Heliyon 10 (2024) e30586

Contents lists available at ScienceDirect

Heliyon
journal homepage: www.cell.com/heliyon

Research article

Dynamics in digital finance and its impact on SME financing

Wang Jun a, b, c, Xin Qian Ran a, *
a
Sichuan University Business School, Sichuan, Chengdu, China
b
Sichuan Key Laboratory of Technology Finance and Mathematical Finance, Chengdu, China
c
Financial Security Research Center of Business School of Sichuan University, Chengdu, China

A R T I C L E I N F O A B S T R A C T

Keywords: Employing the Complex Adaptive System (CAS) framework, we analyzed the evolution of digi­
Digital transformed finance tally transformed financial networks, underscoring the pivotal role of adaptation and learning
Financing constraints among network participants. Government funds supporting banks and platforms in digital
Multi-agent simulation
transformation prove more effective than direct subsidies to SMEs. Achieving sustainable
COVID-19
development requires an optimal balance between investment and efficiency due to the impact of
digitalization costs on constructing digital financial networks. The study advocates for a long-
term microcredit financing mode tailored to SMEs, aiding them in maintaining financial stabil­
ity. Additionally, intermediary nodes within the network, such as credit rating companies, can
facilitate information sharing, thereby fostering the growth of digital financial networks. This
heightened information sharing enhances the payoff for each subject and reduces financing
constraints. Furthermore, the study establishes that digital financial networks can alleviate the
repercussions of unforeseen events, such as the COVID-19 pandemic. Finally, empirical analysis of
the Beijing University Inclusive Finance Index data and financial records of Chinese SMEs listed
on the New Third Board from 2016 to 2018 confirms that digital finance can alleviate the
financing constraints of SMEs. In summary, the progression of digital financial networks is
intricate, necessitating a thorough comprehension of diverse factors. The government is
encouraged to guide the enhancement of relevant systems and regulations, bolstering supervision,
and facilitating the high-quality development of digital finance networks.

1. Introduction

In the context of technology and industrial transformation, the development of digital finance has become an important engine for a
country’s high-quality financial development [1,2]. Digital finance broadly encompasses the utilization of digital technologies by both
traditional financial institutions and internet-based companies to actualize innovative financial models for activities such as financing,
payment, and investment [3]. Positioned as a technology-driven financial innovation, the objective of digital finance is to enable or
enhance financial products and operational processes, thereby reshaping business landscapes through technological interventions.
Blockchain technology furnishes essential support for information sharing, while the evolution of cloud computing facilitates
enhanced connectivity among financial entities. Digital finance proves adept at mitigating information and transaction costs, fostering
streamlined information sharing across enterprises, mitigating information asymmetry, and enhancing the accessibility of financial
services [4]. As digital technology continually evolves within financial institutions, the ambit and breadth of financial services witness

* Corresponding author.
E-mail address: [email protected] (X.Q. Ran).

https://doi.org/10.1016/j.heliyon.2024.e30586
Received 24 June 2023; Received in revised form 21 April 2024; Accepted 29 April 2024
Available online 6 May 2024
2405-8440/© 2024 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
W. Jun and X.Q. Ran Heliyon 10 (2024) e30586

continuous expansion. This progression assumes critical importance as digital finance endeavors to transcend the constraints of the
traditional financial landscape in SMEs finance, particularly by addressing the “eighty-twenty rule,1" and paying heed to the dynamics
of the “long-tail effect.2"
According to research findings, SMEs contribute significantly to China’s economic landscape, accounting for over 50 % of the
country’s tax revenue, more than 60 % of GDP, over 70 % of technological innovation, more than 80 % of urban labor employment and
over 90 % of the number of total enterprises, establishing themselves as a vital driving force for national economic and social
development [5]. The development of SMEs is crucial to improving high-quality and rapid economic growth, ensuring the integrity of
the production system and stable employment. However, SMEs are more susceptible to financing constraints in compared to large
enterprises due to their modest scale and limited risk resilience [6,7]. When confronted with unforeseen external risks, they often
encounter challenges in sustaining operations due to disruptions in the capital chain and their relatively short life cycle. For instance,
the COVID-19 pandemic has adversely impacted 85 % of SMEs, with their cash reserves lasting less than three months, rendering them
financially vulnerability [8].
In the financial market, SMEs represent the “long-tail” segment of capital demand, yet they are often overlooked by the traditional
financial system and are not the primary focus of credit rationing. Consequently, the challenge of securing financing stands as a
significant impediment to the development of SMEs. In recent years, digital big data platforms have emerged as facilitators, enabling
SMEs to access necessary financial services with greater convenience and speed. This, in turn, enhances the efficiency of allocating
traditional financial resources to some extent, thereby mitigating the financing constraints faced by enterprises [9,10]. While efforts
have been made to address the information and cost barriers associated with SME financing, both banks and financial platforms are
encouraging the flow of financial resources to small and medium-sized enterprises in regions not adequately covered by traditional
finance. This proactive approach aligns with the objectives of digital finance. However, the promotion of digital finance’s development
remains in the exploratory stage, and the challenges of arduous and costly financing for SMEs persist. Particularly in the face of
unpredictable events, the risk of financial vulnerability becomes more pronounced, amplifying the need for financial support among
SMEs. Paradoxically, obtaining such support becomes even more challenging during these times of heightened financial uncertainty.
While numerous scholars have investigated the effects of digital finance on SME financing constraints, most concentrate on the
outcomes of the influence. Due to the absence of specific, accurate data, in-depth research on the development process has been
challenging. This paper delves into the developmental trajectory of digital finance via a digital finance network evolution model.
Employing simulation methods to overcome the limitation of precise data, the paper offers more targeted recommendations for the
establishment and advancement of digital finance.
How can the advancement of digital finance be accelerated, and what role does the government play in facilitating this process? To
what extent can digital finance effectively alleviate the financing constraints faced by SMEs? In times of unforeseen events, can digital
finance play a pivotal role in supporting SMEs during critical periods? What insights can be gleaned for the formulation of government
policies? This paper endeavors to address these vital questions by developing a digital financial network evolution model based on the
principles of a complex adaptive system (CAS). Through this framework, our goal is to explore the developmental trajectories of digital
financial networks and their adaptive responses to unpredictable events.

2. Related literature

In this paper, our primary focus is to examine whether the evolution of digital financial networks can alleviate the financing
constraints faced by SMEs. We explore the internal mechanisms of this development through the lens of complex networks. The
relevant literature can be summarized in the following aspects:
Firstly, our study is connected to the literature on the financing constraints of SMEs. Numerous studies have indicated that the
financing challenges faced by SMEs primarily stem from two factors: firstly, the absence of collateralized assets that SMEs can offer,
impeding their ability to effectively showcase their creditworthiness to banks [11,12]. Wang X. and Zhang J. established the credit
rationing model of endogenous collateral and enterprise equilibrium, and demonstrated that SMEs cannot obtain bank loans when
their asset size is less than the critical value [13]. Second, there is a high degree of information asymmetry between SMEs and banks,
which leads to the inability of banks to rationalize credit to SMEs [14]. The low transparency of SMEs and the low standard of financial
statements make it difficult for banks to accurately judge the operation of SMEs and their future profit prospects [15]. It is easier for
SMEs to establish relationship loans with small and medium-sized banks [16,17]. Due to the lack of competition in the state-owned
banking system, financial policies such as maintaining the existing monopoly position of banks make it easy for banks to discrimi­
nate against SMEs with larger loan risks, thus making it difficult for SMEs to obtain financing through large state-owned banks [18].
Liu et al. confirmed the conclusion of that large banks are ’ naturally unsuitable for SMEs ’, using data from 2006 to 2011 at the county
level [19].
Our analysis is also related to literature exploring the development effect of digital finance. The long-term supply shortage of
traditional financial services seriously restricts the transformation of economic structure and high-quality development [3], which also

1
Eighty-twenty rule, also known as the Pareto principle or Pareto law, is a description of an economic principle and social phenomenon. This rule
states that, in many cases, 80 % of the results often come from 20 % of the causes or 20 % of the population. It indicates that in a given system or
group, a small number of important factors or individuals play a dominant role in the overall outcome.
2
Long-tail effect refers to the phenomenon where the total sum of a few niche products or services in sales or consumption can be as important as
the total sum of a few popular products or services, because the demand for these niche products and services is scattered.

