Research on the Spatial Correlation and Influence Factors of Regional Internet Finance in China

Abstract

Based on the Internet Finance development index developed by Peking University in investigating 31 provinces, we use social network analysis to investigate the spatial correlation and influencing factors of the Internet Finance of China’s provinces. The research shows that the spatial correlation of Internet Finance of China’s provinces has significant characteristics of a regional gradient. The correlation among their correlation is moderate, more closely related, smooth, or weak. All provinces gather to form four plates; they play their respective functional advantages and have different statuses, functions, and roles. The differences between provinces in industrial structure, degree of marketization, infrastructure, degree of informatization, and geographical distance significantly affect the spatial correlation of Internet finance development in China, and the degree of influence decreases in turn. Our results might positively affect policymakers in promoting the coordinated development of regional Internet finance in China.

Keywords

Internet finance spatial correlation social network analysis method overflow effect influencing factors

Introduction

Internationally, Internet finance, financial technology, and digital finance have similar meanings, and to some extent, they can be used interchangeably (Shen & Huang, 2016; World Bank Group, 2018; P. Xie et al., 2016). However, the pace of Internet technology development has not stopped. The Internet has not only effectively reduced the operating costs of traditional financial institutions but also gradually actualized its spirit of “openness, equality, cooperation and sharing” into traditional financial formats, which has had an essential influence on the development of the financial industry from both supply and demand sides. This has resulted in a subversive revolution in the traditional financial industry, effectively solving the financial needs of low-income groups (Hasan et al., 2021). Recently, Chinese Internet information communication technology, such as mobile payment, social networks, big data, blockchain, and cloud computing, has advanced by leaps and bounds rapidly. This technology deeply integrates with traditional finance, and new patterns or forms of Internet finance emerge, resulting in an explosive influence on traditional finance (Chen et al., 2017). Therefore, in the process of standardizing the development of Internet finance, China should give full play to the unique advantages of Internet finance, to give full play to the unique advantages of Internet finance, to allocate resources and factors of production more effectively, and to provide a robust and new driving force for the innovative development of regional Internet finance and the formation of new growth poles of the regional economy. Internet finance has not only increasingly become a new engine for driving the innovation and development of the Chinese economy but has also played an irreplaceable role in promoting supply-side structural reform of the financial sector, accelerating economic structural transformation and upgrading, and innovating regional economic development models (Guo et al., 2016; Xie & Zou, 2013). New-born Internet finance has thus attracted great interest from academia, industry, and regulatory bodies and has increasingly become a hot topic of wide concern among scholars.

Therefore, many studies on Internet finance have proliferated, and the research results cover the origin, evolution, essential features, new modes, new patterns, and influence on traditional finance and the economy. For example, the relevant studies on the comparative analysis of Internet finance and traditional finance emphasize the characteristics of Internet finance, such as low-cost and high-efficiency (Li, 2015; Xie & Zou, 2012). Meanwhile, some scholars have demonstrated and analyzed in detail the characteristics of Internet finance, such as promoting the availability of financial resources, enhancing the symmetry of transaction information, and demediating the allocation of resources (Gong, 2013; Jia & Feng, 2014). Meanwhile, they further discussed the significant role of Internet finance in promoting financial inclusiveness and stability (Ozili, 2018). Theoretically, Internet finance can break through traditional constraints of geographic space and provide low-cost and highly convenient financial services to relatively remote areas. It shows that Internet finance has some kind of development characteristic beyond geographic space. From this perspective, regional Internet finance development has no apparent correlation with the geographical location of the area and the development level of the surrounding area. However, Internet finance still belongs to the finance category, and its development cannot be divorced from the general laws of economic development. The development degree of Internet finance for a certain region still depends on the development level of local Internet information technology and the traditional financial industry, which shows that the development of Internet finance cannot wholly transcend geographical space.

Thus, based on an in-depth understanding of the spatial distribution of Internet finance development and scientifically and effectively promoting the innovation of Internet finance, it is particularly necessary for us to discuss the spatial correlation characteristics of regional Internet finance and the correlation relations such as mutual learning, interactive communication, and complementary advantages among regions.

