Evaluating the Reputation of Internet Financial Platforms in China: A Sustainable Operations Perspective

Abstract

In China, many Internet financial platforms (IFPs) are grappling with sustainability challenges due to elevated default rates, which have triggered widespread investor anxiety. To evaluate the sustainability practices of these platforms, we propose a reputation evaluation model designed to rank IFPs based on their sustainability. The economic sustainability of an IFP is decomposed into three components: scale strength, capital liquidity, and sustainable operating capability. Through an analysis of the correlation relationships between various indicators, we have identified nine significant indicators. Mathematical models are established to quantify these nine indicator variables. Subsequently, the score values of each indicator are integrated to establish a reputation evaluation model utilizing the weighted geometric mean method. Furthermore, the reputation evaluation values for 18 Chinese IFPs were calculated using the developed model, and the sustainability of the platforms was ranked according to the reputation evaluation value. A comparative analysis was also conducted between the sustainable rankings proposed in this study and the development rankings of the “Home of Online Loans” (HOL). The results reveal that our model effectively considers both the current operational strength and the sustainable development capability of the platform. It successfully identifies platforms with poor sustainability, assisting investors in making more informed decisions. Simultaneously, this study identifies key indicators influencing the sustainability of IFPs, providing valuable insights for managers seeking to enhance the sustainable operational levels of their platforms.

1. Introduction

Internet financial platforms (IFPs) represent a novel financing framework resulting from the fusion of the Internet and finance. These platforms encompass a range of services including small loans, Internet banking, supply chain finance, crowdfunding, and peer-to-peer (P2P) lending [1,2,3]. The emergence of IFPs has not only effectively broadened the scope of financial services but has also promoted the sustainable development of the financial sector. Specifically, IFPs are instrumental in mobilizing and efficiently allocating financial resources [4], thereby opening new opportunities for sustainable entrepreneurs [5], contributing to poverty reduction, and promoting economic sustainable development [6]. Despite these positive aspects, recent years have seen IFPs grappling with a liquidity crisis, with a significant number facing sustainability challenges, primarily due to excessive defaults [7,8]. As of the beginning of 2021, over 6000 P2P lending platforms had declared bankruptcy in China [9], highlighting a pressing issue: the sustainable operations of IFPs are under considerable strain in the Chinese context. Against this backdrop, it becomes imperative and intriguing to delve into the measurement and attainment of sustainability for IFPs.

According to the World Business Council for Sustainable Development (WBCSD) definition, sustainability is the ability to meet people’s needs today without compromising the ability of future generations to meet their own needs [10,11]. From an enterprise perspective, sustainability is achieving corporate goals and adding long-term value to shareholders by integrating economic, environmental, and social opportunities into corporate strategy [12,13]. However, determining whether an IFP is “sustainable” poses a complex and challenging problem. The operating boundaries of IFPs are often blurred, making it difficult to clarify their evaluation indicators of sustainability. Furthermore, there is no unified and universally agreed method for assessing the sustainability of IFPs. Reputation plays a positive role in promoting the sustainable operation of IFPs. A good reputation can attract more investors [14,15], improving the financial sustainability of the platform. Previous studies have proven that reputation can establish a trust mechanism in the P2P lending market, reduce the defaults of borrowers, and promote the sustainable operation of P2P lending platforms [16,17]. Additionally, various third-party credit rating agencies have attempted to measure the sustainability of IFPs based on reputation. For instance, a sustainable development index of P2P lending platforms was created by the “Home of Online Loans” (HOL) [18], and a sustainability ranking list of P2P lending platforms was launched by “Wang Dai Tian Yan” (WDTY) [19]. Therefore, this paper aims to construct a reputation evaluation model to rank the sustainability of IFPs.

The reputation issues of specific IFPs, such as P2P lending and crowdfunding, have increasingly garnered attention in recent years, though research in this area remains in its early stages. The existing literature predominantly focuses on the reputation problems related to borrowers and investors within IFPs, examining primarily mechanisms and factors influencing reputation. However, there is a notable scarcity of research addressing the evaluation of the reputation of IFPs. For example, a social reputation loss model for P2P platform borrowers who lost contact has been proposed [20]. Still, this model has primarily been applied to the social reputation evaluation of P2P lending borrowers and has not been extended to the reputation evaluation of IFPs. Furthermore, many existing reputation evaluation models rely on financial indicators such as transaction amounts, loan volumes, and total assets. Empirical methods are commonly employed to investigate the relationship among these indicators. However, these studies often fail to establish a clear connection between financial performance and sustainability. Additionally, most reputation evaluation studies are often grounded in the current operational strength of IFPs, with less emphasis on their sustainable development capabilities. This oversight means that the current ranking systems may not adequately reflect the sustainable operational risks associated with IFPs.

Overall, according to the overview of the reputation problem of IFPs, the research questions and gaps that motivate this research can be distilled as follows: (1) How should one rank the sustainability of IFPs using platform reputation indicators? (2) Which reputation indicators genuinely reflect of sustainability of IFPs? (3) In what manner can a reputation evaluation model be developed to comprehensively consider both the operating strength and sustainable development ability of IFPs?

To address the above-mentioned issues, three main tasks are implemented as follows. Firstly, the sustainability of IFPs is decomposed into three components: scale strength, capital liquidity, and sustainable operating ability. The selected reputation evaluation indicators are further categorized into current operating strength and sustainable development capability. Secondly, this paper establishes a comprehensive reputation evaluation model by quantifying nine indicator variables through mathematical models and integrating the score values of each index using the weighted geometric mean method (WGMM). Subsequently, a mathematical analysis of the model is presented. Thirdly, an empirical analysis based on transaction data from 18 P2P lending platforms on the Home of Online Loans (HOL) verifies the model’s effectiveness and applicability in assessing the reputation and sustainability of IFP operations.

The contributions of this work can be outlined through the following three key points. (1) Compared to the previous literature, which mainly focuses on the reputation mechanism, factors influencing reputation, and IFPs’ borrower reputation, this study expands the research on the reputation evaluation of platforms from a new perspective of sustainable operations. (2) The mathematical model, the WGMM, and an empirical study are combined for evaluating the reputation of IFPs in this paper, which expands our approach to assess platform reputation. (3) The study draws interesting conclusions from the empirical analysis. The results reveal a positive correlation between the sustainability of the IFPs and factors such as loan balance, monthly transaction volume, platform background, average rate of return, and average loan term. Conversely, negative correlations are observed with loan amount per capita. Furthermore, the fluctuation of sustainability is identified as being primarily influenced by the capital flow rate. These insights contribute to a deeper understanding of the dynamics impacting the sustainability of IFPs.

The rest of this paper is organized as follows: Section 2 provides an extensive review of the existing literature, focusing on three major research directions: the sustainability of IFPs, the mechanism of and factors influencing reputation, and the evaluation methods of reputation. Section 3 introduces the selection process of reputation evaluation indicators, offering a detailed classification of these indicators. In Section 4, the reputation evaluation model for sustainable operations of IFPs is established and discussed. In Section 5, empirical data are used to test the effectiveness and practicability of the model. Additionally, it analyzes the sustainability of IFPs based on the findings. Section 6 concludes the research.

2. Literature Review

Aligned with the title and structure of this paper, the related work is reviewed through three research streams: the sustainability of IFPs, the mechanism and factors influencing reputation, and the reputation evaluation methods.

First, the sustainability of IFPs takes center stage in this study. Research on the reputation of P2P lending platforms has shown that lenders’ trust promotes the sustainable development of P2P lending [21]. The role of platform providers in fostering the sustainable growth of P2P lending as an alternative financial form is emphasized [22]. You et al. designed an interest coordination mechanism among P2P lending participants to encourage sustainable lending practices [23]. Simultaneously, a conceptual model of the P2P lending process is constructed from the perspective of sustainable operation [24]. Moreover, an innovative sustainable business model for P2P sharing of underutilized assets facilitated by digital platforms is examined by Piscicelli et al. [25]. According to Yu and Shen [26], it is noted that the stringent regulations of P2P lending platforms limit their brokerage role, potentially jeopardizing their commercial sustainability and affecting the sharing economy’s openness and inclusivity. In the same context, Wang et al. proposed a social co-governance model for the P2P lending market aimed at protecting investor interests, which could bolster the sustainable development of the FinTech industry [27]. Additionally, Gupta and Shivnani [28] identified major challenges to the sustainable operation of P2P lending platforms, focusing on security concerns affecting both investors and borrowers. Gao et al. noted that liquidity issues and “bank-run” crises are significant contributors to defaults among P2P lending platforms in China [29]. Chen et al. argued that fostering consumer loyalty is crucial for the sustainable management of Internet consumer finance platforms [30]. Regarding the sustainability of crowdfunding, a five-dimensional framework to conceptualize the meaning of crowdfunding for sustainability is proposed by Petruzzelli et al. [31]. Moreover, crowdfunding is proven to provide unique opportunities for sustainable entrepreneurs [32]. As Liang et al. [33] pointed out, the success of crowdfunding projects has promoted the sustainable development of growing companies.