2
W. Jun and X.Q. Ran Heliyon 10 (2024) e30586

provides opportunities for the leapfrog development of digital finance to some extent. With the help of information technologies such
as big data, cloud computing, blockchain and artificial intelligence, digital finance improves the problems of high risk premium and
operating cost caused by information asymmetry in traditional financial models [20]. Furthermore, the development of digital finance
exhibits significant spatial autocorrelation between different regions, with strong spatial dependence characteristics and positive
spatial spillover effects among different areas [21]. The development of digital finance will affect the debt structure and risk-taking
behavior of commercial banks [22,23]. Compared with traditional financial service modes, digital finance has advantages such as
cost reduction and efficiency improvement [24]. The emergence of digital financial platforms provides more credit vouchers [25],
which can effectively improve financial inclusion and make credit no longer dependent on asset prices [26].
Our paper also contributes to the literature on application of complex networks. The most significant characteristic of complex
network theory is its complex adaptive system (CAS), which involves unpredictable behaviors in adaptive and interactive systems [27].
This is a dynamic system that responds to constantly changing conditions, which in turn influence the behaviors of embedded par­
ticipants [28]. The complex network theory has been used by Karna and Buchmann et al. to empirically analyze the evolution of the
network through the Bengalol cluster and the German automobile industry respectively [29]. With the development of complexity and
network theory, simulation has become a research method closer to reality [30]. Baum et al., Konig et al., proposed an evolutionary
model of a knowledge exchange network and studied the properties of the network through computer simulation [31,32]. Based on the
multi-agent simulation method, the evolution of the network is studied, and its evolution law and characteristics are analyzed and
discussed [33]. Purchase et al. studied how the adaptive behavior of network agents promotes network evolution [34]. Xu Y.L. and Yu
L. incorporated the network model into the analysis of the evolution of science and technology finance, proposed the concept of science
and technology finance network, and explored the stage characteristics of science and technology finance network in different
evolutionary stages [35]. G. B. Korovin developed an agent-based digital transformation model of regional industrial complex and used
the simulation model to study the individual behaviors of agents [36]. Huang et al. discovered that the significance of borrowers in the
digital financial ecological network can assist in predicting their likelihood of defaulting [37]. Additionally, Gambacorta et al. found
that users with a higher level of importance within the network are able to obtain larger loan amounts [38]. Yaniv Proselkov et al.
created an agent-based supply network simulation model to analyze the financing behavior among enterprises [39].
In summary, the application of complex network theory to investigate the development and evolution of digital finance, partic­
ularly in conjunction with simulation methods, has been limited in scholarly literature. Consequently, this paper addresses this gap by
examining the dynamics from the perspectives of capital suppliers and capital demand. Employing the multi-agent simulation method,
we construct and simulate a dynamic and multi-agent complex network to analyze the impact of the digital financial network on the
financing constraints of SMEs.

3. Model setting

Various financial institutions offer distinct forms of financial support to SMEs. This paper initially categorizes financial institutions
into two groups: banks and financial platforms, representing traditional and emerging alternative financial institutions, respectively.
While the government doesn’t directly engage in the financing activities of SMEs, it plays a crucial role in the digital financial network
by providing indirect guidance and support. In the simulation model, the government steers enterprises, banks, and other relevant
actors through a set of policies aimed at fostering the development of digital financial networks. Furthermore, in the face of unforeseen
events, banks and financial platforms within the digital financial network do not shut down financing channels but rather tighten
lending policies. In order to simplify the simulation model, four types of agents are set, namely government Agent, SME Agent, bank
Agent and financial platform Agent. The intermediary node is put into the network environment for investigation.
Digital financial network is a binary tuple composed of a set N = N1 ∪ N2 ∪ N3 ∪ N4 and E = ei ∪ eig ∪ eij , denoted as G = (N, E).
Here, E represents the set of connections between elements in the set N， N1 is the government node (there is only one), N2i is the ith
SMEs node, N3j is the jth bank node, and N4g is the gth financial platform node (platform node for short). ei is the side connecting the ith
enterprise node and the government node; eij is the side connecting the ith enterprise node and the jth bank node; eig is the side

Fig. 1. Interaction rules of agents.

3
W. Jun and X.Q. Ran Heliyon 10 (2024) e30586

connecting the ith enterprise node and the gth platform node. Each node has different attributes: (1) The attributes of government
Agent mainly provide policy support for SME Agent, and provide policy support for digital bank Agent and financial platform Agent;
(2) The attributes of SME Agent include profitability, corporate reputation, enterprise scale, number of employees, industry of the
enterprise, etc. (3) The attributes of bank agent and financial platform Agent mainly include: initial capital, enterprise scale, digital
transformation cost, preferred selection ratio, etc.
The interaction rule of the Agent is the key to realizing the simulation experiment model. According to the digital financial network
evolution simulation process, as shown in Fig. 1, the interaction rules of various agents are as follows: Banks and financial platforms
are entering the digital financial network through digital transformation to provide the necessary operational funds for small and
micro enterprises. Banks offer loans to small and micro enterprises and charge interest, while financial platforms generate investment
returns through investments or loans. However, if small and micro enterprises fail in their operations, banks and financial platforms
may face the risk of investment failure. In such cases, small and micro enterprises, banks, and financial platforms may lose some value.
Therefore, when the value of individuals in the network falls below zero, they will exit the network.

3.1. SME financing and their own growth rules

Combined with the accounting balance sheet structure of the enterprise, and refer to Wang (2021), use θ to represent the SME [40],
the book value, asset value and liability value of the ith SME at time t (iterative tick) are defined as BVt (θi ), AVt (θi ) And DVt (θi ).
Therefore, book value is the difference between asset value and liability value, such as equation (1)：
BVt (θi ) = AVt (θi ) − DVt (θi ) (1)
We define AVt (θi ) as the assets held by SME(θi ) at time t. The non-exhaustive list includes cash-holding items such as accounts
receivable and financial investments. DV is equal to the company’s outstanding debt; therefore, when the book value is below zero, in
the accounting default state and the risk increases, as follows:
BVt (θi ) < 0
Although they are insolvent from the perspective of accounting balance sheet, SMEs often have a potential for future growth in
combination with the actual situation of enterprises. Currently, enterprises still have the operation ability and certain repayment
ability. Specifically, the intrinsic value of an SME is defined as IVt (θi ) different from its book value. This difference can be demonstrated
by the company’s franchise value (The operational value of the existence of the enterprise), which is recorded as FVt (θi ). This results is
shown in equation (2):
IVt (θi ) = BVt (θi ) + FVt (θi ) (2)
Thus, we define an enterprise as being in a state of economic default at a given time t when the intrinsic value is lower than zero. At
this time, the enterprise exits in a state of bankruptcy and exits the network (agents participating in the network are eliminated), as
follows:
IVt (θi ) = < 0
In order to better understand the concept of a firm’s franchise value, we can look forward to the future accounting earnings of SMEs.
Specifically, we define the profit, income and expenditure of a given accounting year [ t, t + 1 ] as, Profit[t, t + 1], Income[t, t +1] and
Expenditure[t, t + 1]. Income comes from normal business operations, not an exhaustive list of expenses that would include product
costs, office costs, salaries, employee benefits and so on. Thus, the expected profit of SME is shown in equation (3):
Profit[t, t + 1] = Income[t, t + 1] − Expenditure[t, t + 1] (3)
Therefore, the franchise value can be regarded as the discounted value of future (pre-tax) profit, shown in equation (4):
∑
∞
Profit[t + j, t + j + 1]
Franchise Valuet (θi ) = (4)
j=0 (1 + R(θi ))j

Where R(θi ) is the discount rate of future cash flow. To simplify the model, we define R(θi ) = R0 , where R0 is the risk-free rate.
When SMEs have insufficient funds, they will seek financing, that is, when the expenditure of [t,t+1] in a future operating period
exceeds the book value of the enterprise, they need to seek financing from financial institutions, as follows:
Expenditure[t,t+1] (θi ) > BVt (θi )

