Coefficient . | P Value . |
Lagrange multiplier (LM) test no spatial lag | 4.311 | 0.038 | 4.510 | 0.034 |
Robust LM test no spatial lag | 11.777 | 0.001 | 12.548 | 0.000 |
LM test no spatial error | 8.687 | 0.003 | 9.159 | 0.002 |
Robust LM test no spatial error | 16.152 | 0.000 | 17.197 | 0.000 |
The results of the model selection and Hausman tests are presented in columns (1) and (2) of Table 4. The random effects and fixed effects of Models (4) and (5) are tested, and the results of the Hausman test show that the null hypothesis is rejected at the 1% level. Overall, the conclusion is that the spatial Durbin fixed effects model is the best fit.
Model selection and Hausman test
. | (1) . | (2) . | ||
---|---|---|---|---|
Test . | Coefficient . | P Value . | Coefficient . | P Value . |
Wald spatial lag test | 78.848 | 0.000 | 90.577 | 0.000 |
LR spatial lag test | 89.601 | 0.000 | 102.587 | 0.000 |
Wald spatial error test | 70.984 | 0.000 | 82.579 | 0.000 |
LR spatial error test | 83.737 | 0.000 | 96.345 | 0.000 |
Hausman test | 95.496 | 0.000 | 134.635 | 0.000 |
. | (1) . | (2) . | ||
---|---|---|---|---|
Test . | Coefficient . | P Value . | Coefficient . | P Value . |
Wald spatial lag test | 78.848 | 0.000 | 90.577 | 0.000 |
LR spatial lag test | 89.601 | 0.000 | 102.587 | 0.000 |
Wald spatial error test | 70.984 | 0.000 | 82.579 | 0.000 |
LR spatial error test | 83.737 | 0.000 | 96.345 | 0.000 |
Hausman test | 95.496 | 0.000 | 134.635 | 0.000 |
4.2. Spatial Durbin model
Table 5 shows the results of the estimation of the spatial Durbin model. As shown in Table 5(a), the spatial lag coefficient ρ of green economic efficiency confirms the positive spatial effect of green economic efficiency; that is, the improvement of green economic efficiency in the region will promote the green economic efficiency of neighboring regions. The above results verify the spillover effect in Hypothesis 1. Table 5(a) also shows that the coefficient of digital financial inclusion is positive, indicating that the use of digital financial inclusion promotes the efficiency of the green economy, which validates Hypothesis 1. In addition, the region’s industrial structure, innovation capacity, and environmental governance have a significant promoting effect on green economic efficiency, while regional financial development and population density have a negative effect. In contrast, the population density, industrial structure, innovation capacity, and environmental governance of neighboring areas have a significant negative impact on the region’s green economic efficiency.
Spatial Durbin model
Variable . | dfi . | rc . | rfd . | pd . | inst . | ic . | eg . | dfi × rc . | ρ . | R2 . | |
---|---|---|---|---|---|---|---|---|---|---|---|
(a) | Coefficient | 0.0492** (2.3497) | 0.0074*** (2.9147) | −0.0331*** (−3.6312) | −0.0113** (−2.1686) | 0.2200*** (4.3202) | 0.0227** (2.3677) | 0.1222*** (7.6490) | 0.1430*** (3.1322) | 0.9022 | |
(b) | Coefficient | 0.0508** (2.4221) | 0.0139*** (3.0876) | 0.0349*** (−3.8301) | 0.0114** (−2.1287) | 0.2303*** (4.5343) | 0.0337*** (2.6082) | 0.1237*** (7.7716) | −0.004* (−1.7266) | 0.1450*** (3.1818) | 0.903 |
Variable | W × dfi | W × rc | W × rfd | W × pd | W × inst | W × ic | W × eg | W × dfi × rc | Log−L | σ2 | |
(a) | Coefficient | −0.0291 (−1.3399) | 0.0070 (0.5344) | −0.0249 (−0.9672) | −0.0276** (−2.2136) | −0.3662*** (−3.1466) | −0.1661*** (−6.4872) | −0.0900*** (−4.1454) | 2914.5497 | 0.0029 | |
(a) | Coefficient | −0.0374* (−1.7212) | 0.0513** (−2.4393) | −0.0129 (−0.4969) | 0.0247** (−1.9783) | 0.3545*** (−3.0541) | −0.289*** (−6.6031) | 0.0798*** (−3.6502) | 0.0277*** (3.5814) | 2922.6469 | 0.0029 |
Variable . | dfi . | rc . | rfd . | pd . | inst . | ic . | eg . | dfi × rc . | ρ . | R2 . | |
---|---|---|---|---|---|---|---|---|---|---|---|