Currently, there are few studies on the spatial correlation of regional Internet finance in China. For example, the Dagum Gini coefficient decomposition method and spatial panel data regression convergence model are used to analyze the regional differences and evolution trend of Internet finance among eight Chinese urban agglomerations. This indicates that the overall development difference of Internet finance among the eight major urban agglomerations shows a trend of gradual decline, and the main reason for the difference among the urban agglomerations of Internet finance is the difference between regions (Liu et al., 2017). Moran’s I index and a spatial econometric model are applied to analyze the spatial agglomeration effect of Internet finance at the national municipal level, which shows that the development of Internet finance in China presents a certain regional agglomeration effect and a strong positive spatial agglomeration effect (Guo et al., 2017). The secondary data analysis method is systematically used to analyze the promotion status of Internet finance to Chinese Inclusive finance and the development status of Internet finance in each province, indicating that there is a development inequality between the most developed region and the least developed region (Arif et al., 2020). At the same time, they further theoretically analyzed how Internet finance played an essential role in promoting the development of Chinese Inclusive Finance; then, they further verified the influence of Internet finance by using the qualitative sampling review method (Hasan et al., 2020). The β-conditional convergence and log-T regression test are applied to analyze the convergence characteristics of Internet finance development in 335 Chinese prefecture-level cities. It is found that there are no convergence characteristics in the development level of Internet finance in China on the whole, and there are significant differences in the development level and growth rate of Internet among the seven aggregates (Bai et al., 2021).

The above literature lays an important foundation for our further research, but there are still three deficiencies. First, the spatial measurement method can verify the spatial correlation of Internet finance among regions, but the complex network relationship and role characteristics of each region in the Internet Financial correlation network cannot be fully reflected; thus, it is difficult to grasp the spatial correlation characteristics between regions. Second, research on the spatial correlation of regional Internet finance development is still lacking at the regional level. Third, traditional measurement methods only analyze the spatial aggregation characteristics of regional Internet finance from the proximity of geographical or economic features, which is not enough to fully explain what factors lead to the unbalanced development of regional Internet finance. Moreover, traditional methods are even more unable to judge whether the relationship between the spatial correlation of regional Internet finance development and its influencing factors is tenable.

As Internet finance is a new model with high integration of Internet and traditional finance, its development in various regions is characterized by the overlapping differences and complexity of Internet and traditional finance. The development of regional Internet finance shows differences and spatial imbalance, and there are intricate spatial relationships between them. The social network analysis (SNA) method is a new research paradigm to describe the interaction and spatial structure characteristics between regions, so the SNA method is very suitable for analyzing the spatial interaction of regional economies. Therefore, based on the network perspective and relationship data, we will use the SNA method to analyze the spatial correlation network characteristics of regional Internet finance development, and then the influencing factors of the spatial correlation will be further analyzed by using the QAP analysis method.

Methodology

Determination of Spatial Correlation

We use the Pearson correlation coefficient to measure the correlation distance between provinces (Araujo & Loucā, 2007; Spelta & Araújo, 2012; Tola et al., 2008) to determine the connections in the spatial correlation network of regional Internet finance development. Sequences of $X_{m} (t) = [X_{m} (1), X_{m} (2), X_{m} (3), L, X_{m} (t - 1), X_{m} (t)]$ and $X_{n} (t) = [X_{n} (1), X_{n} (2), X_{n} (3), L, X_{n} (t - 1), X_{n} (t)] (m = 1, 2, L, 31; n = 1, 2, L, 31)$ represent the index series of Internet finance development of province m and province n at time t, respectively. The correlation coefficient between province m and province n can be expressed as follows:

C_{mn} = \frac{| X_{m} \cdot X_{n} | - | X_{m} | \cdot | X_{n} |}{\sqrt{(X_{m}^{2} - {| X_{m} |}^{2}) - (X_{n}^{2} - {| X_{n} |}^{2})}}

(1)

In equation (1), C_mn is the correlation coefficient of the $X_{m}$ sequence and $X_{n}$ sequence, the value range of C_mn is [−1, 1], and $| X_{m} |$ is the expected value of X_m. We can obtain the measured correlation distance of the Internet finance development space between province m and province n by the correlation coefficient C_mn.

D_{mn} = \sqrt{2 (1 - C_{mn})}

(2)

In equation (2), $D_{mn} \in [0, 2]$ is the distance between provinces, and it verifies three properties that must be satisfied by Euclidean distance measures.

According to equation (2), we construct the correlation distance matrix D of Internet finance development among provinces. When the distance $D_{mn}$ between the two provinces is 0, it shows that Internet finance development has the highest correlation between the two provinces. Generally, the correlation intensity of Internet finance development between m province and n province can also be expressed by the reciprocal of the correlation distance.