The financing cost of SMEs is C (the financing cost of the bank is different from that of the platform). Meanwhile, the successful
financed SME will increase their income due to the expansion of input, and the income is Income∗ [t,t + 1], The expression of profit after
financing is shown in equation (5):
Profit∗ [t, t + 1] = Income∗ [t, t + 1] − Expenditure[t, t + 1] − C (5)
When Profit [t, t + 1] > 0, it indicates the success of business activities, and each successful activity can be regarded as the cor­
∗

responding improvement of enterprise profitability. After a round of business activities, the successful SME will continue to carry out
new business activities to maintain their advantages in the fierce market competition. In order to gain a place in the market, the failed

4
W. Jun and X.Q. Ran Heliyon 10 (2024) e30586

enterprises will still choose to continue operating activities, and their profitability will decline slightly. Referring to the measurement
index of network average performance by Cao Xia et al. (2015) [41], using the logarithm of the weighted average of all SMEs in the
network, the income is set as Profit(θi ) = Profit∗ [t, t +1] or Profit[t,t + 1], and the formula for calculating the average income coefficient
is shown in equation (6):
⎛ ⎛ ⎞ ⎞
∑ Profit(θi ) ⎠
U(θ) = log ⎝ ⎝ ∑ ∗ Profit(θi )⎠ (6)
i∈n
Profit(θi )]
i∈n

In this paper, WW index (White & Wu, 2006) is adopted to measure the financing constraint level of SMEs [42], CFit is the ratio of
cash flow to total assets, DIVPOSit is the 0–1 variable to measure whether dividend is paid (i.e., 1 when the platform gains investment
income), TLTDit is the ratio of liabilities to total assets, LNTAit is the natural logarithm of total assets, ISGit is the industry sales growth
rate. SGit is the sales growth rate of the enterprise, and the coefficient vector b is given by White & Wu. The larger the WW index, the
higher the financing constraint of the enterprise. The calculation model is shown in equation (7):
WWit = b1 CFit + b2 DIVPOSit + b3 TLTDit + b4 LNTAit + b5 ISGit + b6 SGit (7)
3
Firms are affected by two processes: (1) intrinsic value following geometric Brownian motion; (2) Book value liquidity crunch due
to unforeseen events such as COVID-19 (Wang, 2021) [40]. In other words, SME i experiences a liquidity crunch, Although the en­
terprise has positive intrinsic value, it will go bankrupt due to a negative book value (the value will disappear when the company closes
its business due to a liquidity crunch).

3.2. Connection rules and payoffs of banks and financial platforms

Banks and financial platforms enter the digital financial network through digital transformation, which incurs certain expenses
known as digital transformation costs. These costs consist of fixed costs and variable costs. The size of the entity itself influences
variable costs. Therefore, the digital transformation cost D can be represented as D = p * Size + C, where p is the coefficient of digital
transformation cost, and C is the fixed cost of digital transformation. Usually, the cost of digital transformation for banks mainly
includes expenses for personnel training and the purchase of equipment, software, and other related costs. Meanwhile, due to the
nature of their business, financial platforms typically engage in independent research and development of digital technology, leading
to significant upfront investment in research and development, resulting in a relatively high fixed cost of digital transformation. If the
R&D proves successful, revenue is generated through partnerships with banks that utilize digital technology.
When enterprises seek financing from banks and financial platforms, these institutions consider key factors such as the profitability,
credibility, and government information of the enterprises. In real life, there is often information asymmetry between small and
medium-sized enterprises (SMEs) and financial institutions. Compared to financial platforms, banks face more pronounced infor­
mation asymmetry, which is why banks typically prefer to collaborate with larger enterprises that have greater information trans­
parency. SMEs have a low degree of information disclosure, which leads to greater risks in their business activities. Banks need to spend
more time and energy when underwriting loans to SMEs, resulting in large opportunity costs. However, financial platforms are more
professional in this aspect. They generally pursue higher risks and greater returns. In summary, banks adhere to stricter selection
criteria for preferred connections, whereas financial platforms exhibit a higher incidence of preferred connections when selecting
investment targets. At the same time, SMEs are rated at different levels according to different profitability (The ability of a company to
generate profits). The higher the level, the greater the initial credit, and the credit also increases with the repayment of enterprises.
Some other informal connections, such as government policy support, can also improve the success rate of financing of SMEs. Optimal
connection of financial institutions is to choose enterprises with high credit to establish connection relationship. The expression of
profitability Profit Rate(θi ) is shown in equation (8):
Profit[t,t+1] (θi )
Profit Rate[t,t+1] (θi ) = (8)
Expenditure[t,t+1] (θi )

Financing activities for SMEs pose risks to banks and financial platforms. Typically, businesses may default on their debts or
become unable to repay them due to bankruptcy, resulting in losses for the banks and financial platforms. These losses can be
characterized using a loss curve S(X) = (1 − w1 ) X, where w1 represents the yield rate on the outstanding loan amount X. The income
generated from non-default cases can be represented by an income curve S(X)∗ = w2 (L − X), where w2 represents the yield rate in non-
default instances. Therefore, the net income curve ΔS(X) is shown in equation (9):

ΔS(X) = S(X)∗ − S(X) = w2 (L − X) + w1 X − X (9)

Generally, when the value of banks and financial platforms is negative, that is, ΔS(X) < 0, banks and platforms will consider exiting
the digital financial network. However, due to the potential value of enterprises, there is expected income. Therefore, when the lending
capital pool of SMEs is less than 0, banks and financial platforms will withdraw from the network. Given the disparity in data volume

3
Liquidity crunch refers to a situation in the financial market where the available liquid funds decrease, leading to tightness in fund supply and
stricter borrowing conditions, making it more difficult for businesses and individuals to access credit and obtain funds.

5
W. Jun and X.Q. Ran Heliyon 10 (2024) e30586

and data processing capacity between banks and platforms, platforms possess more private data and a more advanced data processing
capability. Consequently, they adopt distinct lending policies for SMEs, leading to varied risks and benefits. Banks tend to implement
more cautious lending policies, resulting in lower risk and corresponding returns, whereas platforms exhibit the opposite trend.

4. Simulation results

The evolutionary process of the digital financial network is simulated by using NetLogo.4 In the established model, all subjects are
randomly distributed on the track, and the network scale, financing constraints, and capital changes of the digital financial network are
taken as the investigation objects of the network. The network analysis of government fund allocation, digital transformation cost and
the information disclosure degree effect on the evolution of digital financial network, explore the evolution regularity and charac­
teristics of digital financial network. Table 1 defines the initial values for each attribute.

4.1. Simulation of different policies

Currently, numerous government policies aim to promote the development of SMEs. Direct incentives targeting SMEs, such as tax
benefits and financial support, are one type of policy. Another type involves indirect assistance to help SMEs access financing, such as
fostering digital finance and facilitating tax interactions.5 However, the effectiveness of government policies varies, and not all policies
have the same impact. Due to limited resources, the government needs to rationally use resources to maximize the utility of resources
when conducting macroeconomic regulation. Experiment A1: The government provides financial support to SMEs, but the policy has
weak effectiveness and can only slightly affect the connection rules; Experiment A2: the government provides financial support for
SMEs, and the strong effectiveness of government policies can affect the connection rules; Experiment A3: The government subsidizes
the cost of digitally transformed banks and platforms; experiment A4: The government provides financial support to SMEs while
affecting the connection rules, and subsidizes the cost of digitally transformed banks and platforms. Experiment A2 and Experiment A3
mainly consider whether to subsidize the cost of digital transformation of banks and platforms or directly provide policy subsidies to
SMEs to help their development under limited government resources.
Figs. 2 and 3 show the more effective the policy is, the more beneficial the development of the digital financial network will be. The
government can play a guiding role in developing the digital financial network. In the case of limited resources, using limited resources
as cost subsidies for the digital transformation of banks and financial platforms is more conducive to the development of digital
financial networks than direct subsidies to SMEs. At the same time, from the perspective of average income, it is also conducive to
improving the average income of SMEs. As shown in Fig. 4, SMEs’ financial constraint index results are not easy to analyze, so the
conversion into probability distribution will facilitate the analysis. The greater the influence of policies on the preferred connection
rate of banks and financial platforms, the more the financing constraints of SMEs can be reduced. When the government provides
financial support to SMEs and cost subsidies to banks and platforms, it is most conducive to the development of digital financial
networks, improving the average income of banks and financial platforms and reducing the financing constraints of SMEs.