Second, the mechanism of and factors influencing reputation are investigated in the existing literature. An effective reputation mechanism has been demonstrated in the P2P lending market through the analysis of transaction data from 78,000 borrowers on the P2P lending platform called “renrendai.com” [34]. Yang and Lee found that reputation significantly influences the trust of lenders on P2P lending platforms [35]. Moreover, the reputation of platforms is proven to play a direct and indirect role in investors’ investment choices [36]. As Kuwabara et al. [37] pointed out, reputation exhibits a curve effect on the success rate of P2P lending. Davies and Giovannetti suggested that reputation plays a crucial role in the success of crowdfunding projects [38]. Similarly, Li et al. [39] found that the perceived reputation of project initiators significantly influences individuals’ donation intentions in crowdfunding projects. Wang et al. [40] further noted that a negative reputation evaluation of Internet financing platforms can more easily raise concerns among potential investors. Additionally, the network reputation of the circle of friends is found to restrain the default behavior of P2P lending borrowers [41]. In the previous studies, the direct reputation indicators are often adopted to construct a reputation evaluation indicator system. For example, third-party credit rating agencies, such as the HOL and WDTY, calculate the reputation level of IFPs based on indicators such as transaction volume per month, average rate of return, platform background, leverage ratio, net capital inflow, and loan dispersion. Xia et al. found that basic features, capital security, operations management, and social networks have a significant impact on identifying problematic platforms [42]. Moreover, other factors such as capital inflow [43], the capital adequacy ratio [44], and information transparency [45] are considered to have a significant influence on the reputation of IFPs.

Third, this study also contributes to the literature on the evaluation method of reputation. For instance, Pang and Yang proposed a social reputation loss model for the disconnection of P2P platform borrowers [20]. You et al. proposed an improved fuzzy evaluation approach (IFEA) for the reputation evaluation of IFPs [9]. Moreover, Deng et al. designed a reputation model to dynamically assess borrower reputation to help lenders and improve the allocative efficiency in P2P lending markets [46]. Additionally, Zong et al. presented a reputation evaluation model which combines backpropagation neural networks with a point reputation-weighted directed network model to resist single-point attacks [47]. Panagopoulos et al. proposed a robust reputation system consisting of new reputation measurement and attack prevention mechanisms [48]. Feng et al. presented a hierarchical and configurable reputation evaluation method based on collaborative filtering (CF) for evaluating the service reputation of cloud manufacturing enterprises on a CMfg service platform [49]. Shi et al. calculated the reputation value of P2P lending platforms based on the beta function, with their results indicating that the reputation of P2P lending platforms plays a direct and indirect role (through credit enhancement information) in investors’ investment choices [36]. Other methods, such as a reputation assessment approach for cloud services based on a hybrid deep learning model [50], a dynamic reputation evaluation method based on information behavior [51], and a second-order reputation evaluation model [52], have also been proposed.

To sum up, although the aforementioned literature has discussed reputation evaluation problems of IFPs from different perspectives, there are still limitations that need to be addressed. The following conclusions can be summarized:

(1) Prior research on the sustainability of IFPs has been carried out in three aspects: mechanisms and factors to promote sustainable lending [21,22,23,32,33], concepts or business models of sustainable operations [24,25,27,31], and factors influencing the sustainability of IFPs [26,28,29,30]. However, few studies have focused on ranking the reputation of the sustainable operation of platforms. In practice, a sustainable reputation ranking of IFPs can help investors effectively distinguish between high-risk and low-risk platforms in online lending. Hence, further research is needed to explore quantitative methods that can measure the sustainability of IFPs.

(2) The majority of the existing literature on the reputation of IFPs focuses on the reputation mechanism [34,35,36,37,38,39,40,41], factors influencing reputation [42,43,44,45], and borrower reputation measurement [20,46], rather than reputation evaluations of IFPs. Moreover, the existing reputation evaluation indicator system only focuses on the factors influencing current operating strength. Yet, the indicators influencing the sustainable operating capability of IFPs have been ignored in previous studies. These gaps prompt our focus on constructing a reputation evaluation indicator system that accounts for the current operating strength and sustainable operation capability of IFPs.

(3) Empirical approaches [36] or quantitative research such as mathematical models [20,46,51,52], machine learning techniques [47,48,49,50], and fuzzy evaluation approaches [9] are popularly used to assess reputation in the existing literature. In these studies, Pang and Yang [20] and Deng et al. [46] present reputation evaluation models of P2P lending borrowers, but these have not been extended to the reputation evaluation of IFPs. Zong et al. [47], Wang and Panagopoulos et al. [48], Feng et al. [49], and Li et al. [52] do not carry out empirical analyses. Moreover, the WGMM used to establish reputation evaluation models is often ignored in previous studies. In this paper, we combine the mathematical model, the WGMM, and an empirical study to evaluate the reputation of IFPs.

To fill the aforementioned gaps, this study aims to rank the sustainability of IFPs based on reputation evaluation value. The detailed steps of the evaluation process are shown in Figure 1: (i) Conducting an association analysis and selecting reputation indicators; (ii) Establishing mathematical models to quantify reputation evaluation indicators; (iii) Establishing a reputation evaluation model utilizing the WGMM; and (iv) Calculating the reputation evaluation value of IFPs based on data from 18 P2P lending platforms. On the one hand, the results of this paper assists in determining whether a platform’s operation is “sustainable”, aiding investors in making informed and rational investment decisions. On the other hand, it identifies key indicators influencing the sustainability of IFPs, providing both a theoretical foundation and empirical reference for enhancing a platform’s sustainable operational level and competitiveness.

3. The Association Analysis and Selection of Reputation Indicators

3.1. The Association Analysis of Reputation Indicators

From HOL, WDTY, and the relevant literature [43,44,45,53,54], we selected 18 reputation evaluation indicators. These include the net capital inflow, monthly investment number, monthly transaction volume, average loan term, loan amount per capita, average rate of return, monthly number of loans, operating time, loan balance, pending amount, full bidding time, registered capital, investment amount per capita, background of the platform, capital flow rate, sustainable growth rate, capital adequacy ratio, and information transparency. Subsequently, we have categorized the relationship between reputation indicators into two types: the business logical association relationship and calculation association relationship. The business logical association relationship refers to the connection of, coordination of, or supporting relationship between indicators based on business characteristics. Meanwhile, the calculation association relationship refers to the calculation relationship established according to the definition of the calculation formula. For instance, the indicator “loan balance” is a component of the “capital adequacy ratio” calculation formula. Then changes in indicator “loan balance” directly impact the “capital adequacy ratio”. Therefore, from the perspective of the calculation formula, the “loan balance” influences the “capital adequacy ratio”. Finally, the indicators’ business logical association relationship and calculation association relationship are clarified. Based on the presence or absence of associations between the indicators, an association model among the reputation indicators of IFPs is established, as shown in Figure 2.

According to the correlation relationship among the above indicators in Figure 2, the results show that 14 relationship chains impact the reputation indicators of IFPs, as shown in

Table 1. The relationship chains of reputation indicators of IFPs.

Number	Relationship Chain
1	Full bidding time → average loan term → scale strength of platform → reputation of IFPs
2	Monthly transaction volume → scale strength of platform → reputation of IFPs
3	Background of platform → scale strength of platform → reputation of IFPs
4	Average rate of return → scale strength of platform → reputation of IFPs
5	Monthly number of loans → loan dispersion → capital liquidity → reputation of IFPs
6	Loan amount per capita → loan dispersion → capital liquidity → reputation of IFPs
7	Monthly investment number → loan dispersion → capital liquidity → reputation of IFPs
8	Investment amount per capita → loan dispersion → capital liquidity → reputation of IFPs
9	Net capital inflow → capital flow rate → capital liquidity → reputation of IFPs
10	Operating time → sustainable growing rate → sustainable operating ability → reputation of IFPs
11	Amount to be recovered → capital adequacy ratio → sustainable operating ability → reputation of IFPs
12	Loan balance → capital adequacy ratio → sustainable operating ability → reputation of IFPs
13	Registered capital → capital adequacy ratio → sustainable operating ability → reputation of IFPs
14	Information transparency → sustainable operating ability → reputation of IFPs

.

To identify the key indicators influencing the reputation of IFPs, the following detailed steps were taken: (i) If a relationship chain does not intersect with other chains, the first indicator of the chain is considered as the key indicator. (ii) If a relationship chain intersects with other chains, the first intersected indicator is selected as the key indicator. It is worth noting that indicators such as platform scale strength, capital liquidity, and sustainable operation ability cannot be directly quantified and only hold representational significance. Consequently, they are not utilized as the intersected indicators in this analysis. According to the two steps, an analysis tree consisting of 14 relationship chains of reputation evaluation indicators is obtained, as shown in Figure 3.

Using the selection of monthly transaction volume and loan dispersion as examples, we can explain the process of the key indicator selection. There are two interconnected chains linking the monthly transaction volume. Specifically, these two chains (“monthly transaction volume → scale strength of platform → reputation of IFPs” and “monthly transaction volume → sustainable growth rate → sustainable operating ability → reputation of IFPs”) intersect with the monthly transaction volume indicator. The monthly transaction volume can be directly quantified, so it is chosen as the key indicator. Similarly, the other eight indicators (the loan dispersion, background of the platform, average loan term, capital flow rate, average rate of return, sustainable growing rate, capital adequacy ratio, and information transparency) are also acquired. The following sections will explain the specific meaning of these nine key indicators.