J_{mn} = \frac{1}{D_{mn}}

(3)

In equation (3), $J_{mn}$ is the correlation intensity of Internet finance between province m and province n. In the intensity matrix $J (m, n) = (J_{mn})_{N \times N}$ , each element determines the edges of the spatial spillover network of Internet finance development, and each province is a node in the network. To maintain the accuracy of the overall analysis of the spillover correlation network, it is necessary that we set a threshold to binarize each element in the correlation intensity matrix. Generally, the mean value of each row of elements in the intensity matrix $J (m, n)$ is set as the threshold, and the intensity matrix $J (m, n)$ can be transformed into a 0–1 asymmetric matrix ${\bar{J (m, n)}}_{31 \times 31}$ . The specific binarization method is as follows:

\bar{J (m, n)} = {\begin{matrix} 1 & J_{mn} \geq \sum_{n = 1}^{N} J_{m} . / N \\ 0 & J_{mn} < \sum_{n = 1}^{N} J_{m} . / N \end{matrix}

(4)

In equation (4), $J_{mg}$ and $N$ are the sum value of the elements in each row and the number of elements in the matrix, respectively.

Social Network Analysis Method

After calculating the correlation matrix between provinces according to equations (1) to (4), the directed spatial correlation network can be drawn by using the SNA method. Then, the structural characteristics of the correlation network and the influencing factors of the spatial correlation can be analyzed. The SNA method is used to describe the relationship between different variables in society and the economy. Its application has gradually expanded from sociology to economics and management science (Barnett, 2011; Everett, 2002; Scott, 2013).

Correlation Analysis

Network correlation is usually measured by indexes such as network density, correlation degree, efficiency, and hierarchy degree. The indexes reflect the tightness, structure, and stability of the correlation network, and the values of these indexes are all between [0, 1]. Network density is used to measure the scale of the entire correlation network; the greater the network density is, the stronger the network correlation. The network correlation degree refers to the degree of interconnection between nodes in the network and measures the robustness of the network. If the correlation degree is equal to 1, it indicates that the degree of interconnection between nodes in the network is high and the robustness of the network is good. In contrast, at least one node is excluded from the network. Network hierarchy is used to measure the degree of asymmetric accessibility between nodes in the network; the greater the hierarchy, the more pronounced the network hierarchy, the greater the location difference in each node, the fewer nodes at the core of the network, and the more nodes at the edge. Network efficiency is a measure of the interrelated transmission channels between nodes, the greater the network efficiency, the fewer the communication channels between nodes, and the worse the stability of the network.

Block Model Analysis

Block model analysis within the SNA method is used to analyze the location of each node in the network. It is a descriptive algebraic analysis of social roles. According to certain partition criteria, all nodes are divided into several plates, and each node in each plate has structural equivalence (White, Boorman & Breiger, 1976) to analyze the connection mode between the plates and the status, function, and role in the correlation network. “Plate” stands for dividing all nodes in the associated network into several plates according to certain division criteria, analyzing the regional position of each node in the correlation network, and exploring the role of the divided plates in the network. Nodes are usually divided into the following four types.

(1) Two-way overflow plate: the members of this plate send more connections both within and outside the plate but receive fewer connections that are sent from outside the plate.

(2) Major beneficiary plate: this plate has a high ratio of internal connections and a low ratio of external connections. In extreme cases, the correlation coefficient of both the inner plate and the outer plate is zero.

(3) Broker plate: this plate is a bridge between other plates. The internal members of this plate have fewer connections to each other and have more connections with the external members. They both send and receive the connections to/from the outside of the plate.

(4) Net overflow plate: the plate sends more connections to the outside but fewer connections to the inside and receives fewer connections from the outside.

Quadratic Assignment Procedure Analysis Method

The quadratic assignment procedure (referred to as QAP) nonparametric analysis method is an analysis method used to compare the similarity of each element in two relationship matrices. To say, it compares the pair of elements of the matrix, gives the correlation coefficient between the two matrices, and conducts nonparametric tests on the coefficient. It is based on replacing the matrix (Everett, 2002). QAP analysis includes correlation analysis and regression analysis. QAP correlation analysis is used to determine whether a correlation exists between two relationship matrices by calculating the correlation coefficient between the two relationship matrices, and then a nonparametric test is conducted on the correlation coefficient.

The principle of the QAP regression analysis method is basically the same as that of the QAP correlation analysis method. It is used to study the regression relationships between a relationship matrix and multiple relationship matrices and to evaluate whether the judgment coefficient of the regression model is significant. Compared with the traditional parameter statistical measurement method, the QAP analysis method has at least two advantages. First, the assumption of independent variables in traditional statistical measurement methods is not considered, and the estimation failure or instability caused by high correlation or multicollinearity between independent variables is solved. Second, the traditional parameter statistical measurement method cannot make an accurate and reliable test on whether the relationship between relationship data is established. This method is more robust and applicable. Therefore, we use the QAP analysis method to explore the influencing factors of the spatial correlation of regional Internet finance in China.