4.2. Simulation of different loan rates and digital costs of banks and platforms

The cost of digital transformation will affect whether banks and platforms can enter the digital financial network, and the loan
interest rate is the main index affecting the return of banks and financial platforms, and the return is a major factor to measure whether
it is worthwhile to enter the digital financial network. Experiment B1: Low cost and low-interest rate; Experiment B2: High cost and
low-interest rate; Experiment B3: High cost and high-interest rate; Experiment B4: Low costs and high-interest rates.
As shown in Figs. 5 and 6, by comparing assumptions B1 and B4 with assumptions B2 and B4, it is found that the greater the cost of
digital transformation, the smaller the number of nodes in the digital financial network, and the greater the financing constraints of
SMEs. Meanwhile, under the circumstance of high digital transformation cost, the lower the loan interest rate of banks and platforms,
the larger the scale of the digital financial network. Mainly because low interest rate can reduce the financing constraints of SMEs to
some extent, and stimulate more SMEs to enter the digital financial network for loan financing by reducing the loan interest rate to
expand the scale of digital financial networks. However, when the cost of digital transformation is low, the loan interest rate of banks
and platforms has little influence on the size of digital financial networks, and the average income curve of SMEs is highly overlapped.
It shows that the cost of digital transformation is the main factor affecting the scale of digital financial networks. Moreover, it can be
found from the average bank income chart that in the case of high cost, the average income in the early stage is 0, which indicates that
fewer banks enter the digital financial network due to the problems of digital capability and cost. In the later stage, under the influence
of competitors, it gradually made digital transformation and entered the digital financial network. However, the early digitalization
cost has little impact on the average revenue of financial platforms, because financial platforms can also obtain part of the digital
revenue through cooperation with banks. As shown in Fig. 7, the financing constraints of Experiment B2 and Experiment B4 are
roughly the same.

4
NetLogo is a programming language and integrated development environment for modeling and simulating complex systems.
5
Facilitating tax interactions refers to the mechanism of information sharing and collaboration between banks and tax departments. This
mechanism can help banks better understand the tax situation of enterprises, thus more accurately assessing their credit and risks.

6
W. Jun and X.Q. Ran Heliyon 10 (2024) e30586

Table 1
The initial value of each variable.
Variable Description Initial Value

num-governments Number of governments 1

num-firms Number of SMEs 25
num-banks number of banks 2
num-platforms number of platforms 2
capital-size SME funds [1,5]
bank-size Bank capital pool [10,20]
platform-size Platform capital pool [20,40]
support-size Government subsidy 5
success-rate Operation success rate [1,10]
bank-choice-rate Bank preferred connection rate 40 %
platform-choice-rate Platform preferred connection rate 60 %
C Digital fixed cost 5
ρ Digital cost coefficient 5%
R0 = w2 Risk-free rate 3%
w1 The yield rate on outstanding loan amoun 0

Fig. 2. Network size (A panel) and average bank revenues (B panel) under different policies.

Fig. 3. The average revenues of financial platforms (A panel) and SMEs (B panel) under different policies.

4.3. Simulation of different repayment terms

If the repayment period is extended, SMEs face reduced pressure in repaying the loan. They do not need to worry about the outflow
of a large amount of funds in the short term, and the capital chain may break and lead to the bankruptcy of enterprises. But the longer
the repayment term, the more interest will be paid. For banks and platforms, the longer the repayment period, the greater the

7
W. Jun and X.Q. Ran Heliyon 10 (2024) e30586

Fig. 4. Financing constraints of different policies (A panel is financing constraint index, B panel is index distribution).

Fig. 5. Network size (A panel) and average bank revenues (B panel) at different interest rates and costs.

Fig. 6. The average revenues of financial platforms (A panel) and SMEs (B panel) with different interest rates and costs.

possibility of bad debts and the higher the risk. Experiment C1: short-term micro-credit (1 year, bank interest rate: 1 %, scale:1;
platform interest rate: 2 %, scale: 2); Experiment C2: short-term large loans (1 year, bank interest rate: 1 %, scale:2; platform interest
rate: 2 %,scale: 4); Experiment C3: Long-term micro-credit (5 years, bank interest rate: 2 %, scale 1; platform interest rate: 4, scale: 2);
Experiment C4: long-term large loan (5 years, bank interest rate: 2 %, scale: 2, platform interest rate: 4 %, scale: 4).
As shown in Figs. 8 and 9, in the early stage of the development of digital financial networks, long-term large loans are too risky for

8
W. Jun and X.Q. Ran Heliyon 10 (2024) e30586

Fig. 7. The probability distribution of financing constraint index under different interest rate and cost.

banks. Meanwhile, large loans increase the loan repayment pressure of SMEs, which is easy to break the capital chain and lead to the
bankruptcy of enterprises, thus forming bad debts of banks, increasing the risk of bank lending, and not conducive to the construction
of digital financial network. The growth of network nodes of short-term and long-term micro-loans is basically the same, but from the
average income of financial platforms and banks, the average income of long-term micro-loans is the largest. Meanwhile, long-term
micro-loans in the early and middle stages make the average income of SMEs higher than that of short-term micro-loans. In the
late period of the development of digital financial networks, the average income curves of SMEs with long-term and short-term small
loans are basically the same. As shown in Fig. 10, The financing constraints of short-term large loans and long-term small loans for
SMEs are roughly the same, slightly larger than short-term small loans. However, it is obvious that the financing constraint of long-
term large loans is the greatest. Comprehensive, long-term financing alleviates the repayment pressure of SMEs, maintains the
excellent cash flow of SMEs, and avoids the risk of bankruptcy. At the same time, small loans ensure the risk control of banks and
financial platforms, maximize the benefits of each subject in the network, which is conducive to the development of digital financial
networks and alleviates the financing constraints of SMEs.

4.4. Simulation of different information sharing degree

Due to imperfect financial statements and limited information disclosure by SMEs, banks and financial platforms lack a compre­
hensive understanding of their financial and operational conditions. Consequently, this lack of clarity increases financing costs. One of
the advantages of a digital financial network is its high information transparency, which can alleviate information asymmetry between
banks, financial platforms, and SMEs to a certain extent. The intermediary node in the digital financial network serves as the primary
provider of information, and the development of these intermediary nodes reflects the extent of information sharing. In this paper,
random connection and preferred connection are used to reflect information symmetry. Random link refers to that banks and financial
platforms have no access to all aspects of SMEs’ information, resulting in high financing risk. Preferred connection involves

Fig. 8. Network size (A panel) and average bank revenues (B panel) under different loan terms.

9
W. Jun and X.Q. Ran Heliyon 10 (2024) e30586

Fig. 9. The average revenues of financial platforms (A panel) and SMEs (B panel) under different loan terms.

Fig. 10. Probability distribution of financing constraint index for different loan terms.

symmetrical information, enabling banks and platforms to comprehend SMEs’ operational status, lending, and credit information. This
allows them to select enterprises with favorable operating conditions and minimal risks for financing. Experiment D1: The preferred
connection rate of banks and financial platforms is 0, that is, the digital financial network information is not open, and there is no

Fig. 11. Network size (A panel) and average bank revenues (B panel) under different information sharing.

10
W. Jun and X.Q. Ran Heliyon 10 (2024) e30586

information sharing among various subjects; Experiment D2: The preferred connection rate of banks and financial platforms is 50 %,
that is, there is certain information sharing among each subject, and at the same time they retain some information. Experiment D3:
The preferred connection rate of banks and financial platforms is 100 %, that is, all subjects in the digital financial network share
complete information and achieve complete transparency and openness.
As shown in Figs. 11–13, it can be seen that information sharing is conducive to the expansion of the scale of digital financial
networks, while the average returns of banks and financial platforms increase, and the financing constraints of SMEs decrease.
However, by comparing the average return of the financial platform in experiment D2 and experiment D3, it can be found that the
average return of the previous experiment D2 is higher than that of experiment D3. This may be due to the unsound digital financial
network in the early stage, which leads to the deviation of the information collection and processing of the financial platform and the
relatively high investment risk. Thus, the average return of the optimal connection rate of 100 % in the early stage is less than the
optimal connection rate of 50 %.In summary, the preferred connection rate of banks and financial platforms should comprehensively
consider the degree of information sharing and the perfection of digital financial network.

