Loan Evaluation Research Articles

A Study of Debt Problems in the Chinese Real Estate Industry

This paper will focus on real estate industry in China, uses Evergrande group as a case to compare with another Chinese real estate company Vanke and a real estate company of United State which is AvalonBay Communities, Inc. This paper will introduce the background of real estate industry in China, and base on the case which is bankruptcy of Evergrande Group to throughout the causes find out the issues of real estate in China, in causes is discuss of three factors: Industry characteristics, financing channels, loan evaluation, analysing what make Evergrande can owe a massive $2.4 trillion in debt, and analysing the United States real estate industry not usually having such high leverage. Essay will tell the strategy of real estate companies in China usually do. The strategy led to Evergrande group bankruptcy and compare with Vanke uses what strategy avoid debt crisis. Finally, conclude the chaos of real estate industry in China reflected by Evergrande group, recognise the issues of Evergrande group and prevent the significant consequence occur again.

Factors Influencing the Use of Property Titles as Security for Lending in Rural Kenya: A Case Study of Kivani Location, Machakos County

The development of the world's economies, especially in emerging countries, depends significantly on land. Despite being a crucial tool for obtaining loans, the use of land as collateral is relatively low in rural Kenya. This has largely been attributed to lack of formal ownership documents attributable to poor titling, which has subsequently hindered investments and development in these areas. The research gap, therefore, lies in the need to understand and address the reasons behind the low uptake of land as collateral in rural Kenya, despite its potential as a valuable asset and its significant role in rural livelihoods and development. This study set out to establish factors influencing the use of property titles as loan security in a rural setting using Machakos County as a case study. Employing a descriptive research design, the study targeted 724 households in Kivani and 30 financial institution officers. Data was collected using questionnaires and topical discussions, achieving response rates of 96% and 100%, respectively. Findings reveal that 67 percent of respondents have never used land titles as collateral. Several factors were identified: social relations, lending conditions, and loan pricing. Social relations factors included communal land ownership, land succession problems, cultural norms, and sentimental attachment to land. Lending conditions encompassed land valuation processes, borrower creditworthiness requirements, collateral quality, and the need for family consent. Loan pricing factors involve the costs and fees associated with credit financing. The study recommends increasing awareness of individual land ownership, exploring alternative collateral options, establishing insurance markets for collateral assets, and improving loan evaluation processes to better support rural residents. These measures aim to enhance access to financing, thereby promoting rural development

The Machine Learning Methods For Micro-Credit Scoring: The Case Of Micro-Financing In Mongolia

As a result of growing digital technologies in the financial sector, the traditional slow lending process is being replaced by fast and easy digital lending systems that can make decisions in real time. Both lenders and borrowers have experienced the benefits of digital lending, the activities of microfinance institutions have expanded rapidly and the volume of digital microloans has increased significantly worldwide, including Mongolia. At the same time with the growing volume of digital microloans in Mongolia, the rationality of credit risk management has been becoming more critical. Credit quality is the most important factor in optimal credit risk management. It depends on determining the customer's creditworthiness and making accurate credit decisions. This research focuses on a credit scoring system to improve the digital loan evaluation system of the Mongolian microfinance institute. This study aims to contribute to the development of possible credit scoring systems for Mongolian microfinance institutions by comparing several machine-learning approaches based on loan datasets of a non-banking microfinance institute in Mongolia. The result shows the ensemble methods Random Forest and XGBoost Tree's accuracies are higher than other machine learning models for the microloan borrowers' repayment status prediction.

Reintermediation in FinTech: Evidence from Online Lending

AbstractWe document the unique structure of the peer-to-peer lending market. Originally designed as decentralized, the market has become highly, but not fully, reintermediated. The platforms’ software now performs essentially all tasks related to loan evaluation, whereas most lenders are passive and automatically fund most applications on offer. Yet unlike banks, and in contrast to theories predicting full reintermediation, the platforms provide detailed loan information, and some active loan pickers coexist with passive investors. We argue that while intermediation attracts unsophisticated passive investors, transparency in the presence of active investors resolves the lending platform’s moral hazard problem inherent in intermediated markets.

An Optimized BP Neural Network Model and Its Application in the Credit Evaluation of Venture Loans.

With the rapid development of entrepreneurship loans in China, the construction of a credit evaluation system of risk loans has become an important financial safeguard measure. This paper mainly studies the following three aspects. Firstly, in view of the subjective factors in the approval process of venture loans, based on the credit evaluation system of commercial banks and the data characteristics of venture loans, a credit evaluation system based on venture loans is constructed. Secondly, the randomized uniform design method is used to improve the population initialization method to realize the uniform distribution of the individual population. Finally, aiming at the problem of low efficiency of venture loan audit, this paper proposes an optimized BP neural network to evaluate the venture loan. Especially, through data processing, a credit index system is constructed, and then the optimized BP neural network model is determined in parameters. The model contains 15 input nodes, 1 hidden layer, and 2 output layers. Finally, the simulation shows that the optimized BP neural network model has obvious advantages in the loan evaluation. This paper includes the development status of credit evaluation of venture loans is empirically studied by using an optimized BP neural network model of nonexpected output.