Empirical Analysis of the Spatial Correlation of Internet Finance

Sample Selection and Data Description

Considering the availability and reliability of data, we select the Peking University Internet Finance Development Index of 31 provinces (autonomous regions and municipalities) as sample data from January 2014 to March 2016 to measure and analyze the spatial correlation network of regional Internet finance in China (Research Group of Internet Finance Research Center of Peking University, 2016), and the specific sample size is 27 monthly data. The sample data are from the website of the Internet Finance Research Center of Peking University (http://idf.pku.edu.cn/).

There are four reasons to choose the sample data. First, this period is an important period of the fastest development of Internet finance, frequent risk events, and equal emphasis on regulation in China. Second, the index is compiled according to the representativeness, operability, independence, and expansibility principles. Third, the index is based on data from Ant Financial Service Group data and a representative cross-section of Internet finance companies. Fourth, the index covers the regional level, which can provide reliable data support for studying spatial correlation relationships and distribution characteristics of Internet finance between different provinces. The development indexes of Internet finance between different provinces are comparable.

Correlation Analysis

The correlation strength matrix of Internet finance among provinces can be calculated according to equations (1) to (4), on which the directed network diagram of spatial associations of Internet finance can be drawn (see Figure 1). The nodes of the network diagram represent 31 provinces. The edge lines between nodes represent the correlation strength among the provinces. There are no isolated provinces in the correlation network. The maximum number of possible directional relationships between provinces is 930, and the number of actual relationships is 478.

Figure 1.

The directed network of spatial correlation of Internet finance in China.

The overall density of the spatial correlation network of Internet finance development among the 31 provinces is 0.5140, indicating that the degree of correlation among the provinces is moderate, and there is still room for improvement in the correlation and coordination among the provinces. Therefore, it is necessary to promote the rational flow of element resources of Internet finance development and improve the optimal allocation efficiency and the linkage effect among provinces of element resources. The correlation degree of the network is 1, indicating that each province is interconnected and fully interconnected in the correlation network. The network efficiency is 0.3425, indicating that the correlation between the provinces in the correlation network is relatively close, and the stability of the network structure is relatively good. The network hierarchy is 0.0645, indicating that the hierarchy structure among provinces is weak, and the correlation and communication channels among the provinces are smooth.

Block Model Analysis

The block model analysis is used to explore the spatial form and cluster structure characteristics of the correlation network of regional Internet finance in China, which is helpful to deeply reveal the position, function, and role of each plate or province in the correlation network. In Figure 1, developed eastern coastal provinces are mainly distributed on the right end of the network diagram, central provinces are mainly distributed in the middle, and western provinces are mainly distributed on the left end. This indicates that Chinese Internet finance development presents obvious characteristics of regional gradient development of “East-Central-West.”

According to the convention of parameter settings in the literature, when we use block model analysis, the maximum depth of segmentation is 2, and the concentration standard is 0.2 in the Ucinet software package. We divide the spatial network of Chinese Internet finance into four types of plates (see Table 1 and Figure 2); the number of relationships between provinces within the plates is 181, and the number of relationships between the plates is 295. Plate I includes seven provinces, Beijing, Shanghai, Guangdong, Zhejiang, Jiangsu, Tianjin, and Hainan, within the eastern region. Plate I can be classified as a two-way overflow plate because it sends out more relationships and receives fewer relationships. Plate II includes seven provinces, such as Fujian, Hebei, Liaoning, and Shandong within the more developed eastern region and Jilin, Heilongjiang, and Hunan within the central region. This plate sends out relatively few relationships to the provinces outside the plate but receives relatively more relationships from the provinces outside the plate. Meanwhile, there is a relatively high proportion of relationships inside the plate, so plate II can be classified as the major beneficiary plate.

Table 1.

Plate Role of Spatial Correlation of Regional Internet Finance in China.

Plate	Internal relationship	Receive relationships	Overflow relationships	Number of members in each plate	Expected internal relationship ratio (%)	Actual Internal relationship ratio (%)	Plate role
Plate I	33	14	75	7	20.00	30.56	Two-way overflow
Plate II	41	108	59	7	20.00	41.00	Major beneficiary
Plate III	53	143	76	8	23.33	41.09	Broker
Plate IV	54	30	85	9	26.67	38.85	Two-way overflow

Figure 2.