5. Impact of emergencies (COVID-19) and mitigation measures

Enterprises are required to uphold a designated level of cash reserves to sustain their day-to-day operations. These operating costs
encompass expenditures like employee salaries, office operational costs, and outstanding loan interest payments. Specifically, we
express a business’s cash level as the number of months a company can maintain daily operation until it runs out of cash. Therefore, we
represent the cash level of the i SME within a given time t as Casht (θi ), that is, the cash level can maintain the operation of the en­
terprise for Casht (θi ) cycles. We define when cash based accounting book value falls to zero and liquidity crunch occurs at t = T:
BV(θi )|t=T < 0

The random variable T is the number of months of economic disruption caused by COVID-19. Wang (2021) proposed that the
number of months of economic disruption caused by COVID-19 follows the Weibull distribution of shape parameters and proportional
parameters k and β, where k > 0 and β > 0 [40]. In this paper, we learn from its treatment, in which the parameters of Weibull
distribution depend on the strength and effectiveness of various containment measures implemented. These measures may include
government strategies and international efforts to contain the COVID-19 pandemic. From the government ‘s perspective, these policy
objectives can be largely attributed to: (1) stabilizing SMEs to reduce unemployment, (2) alleviating the financial pressure of SMEs
through government tax incentives. q(θi ) is defined as the probability of liquidity crunch in SME i, the formula is as (10):
⎧ ( )k− 1 ( ( )k )
⎨ k ∗ Cash0 (θi )
⎪
∗ exp −
Cash0 (θi )
,
q(θi ) = β β β Cash0 (θi ) ≥ 0 (10)
⎪
⎩
0, Cash0 (θi ) < 0
Assuming that SME will successively declare bankruptcy when they encounter a liquidity crunch, that is, when their cash operation
is not enough to survive the economic disruption caused by COVID-19, as follows:
Cash0 (θi ) < Casht=T (θi )
The comprehensive loss (the value of SMEs) caused by the bankruptcy of SMEs to the whole society is A times of the franchise value
of an enterprise. The expected losses when the SME closes are shown in equation (11):

Fig. 12. Average revenues of financial platforms (A panel) and SMEs (B panel) under different information sharing.

11
W. Jun and X.Q. Ran Heliyon 10 (2024) e30586

Fig. 13. Financing constraint index probability distribution of different information sharing.

( )k− 1 ( ( )k )
k Cash0 (θi ) Cash0 (θi )
Expected Loss(θi ) = A ∗ ∗ ∗ exp − (11)
β β β

Suppose that the government grants emergency funds S to enterprises suffering from liquidity crunch, helping some SMEs through
the crisis period, thus reducing social losses. Therefore, the loss reduction caused by the emergency fund S is expressed as its value V,
and the formula is shown in equation (12):
( )k− 1 ( ( )k )
k Cash0 (θi ) Cash0 (θi )
Value(θi ) = A ∗ ∗ ∗ exp −
β β β

( )k− 1 ( ( )k )
k Cash0 (θi ) + S Cash0 (θi ) + S
− A∗ ∗ ∗ exp − (12)
β β β

According to the simulation results, it is assumed that Cash0 (θi ) = 2, β = 3, S ∈ [1,6], k = 1 or 2, A = 2 or 2.5. An optimal emergency
fund is obtained by solving the results. As shown in Fig. 14, when k = 1, the optimal emergency fund is about S = 3. When k = 2, the
optimal emergency funding is around S = 2.
At the same time, it can be found that the greater the k, the smaller the emergency funds required, the higher the value created by
the emergency funds, Moreover, when k = 1, the difference between the value created by the emergency fund under the conditions of
A = 2 and A = 2.5 is about 0.1, but when k = 2, the difference between the value created by the emergency fund is about 0.2. In other
words, the greater the k, the greater the impact of corporate social value multiple A on the value of emergency funds.
In the event of an emergency, SMEs may find themselves in a situation where they have no operating income but still incur
operating costs. SMEs declare bankruptcy when they experience a liquidity crunch, unable to sustain their operations due to

Fig. 14. Value of emergency subsidy with shape parameters, and with k = 1 (left panel) and k = 2 (right panel).

12
W. Jun and X.Q. Ran Heliyon 10 (2024) e30586

insufficient cash flow amidst the economic disruption caused by an emergency. Due to the bankruptcy of SMEs, banks may not be able
to recover their outstanding loans, the investment income and bonds of financial platforms cannot be realized, the government loses all
future tax revenues and the economy may generate indirect losses due to unemployment and reduced consumption. It will have a huge
impact on the whole society, further exacerbating the socio-economic impact of an emergency. If the bank and financial platform
protect their own circumstances, according to the previous credit of SMEs financing, to help them survive the economic disruption and
reduce social economic losses. Experiment E1: When the digital financial network encountered an emergency at the initial stage, i.e.,
ticks = 75 digital financial network encountered the impact of the COVID-19; Experiment E2: The digital financial network
encountered an emergency during the development period, i.e., ticks = 150 digital financial network encountered the impact of the
COVID-19; Experiment E3: The digital financial network encountered an emergency at the maturity stage, i.e., ticks = 225 digital
financial network encountered the impact of the COVID-19.
As shown in Figs. 15–17, the more perfect the digital financial network is, the better it can cope with the impact of emergency
events. The scale of digital financial network, the average income of each subject and the financing constraints of SMEs are not affected
by external emergencies in the mature stage. However, in the initial and development stages, the scale of the digital financial network
will be seriously affected, which will be detrimental to its subsequent development, but also affect the average income of each subject,
increasing the financing constraints of SMEs. Because financial platforms and banks can reasonably evaluate the operational capa­
bilities and credibility of SMEs through historical transaction data, they can invest and lend to potential enterprises with historical
repayment records to help them survive the liquidity crunch. At the same time, to a certain extent, it prevents the failure of SME
bankruptcy banks to recover their outstanding loans and the investment income and bonds of financial platforms from being realized,
reducing sunk costs and increasing returns. For SMEs, the digital financial network reduces the cost of financing and improves the
availability of financing. If the digital financial network encounters the impact of emergency events at the early stage, the financial
platform and banks do not have enough information for the enterprise, and the financing and lending risks of the financial platform and
banks are high, which leads to the tightening of the connection rate after weighing the advantages and disadvantages. As a result, many
SMEs are unable to obtain financing, Thus cannot successfully survive the tightening period.

6. Further analysis

To empirically validate the theoretical model with real-world data, this study selected financial data disclosed by SMEs listed on
China’s New Third Board from the Guotai An database, as well as the Peking University Inclusive Finance Index to verify the impact of
digital finance on SMEs’ financing constraints. Considering an adequate sample size, the study set the time span from 2016 to 2018 and
focused on SMEs that operated normally during this period. To ensure the validity of the sample data, further filtering was conducted
as follows: (1) Exclusion of financial listed companies; (2) Exclusion of ST and *ST companies, as well as companies delisted, facing
operational abnormalities, or with missing data during the period; (3) To avoid the influence of extreme values, truncation was applied
at the 1 % level. The study ultimately obtained 13,665 sample observations.
This article selects the digital finance index (DIFI) and two sub-indices, namely usage depth (Depth) and digitization level (Digital),
as explanatory variables. Since its publication, this index has been widely used by scholars engaged in related research and exhibits
strong reliability. The dependent variable is the WW index proposed by White & Wu [42]. Other relevant variables in this article are
presented in Table 2.
Model (13) is used to verify whether digital finance significantly alleviates financing constraints for SMEs:
WWit = β0 + β1 DIFIit + β2 FINAit + β3 CFAit + β4 ROAit + Tt + Ii + εit (13)
In the model, i represents the individual enterprise, t represents the year, εit denotes the random error term, β1 -β4 represents the
coefficients of corresponding variables, Tt indicates time effects control, and Ii represents industry effects control. FINA, CFA, and ROA

Fig. 15. Network size (A panel) and average bank revenues (B panel) with emergency events in different periods.

13
W. Jun and X.Q. Ran Heliyon 10 (2024) e30586

Fig. 16. The average revenues of financial platforms (A panel) and SMEs (B panel) with emergency events in different periods.

Fig. 17. Financing constraint index probability distribution of emergency events in different periods.