3.2. The Selection of Reputation Indicators

3.2.1. Reputation Indicators of the IFPs’ Scale Strength

The monthly transaction volume refers to the amount of loans the IFP trades in the current month. The larger the monthly transaction volume of loans on the platform, the more transactions occur, enhancing the platform’s competitiveness and amplifying its impact on reputation. The background of a platform refers to the type of social capital force behind the IFP. The stronger the social capital force supporting the platform, the more abundant the platform’s fund and the stronger its anti-risk capability. Furthermore, the average rate of return is the market performance of matching transactions between wealth managers and borrowers. It directly affects the amount and enthusiasm of investors. If the platform’s average rate of return is too low, it results in a low level of investment attraction and a low reputation. Moreover, the average loan term represents the liquidity of the platform’s funds. The longer the loan term, the stronger the capital turnover capacity of the platform. Additionally, a longer loan term contributes to a higher reputation for the platform.

3.2.2. Reputation Indicators of IFPs’ Capital Liquidity

The capital flow rate refers to the ratio of inflow and outflow of funds from financial lending platforms. As Teubner et al. [43] pointed out, the capital flow rate is closely related to the operating strength of enterprises. Enhanced capital liquidity signifies more robust cash flow for the platform, bolstering its ability to resist risks. Loan dispersion refers to the degree of decentralization of loan amounts on IFPs. As Cao et al. [54] highlighted, a better loan dispersion reflects the anti-risk capability of the IFP. A greater loan dispersion on the platform results in smaller per capita loan amounts. This diminishes systemic risk caused by the bad debts of individual borrowers, thereby strengthening the anti-risk capability of the platform.

3.2.3. Reputation Indicators of the IFPs’ Sustainable Operating Ability

The sustainable growth rate refers to the sustainable increase in the transaction amount of an IFP, reflecting the stable growth of the platform. The higher the sustainable growth rate of the IFP, the better the platform’s reputation. The capital adequacy ratio denotes the ratio of the platform’s capital to venture capital. This ratio serves as a reflection of the platform’s reputation regarding its anti-risk capability. As indicated by [44], a higher capital adequacy ratio strengthens the platform’s ability to resist risks. Information transparency represents the degree of information disclosure by IFPs. Compliance-oriented IFPs are typically inclined to disclose more information, showcasing the sustainability and safety of platform operations. Conversely, problematic platforms often resist disclosure or provide misleading information to conceal business risks rooted in speculative psychology [45].

Based on the analysis above, nine reputation evaluation indicators have been identified, including the monthly transaction volume, background of the platform, average rate of return, average loan term, capital flow rate, loan dispersion, capital adequacy ratio, sustainable growing rate, and information transparency. According to the characteristics of reputation indicators, this paper divides the above nine indicators into two categories. The first group consolidates indicators reflecting the scale strength and capital liquidity of IFPs and incorporating the monthly transaction volume, background of the platform, average rate of return, average loan term, capital flow rate, and loan dispersion to represent the current operating strength of IFPs. The second group, comprising the sustainable growth rate, capital adequacy ratio, and information transparency, is adopted to represent the sustainable development ability of IFPs. The structure of reputation evaluation indicators for the sustainability of IFPs is illustrated in Figure 4.

4. Reputation Evaluation Model of IFPs

4.1. Model Assumptions and Parameter Setting

In light of the above, three appointments and two reality-based assumptions are outlined as follows:

Appointment 1.

The reputation evaluation value of the IFP at the initial moment is 0.

Appointment 2.

The reputation evaluation value of the IFP with delayed payment is 0.

Appointment 3.

The reputation evaluation value of the IFP that has run off is 0.

Assumption 1.

Since investors are profit-seeking and only invest in IFPs that obtain a profit, this paper assumes that the IFPs studied are all loss-free platforms with non-negative average returns.

Assumption 2.

Since the national law stipulates that the annual interest rate agreed upon by the lender and borrower shall not exceed 36% [55], and any interest agreement exceeding this limit shall be considered invalid, this paper assumes that all the IFPs studied comply with the provisions of national law, and the average rate of return shall not exceed 36%.

According to the aforementioned appointments and assumptions, the parameters involved in the reputation evaluation model of IFPs are defined, as shown in

Table 2. Symbols and descriptions of the parameters.

Parameter	Description	Units of Measure
i	i represents an IFP.	/
t	$t \mathrm{represents} \mathrm{time} (\mathrm{time} \mathrm{step} \mathrm{is} \mathrm{month}), t\ge 0$.	month
$B(i,t)$	The monthly transaction volume for the IFP i at time t.	10,000 CNY
$V(i,t)$	The number of loans of the IFP i at time t.	/
$A(i,t)$	The pending amount for the IFP i at time t.	10,000 CNY
$Z(i,t)$	The loan balance for the IFP i at time t.	month
$L(i,t)$	The average loan term for the IFP i at time t.	month
$T(i,t)$	The operating time of the IFP i by the end of time t.	month
$C(i,t)$	The registered capital of the IFP i at time t.	10,000 CNY
$E(i,t)$	The loan amount per capita of the IFP i at time t.	10,000 CNY
$W(i,t)$	The turnover rate of monthly transaction volume for the IFP i at time t.	/
$G(i,t)$	The capital adequacy ratio of the IFP i at time t.	/
$F(i,t)$	The background of the IFP i at time t.	/
$D(i,t)$	The loan dispersion of the IFP i at time t.	/
$H(i,t)$	The capital flow rate of the IFP i at time t.	/
$Y(i,t)$	The information transparent of the IFP i at time t.	/
$N(i,t)$	The average rate of return for the IFP i at time t.	/
$Q(i,t)$	The score of average loan term for the IFP i at time t.	/
$R(i,t)$	The value of reputation evaluation of the IFP i at time t.	/

.

4.2. Calculation of Reputation Evaluation Indicators

(1) The monthly transaction volume score

The monthly transaction volume score of IFPs can be measured by the ratio of the monthly transaction volume of IFP $i$ to the mean transaction volume of all IFPs at time $t$. Then, the monthly transaction volume score $W(i,t)$ of IFPs is calculated as follows:

$$W(i,t)={\displaystyle \frac{B(i,t)}{{\mu }_t}}$$

(1)

where ${\mu }_t$ denotes the average value of the monthly transaction volume of all IFPs at time t. In the specific calculation, the average value of the monthly transaction volume of all the platforms listed in the “Home of Online Loans” is taken as the value of ${\mu }_t$.

(2) The background score of platforms

According to the differences in capital forces behind IFPs, we subdivide IFPs into five types: private sector, banking sector, state-owned sector, listed sector, and venture capital sector. The background score $F(i,t)$ of IFPs is defined as follows:

$$F(i,t)=\{\begin{array}{ll}a& \mathrm{state}\text{-}\mathrm{owned} \sec \mathrm{tor} \mathrm{or} \mathrm{banking} \sec \mathrm{tor}\\ b& \mathrm{listed} \sec \mathrm{tor}\\ c& \mathrm{venture} \mathrm{capital} \sec \mathrm{tor}\\ d& \mathrm{private} \sec \mathrm{tor}\end{array}$$

(2)

where the IFPs of the banking sector and state-owned sector possess relatively reliable capital, they exhibit a strong anti-risk ability. Consequently, the platform investment security is relatively high, and the background score of the platform at time $t$ is set to $F(i,t)=a$. In the case of the listed platforms, given their status as public companies, their operations adhere to relatively normative standards. The background score of the platform at time $t$ is set to $F(i,t)=b$. As for the venture capital platforms, there exist restrictions imposed by foreign capital, coupled with specific requirements regarding the operational standards of these platforms. This leads to a relatively high safety guarantee coefficient, and the background score of these platforms at time $t$ is set to $F(i,t)=c$. Private IDPs, on the other hand, feature a relatively low investment threshold and poor operational compliance, resulting in higher associated risks. The background score of the platform at time $t$ is set to $F(i,t)=d$. $a,b,c,d$ have constant values, and $a>b>c>d$.

(3) The score of average rate of return

We categorize the average income range of IFPs in the HOL into the following intervals: below 8%, 8%–10%, 10%–12%, 12%–16%, and above 16%. According to assumption 1 in Section 2, the IFP considered in this paper operates as a non-loss platform, implying that the average rate of return should be greater than 0. Assumption 2 also indicates that the upper limit of the average rate of return complies with national laws, ensuring it does not exceed 36%. Consequently, the average rate $N(i,t)$ of return of an IFP is expressed as follows:

$$N(i,t)=\{\begin{array}{cc}{x}_1(t)& 0\le M(i,t)<8\%\\ x_2(t)& 8\%\le M(i,t)<10\%\\ x_3(t)& 10\%\le M(i,t)<12\%\\ x_4(t)& 12\%\le M(i,t)<16\%\\ x_5(t)& 16\%\le M(i,t)<36\%\end{array}$$

(3)

where x₁(t), x₂(t), x₃(t), x₄(t), and x₅(t) are the average rates of return of all IFPs in the range 0–8%, 8%–10%, 10%–12%, 12%–16%, and 16%–36%. In the specific calculation, the average rates of return of all the platforms on the HOL in the range of 0–8%, 8%–10%, 10%–12%, 12%–16%, and 16%–36% are taken as the values of x₁(t), x₂(t), x₃(t), x₄(t), x₅(t), respectively.