The correlation relationships between the plates in the spatial correlation network of Internet finance.

Plate III includes eight provinces, such as Anhui, Jiangxi, and Shanxi within the central region and Shaanxi, Inner Mongolia, Tibet, Xinjiang, and Yunnan within the western region. This plate has a relatively low proportion of internal relationships. It receives the relationships from outside the plate and sends relationships to outside of the plate; thus, plate III can be classified as a broker plate. Plate IV includes nine provinces, such as Henan and Hubei within the central region and Sichuan, Chongqing, Guangxi, Guizhou, Gansu, Ningxia, and Qinghai within the western region. This plate sends more relationships to outside of provinces but receives fewer relationships from outside of the provinces, so plate IV can be classified as a two-way overflow plate. Therefore, the provinces are divided into different plates in the spatial correlation network of Internet finance. The spatial distribution of the provinces in each plate has obvious regional characteristics, each province or plate shows significant gradient correlation characteristics, they have different statuses, functions, and roles, and they mutually influence, mutual interaction, and interconnectedness with each other.

To deeply analyze the correlation relationships among various plates, we should investigate the density matrix and image matrix among the plates. Thus, the overall density (0.5140) of the spatial correlation network of Internet finance development and the network density matrix among each plate are first calculated. The overall density of the correlation network is used as the critical reference. If the density between plates in the density matrix is greater than the overall network density, the corresponding plate density is replaced by 1, which means that there is a strong spatial correlation among plates. Otherwise, we replace 0, and we obtain the image matrix among the plates (see Table 2).

Table 2.

Density Matrix and Image Matrix of Plates.

Plate	Density matrix				Image matrix
Plate	Plate I	Plate II	Plate III	Plate IV	Plate I	Plate II	Plate III	Plate IV
Plate I	0.786	0.878	0.554	0.016	1	1	1	0
Plate II	0.204	0.976	0.839	0.032	0	1	1	0
Plate III	0.071	0.804	0.946	0.575	0	1	1	1
Plate IV	0.000	0.317	0.903	0.750	0	0	1	1

As shown in Table 2, first, the elements of the main diagonal in the plate image matrix are all 1, indicating that the provinces within each plate show obvious interrelationships. Second, the provinces within plate I take the lead in transferring the innovative development of Internet finance to plate II. Third, plate II further transfers to plate III, and plate II receives the overflow of development from plate I and plate III. Fourth, plate III not only receives development overflow from the other three plates but also transfers its development mode to the provinces within plates II and IV, and it plays a transmission role as a bridge. Plate IV is mainly correlated to plate III. It can be seen that all plates and provinces generally have their respective functional advantages in the correlation network of Internet finance development, and the correlation relationships show a “gradient” feature.

Analysis of Influence Factors of Spatial Correlation of Internet Finance

Model Setting and Index Selection

The spatial correlation of regional Internet finance in China has a significant characteristic of regional gradient development. Each province or plate has different positions, functions, and roles in the correlation network of Internet finance. The overflow and receiving relationships are different, which may be related to the endowment differences of Internet and traditional finance in different regions. Therefore, based on the influence characteristics of the development of the Internet and traditional finance, we establish relationship model (5) by selecting many different indicators that influence the development of Internet finance.

\begin{matrix} \bar{J (m, n)} = f (SD (m, n), DK (m, n), DC (m, n), JC (m, n), \\ ED (m, n), RG (m, n), SC (m, n), XC (m, n)) \end{matrix}

(5)

In equation (5), SD means geographical distance among provinces; DK means difference matrix of traditional finance development, and this variable DK is represented by credit balance of financial institutions as a proportion of regional GDP; DC means difference matrix of industrial structure, and this variable DK is represented by the proportion of tertiary industry in GDP; JC means infrastructure, and this variable JC is represented by the proportion of total postal and telecommunications services in regional GDP; ED means difference matrix of human capital; RG means difference matrix of economic development, and this variable RG is represented by regional per capita GDP; SC means difference matrix of marketization degree, and this variable SC is represented by government expenditures as a proportion of regional GDP; XC means difference matrix of the degree of informatization, and this variable XC is represented by the number of Internet users out of 100. Human capital is represented by the proportion of high school, technical secondary school, and university graduates in the total population multiplied by the corresponding years of education. According to the China Statistical Yearbook, the years of education for the above three types of education are 12 years of education for senior high school and technical secondary school and 16 years of education for junior college or above. The sample data of all explanatory variables is a 31 × 31 dimensional difference relationship matrix, so the sample size is eight difference matrices.