(a) | Coefficient | 0.0492** (2.3497) | 0.0074*** (2.9147) | −0.0331*** (−3.6312) | −0.0113** (−2.1686) | 0.2200*** (4.3202) | 0.0227** (2.3677) | 0.1222*** (7.6490) | 0.1430*** (3.1322) | 0.9022 | |
(b) | Coefficient | 0.0508** (2.4221) | 0.0139*** (3.0876) | 0.0349*** (−3.8301) | 0.0114** (−2.1287) | 0.2303*** (4.5343) | 0.0337*** (2.6082) | 0.1237*** (7.7716) | −0.004* (−1.7266) | 0.1450*** (3.1818) | 0.903 |
Variable | W × dfi | W × rc | W × rfd | W × pd | W × inst | W × ic | W × eg | W × dfi × rc | Log−L | σ2 | |
(a) | Coefficient | −0.0291 (−1.3399) | 0.0070 (0.5344) | −0.0249 (−0.9672) | −0.0276** (−2.2136) | −0.3662*** (−3.1466) | −0.1661*** (−6.4872) | −0.0900*** (−4.1454) | 2914.5497 | 0.0029 | |
(a) | Coefficient | −0.0374* (−1.7212) | 0.0513** (−2.4393) | −0.0129 (−0.4969) | 0.0247** (−1.9783) | 0.3545*** (−3.0541) | −0.289*** (−6.6031) | 0.0798*** (−3.6502) | 0.0277*** (3.5814) | 2922.6469 | 0.0029 |
T values are in parentheses.
***, **, and * denote significance at the 0.01, 0.05, and 0.1 levels, respectively.
Table 5(b) shows the direct impact of regional competition on green economic efficiency and its moderating effect between digital financial inclusion and green economic efficiency. On the one hand, green economic efficiency benefits from regional competition. On the other hand, the negative interaction term suggests that the positive effect of digital financial inclusion on green economy efficiency is diminished by the effect of regional competition. The above conclusions further validate Hypothesis 2.
4.3. Effect on use depth
Digital financial inclusion mainly changed the business scope of insurance, credit, and payment services. Due to differences in the number of users, transaction volume, and transaction amount, each service has an impact on green economic efficiency. Therefore, we use a subcategory index that further divides the digital financial inclusion index into the insurance index, financing index, and payment index according to Guo et al. (2020). These indices include the indicators of total actual usage (number of people using these services per 10,000 users) and active usage (number of transactions per capita and amount of transactions per capita). For example, the digital insurance index includes the number of insured users per 10,000 users, the amount of insurance per capita, and the number of insurance per capita.
Columns 1–3 in Table 6 are the estimation results of digital insurance, credit, and payment services, respectively. Different types of digital finance services have quite different impacts on green economic efficiency.
Classification index model
Variable . | (1) . | (2) . | (3) . |
---|---|---|---|
dfi_ins | 0.0146** (2.0844) | ||
dfi_cre | 0.0351* (1.8084) | ||
dfi_pay | 0.0037 (0.2439) | ||
rc | 0.0159*** (3.9546) | 0.0116** (2.4585) | 0.0132*** (3.2905) |
rfd | −0.0346*** (−3.7924) | −0.0338*** (−3.7359) | −0.0341*** (−3.7257) |
pd | −0.0141*** (−2.6650) | −0.0110** (−2.0732) | −0.0128** (−2.3773) |
inst | 0.1897*** (3.7278) | 0.2484*** (4.8788) | 0.2366*** (4.6594) |
ic | 0.0409*** (3.5797) | 0.0280** (2.2321) | 0.0349** (2.5592) |
eg | 0.1206*** (7.5573) | 0.1264*** (7.9810) | 0.1225*** (7.7313) |
dfi_ins * rc | −0.0024** (−2.5590) | ||
dfi_cre * rc | −0.003 (−0.8089) | ||
dfi_pay * rc | −0.0032* (−1.6678) | ||
W* dfi_ins | −0.0113 (−1.5654) | ||
W* dfi_cre | −0.0082 (−0.3790) | ||
W* dfi_pay | 0.0150 (0.9396) | ||
W*rfd | −0.0026 (−0.1032) | −0.0195 (−0.7672) | −0.0099 (−0.3971) |
W*pd | −0.0295** (−2.3469) | −0.0213* (−1.7117) | −0.0211* (−1.6691) |
W*inst | −0.4781*** (−4.3836) | −0.2444** (−2.0083) | −0.1994 (−1.5726) |
W*rc | 0.0196 (1.0775) | −0.0378** (−2.0495) | −0.0375** (−2.09) |
W*ic | −0.1494*** (−4.3552) | −0.3045*** (−7.5339) | −0.3291*** (−6.8233) |
W*eg | −0.1130*** (−5.5269) | −0.0822*** (−3.8978) | −0.1256*** (−6.1619) |
W* dfi_ins *rc | 0.0001 (0.0189) | ||
W* dfi_cre *rc | 0.0452*** (4.0740) | ||