(4) The score of average loan term

Since interest is typically calculated when determining the annual interest rate, the average loan term score is measured by the ratio of the platform’s average loan term at a given time $t$ to 12 (as there are 12 months in a year). Therefore, the score $Q(i,t)$ for the average loan term of IFP is as follows:

$$Q(i,t)={\displaystyle \frac{L(i,t)}{12}}$$

(4)

(5) The capital flow rate

As the fund flow situation of the platform is determined by the number of bids and borrowing activities, we gauge the capital liquidity of the platform using the ratio of the loan balance of the IFP to the sum of the loan balance and monthly transaction volume at a given time $t$. Then the capital flow rate $H(i,t)$ of the IFP is calculated using the following equation:

$$H(i,t)={\displaystyle \frac{Z(i,t)}{B(i,t)+Z(i,t)}}$$

(5)

(6) The loan dispersion score

We use the ratio of the average value of loan amount per capita to the loan amount per capita to measure the loan dispersion scores. Then the loan dispersion $D(i,t)$ of the IFP is as follows:

$$D(i,t)={\displaystyle \frac{e_t}{E(i,t)}}$$

(6)

where $e_t$ denotes the average loan amount per capita of all IFPs at time t. In the specific calculation, the average loan amount per capita of all platforms listed by the HOL is taken as the value of $e_t$.

(7) The capital adequacy ratio

Here, we calculate the capital adequacy ratio of the IFP by taking the ratio between the sum of the registered capital and loan balance and the sum of the registered capital, loan balance, and the amount to be collected. Therefore, the capital adequacy ratio $G(i,t)$ of the IFP is presented as follows:

$$G(i,t)={\displaystyle \frac{C(i,t)+Z(i,t)}{C(i,t)+Z(i,t)+A(i,t)}}$$

(7)

(8) The sustainable growing rate

The sustainable growth rate of the IFP is measured by taking the ratio of the sum of changes in the monthly transaction volume to the sum of the monthly transaction volume from the inception of the IFP. Accordingly, the sustainable growing rate $S(i,t)$ of the IFP can be expressed as follows:

$$S(i,t)={\displaystyle \frac{{\displaystyle \sum _{k=1}^{T(i,t)}\left[B(i,k)-B(i,k-1)\right]}}{{\displaystyle \sum _{k=1}^{T(i,t)}B(i,k)}}}$$

(8)

where ${\displaystyle \sum _{k=1}^{T(i,t)}\left[B(i,k)-B(i,k-1)\right]}$ represents the sum of the changes in monthly transaction volume since the IFP i started its operation. ${\displaystyle \sum _{k=1}^{T(i,t)}B(i,k)}$ represents the sum of the monthly transaction volume of the IFP i.

(9) The information transparency score

Rupert Taylor pointed out that in the era of information disclosure transparency 2.0, information disclosure should align with the four Cs: comparable, consistent, consumable, and credible. Thus, the 4C theory is adopted to classify the information disclosure quality of IFP into four levels: 4C, 3C, 2C, and C. Subsequently, the information transparency score $Y(i,t)$ of IFPs is shown as follows:

$$Y(i,t)=\{\begin{array}{c}\begin{array}{ccc}k& & 4C\end{array}\\ \begin{array}{ccc}l& & 3C\end{array}\\ \begin{array}{ccc}m& & 2C\end{array}\\ \begin{array}{ccc}n& & C\end{array}\end{array}$$

(9)

The 4C level represents the highest tier, and the score of information transparency of the IFP evaluated at time $t$ is set to $Y(i,t)=k$. Following this, the 3C level is $Y(i,t)=l$. Thirdly, the 2C level is $Y(i,t)=m$. Finally, the C level is $Y(i,t)=n$, where $k>l>m>n$.

4.3. Construction of Reputation Evaluation Model

Considering the historical data and development process of IFPs, the calculation of the reputation evaluation value is divided into three distinct situations:

(1) If the IFP starts to operate at time $t$ = 0, then the monthly transaction volume, average loan term, loan dispersion, and average rate of return of the platform are all 0. In this initial stage, the IFP has not undergone any operational processes to accumulate a reputation. Following Convention 1, it is evident that the reputation evaluation value of the IFP at this time is 0, where the following is true:

$$R(i,0)=0$$

(10)

(2) If the IFP encounters issues, such as delayed payments, website closure, platform suspension, and road running at time $t$, then, assuming the delay period is $\tau$ ($\tau >0$), as per Convention 2, it is established that the reputation evaluation value of the IFP at this time is 0, where the following is true:

$$R(i,t+\tau )=0$$

(11)

(3) If the IFP is operating normally at time $t$, the reputation evaluation value of the IFP is determined by nine indicators: the monthly transaction volume, sustainable growing rate, capital adequacy ratio, platform background, loan dispersion, capital flow rate, information transparency, average rate of return, and average loan term. As analyzed in Section 3, a strong correlation exists between the reputation evaluation value of IFPs and the above nine indicators. The alteration of any indicator significantly impacts the evaluation results. In this type of evaluation, the score weight becomes meaningless and does not affect the ranking of the evaluation results. Moreover, due to the close correlation among the evaluation indicators, they must develop in coordination, exemplifying the characteristic of a typical geometric average.

Based on the correlation characteristics of the reputation evaluation indicators of IFPs, we construct a reputation evaluation model using the geometric mean value method. Due to the varying emphasis on the platform’s operating strength and sustainable development ability in different periods, we assign weights to the evaluation indicators representing these aspects. This leads to the formulation of the reputation evaluation model of the IFP as shown in Equation (12), where $\alpha$ represents the weight quantifying the evaluation indicators of the IFP’s operating strength, and $\beta$ represents the weight quantifying the evaluation indicator of the IFP’s sustainable development ability, where $0\le \alpha ,\beta \le 1$.

Accordingly, the reputation evaluation model of the IFP can be calculated as follows:

$$R(i,t)=\{\begin{array}{cc}0& S(i.t)\le 0\\ \sqrt[(6\alpha +3\beta )]{{\left[W(i,t)D(i,t)H(i,t)F(i,t)N(i,t)Q(i,t)\right]}^{\alpha }\times {\left[S(i,t)Y(i,t)G(i,t)\right]}^{\beta }}& S(i,t)>0\end{array}$$

(12)

When $S(i,t)>0$, we substitute Equations (1) and (4)–(8) into Equation (12). Then the reputation evaluation model of the IFP can be further expressed as follows:

$$R(i,t)=\sqrt[(6\alpha +3\beta )]{{\left[{\displaystyle \frac{e_{t}B(i,t)Z(i,t)F(i,t)N(i,t)L(i,t)}{12{\mu }_{t}E(i,t)\left[B(i,t)+R(i,t)\right]}}\right]}^{\alpha }\times {\left[{\displaystyle \frac{B(i,T(i,t))-B(i,0)}{{\displaystyle \sum _{k=1}^{T(i,t)}B(i,k)}}}\times {\displaystyle \frac{\left[C(i,t)+Z(i,t)\right]Y(i,t)}{C(i,t)+Z(i,t)+A(i,t)}}\right]}^{\beta }}$$

(13)

In Equation (13), the background score of the platform $F(i,t)$ is assigned in sections according to the type of social capital force of the IFP. In the reputation evaluation model calculation process, values are directly assigned according to the subsection rule of Equation (2). Therefore, Equation (2) is not substituted into Equation (13). Similarly, the average rate of return score is assigned in segments according to the average return interval, with the subsection rule outlined in Equation (3). The information transparency score $Y(i,t)$ is assigned according to the quality of information disclosure of IFPs, and the subsection rule is shown in Equation (9). Consequently, Equations (3) and (9) are not substituted in the equation of the reputation evaluation model (13). In the following model discussion, $F(i,t)$, $Y(i,t)$, and $N(i,t)$ are discussed in more detail.

Following the discussion and analysis of the weights $\alpha$ and weights $\beta$ in Equation (13), which represents the reputation evaluation model of an IFP, three application cases of the model are summarized as follows:

Case 1.

When  $\alpha >0,\beta =0$, the reputation evaluation value of the IFP solely depends on the reputation evaluation indicators that represent the current operating strength of the platform. In this scenario, the reputation evaluation value of IFPs is only related to six reputation evaluation indicators, namely, the monthly transaction volume, background of the platform, average rate of return, average loan term, capital flow rate, and loan dispersion.

When $S(i,t)>0$, $\alpha >0,\beta =0$, substituting $\beta =0$ into Equation (13), the reputation evaluation model of the IFP in case 1 can be presented as follows:

$$R(i,t)=\sqrt[6]{{\displaystyle \frac{e_{t}B(i,t)Z(i,t)F(i,t)N(i,t)L(i,t)}{12{\mu }_{t}E(i,t)\left[B(i,t)+Z(i,t)\right]}}}$$

(14)

From the properties of Equation (14), proposition 1 can be obtained.