Table 2
Variable definitions.
Variable type Variable name Variable symbol Variable meaning

Explained variable Financing constraint WW WW Index

Explanatory variables Digital finance index DIFI Peking University Digital Financial Inclusion Index
Usage depth Depth
Digitization level Digital
Control variable Financing cost FINA Finance expenses/total liabilities
Net operating cash flow CFA Operating cash flow/total assets
Return on assets ROA Net profit/average assets
Enterprise age AGE Listing time
Industry of affiliation Industry 1 represents that the company belongs to this industry, and 0 represents that it does not.
Region of affiliation Region 1 represents that the company belongs to this region, and 0 represents that it does not.

represent financing costs, operating net cash flow, and asset returns, respectively. If the coefficient of DIFI (Digital Finance Index) is
positive, it indicates that with better development of digital finance, higher WW values correspond to greater financing constraints for
enterprises. Conversely, if the coefficient is negative, it implies that the development of digital finance can alleviate financing
constraints.
To further examine the impact of depth and digitization on financing constraints faced by SMEs, the following validation models
were obtained by replacing the overall index with two secondary indexes of depth and digitization respectively based on Model (13):
WWit = β0 + β1 Depthit + β2 FINAit + β3 CFAit + β4 ROAit + Tt + Ii + εit (14)

14
W. Jun and X.Q. Ran Heliyon 10 (2024) e30586

WWit = β0 + β1 Digitalit + β2 FINAit + β3 CFAit + β4 ROAit + Tt + Ii + εit (15)

6.1. Empirical result analysis

According to Table 3, the mean of the dependent variable WW is − 0.344, with a maximum value of 0.149 and a minimum value of
− 39.51. This indicates that the sampled enterprises are subject to varying degrees of financing constraints. Among the explanatory
variables DIFI, Depth, and Digital, Digital has the highest mean, minimum value, and maximum value, and the lowest standard de­
viation. This suggests that in the digital inclusive finance index, digitalization has the best development and the least internal
variation.
To verify if there is excessive correlation among the variables, this study conducted a correlation analysis of the main variables, and
the results are presented in Table 4. It can be observed that DIFI is significantly negatively correlated with WW at a 1 % significance
level, indicating that digital finance can alleviate financing constraints for SMEs. Typically, when the absolute value of the correlation
coefficient exceeds 0.8, the multicollinearity among variables becomes more prominent. In this study, the maximum absolute cor­
relation coefficient among variables is 0.511, suggesting a low level of multicollinearity among the variables.
Before conducting the regression analysis, we used the Hausman test to confirm that a fixed effects model should be used. Firstly,
we regressed model (13) using a fixed effects model that controls for time and industry effects to examine the impact of the overall
digitization index on the financing constraints of small and medium-sized enterprises (SMEs). Then, we replaced DIFI with Depth and
Digital as the explanatory variables and conducted regression analysis on models (14) and (15) respectively. The regression results are
shown in Table 5.
Based on the regression results of model (13), it can be observed that DIFI is negatively correlated with WW at a significant level of
1 %. The coefficient of the variable is − 0.003, indicating that, holding other factors constant, a one-unit increase in DIFI leads to a
decrease of 0.003 units in WW. Since a smaller WW value indicates a lower level of financing constraints faced by enterprises, this
validates that the development of digital finance significantly alleviates the financing constraints of SMEs. In models (14) and (15),
Depth is significant at the 5 % level, while Digital has a significant relationship with WW at the 1 % level. This further verifies the
significant alleviating effect of both dimensions of digital finance on financing constraints. Moreover, considering the magnitude of the
coefficients, the digitization degree has the most significant effect on alleviating financing constraints for enterprises. In conclusion,
developing digital finance is beneficial in mitigating the financing constraints faced by small and medium-sized enterprises.

6.2. Analysis of different industries

In order to explore the differential impact of digital finance on the financing constraints of SMEs in different industries, a sectoral
analysis was conducted on the top five industries by proportion, including manufacturing, information technology services, leasing,
wholesale and retail, and scientific research and technical services.
According to the regression results in Table 6, it can be observed that digital finance significantly impacts SMEs’ financing in the
manufacturing and information technology services sectors at a level of 1 %. However, it does not significantly influence the leasing,
wholesale and retail, scientific research and technical services industries. This indicates that the impact of digital finance varies across
different industries.

7. Discussion

The emergence of digital finance, known as Fintech, has introduced a more diversified business model for financial institutions and
significantly eased financing constraints for SMEs. While existing literature mainly focuses on examining various emerging digital
financial models, there remains a gap in research regarding the dynamic evolution among stakeholders within the digital finance
landscape. This study addresses this gap by constructing a digital finance network evolution model based on Complex Adaptive System
(CAS) theory. Utilizing simulation to emulate the financing behavior of financial institutions towards SMEs, it investigates the
evolutionary process of digital finance networks.

Table 3
Decriptive statistics.
Variables (1) (2) (3) (4) (5)

N mean sd min max

WW 13,665 − 0.344 0.402 − 39.51 0.149

DIFI 13,665 2.936 0.423 2.004 3.777
Breadth 13,665 2.732 0.427 1.672 3.539
Depth 13,665 3.003 0.544 1.727 4.004
Digital 13,665 3.487 0.423 2.931 4.403
FINA 13,665 − 0.00598 1.125 − 125.1 0.722
CFA 13,665 − 0.000879 0.164 − 5.184 1.401
ROA 13,665 0.0246 0.159 − 6.479 0.970
AGE 13,665 1.743 1.367 0 12

15
W. Jun and X.Q. Ran Heliyon 10 (2024) e30586

Table 4
Correlation analysis.
WW DIFI FINA CFA ROA AGE

WW 1
DIFI − 0.057a 1
FINA 0.00400 − 0.0140 1
CFA 0.018b − 0.081a 0.00600 1
ROA − 0.0120 − 0.081a − 0.00800 0.511a 1
AGE − 0.052a 0.481a 0.00300 − 0.021b − 0.067a 1
c
p < 0.1.
a
p < 0.01.
b
p < 0.05.

Table 5
Regression result.
VARIABLES (13) (14) (15)

WW WW WW
c
DIFI − 0.003
(-1.89)
Depth − 0.003b
(-2.51)
Digital − 0.009a
(-3.80)
FINA 0.000 0.000 0.000
(0.70) (0.70) (0.70)
CFA − 0.009a − 0.009a − 0.009a
(-3.17) (-3.14) (-3.12)
ROA − 0.097a − 0.097a − 0.097a
(-27.27) (-27.26) (-27.27)
AGE − 0.004a − 0.004a − 0.004a
(-11.06) (-11.11) (-10.93)
Time effect YES YES YES
Industry effect YES YES YES
Constant − 0.316a − 0.316a − 0.294a
(-78.92) (-95.35) (-38.95)
Observations 13,665 13,665 13,665
R-squared 0.515 0.516 0.516

t-statistics in parentheses.
a
p < 0.01.
b
p < 0.05.
c
p < 0.1.

In line with Gambacorta et al. (2019), our research confirms that enterprises with solid profitability and good credit encounter
fewer financing constraints within the network [38]. What distinguishes this study is its utilization of simulation to explore whether
digital finance can mitigate external shocks, such as the COVID-19 pandemic, on SMEs. Our findings indicate that the more mature the
digital finance network, the better it can withstand liquidity squeezes resulting from emergency events.
Moreover, this study empirically validates the conclusions derived from the simulation using the Peking University Inclusive
Finance Index. The results demonstrate that digital finance indeed alleviates financing constraints for SMEs. However, we also uncover
that digital finance exacerbates financing constraints for the manufacturing industry. This may be attributed to the operational model
of traditional manufacturing industries, prioritizing business secrets and intellectual property, leading to a cautious approach towards
data disclosure. In contrast, digital finance platforms have higher requirements for data transparency, resulting in conflicting interests
between the two sectors and exacerbating financing constraints for the manufacturing industry.

8. Conclusion

Commencing with the analysis of capital flow and information flow, this paper establishes a digital financial network comprising
two primary nodes: the capital demand main body (SMEs) and the capital supply main bodies (government, bank financial platform).
Employing NetLogo simulation for the digital financial network, we investigate the influence of digital finance on the financing
constraints of SMEs, the validity and authenticity of the model are verified by an empirical study We explore the cost implications of
constructing a digital financial network and its responsiveness to emergencies. The research yields the following conclusions and
recommendations.