To express the above economic characteristic variables as a 31 × 31 dimensional difference matrix among provinces, first, we calculate the mean value of the corresponding variable factors during the sample period. Second, we use the absolute value of the difference between the actual values and average value of each province to construct the difference matrix of each variable. Finally, QAP correlation analysis and regression analysis are conducted on the relationship between the relationship matrix of the spatial correlation of Internet finance development and the distance matrix and the difference matrixes of the economic variable influences. The geographical distance between provinces is calculated by using Google Maps. Other variables’ data are obtained from the China Statistical Yearbook and the People’s Bank of China website.

QAP Correlation Analysis

Table 3 shows the correlation analysis results between the regional correlation of Internet finance development and its influencing factors. The mean value of the correlation coefficient refers to the average value of the correlation coefficient between pairs of elements in the pairwise relationship matrix calculated according to 5,000 random permutations. The actual correlation coefficient is calculated by using the correlation relationship matrix and the influence factors’ difference matrixes of Internet finance development among provinces after random substitution. The maximum and minimum values are the maximum and minimum values of the actual correlation coefficients, respectively.

Table 3.

The Results of Correlation Analysis Between the Relationship Matrix and Difference Matrixes of Influence Factors.

Factor variance	Actual correlation coefficient	Significance level (p value)	Mean of correlation coefficient	SD	Minimum value	Maximum value
SD (m, n)	−0.096	.061	−0.001	0.061	−0.216	0.203
DK (m, n)	0.010	.441	−0.000	0.077	−0.237	0.252
DC (m, n)	−0.170	.012	−0.001	0.078	−0.256	0.251
JC (m, n)	0.134	.043	−0.001	0.077	−0.233	0.246
ED (m, n)	0.082	.151	0.000	0.077	−0.267	0.265
RG (m, n)	0.065	.213	−0.001	0.077	−0.232	0.263
SC (m, n)	0.135	.043	0.001	0.077	−0.263	0.258
XC (m, n)	−0.116	.071	−0.002	0.078	−0.301	0.246

The QAP correlation analysis results show that the spatial correlation of regional Internet finance development is significantly negatively correlated with the differences in geographical distance, industrial structure, and degree of informatization at least at the level of 10%. This indicates that the increasing differences in various factors hinder the flow, overflow, and promotion of the innovative development of Internet finance among provinces, making the “gradient” feature of regional correlation increasingly obvious. Additionally, the spatial correlation of regional Internet finance is significantly positively correlated with the differences in infrastructure and marketization degree at least at the 5% level. This shows that the increase in differences in infrastructure and informatization among provinces is conducive to the expansion of Internet finance development among provinces, thus weakening the “gradient” difference characteristics of the spatial correlation of Internet finance development. However, the differences in human capital, economic development, and the development of traditional finance have no significant influence on the spatial correlation of regional Internet finance development at the level of 10%, and the influence degree is smaller than that of the other five indicators.

QAP Regression Analysis

Table 4 shows the regression results of the correlation matrix of regional Internet finance development and influence factors’ difference matrixes. Probability 1 and probability 2 represent the probability that the absolute value of the judgment coefficient after random replacement of the matrixes is no less than and no greater than the observed judgment coefficient in the table, respectively.

Table 4.

The Regression Results of the Spatial Correlation Matrix and Difference Matrixes of the Influencing Factors.

Variable	Unstandardized regression coefficient	Standardized regression coefficient	Significance level	Probability 1	Probability 2
Intercept	0.601	0.000
SD (m, n)	−0.001	−0.096	0.071	0.929	0.071
DC (m, n)	−2.259	−0.184	0.004	0.997	0.004
JC (m, n)	0.367	0.005	0.000	0.000	1.000
SC (m, n)	1.957	0.123	0.044	0.044	0.956
XC (m, n)	−0.011	−0.064	0.083	0.818	0.183

Table 4 shows that the regression coefficients of the differences in geographic distance, the degree of informatization and industrial structure are significantly negative at least at the level of 10%. The regression coefficients of the differences between infrastructure and the degree of marketization are significantly positive at the 5% level. Moreover, the ordering of absolute values of all variable factors’ coefficients is consistent with the results of QAP correlation analysis. This indicates that the differences in variable factors, such as geographical distance, industrial structure, and the degree of informatization, infrastructure, and degree of marketization, have different degrees of influence on the spatial correlation of regional Internet finance development in China, thus changing the “gradient” characteristics of spatial correlation. The differences in industrial structure, the degree of marketization, infrastructure, the degree of informatization, and geographical distance have a decreasing influence on the formation of the spatial correlation in turn.