Proposition 1.

When  $\alpha >0,\beta =0$, the sustainability of the IFP is positively correlated with the loan balance, monthly transaction volume, background of the platform, average rate of return, and average loan term and negatively correlated with the loan amount per capita.

Proof. 

Due to $B(i,t)$, $E(i,t)$, $Z(i,t)$, $F(i,t)$, $N(i,t)$, $L(i,t)$, ${\mu }_t$, and $e_t$ are all greater than 0,

thus $\sqrt[6]{{\displaystyle \frac{e_{t}B(i,t)Z(i,t)F(i,t)N(i,t)L(i,t)}{12{\mu }_{t}E(i,t)\left[B(i,t)+Z(i,t)\right]}}}>0$.

Since $t>0$, $\alpha >0$, ${\displaystyle \frac{\partial (R(i,t))}{\partial (F(i,t))}}={\displaystyle \frac{1}{6}}F{(i,t)}^{-{\displaystyle \frac{5}{6}}}\times \sqrt[6]{{\displaystyle \frac{e_{t}B(i,t)Z(i,t)F(i,t)N(i,t)L(i,t)}{12{\mu }_{t}E(i,t)\left[B(i,t)+Z(i,t)\right]}}}>0$.

Similarly, ${\displaystyle \frac{\partial (R(i,t))}{\partial (Z(i,t))}}>0$, ${\displaystyle \frac{\partial (R(i,t))}{\partial (B(i,t))}}>0$, ${\displaystyle \frac{\partial (R(i,t))}{\partial (N(i,t))}}>0$, ${\displaystyle \frac{\partial (R(i,t))}{\partial (L(i,t))}}>0$, ${\displaystyle \frac{\partial (R(i,t))}{\partial (E(i,t))}}<0$.

Therefore, at this point, the sustainability of the IFP $R(i,t)$ is positively correlated with $Z(i,t)$, $B(i,t)$, $F(i,t)$, $N(i,t)$, and $L(i,t)$ and negatively correlated with $E(i,t)$. Proposition 1 is thus proven. □

Case 2.

When  $\alpha =0,\beta >0$, the reputation evaluation value of the IFP only depends on the sustainable development ability of the platform. The reputation evaluation value of IFPs is only related to the sustainable growing rate, capital adequacy ratio, and information transparency.

When $S(i,t)>0$, $\alpha =0,\beta >0$, substituting $\alpha =0$ into Equation (13), the reputation evaluation model of the IFP in case 2 can be presented as follows:

$$R(i,t)=\sqrt[3]{{\displaystyle \frac{B(i,T(i,t))-B(i,0)}{{\displaystyle \sum _{k=1}^{T(i,t)}B(i,k)}}}\times {\displaystyle \frac{\left[C(i,t)+Z(i,t)\right]Y(i,t)}{C(i,t)+Z(i,t)+A(i,t)}}}$$

(15)

From the properties of Equation (15), proposition 2 can be obtained.

Proposition 2.

When  $\alpha =0,\beta >0$, the sustainability of the IFP is positively correlated with the registered capital, loan balance, and information transparency.

Proof. 

Due to the fact that $S(i,t)$, $Y(i,t)$, and $G(i,t)$ are all greater than 0,

$$\sqrt[3]{{\displaystyle \frac{B(i,T(i,t))-B(i,0)}{{\displaystyle \sum _{k=1}^{T(i,t)}B(i,k)}}}\times {\displaystyle \frac{\left[C(i,t)+Z(i,t)\right]Y(i,t)}{C(i,t)+Z(i,t)+A(i,t)}}}>0$$

Since $t>0$, $\alpha >0$, $C(i,t)$, and $R(i,t)$ are all greater than 0,

$${\displaystyle \frac{\partial (R(i,t))}{\partial (Y(i,t))}}={\displaystyle \frac{1}{3}}Y{(i,t)}^{-{\displaystyle \frac{2}{3}}}\times \sqrt[3]{{\displaystyle \frac{B(i,T(i,t))-B(i,0)}{{\displaystyle \sum _{k=1}^{T(i,t)}B(i,k)}}}\times {\displaystyle \frac{\left[C(i,t)+Z(i,t)\right]Y(i,t)}{C(i,t)+Z(i,t)+A(i,t)}}}>0$$

Similarly, ${\displaystyle \frac{\partial (R(i,t))}{\partial (C(i,t))}}>0$, ${\displaystyle \frac{\partial (R(i,t))}{\partial (Z(i,t))}}>0$.

Accordingly, the sustainability of the IFP $R(i,t)$ is positively correlated with $Y(i,t)$, $C(i,t)$, and $Z(i,t)$. Proposition 2 is thus proven. □

Case 3.

When  $\alpha >0,\beta >0$, the reputation value of the IFP is determined by the platform’s current operating strength and sustainable development ability. The reputation evaluation value of IFPs is influenced by nine evaluation indicators, including the monthly transaction volume, background of the platform, average rate of return, average loan term, capital flow rate, loan dispersion, capital adequacy ratio, sustainable growing rate, and information transparency.

When $S(i,t)>0$, $\alpha >0,\beta >0$, the reputation evaluation model of the IFP in case 3 can be calculated using the following equation:

$$R(i,t)=\sqrt[(6\alpha +3\beta )]{{\left[{\displaystyle \frac{e_{t}B(i,t)Z(i,t)F(i,t)N(i,t)L(i,t)}{12{\mu }_{t}E(i,t)\left[B(i,t)+Z(i,t)\right]}}\right]}^{\alpha }\times {\left[{\displaystyle \frac{B(i,T(i,t))-B(i,0)}{{\displaystyle \sum _{k=1}^{T(i,t)}B(i,k)}}}\times {\displaystyle \frac{\left[C(i,t)+Z(i,t)\right]Y(i,t)}{C(i,t)+Z(i,t)+A(i,t)}}\right]}^{\beta }}$$

(16)

From the properties of Equation (16), proposition 3 can be obtained.

Proposition 3.

When  $\alpha >0,\beta >0$, the sustainability of the IFP is positively correlated with the loan balance, monthly transaction volume, background of the platform, average rate of return, average loan term, registered capital, and information transparency and negatively correlated with the loan amount per capita.

Proof. 

Since $B(i,t)$, $n(i,t)$, $R(i,t)$, $F(i,t)$, $N(i,t)$, $L(i,t)$, ${\mu }_t$, $S(i,t)$, $Y(i,t)$, and $G(i,t)$ are all greater than 0,

$$\sqrt[(6\alpha +3\beta )]{{\left[{\displaystyle \frac{e_{t}B(i,t)Z(i,t)F(i,t)N(i,t)L(i,t)}{12{\mu }_{t}E(i,t)\left[B(i,t)+Z(i,t)\right]}}\right]}^{\alpha }\times {\left[{\displaystyle \frac{B(i,T(i,t))-B(i,0)}{{\displaystyle \sum _{k=1}^{T(i,t)}B(i,k)}}}\times {\displaystyle \frac{\left[C(i,t)+Z(i,t)\right]Y(i,t)}{C(i,t)+Z(i,t)+A(i,t)}}\right]}^{\beta }}>0.$$

Since $t>0$, $\alpha >0$, and $\beta >0$,

$${\displaystyle \frac{\partial (R(i,t))}{\partial (N(i,t))}}={\displaystyle \frac{\alpha }{6\alpha +3\beta }}N{(i,t)}^{-{\displaystyle \frac{5\alpha +3\beta }{6\alpha +3\beta }}}\times \sqrt[(6\alpha +3\beta )]{\begin{array}{l}{\left[{\displaystyle \frac{e_{t}B(i,t)Z(i,t)F(i,t)N(i,t)L(i,t)}{12{\mu }_{t}E(i,t)\left[B(i,t)+Z(i,t)\right]}}\right]}^{\alpha }\times \\ {\left[{\displaystyle \frac{B(i,T(i,t))-B(i,0)}{{\displaystyle \sum _{k=1}^{T(i,t)}B(i,k)}}}\times {\displaystyle \frac{\left[C(i,t)+Z(i,t)\right]Y(i,t)}{C(i,t)+Z(i,t)+A(i,t)}}\right]}^{\beta }\end{array}}>0$$

Similarly, ${\displaystyle \frac{\partial (R(i,t))}{\partial (Z(i,t))}}>0$, ${\displaystyle \frac{\partial (R(i,t))}{\partial (B(i,t))}}>0$, ${\displaystyle \frac{\partial (R(i,t))}{\partial (F(i,t))}}>0$, ${\displaystyle \frac{\partial (R(i,t))}{\partial (L(i,t))}}>0$, ${\displaystyle \frac{\partial (R(i,t))}{\partial (C(i,t))}}>0$, ${\displaystyle \frac{\partial (R(i,t))}{\partial (Y(i,t))}}>0$, ${\displaystyle \frac{\partial (R(i,t))}{\partial (E(i,t))}}<0$.