16
W. Jun and X.Q. Ran Heliyon 10 (2024) e30586

Table 6
Analysis of different industries.
VARIABLES (1) (2) (3) (4) (5)

Manufacturing Information Leasing Wholesale Scientific

a a
DIFI 0.005 − 0.009 − 0.006 0.001 − 0.004
(3.53) (-4.05) (-1.41) (0.18) (-0.88)
FINA 0.057a 0.014a − 0.000 0.102a 0.000
(5.78) (3.83) (-0.25) (5.25) (0.06)
CFA − 0.008b − 0.000 − 0.009 − 0.003 − 0.001
(-2.36) (-0.02) (-1.22) (-0.34) (-0.05)
ROA − 0.116a − 0.096a − 0.100a − 0.095a − 0.094a
(-25.82) (-20.10) (-10.87) (-8.15) (-6.90)
AGE − 0.004a − 0.004a − 0.005a − 0.006a − 0.005a
(-10.67) (-8.33) (-3.55) (-3.57) (-3.55)
Time effect YES YES YES YES YES
Constant − 0.329a − 0.175a − 0.294a − 0.329a − 0.232a
(-94.73) (-29.05) (-24.63) (-28.05) (-18.79)
Observations 6474 2943 792 687 630
R-squared 0.449 0.845 0.272 0.441 0.726

t-statistics in parentheses.
c
p < 0.1.
a
p < 0.01.
b
p < 0.05.

(1) Government financial support can accelerate the construction of digital financial networks, ease the financing constraints of
SMEs, and promote the development of SMEs. Moreover, the study found that the government strongly supports banks and
financial platforms to carry out digital transformation, reducing their transformation costs, which is more conducive to the
financing of SMEs and promoting the development of digital financial networks. Through the simulation results, it is found that
many banks and financial platforms will not choose to enter the digital financial networks if the transformation cost is very high
in the early stage, even if stimulated by high returns in the later stage. This may be due to the following reasons: First, for some
banks and financial platforms, there is insufficient capital or technical capacity, and these banks and financial platforms will
initially choose to wait and see; Second, The early investment of the financial platform is too large, which leads to the extension
of the fund withdrawal front and increases the risk of transformation; Finally, banks and financial platforms that venture into
digital financial networks are still in a capital shortage in the early stage because most of their funds are used for digital
transformation, and the funds that can be used to invest in SMEs are limited. Therefore, even if the rate of return is high, there
are not enough funds for investment.
(2) When considering the scale of lending to SMEs, banks and financial platforms should focus on long-term small loans. Long-term
micro-credit is more conducive to the development of SMEs, but it also can guarantee the profits of banks and financial plat­
forms. It is also conducive to expanding and developing the scale of digital financial networks. It is a mutually beneficial and all-
win financing method. Long-term financing not only relieves the loan repayment pressure of small and micro enterprises,
maintains their good cash flow and avoids bankruptcy risk, but also guarantees the risk control of banks and financial platforms
themselves and maximizes the income of each subject in the network.
(3) Banks and financial platforms should strengthen cooperation, which reduces the digitalization cost of banks, increases the
income of financial platforms, and enables mutual information sharing. The more transparent information digital financial
networks, the more conducive to the financing of SMEs. The high degree of information asymmetry between banks and financial
platforms and SMEs makes it impossible for banks and financial platforms to rationalize credit to avoid risks. Due to the lack of
mortgaged assets, SMEs cannot effectively provide their credit type to the bank, leading to their financing difficulties. However,
the digital financial network can help banks and financial platforms obtain the operating information of SMEs at low cost,
greatly reduce the information asymmetry, improve the ability to predict the default risk of borrowers, and increase the
availability of SME bank credit.
(4) Developing a digital financial network is advantageous in mitigating the impact of emergency events on the socio-economic
landscape. The more mature the network development, the better it can withstand sudden changes in the external environ­
ment. Simulation results indicate that when a mature digital financial network faces the impact of an event like the COVID-19
pandemic, the scale of the network and the average income of each main body are basically not affected. With a mature network
infrastructure, banks and financial platforms have established trust bridges for small and micro enterprises. Consequently, when
emergencies occur, these enterprises can borrow externally to overcome the liquidity crunch caused by the events, thus alle­
viating the impact of COVID-19 on them.
(5) The government can encourage the development of digital financial networks by providing tax incentives and financial support
to technology innovation enterprises, thereby promoting digital transformation and expanding network scale. It also needs to
pay attention to the potential risks digital financial networks pose regarding cybersecurity threats, data privacy, and regulatory
challenges. The government should play a guiding and supervisory role in this aspect. Strengthen network security, establish
perfect network security protection measures, and use technical means such as data encryption, identity identification, and

17
W. Jun and X.Q. Ran Heliyon 10 (2024) e30586

firewalls to ensure the security of user data. Protecting user privacy requires the government to establish corresponding privacy
protection laws and regulations and regulatory agencies to ensure compliance with relevant standards and regulations in data
collection and processing, thus safeguarding users’ privacy rights. Strengthening cooperation among regulatory agencies in
different countries and regions is necessary to establish common regulatory standards and regulations, promoting the healthy
development of digital financial networks. Platforms and the government should also focus on user education and training,
raising awareness of network security and privacy protection, and helping users better protect their interests.

We acknowledge a limitation in the current model, focusing solely on the borrowing mode for SMEs. We recognize the need for
expansion and improvement to include more comprehensive and accurate representations of different financing modes employed by
financial platforms. For instance, the interactions among various entities in the model and the evolution of the digital financial network
when SMEs obtain financing through issuing stocks or converting debt into equity warrant further exploration. This aspect requires
additional refinement and will be the focal point of our future efforts.

Data availability statement

Data will be made available on request.

CRediT authorship contribution statement

Wang Jun: Writing – review & editing. Xin Qian Ran: Writing – review & editing, Writing – original draft.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing in­
terests:Wang Jun reports financial support was provided by Sichuan Science and Technology Plan Project. Wang Jun reports financial
support was provided by Humanities and Social Sciences Research project of Ministry of Education of China. If there are other authors,
they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the
work reported in this paper.