W* dfi_pay *rc | 0.0236*** (3.7351) | ||
Ρ | 0.1630*** (3.6057) | 0.1240*** (2.7043) | 0.1430*** (3.1411) |
R2 | 0.9019 | 0.9039 | 0.9032 |
Log-L | 2911.5823 | 2931.2503 | 2924.47 |
σ2 | 0.0029 | 0.0029 | 0.0029 |
Variable . | (1) . | (2) . | (3) . |
---|---|---|---|
dfi_ins | 0.0146** (2.0844) | ||
dfi_cre | 0.0351* (1.8084) | ||
dfi_pay | 0.0037 (0.2439) | ||
rc | 0.0159*** (3.9546) | 0.0116** (2.4585) | 0.0132*** (3.2905) |
rfd | −0.0346*** (−3.7924) | −0.0338*** (−3.7359) | −0.0341*** (−3.7257) |
pd | −0.0141*** (−2.6650) | −0.0110** (−2.0732) | −0.0128** (−2.3773) |
inst | 0.1897*** (3.7278) | 0.2484*** (4.8788) | 0.2366*** (4.6594) |
ic | 0.0409*** (3.5797) | 0.0280** (2.2321) | 0.0349** (2.5592) |
eg | 0.1206*** (7.5573) | 0.1264*** (7.9810) | 0.1225*** (7.7313) |
dfi_ins * rc | −0.0024** (−2.5590) | ||
dfi_cre * rc | −0.003 (−0.8089) | ||
dfi_pay * rc | −0.0032* (−1.6678) | ||
W* dfi_ins | −0.0113 (−1.5654) | ||
W* dfi_cre | −0.0082 (−0.3790) | ||
W* dfi_pay | 0.0150 (0.9396) | ||
W*rfd | −0.0026 (−0.1032) | −0.0195 (−0.7672) | −0.0099 (−0.3971) |
W*pd | −0.0295** (−2.3469) | −0.0213* (−1.7117) | −0.0211* (−1.6691) |
W*inst | −0.4781*** (−4.3836) | −0.2444** (−2.0083) | −0.1994 (−1.5726) |
W*rc | 0.0196 (1.0775) | −0.0378** (−2.0495) | −0.0375** (−2.09) |
W*ic | −0.1494*** (−4.3552) | −0.3045*** (−7.5339) | −0.3291*** (−6.8233) |
W*eg | −0.1130*** (−5.5269) | −0.0822*** (−3.8978) | −0.1256*** (−6.1619) |
W* dfi_ins *rc | 0.0001 (0.0189) | ||
W* dfi_cre *rc | 0.0452*** (4.0740) | ||
W* dfi_pay *rc | 0.0236*** (3.7351) | ||
Ρ | 0.1630*** (3.6057) | 0.1240*** (2.7043) | 0.1430*** (3.1411) |
R2 | 0.9019 | 0.9039 | 0.9032 |
Log-L | 2911.5823 | 2931.2503 | 2924.47 |
σ2 | 0.0029 | 0.0029 | 0.0029 |
T values are in parentheses.
***, **, and * denote significance at the 0.01, 0.05, and 0.1 levels, respectively.
First, all services have a positive impact on the green economic efficiency of the region, but they differ in their significance level and coefficient. The insurance index and the credit index are significant at the 5% and 10% levels, respectively. The estimated coefficient of the payment index is positive but insignificant. What is more, the coefficient of the credit index is greater than that of the insurance index, showing that digital credit services contribute most to the local green economy. Digital finance mainly provides the general public or micro and small enterprises with more convenient financial services, lower financing costs, and higher financing efficiency from the production side, which improves their ability to carry out green product innovation and green process innovation, and therefore, its role in promoting local green economy efficiency improvement is the greatest. Digital insurance and digital payments, on the other hand, mainly provide personal financial services on the consumer side, and their contribution to local green economy efficiency gains is relatively small.
Second, the interaction term indicates that the interaction of any digital financial service and regional competition is negatively related to green economic efficiency. The coefficients of both insurance services and payment services are significant, but the former is greater. The coefficient of financing service is not significant. This means that regional competition has mainly crowded out financial resources for digital payments and digital insurance on the individual consumption side, diminishing the role of digital payments and digital insurance for green economy efficiency, while regional competition has not affected the positive role of digital credit, such as production loans and business loans on the production side for green economy efficiency. We conclude that the negative impact of digital financial inclusion and regional competition is mainly attributed to digital payment services.