Therefore, the sustainability of the IFP $R(i,t)$ is positively correlated with $Z(i,t)$, $B(i,t)$, $F(i,t)$, $N(i,t)$, $L(i,t)$, $Y(i,t)$, and $C(i,t)$ and negatively correlated with $E(i,t)$. Proposition 3 is thus proven. □

5. Empirical Analysis

5.1. Data Collection

This paper utilizes transaction data from P2P lending platforms as empirical evidence to validate the proposed reputation model. The octopus data collector is used to capture monthly transaction data from 18 P2P lending platforms spanning from January to June 2019 [9] sourced from the “Home of Online Loans” (HOL). The data include the net capital inflow, monthly transaction volume, per capita borrowing amount, monthly number of loans, operation time, loan balance, registered capital, per capita investment amount, platform background, platform yield rate, average borrowing term, and other information. Simultaneously, we checked the information releases of the 18 P2P lending platforms and rated the information transparency of the platforms according to the 4C theory. Then, following the reputation evaluation model, we selected the indicators’ data involved in the model, resulting in the compilation of the monthly indicator dataset for the 18 specific P2P lending platforms. Additionally, we calculated the average monthly transaction volume for all P2P lending platforms from January to June 2019, revealing values of 253.3 million CNY, 185.77 million CNY, 268.82 million CNY, 242.67 million CNY, 287.06 million CNY, and 281.08 million, CNY respectively. These figures were derived from the monthly transaction volume data of all P2P lending platforms published by the HOL.

Table 3. The operation data of 18 P2P lending platforms in June 2019.

Platfom	Monthly Transaction Volume	Number of Bids	Number of Loans	Per Capita Loan Amount	Loan Balance	Registered Capital	Platform Background	Information Transparency	Average Rate of Return	Average Loan Term
PF1	430,763.44	72,234	537,613	0.8	1,409,895	11,000	state-owned	4C	11.96%	15.85
PF2	354,094.21	96,282	273,399	1.3	1,358,387	20,000	listed	4C	7.21%	10.55
PF3	205,577.7	100,618	436,536	0.47	10,111,352	10,000	listed	4C	8.51%	10.40
PF4	231,358.6	129,768	201,300	1.15	1,406,186	5000	venture	3C	10.68%	27.69
PF5	188,982.2	29,335	530,320	0.36	364,953.2	10,000	private	3C	8.34%	7.72
PF6	143,406.1	29,440	21,316	6.73	1,296,786	10,000	listed	4C	8.59%	6.23
PF7	145,526.6	10,746	7969	18.26	155,927.5	655	venture	3C	8.89%	2.1
PF8	104,485.2	54,202	43,022	2.43	1,100,569	1065	listed	2C	8.68%	12.68
PF9	59,294.21	19,387	18,293	3.24	349,600.3	2942	private	3C	10.93%	16.69
PF10	48,248.77	15,665	49,890	0.97	776,060.3	30,000	listed	3C	9.13%	7.56
PF11	41,077.48	15,541	20,000	2.05	177,490.9	1212	venture	3C	9.56%	16.93
PF12	21,708.92	3789	2755	7.88	131,954.85	14,150	state-owned	3C	8.61%	6.62
PF13	14,998.38	1114	32,188	0.47	76,844.92	1000	state-owned	3C	7.08%	3.9
PF14	17,209.06	5374	551	31.23	113,913.17	11,835	state-owned	2C	9.03%	6.44
PF15	24,905.28	3976	2404	10.36	114,292.6	50,000	venture	C	8.95%	5.78
PF16	4576.81	821	249	18.38	53,459.55	5000	listed	2C	9.29%	7.14
PF17	25,444.25	7762	4858	5.24	159,344.81	30,000	listed	2C	8.10%	7.32
PF18	7300.36	2594	805	9.07	40,605.19	10,000	private	2C	8.67%	18.56

Explanatory note on units of measure: Monthly transaction volume (10,000 CNY), Registered capital (10,000 CNY), Average loan term (month).

shows the operational indicator data of the 18 P2P lending platforms in June 2019.

5.2. Reputation Evaluation Value Calculation and Sustainability Ranking

In this paper, the IFP reputation evaluation model primarily addresses three cases: (1) The reputation evaluation value exclusively accounts for the influence of indicators representing the current operational strength of the platform, with α = 1 and β = 0. (2) The reputation evaluation value exclusively considers the influence of the indicators representing the sustainable development ability of the platform, with α = 0 and β = 1. (3) The reputation evaluation value is considered to be influenced by the combination of the platform’s current operational strength and sustainable development ability, with α = 1 and β = 1.

The parameters were set as follows: a = 1.0, b = 0.75, c = 0.5, d = 0.25, k = 1.0, l = 0.75, m = 0.50, and n = 0.25. Here, x₁(t), x₂(t), x₃(t), x₄(t), and x₅(t) represent the average rate of return of all platforms on the HOL with the following range: 0–8%, 8%–10%, 10%–12%, 12%–16%, and 16%–36%. The data from the 18 P2P lending platforms in Table 3 were substituted into the quantitative formulas of the nine indicators to calculate the quantitative values of evaluation indicators, such as the monthly transaction volume, background of the platform, average rate of return, average loan term, capital flow rate, loan dispersion, capital adequacy ratio, sustainable growing rate, and information transparency of the P2P lending platforms. Then, the quantitative results of the nine major impact assessment indicators were set up for each of the following scenarios: (α = 1, β = 0), (α = 0, β = 1), and (α = 1, β = 1). This process resulted in the creation of the reputation evaluation value sequence table for the 18 P2P lending platforms in June 2019 under the three cases, as presented in

Table 4. The reputation evaluation value of 18 P2P lending platforms in three cases.

Platform	Monthly Transaction Volume	Sustainable Growing Rate	Capital Adequacy Ratio	Platform Background	Loan Dispersion	Capital Flow Rate	Information Transparent	Average Rate of Term	Average Loan Term	Reputation Evaluation Value 1	Reputation Evaluation Value 2	Reputation Evaluation Value 3
PF1	15.33	4.08%	51.17%	1.00	1.25	0.13	1.00	10.93%	1.32	0.982	0.275	0.715
PF2	12.60	5.65%	49.53%	0.75	0.77	0.35	1.00	7.21%	0.88	0.845	0.304	0.601
PF3	7.31	0.12%	49.35%	0.75	2.13	0.23	1.00	9.13%	1.28	0.978	0.085	0.450
PF4	8.23	7.03%	49.57%	0.5	0.87	0.64	0.75	10.93%	2.31	0.954	0.297	0.647
PF5	6.72	7.64%	46.37%	0.25	2.78	0.06	0.75	9.13%	0.64	0.732	0.298	0.543
PF6	5.10	5.65%	12.90%	0.75	0.15	1.38	1.00	9.13%	0.52	0.492	0.194	0.361
PF7	5.18	8.15%	50.18%	0.50	0.05	1.35	0.75	9.13%	0.18	0.325	0.313	0.318
PF8	3.72	4.73%	48.26%	0.75	0.41	1.26	0.50	9.13%	1.81	0.740	0.225	0.497
PF9	2.11	5.43%	73.25%	0.25	0.31	1.06	0.75	10.93%	2.22	0.578	0.310	0.450
PF10	1.72	4.30%	55.92%	0.50	1.03	0.31	0.75	9.13%	0.63	0.604	0.262	0.457
PF11	1.46	8.80%	52.56%	0.50	0.49	0.78	0.75	9.13%	1.41	0.584	0.326	0.481
PF12	0.77	3.81%	52.50%	1.00	0.13	1.38	0.75	9.13%	0.55	0.398	0.247	0.340
PF13	0.53	3.26%	50.76%	1.00	2.13	0.03	0.75	7.21%	0.33	0.530	0.232	0.402
PF14	0.61	3.85%	52.78%	1.00	0.03	9.75	0.5	9.13%	0.54	0.307	0.217	0.271
PF15	0.89	3.30%	58.65%	0.50	0.10	1.65	0.25	9.13%	0.48	0.338	0.169	0.270
PF16	0.16	2.65%	54.88%	0.75	0.05	3.30	0.5	9.13%	0.60	0.265	0.194	0.236
PF17	0.91	6.03%	57.09%	0.75	0.19	1.60	0.5	9.13%	0.61	0.422	0.258	0.358
PF18	0.26	4.15%	60.36%	0.25	0.11	3.22	0.5	9.13%	1.55	0.307	0.232	0.280

Explanatory note: When α = 1 and β = 0, the result calculated by the reputation model is denoted as reputation evaluation value 1; when α = 0 and β = 1, the result calculated by the reputation model is denoted as reputation evaluation value 2; and when α = 1 and β = 1, the result calculated by the reputation model is denoted as reputation evaluation value 3.

. In Table 4, reputation evaluation value 1 represents the result of the reputation calculation in case 1, reputation evaluation value 2 represents the result of the reputation calculation in case 2, and reputation evaluation value 3 represents the result of the reputation calculation in case 3.

According to the reputation evaluation value for the three cases, as calculated in Table 4, the sustainability of the 18 P2P lending platforms is ranked in descending order of reputation evaluation value. A comparison is made between the sustainability rankings in the three cases and the rankings of the 18 P2P lending platforms by the HOL in June 2019, as shown in

Table 5. Comparison table of reputation ranking of 18 P2P lending platforms.