References

[1] Junkuo Zhang, The goals and strategy path of high-quality development, J. Manag. World 35 (7) (2019) 1–7, https://doi.org/10.19744/j.cnki.11-1235/
f.20190711.001.
[2] Yong JaeShin, In Lee, Fintech: ecosystem, business models, investment decisions, and challenges, Bus. Horiz. 61 (2018) 35–46, https://doi.org/10.1016/j.
bushor.2017.09.003.
[3] yiping Huang, Huang Zhuo, Development of digital finance in China: present and future, China Econ. Q. 17 (4) (2018) 1489–1502. https://link.cnki.net/doi/10.
13821/j.cnki.ceq.2018.03.09.
[4] Jing YuWu, Jian Xiao, Jie Li, The impact of digital finance on household consumption: evidence from China, Econ. Modell. 86 (2020) 317–326, https://doi.org/
10.1016/j.econmod.2019.09.027.
[5] william cong Lin, M.C. Charles, Y.Y. LeeQu, T. Shen, "Valley of Death" and "Exit Trap" fetters Chinese start-ups: financing status and predicament of Chinese
start-ups, Tsinghua Business Review (2019) 34–40.
[6] Yu kun, Li Zho, Zhang Xiaorong, et al., The puzzle of firm investment efficiency: financing constraint hypothesis and monetary policy impact, Econ. Res. J. 49
(5) (2014) 106–120. DOI: CNKI:SUN:JJYJ.0.2014-05-009.
[7] kebin Deng, Zeng Haijian, Financial constraints of Chinese firms: a test of their characteristics and causes, Econ. Res. J. 49 (2) (2014) 47–60, 140.DOI: CNKI:
SUN:JJYJ.0. 2014-02-004.
[8] Zhu wuxiang, Zhang Ping, Li Pengfei, et al., Firm crisis, government support and policy efficiency under the epidemic shock: evidence from two waves of
questionnaire on SMEs, Journal of Management World 36 (4) (2020) 13–26. DOI: CNKI:SUN:GLSJ.0. 2020-04-005.
[9] Laeven Luc, Levine Ross, Michalopoulos Stelios, Financial innovation and endogenous growth, J. Financ. Intermediation 24 (1) (2015) 1–24, https://doi.org/
10.1016/j.jfi.2014.04.001.
[10] xun Zhang, Wan Guanghua, Zhang Jiajia, et al., Digital economy, financial inclusive, and inclusive growth, Econ. Res. J. 54 (8) (2019) 71–86. https://kns.cnki.
net/kcms/detail/11.1081.F.20190819.1737.010.
[11] Vs Helmut-Bester, Screening. Rationing in credit markets with imperfect information, Am. Econ. Rev. 75 (4) (1985) 850–855. https://www.webofscience.com/
wos/alldb/full-record/WOS:A1985APZ2600020.
[12] Y.S. Chan, G. Kanatas, Asymmetric valuations and the role of collateral in loan agreements, J. Money Credit Bank. 17 (1) (1985) 84–95. https://www.
webofscience.com/wos/alldb/full-record/WOS:A1985AAX7800007.
[13] xiao Wang, Zhang Jie, On the bank credit rationing and loan of small and medium-sized enterprises (SMEs), Econ. Res. J. 7 (2003) 68–75, 92.DOI: CNKI:SUN:
JJYJ. 0.2003-07-008.
[14] J.E. Stiglitz, A. Weiss, Credit rationing in markets with imperfect information, Am. Econ. Rev. 73 (1981) 393–410. https://www.webofscience.com/wos/alldb/
full-record/WOS:A1981LU03100011.
[15] Z.Y. Li, Banking structure and the small-medium-sized enterprise financing, Econ. Res. J. 6 (2002), 38-45, 94.DOI: CNKI:SUN:JJYJ.0.2002-06-004.
[16] yifu Lin, Li Yongjun, Promoting the growth of medium and small-sized enterprises through the development of medium and small-sized financial institutions,
Econ. Res. J. 1 (2001) 10–18, 53-93, http://ir.calis.edu.cn/hdl/211010/4926.
[17] yifu Lin, Sun Xifang, Banking structure and economic growth:evidence from Chinese Provincial panel data, Econ. Res. J. 43 (9) (2008) 31–45, https://doi.org/
10.1007/s11459-009-0026-z.
[18] yifu Lin, Sun Xifang, Jiang Ye, Toward A theory of optimal financial structure in economic development, Econ. Res. J. 44 (8) (2009) 4–17.
[19] chang Liu, Liu Chong, Ma Guangrong, Small-and-medium banks and small-and-medium enterprise loans, Econ. Res. J. 52 (8) (2017) 65–77. DOI: CNKI:SUN:
JJYJ.0. 2017-08-006.
[20] Hao Huang, The Formation and challenges of digital financial ecosystem-experience from China, Economist (4) (2018) 80–85, https://doi.org/10.16158/j.
cnki.51-1312/f.2018.04.011.

18
W. Jun and X.Q. Ran Heliyon 10 (2024) e30586

[21] Guo feng, Kong Tao, Wang Jingyi, The spatial agglomeration effect of internet finance——evidence from the financial development index, Stud. Int. Fin. 8
(2017) 75–85, https://doi.org/10.16475/j.cnki.1006-1029.2017.08.008.
[22] Guo pin, Shen Yue, Internet finance, deposit competition, and bank risk-taking, J. Financ. Res. 8 (2019) 58–76. DOI: CNKI:SUN:JRYJ.0.2019-08-004.
[23] Qiu han, Huang Yiping, Ji Yang, The influence of fintech on traditional bank behavior: from the perspective of internet finance, J. Financ. Res. 11 (2018) 17–29.
[24] yu Wu, Li Chengshun, Lixiao Yi-Daichun, The impact of digital finance development on traditional private credit market, Journal of Management World 36 (10)
(2020) 53–64, https://doi.org/10.19744/j.cnki.11-1235/f.2020.0153, 138, 65.
[25] xun Zhang, et al., Digital finance and household consumption:theory and evidence from China, Journal of Management World 36 (11) (2020) 48–63, https://
doi.org/10.19744/j.cnki.11-1235/f.2020.0168.
[26] Huang yiping, Qiu Han, Big tech lending:A new credit risk Management framework, Journal of Management World 37 (2) (2021) 12–21, https://doi.org/
10.19744/j.cnki.11-1235/f.2021.0016. , 50, 2, 16.
[27] Kevin Jdooley, Manus Rungtusanatham, Thomas Ychoi, Supply networks and complex adaptive systems: control versus emergence[J/OL], J. Oper. Manag. 19
(2001) 351–366, https://doi.org/10.1016/s0272-6963(00)00068-1.
[28] Yun sookHong, Thomas Ychoi, Unveiling the structure of supply networks: case studies in Honda, Acura, and DaimlerChrysler[J/OL], J. Oper. Manag. 20 (2002)
469–493, https://doi.org/10.1016/s0272-6963(02)00025-6.
[29] Amit karna, Florian Täube, Petra Sonderegger, Evolution of innovation networks across geographical and organizational boundaries: a study of R & D
Subsidiaries in the B angalore IT cluster, Eur. Manag. Rev. 10 (4) (2013) 211–226, https://doi.org/10.1111/cp.12012.
[30] Enhua Hu, Hong Liu, Study on cluster innovation enterprises and external environments based-on synergy innovation, Scientific Management Research (3)
(2007) 23–26, https://doi.org/10.19445/j.cnki.15-1103/g3.2007.03.007.
[31] Robin Cowan, Nicolas Jonard, A. Joel, C. Baum, Network-independent partner selection and the evolution of innovation networks[J/OL], Manag. Sci. 56 (2010)
2094–2110, https://doi.org/10.1287/mnsc.1100.1229.
[32] S. Battiston, M. Napoletano, F. Schweitzer, Michael D. König, Recombinant knowledge and the evolution of innovation networks[J/OL], J. Econ. Behav. Organ.
79 (2011) 145–164, https://doi.org/10.1016/j.jebo.2011.01.007.
[33] Chen jindan, H.U. Hanhui, Ji Min, Research on endogenous evolution of industrial cluster knowledge network:based on multi-agent simulation, J. Intell. 31 (7)
(2012) 128–133, https://doi.org/10.3969/j.issn.1002-1965.2012.07.026.
[34] Doina Olaru, Sara Denize, Sharon Purchase, Innovation network trajectories and changes in resource bundles[J/OL], Ind. Market. Manag. 43 (2014) 448–459,
https://doi.org/10.1016/j.indmarman.2013.12.013.
[35] Xuyulian Yu-Lang, Research on evolution simulation of regional Sei-tech finance network based on CAS theory, Science and Technology Management Research
40 (3) (2020) 46–56. DOI: CNKI:SUN:KJGL.0.2020-03-008.
[36] G.-B. Korovin, Architecture of the agent-based model for the region’s industrial complex digital transformation, Journal of New Economy 21 (2020) 158–174,
https://doi.org/10.29141/2658-5081-2020-21-3-8.
[37] y Huang, L. Zhang, Z. Li, et al., Fintech credit risk Assessment for SMEs: evidence from China, IMF Working Papers 20 (193) (2020), https://doi.org/10.5089/
9781513557618.001.
[38] l. Gambacorta, Y. Huang, H. Qiu, et al., How do machine learning and non-traditional data affect credit scoring? New evidence from a Chinese fintech firm, BIS
Working Papers 12 (2019).
[39] Yaniv proselkov, Jie Zhang, Liming Xu, Erik Hofmann, Y. Thomas, Dale Rogers Choi, Alexandra Brintrup, Financial ripple effect in complex adaptive supply
networks: an agent-based model, Int. J. Prod. Res. (2023), https://doi.org/10.1080/00207543.2023.2173509.
[40] Jing RongGoh, Didier Sornette, He Wang, Esther YingYang, Shaun ShuxunWang, Government support for SMEs in response to COVID-19: theoretical model
using Wang transform, China Finance Rev. Int. 11 (2021) 406–433, https://doi.org/10.1108/cfri-05-2021-0088.
[41] xia Cao, Liu Guowei, Research on the relationship between network Scale，Connection mechanism of I-U cooperative innovative network and innovative
performance—based on multi-agent simulation and dynamic system theory, Oper. Res. Manag. Sci. 24 (2) (2015) 46–56. DOI: CNKI:SUN:YCGL.0.2015-02-034.
[42] Wu Guojun, Toni M. Whited, Financial constraints risk, Rev. Financ. Stud. 19 (2006) 531–559, https://doi.org/10.1093/rfs/hhj012.

19