Discussion

Internet finance has dramatically changed people’s means of payment and investment. At present, China’s Internet finance has seen rapid development, playing a positive role in promoting the development of inclusive finance, improving the quality and efficiency of financial services, and meeting diversified investment needs, which shows great market space and development potential. Internet finance has created new business models of demand and innovation, including P2P lending, crowdfunding, third-party payment, big data finance, digital currency, etc. The emergence of new models and businesses has injected new growth vitality into China’s economic development and industrial restructuring.

However, there are obvious differences in the element resource endowment of the Internet and financial development among different provinces in China, which makes the development of Internet finance unbalanced among provinces. In the context of the lack of adequate supervision, low entry threshold, and uneven qualifications of participants, the development of Internet finance in China faces new challenges and exposes some problems and hidden risks. In the downturn of the real economy and the rise of financial risks, small and medium-sized enterprises have more difficulties in operating, and the possibility of debt default increases, which leads to a decline in the quality of major assets connected to Internet finance platforms and a rise in the overdue and nonperforming rates. In the process of compliance transformation of Internet finance, some practitioners have tried to continue to operate. Nevertheless, due to irregular operation behaviors such as maturity mismatch, capital pool, and large target value in the early stage, the accumulated risk exposure is considerable, the transformation is difficult, and the exit cannot be smooth, which may cause social problems and financial risk. Internet financial risk has vital stakeholders, crossover, and infectivity; the risk may produce cross-institutional, cross-regional, and cross-market chain reactions in the process of disposing of risks.

With the Chinese government’s crackdown on Internet financial risk, the development of Internet finance in China will enter the stage of standardized development. Although the total amount of Internet financial business accounts for a low proportion of the total amount of financial business, Internet finance business still involves a wide range of people. There are many business models of Internet finance, but the development of major business forms shows a trend of differentiation. The development of Internet finance may have an obvious “catfish effect.” Its innovation in concept, technology, and mode has prompted traditional financial institutions to constantly change their business models and service methods, injecting new impetus into the reform and development of conventional financial institutions.

Conclusion

The SNA method and QAP analysis method are applied to explore the spatial structure features and influencing factors of the correlation network of regional Internet finance development in China. The results show that China’s regional Internet finance development presents an interconnected spatial correlation network, which shows significant “east-middle-west” regionalization and gradient characteristics. In the spatial correlation network, the provinces gather to form four plates with different positions, functions, and roles. The developed eastern coastal regions are gathered together in network plate I, which includes Beijing, Shanghai, Guangdong, Zhejiang, Jiangsu, Tianjin, and Hainan; the central and western regions are gathered together in network plate IV with high development potential, which includes Henan, Hunan, Sichuan, Chongqing, Guangxi, Guizhou, Gansu, Ningxia, and Qinghai; these provinces within plate I and plate IV play the role of “two-way spillover.” The subdeveloped eastern provinces include Fujian, Hebei, Liaoning, and Shandong, and the central provinces with better industrial bases include Jilin, Heilongjiang, and Hunan; these provinces are clustered in network plate II, which plays the role of “broker.” The central and western provinces with high development potential include Anhui, Jiangxi, Shanxi, Shaanxi, Inner Mongolia, Tibet, Xinjiang, and Yunnan; these provinces are clustered in network plate III, which plays the role of “main beneficiary.” Finally, we conducted QAP analysis on the influencing factors of the spatial correlation of Internet finance development and found that the differences in industrial structure, marketization degree, infrastructure, informatization degree, and geographical distance among provinces may have a significant influence on the spatial correlation of Internet finance development, and the influence degree decreases in turn; the differences in human capital, regional economic development, and traditional financial development have no apparent influence on the spatial correlation of Internet finance.

Our findings may have some promising implications for policymakers. Because the development level of Internet finance among provinces is different, where the development level of Internet finance in central regions is high, the development level of Internet finance in central regions is second, and the development level of Internet finance in central regions is low. Therefore, policymakers should pay more attention to the development of Internet finance in the central and western regions. In addition, policymakers should give full play to the advantages of the eastern provinces’ Internet finance development, to guide the eastern regions to promote the development of Internet finance in the central and western regions, to strengthen exchanges and cooperation between different regions or provinces, and to promote coordinated and high-quality development of regional Internet finance in China.