Platform	HOL Ranking	Sustainability Ranking 1	Sustainability Ranking 2	Sustainability Ranking 3	Platform	HOL Ranking	Sustainability Ranking 1	Sustainability Ranking 2	Sustainability Ranking 3
PF1	3	1	7	1	PF10	12	7	8	7
PF2	2	4	4	3	PF11	16	8	1	6
PF3	1	2	18	9	PF12	11	13	10	13
PF4	6	3	6	2	PF13	13	10	12	10
PF5	7	6	5	4	PF14	9	17	14	16
PF6	4	11	15	11	PF15	17	14	17	17
PF7	8	15	2	14	PF16	15	18	16	18
PF8	5	5	13	5	PF17	10	12	9	12
PF9	14	9	3	8	PF18	18	16	11	15

. In this study, the ranking results are considered to be consistent if the difference is less than two (including two). Therefore, sustainability ranking 1 is consistent with the ranking of the HOL for eight platforms, namely PF1, PF2, PF3, PF5, PF8, PF12, PF17, and PF18. Sustainability ranking 2 is consistent with the reputation ranking of the HOL for eight platforms, namely PF2, PF4, PF5, PF12, PF13, PF15, PF16, and PF17. Sustainability ranking 3 is consistent with the ranking of the HOL for six platforms, namely PF1, PF2, PF8, PF12, PF15, and PF17. In comparison, sustainability rankings 1 and 2 are very similar in the HOL ranking, while sustainability ranking 3 exhibits significant differences from the HOL ranking.

A comparative analysis between Table 4 and Table 5 sheds light on the sustainability ranking of the 18 P2P lending platforms for the three cases. Notably, sustainability ranking 3 differs significantly from the HOL ranking for six platforms, specifically PF3, PF6, PF7, PF10, PF11, and PF14. The primary reason for this disparity lies in the fact that the HOL ranking solely considers the current economic strength of the platform, whereas sustainability ranking 3 takes into account not only the current operational strength but also the sustainable development ability of the platform. Specifically, the HOL ranking prioritizes platforms with commendable monthly transaction volumes and capital flow rates, while sustainability ranking 3 emphasizes platforms exhibiting high monthly transaction volumes, capital flow rates, sustainable growing rates, and capital adequacy ratios. Taking PF3 as an illustrative example, the HOL ranks PF3 in first place due to its high monthly transaction volume, good capital flow rate, and high overall economic strength. However, sustainability ranking 3 ranks PF3 in ninth place, influenced by its low capital adequacy ratio and the sustainable growing rate of PF3. They lowers the influence of the monthly transaction volume and capital flow rate on sustainability to some extent. As a result, PF3 ranked lower in sustainability ranking 3. Theoretically, the low capital adequacy ratio and sustainable growing rate indicate PF3’s subpar sustainable development ability and elevated investment risk. In reality, PF3 has announced its gradual withdrawal from the P2P lending business, which further validates the rationale behind sustainability ranking 3. It also shows that it is reasonable to consider the sustainable development ability of the platform in the reputation evaluation of IFPs.

5.3. Fluctuation Analysis of Reputation Evaluation Value

To delve deeper into the dynamics of reputation changes among P2P lending platforms and analyze variations in reputation evaluation values across different platforms, the reputation evaluation values of the 18 P2P lending platforms were computed for the period from January to June 2019 under three distinct conditions using the aforementioned reputation evaluation model. Under these three cases, the temporal evolution of reputation evaluation values for P2P lending platforms is outlined below.

(1) The reputation value of P2P lending platforms only considers the influence of the current operating strength of the platform.

The reputation evaluation values of the 18 P2P online lending platforms, as presented in

Table 6. The reputation evaluation value of 18 P2P lending platforms from January to June 2019 for case 1.

Platform	January	February	March	April	May	June	Reputation Volatility	Reputation Volatility Ranking
PF1	0.983	0.992	0.981	0.999	0.989	0.982	0.0067	16
PF2	0.877	0.849	0.859	0.859	0.841	0.845	0.0132	9
PF3	0.961	0.985	0.954	0.990	0.982	0.978	0.0143	8
PF4	0.968	0.977	0.977	0.985	0.959	0.954	0.0117	11
PF5	0.816	0.863	0.786	0.736	0.719	0.732	0.0565	2
PF6	0.595	0.622	0.589	0.553	0.482	0.492	0.0576	1
PF7	0.323	0.344	0.328	0.339	0.331	0.325	0.0082	14
PF8	0.778	0.799	0.754	0.744	0.748	0.740	0.0234	5
PF9	0.576	0.570	0.570	0.594	0.573	0.578	0.0091	13
PF10	0.541	0.597	0.623	0.629	0.605	0.604	0.0315	4
PF11	0.547	0.584	0.564	0.583	0.575	0.584	0.0149	7
PF12	0.397	0.401	0.404	0.395	0.396	0.398	0.0034	18
PF13	0.637	0.601	0.624	0.639	0.563	0.530	0.0439	3
PF14	0.317	0.313	0.311	0.325	0.315	0.307	0.0061	17
PF15	0.365	0.381	0.356	0.351	0.342	0.338	0.0158	6
PF16	0.291	0.269	0.283	0.290	0.272	0.265	0.0112	12
PF17	0.419	0.435	0.417	0.434	0.428	0.422	0.0076	15
PF18	0.320	0.341	0.319	0.319	0.304	0.307	0.0128	10

Explanatory note: α = 1 and β = 0.

, are visually depicted in Figure 5.

In Figure 5, when the reputation of P2P lending platforms only accounts for the influence of the platforms’ current operating strength, the majority of the platforms exhibit a stable reputation change curve. However, a few platforms, namely PF5, PF6, PF10, and PF13 demonstrate considerable reputation fluctuations, predominantly displaying a downward trend. To comprehend these variations, an analysis of influencing factors, such as the monthly transaction volume, capital flow rate, platform background, loan dispersion, average rate of return, and average loan term, for these four P2P platforms from January to June was conducted. The results reveal that the significant fluctuation in the capital flow rate is the root cause behind the pronounced reputation volatility of these four platforms. The reputation evaluation value of PF5, PF6, PF10, and PF13 experienced rapid increases initially, followed by a sharp decline, and eventually stabilized. This pattern corresponds to the drastic fluctuations in the capital flow rate of these four platforms from February to April. This shows that when the reputation of an IFP only considers the influence of economic strength on the platform, the sustainability fluctuation of the platform is profoundly influenced by the capital flow rate.

(2) The reputation value of P2P lending platforms only considers the influence of the sustainable development capability of the platforms.

The reputation evaluation values of the 18 P2P lending platforms, as presented in

Table 7. The reputation evaluation value of 18 P2P lending platforms from January to June 2019 in case 2.

Platform	January	February	March	April	May	June	Reputation Volatility	Reputation Volatility Ranking
PF1	0.343	0.274	0.290	0.295	0.277	0.275	0.0239	9
PF2	0.354	0.289	0.349	0.328	0.307	0.304	0.0238	10
PF3	0.090	0.103	0.092	0.093	0.088	0.085	0.0059	18
PF4	0.362	0.339	0.338	0.324	0.310	0.297	0.0212	14
PF5	0.367	0.340	0.359	0.295	0.282	0.298	0.0332	4
PF6	0.390	0.363	0.357	0.317	0.183	0.194	0.0821	1
PF7	0.368	0.295	0.351	0.345	0.324	0.313	0.0247	8
PF8	0.268	0.210	0.219	0.142	0.235	0.225	0.0380	2
PF9	0.351	0.290	0.347	0.337	0.312	0.310	0.0220	11
PF10	0.292	0.221	0.276	0.297	0.280	0.262	0.0251	7
PF11	0.390	0.346	0.366	0.334	0.342	0.326	0.0214	13
PF12	0.286	0.264	0.284	0.258	0.254	0.247	0.0146	16
PF13	0.311	0.265	0.325	0.329	0.271	0.232	0.0354	3
PF14	0.239	0.195	0.236	0.240	0.232	0.217	0.0162	15
PF15	0.206	0.192	0.188	0.177	0.178	0.169	0.0120	17
PF16	0.271	0.217	0.280	0.259	0.251	0.194	0.0303	5
PF17	0.272	0.192	0.261	0.275	0.278	0.258	0.0293	6
PF18	0.281	0.284	0.259	0.243	0.229	0.232	0.0216	12

Explanatory note: Set α = 0, β = 1.

, are visually depicted in Figure 6.

In Figure 6, when the reputation of P2P lending platforms exclusively reflects the impact of the platforms’ sustainable development capability, the majority of the platforms exhibit significant fluctuations in their reputation change curves. Notably, the platforms PF5, PF6, PF7, PF8, PF10, PF13, and PF16 experience pronounced reputation oscillations. This study delves into the changes in information transparency, capital adequacy ratio, and sustainable growing rate for these eight P2P lending platforms from January to June. The results reveal that the drastic fluctuations in the reputation evaluation values of these eight platforms are primarily attributed to significant changes in the sustainable growing rate. Specifically, the reputation values of PF5, PF6, PF8, PF13, PF16, and PF17 have decreased rapidly followed by a swift increase, mirroring the drastic fluctuations in their sustainable growing rates from January to March. Similarly, PF7 and PF10 witness a rapid decline in their reputation evaluation value from March to April, aligning with the substantial oscillation in their sustainable growing rate during that period. This observation underscores that when the IFP’s sustainable development capability exclusively influences its reputation, the changes in sustainability are primarily driven by the factor of the sustainable growing rate.