4.4. Robustness test
To test the robustness of these results, we conducted a Hausman test and used a fixed effect model. Estimations are shown in Table S1 and support our conclusions.
5. Discussion
This article proposes a novel objective that explored how user response to easier and cheaper access to digital financial services affects green economic efficiency. Most related studies focus on the impact of digital financial inclusion on economic growth (Mohan, 2006; Kim, 2016) and green innovations (Liu et al., 2021), which discussed the effect of digital financial inclusion from the supply side. Few existing studies have explained the impact of digital financial inclusion on the efficiency of the green economy.
Regional competition has received considerable attention in research and policy because regions are always competing for capital investment, which can be taken away by digital financial inclusion to some degree. Local governments are supposed to respond dramatically to the promotion of digital financial inclusion in central government policy. To avoid possible conflict between capital investment and digital financial inclusion on green economic efficiency, we also figured out major determinants that significantly influence interaction between digital financial instruments and regional competition.
Our research concludes that in this moment, the interaction of digital financial inclusion and regional competition is harmful to green economic efficiency. While Liu et al. (2021)’s conclusions focus on spatial effect of digital financial inclusion and advocate for moderate regional competition, but what is the right amount of regional competition is ambiguous. Our explanation to this deviation is that local financial institutions that regions are competing for failed to offer quality digital financing services right now, local governments and major digital financing service providers are competitors in this case. Their potential conflicts will harm green economic development. When local financial institutions change in terms of online services, interactions between digital financing inclusion and regional competition result in nothing but easier use of digital financial services and higher economic efficiency. We point out, in order to make full use of digital financial inclusion, digital credit instruments should be promoted.
6. Conclusions and implications
This study elucidates the impact and mechanisms of digital financial inclusion on the efficiency of the green economy, considering the moderating role of regional competition, and validates our hypotheses through an empirical analysis based on data from 265 prefecture-level cities in China from 2010 to 2017. Specifically, this article fills a critical gap in the literature on how green economic efficiency responds to different digital financial instruments when use of digital financial services becomes easier and cheaper.
Our main findings are as follows: (1) Digital financial inclusion promotes the green economy which has a significantly positive spillover effect, (2) regional competition is beneficial for green economic development, but interactions between digital financial inclusion and regional competition are harmful, and (3) digital credit services make the greatest positive contribution to the green economy, and digital payment services have the greatest negative effect.
Our findings are of great practical importance for policymakers in response to such results. First, to promote the green economy, local governments should cooperate with neighboring areas instead of engaging in unhealthy competition. Second, developing areas should treat digital financial inclusion with caution and make use of both digital financial inclusion and local financial institutions. Third, digital credit services are the major contributor to the positive effect on the green economy. It will therefore be beneficial to promote digital credit services and to adjust digital payment services to customer preferences.
The uses of digital financial inclusion are continuously growing in quantity and quality. However, due to space limitations, the effect of use coverage was not empirically investigated in this study, and the conclusions and implications listed are only applicable to use depth. Future works should integrate both the service quality and quantity of digital financial inclusion and may produce more generalizable findings.
Data accessibility statement
Data are publicly archived by the China National Bureau of Statistics at http://www.stats.gov.cn/ and by Guo et al. (2020).
Supplemental files
The supplemental files for this article can be found as follows:
Table S1. Docx
Funding
This article is supported by the National Natural Science Foundation of China (grant number: 71863020), the Fundamental Research Funds for the Central Universities (program no. 2722019PY028), and the Specialized Research Fund for Postgraduate Innovation Programs of Shanghai University of Finance and Economics (grant number: CXJJ-2021-382).
Competing interests
The authors declare no conflicts of interest.
Author contributions
Conceptualization: CS and LY; methodology: ZQ; software: LY; investigation: CS, LY, and ZQ; writing—original draft preparation: CS and LY; writing—review and editing: CS, LY, and ZQ; project administration: CS; funding acquisition: CS, LY, and ZQ. All authors have read and agreed to the published version of the manuscript.
References
How to cite this article: Song, C, Qiu, Z, Yang, L. 2023. How does digital financial inclusion affect green economic development? A perspective from regional competitions. Elementa: Science of the Anthropocene 11(1). DOI: https://doi.org/10.1525/elementa.2022.00090
Domain Editor-in-Chief: Alastair Iles, University of California Berkeley, Berkeley, CA, USA
Associate Editor: Yuwei Shi, University of California Santa Cruz, Santa Cruz, CA, USA
Knowledge Domain: Sustainability Transitions
Part of an Elementa Special Feature: Social Entrepreneurship and Sustainability Transitions in China