Limitations of This Research and Future Research Direction

Given the weak timeliness of sample data (the period from 2014 to 2016), it is difficult to obtain very meaningful results through empirical analysis of these sample data. Because of this limitation, even if any different statistical model is used, we cannot find better results. In any case, based on the network perspective and relationship data, this paper is the first to comprehensively analyze the spatial correlation and influencing factors of regional Internet financial development in China. In the future, when the available sample data are updated on time or the sample data with stronger timeliness are remeasured, we will continue to conduct in-depth research on this problem. In addition, exploring the promotion effect of Internet finance development on Inclusive finance is also worth studying in the next step.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Social Science Foundation of Fujian Province (No. FJ2021BF011); National Natural Science Foundation of Fujian Province (Nos. 2022J01893; 2021J05251; 2020J01797); Science and Technology Project of Fujian Education Department (No. JAT200676); Foundation of Wuyi University (No. YJ202012).

ORCID iD

Zhiyi Zhuo

References

Araujo

Loucā

(2007). The geometry of crashes—A measure of the dynamics of stock market crises. Quantitative Finance, 7(1), 63–74.

Arif

M. M. H.

Yajuan

Mahmud

(2020). Regional development of China’s inclusive finance through financial technology. SAGE Open, 10(1), 1–16.

Bai

C. Q.

Yan

Yin

S. G.

Feng

Wei

(2021). Exploring the development trend of Internet Finance in China: Perspective from club convergence. The North American Journal of Economics and Finance, 58, 101505.

Barnett

G. A.

(2011). Encyclopedia of social networks. Sage.

Chen

Luo

(2017). The transition from traditional banking to mobile Internet Finance: An organizational innovation perspective—A comparative study of Citibank and ICBC. Financial Innovation, 3(1), 16.

Everett

(2002). Social network analysis. Textbook at Essex Summer School in SSDA.

Gong

X. L.

(2013). Internet financial model and its influence on traditional banking. Southern Finance, 5, 86–88. (in Chinese).

Guo

Kong

S. T.

Wang

J. Y.

(2016). General patterns and regional disparity of internet finance development in China: Evidence from the Peking University Internet Finance Development Index. China Economic Journal, 9(3), 253–271.

Guo

Kong

S. T.

Wang

J. Y.

(2017). Analysis of spatial agglomeration effect of Internet Finance—Evidence from internet financial development index. International Financial Research, 8, 75–85. (in Chinese).

10.

Hasan

M. M.

Hoque

(2021). How does financial literacy influence on inclusive finance? Financial Innovation, 7, 40.

11.

Hasan

M. M.

Yajuan

Khan

(2020). Promoting China’s inclusive finance through digital financial services. Global Business Review, 23(6),097215091989534.

12.

Jia

Feng

(2014). When financial Internet meets Internet Finance: substitution or integration. Shanghai Finance, 2, 30–35. (in Chinese).

13.

J. Z.

(2015). Thoughts on internet finance. Management World, 7, 1–7 + 16. (in Chinese).

14.

Liu

Wang

Wei

X. M.

(2017). Research on regional difference decomposition and convergence of Internet financial development in China’s eight urban agglomerations. Research on Quantitative Economy and Technical Economy, 34(8), 3–20. (in Chinese).

15.

Ozili

P. K.

(2018). Influence of digital finance on financial inclusion and stability. Borsa Istanbul Review, 18, 329–340.

16.

Research Group of Internet Finance Research Center of Peking University. (2016). Compilation and analysis of Internet financial development index. New Financial Review, 1, 101–129. (in Chinese).

17.

Scott

(2013). Social network analysis. Sage.

18.

Shen

Huang

(2016). Introduction to the special issue: Internet Finance in China. China Economic Journal, 9(3), 221–224.

19.

Spelta

Araújo

(2012). The topology of cross-border exposures: Beyond the minimal spanning tree approach. Physica A: Statistical Mechanics and Its Applications, 391(22), 5572–5583.

20.

Tola

Lillo

Gallegati

Mantegna

R. N.

, et al. (2008) . Cluster analysis for portfolio optimization. Journal of Economic Dynamics and Control, 32(1), 235–258.

21.

White

H. C.

Boorman

S. A.

Breiger

R. L.

(1976). Social structure from multiple networks. Foreign Languages Press.

22.

World Bank Group. (2018). Toward universal financial inclusion in China: Models, challenges, and global lessons.

23.

Xie

Zou

C. W.

(2013). The theory of internet finance. Economist, 8(2), 18–26. (in Chinese).

24.

Xie

Zou

C. W.

Liu

(2016). The fundamentals of Internet Finance and its policy implications in China. Economic Journal, 9(3), 240–252. (in Chinese).

25.

Xie

H. P.

Zou

C. W.

(2012). Research on Internet financial model. Financial Research, 12, 11–22. (in Chinese).