(3) The reputation value of a P2P lending platform is considered to be influenced by the platform’s current operating strength and sustainable development capability.

The reputation evaluation values of the 18 P2P lending platforms, as presented in

Table 8. The reputation evaluation value of 18 P2P lending platforms from January to June 2019 in case 3.

Platform	January	February	March	April	May	June	Reputation Volatility	Reputation Volatility Ranking
PF1	0.763	0.736	0.723	0.739	0.710	0.715	0.0175	10
PF2	0.648	0.593	0.636	0.623	0.601	0.601	0.0203	7
PF3	0.437	0.485	0.438	0.473	0.459	0.450	0.0176	9
PF4	0.718	0.709	0.686	0.688	0.658	0.647	0.0254	5
PF5	0.625	0.633	0.605	0.542	0.526	0.543	0.0430	2
PF6	0.517	0.520	0.499	0.459	0.349	0.361	0.0706	1
PF7	0.337	0.327	0.335	0.341	0.328	0.318	0.0078	16
PF8	0.545	0.512	0.499	0.428	0.508	0.497	0.0351	4
PF9	0.488	0.455	0.483	0.492	0.468	0.450	0.0162	11
PF10	0.440	0.429	0.475	0.490	0.468	0.457	0.0207	6
PF11	0.489	0.491	0.488	0.485	0.484	0.481	0.0033	18
PF12	0.356	0.349	0.359	0.343	0.342	0.340	0.0071	17
PF13	0.501	0.458	0.502	0.512	0.442	0.402	0.0393	3
PF14	0.289	0.268	0.284	0.294	0.284	0.271	0.0093	15
PF15	0.302	0.304	0.288	0.279	0.275	0.270	0.0128	14
PF16	0.284	0.250	0.282	0.279	0.265	0.236	0.0177	8
PF17	0.363	0.331	0.357	0.373	0.371	0.358	0.0136	13
PF18	0.306	0.320	0.297	0.291	0.277	0.280	0.0150	12

Explanatory note: α = 1 and β = 1.

, are visually depicted in Figure 7.

In Figure 7, when evaluating the reputation of P2P lending platforms while considering the joint influence of their current operating strength and sustainable development capability, each P2P lending platform exhibits a unique reputation change curve. The reputation fluctuations of the PF7, PF11, PF12, PF14, PF15, PF17, and PF18 platforms are relatively small. The reputation of the PF4, PF5, PF6, PF8, and PF13 platforms fluctuates greatly. Analyzing the changes in the evaluation indicators of 18 P2P lending platforms from January to June reveals that the platforms with a stable capital flow rate and sustainable growing rate have less fluctuation in their reputation evaluation value. Conversely, platforms with drastic fluctuations in their capital flow rate and sustainable growing rates exhibit more significant fluctuations.

Specifically, the reputation evaluation value of the PF4, PF5, PF6, and PF13 platforms began to rapidly decline from March, coinciding with sharp fluctuations in their sustainable growing rates from March to June. For the same reason, the PF8 platform experienced a decline in reputation evaluation value from March to May, followed by a subsequent rise. This pattern indicates that, when considering the influence of the platform’s current operating strength and sustainable development capability, sustainability fluctuations are primarily influenced by the capital flow rate and sustainable growing rate. Furthermore, the sustainable growth rate has a greater impact on a platform’s reputation volatility than the capital flow rate.

5.4. Analysis of Sustainable Operation of P2P Lending Platforms

The sustainability ranking of P2P lending platforms in June 2019 is depicted in Figure 8. The sustainability of P2P lending platforms of the top five in sustainability ranking 3 is analyzed. In descending order of reputation evaluation value, these platforms are as follows: PF1, PF4, PF2, PF5, and PF8. The detailed sustainability analysis of these five P2P lending platforms is as follows:

In Figure 8, reputation evaluation values 1 and 3 of PF1 rank first, while the reputation evaluation value 2 of PF1 ranks seventh. This suggests that the current operating strength of PF1 is stronger, but the sustainability of PF1 is normal. Reputation evaluation value 1 of PF4 ranks third, reputation evaluation value of PF4 ranks sixth, and reputation evaluation value 3 of PF4 ranks second. This shows that the operating strength of PF4 is strong, but the sustainability of PF4 is normal. Reputation evaluation values 1 and 2 of PF2 rank fourth, and reputation evaluation value 3 of PF2 ranks third. This shows that the operating strength and sustainability of PF2 are both strong. Reputation evaluation values 1, 2, and 3 of PF5 respectively rank sixth, fifth, and fourth. This shows that the operating strength of PF5 is normal, but the sustainability of PF5 is strong. Reputation evaluation values 1 and 3 of PF8 are ranked fifth, and reputation evaluation value 2 of PF8 ranks thirteenth. This indicates that PF8 has a strong operating strength, but its sustainability is normal.

According to the above analysis, the two IFPs of PF2 and PF5 have more sustainable operations. When investors are risk-averse, they pay more attention to the capability of platforms to sustain operations. At this point, platforms such as PF2 and PF5 are more suitable for investors to invest in. Therefore, the sustainability of IFPs is analyzed through the reputation evaluation value ranking, and the analytical results can help investors make better decisions.

6. Conclusions and Future Research

To effectively evaluate the sustainability of IFPs, this paper comprehensively considers the impact of nine indicators, including the monthly transaction volume, background of the platform, average rate of return, average loan term, capital flow rate, loan dispersion, capital adequacy ratio, sustainable growing rate, and information transparency, on the reputation of IFPs. A reputation evaluation model for the sustainable operation of IFPs has been established, presenting relevant mathematical propositions and proofs. The validity and practicability of the model were empirically analyzed using loan transaction data from 18 P2P lending platforms on the Home of Online Loans (HOL). Based on the empirical results, this study draws the following five conclusions:

(1) The sustainability of IFPs can be measured based on the reputation evaluation value derived from the model. Thus, by analyzing key factors influencing IFPs’ reputation, the introduced model ranks IFPs based on their sustainability and identifies those with lower levels of sustainability.

(2) When evaluating IFPs’ sustainability based solely on operational strength, there is a positive correlation with the loan balance, monthly transaction volume, platform background, average rate of return, and average loan term. Conversely, there is a negative correlation with the loan amount per capita. Additionally, the capital flow rate predominantly affects sustainability fluctuations. When considering only the platform’s sustainable development capabilities, sustainability is positively associated with the registered capital, loan balance, and information transparency, with the sustainable growth rate significantly influencing fluctuations. A comprehensive evaluation, which considers both operational strength and sustainable development, reveals that IFPs’ sustainability is positively correlated with the loan balance, monthly transaction volume, platform background, average rate of return, average loan term, registered capital, and information transparency and negatively correlated with the loan amount per capita. Sustainability fluctuations are primarily influenced by both the capital flow rate and sustainable growth rate, with the latter being more dominant.

(3) Integrating indicators of sustainable development into the reputation evaluation model allows for a more comprehensive assessment of both the current operational strength and the sustainable development ability of IFPs. This approach helps in ranking IFPs based on their sustainability, providing investors with valuable insights to make informed decisions and mitigate credit risk in Internet financial investments. Additionally, identifying key indicators affecting IFPs’ sustainability offers a theoretical and empirical foundation for improving platform sustainability and competitiveness.

This study provides an effective method for investors and regulators to identify IFPs with sustainable operating practices, offering valuable insights that can promote the sustainable development of the Internet finance industry. Moreover, the evaluation model developed in this study not only highlights key variables for platform operators to focus on for operational improvements but also underscores the importance of sustaining transaction volume and enhancing information transparency. Thus, for platform operators, prioritizing these factors is crucial for bolstering IFP sustainability. Although designed specifically for Internet finance, the evaluation indicators and models presented in this study are also relevant for regulators overseeing digital finance services. These findings offer important considerations and implications for the digital finance industry, encouraging reflections on its future directions and strategies for improvement.

This study does not come without limitations. Firstly, while this study utilizes nine reputation evaluation indicators derived from the existing literature and the index systems of third-party credit rating agencies, there is a need to explore more representative indicators. Future research could benefit from mining big social network data to identify additional, relevant indicators that could enhance the comprehensiveness of the reputation evaluation model. Secondly, the current study does not quantitatively assess how reputation evaluation results influence investment decisions. Future research should address this gap by collecting detailed investment data from IFP investors and analyzing their risk appetites. Such an analysis would supplement the current studies by quantifying the impact of reputation evaluations on investment decisions. Hence, to further refine the reputation evaluation model, future research could integrate machine learning and artificial intelligence technologies. These advanced methods could optimize the model, making it more robust and adaptive to evolving market conditions.