COSLE: Cost sensitive loan evaluation for P2P lending

The loan evaluation is a fundamental task in peer-to-peer (P2P) lending. Effective loan evaluation can help lenders make informed investment decisions. Existing methods do not consider the return of loans in the core learning stage and thus fail to explore the relationship between the return of loans and their final loan payoff outcomes. In this study, we propose a systematic loan evaluation framework called COst Sensitive Loan Evaluation (COSLE). Specifically, we first develop an instance-aware misclassification cost (IMCO) matrix, which specifies personalized cost for each loan. Then, we present a differential labelling algorithm called DILA cost for assigning node labels and assessing the corresponding cost. By integrating these enhancements into the tree-induction process, we construct a node splitting measurement called COG index. It exploits the relationship between the return information and the final payoff outcome. Additionally, we design the LER evaluation metric to measure the ability of a loan evaluation model to increase the lender’s return. Finally, the COSLE is used to improve popular tree models. Extensive experiments based on the Lending Club dataset show that our COSLE can effectively increase the lender’s return.

Risk assessment for contagion path in complex loan network

Small and medium-sized enterprises (SMEs) loans play an essential role in many aspects: including technological innovation, economic development, employment, and peoples livelihood, etc. In order to meet the loan evaluation criteria of commercial banks, many SMEs choose to guarantee each other to obtain loans, thus forming a complex guarantee network. If the borrower defaults on the loan, the risk will be diffused to its guarantors along with the contagion path, which may lead to systemic risk across the loan networks. This has brought severe challenges to the nations financial security and regulation. Thus, accurately rating the contagion path is an urgent task for systematic risk management in the loan network. Therefore, we present a deep learning-based approach to the risk rating of contagion paths in the bank industry. We leverage the graph neural network and attention mechanism on graph-structured loan behavior to learn high-order representations, which do not require handcraft feature engineering. We demonstrate that our approach outperforms the existing baselines with 2%$\\sim$15% improvements in risk rating and 3.5% in the newly constructed path rating problem. The result demonstrates the effectiveness of our proposed approach, which provides an effective method and theory basis for regulatory commissions and financial institutions to monitor systematic risks in networked-guarantee loans.

A predictive indicator using lender composition for loan evaluation in P2P lending

Most loan evaluation methods in peer-to-peer (P2P) lending mainly exploit the borrowers’ credit information. However, the present study presents the maturity-based lender composition score, which exploits the investment capability of a group of lenders who fund the same loan, to enhance the P2P loan evaluation. More specifically, we extract lenders’ profiles in terms of performance, risk, and experience by quantifying their investment history and develop our loan evaluation indicator by aggregating the profiles of lenders in the composition. To measure the ability of a lender for continuous improvement in P2P investment, we introduce lender maturity to capture this evolvement and incorporate it into the aggregation process. Our empirical study demonstrates that the maturity-based lender composition score can serve as an effective indicator for identifying loan quality and be included in other commonly used loan evaluation models for accuracy improvement.

Price Attractiveness and Price Complexity: Why People Prefer Level-Payment Loans.

The improving sequence effect suggests that in choices between a rising earning and any other sequences, participants prefer the rising earning. Recent studies show that the improving sequence effect also exists in a loan context. As consumers have a strong preference for falling loan profiles, banks may consider to offer loans in which the loan repayments concentrate at the beginning of the loan term. In this paper, we examined the improving sequence effect in context of a car loan with three repayment plans expressed in temporally reframed prices (TRP). By regressing the evaluation of loan profiles on the perceived price attractiveness, price complexity, TRP and the interaction terms, we find that (1) the perceived price attractiveness and price complexity significantly predict the loan evaluation, and they also explain a significant proportion of variance in loan evaluation; (2) the TRP effect interacts with the improving sequence effect. Specifically, with the introduction of TRP, respondents prefer constant profiles over falling profiles. TRP may explain why level-payment loans are still popular in real world, though the improving sequence effect suggests otherwise.

Two-stage rule extraction method based on tree ensemble model for interpretable loan evaluation

The tree ensemble model has been widely employed as a loan evaluation method in credit risk assessment due to its high accuracy and robustness. However, the tree ensemble model is complex and incomprehensible, which restricts its adoption for decision-making in loan evaluation. In this paper, we propose a novel rule extraction method for improving the ensemble model by balancing predictive performance and interpretability in two stages: a local rule extraction method followed by a global rule extraction method. The local method simplifies each rule by removing its redundant constraints, while the global method optimizes the complete rule set based on the multiobjective optimization method. An interpretable rule-based model is extracted from the tree ensemble model via the proposed method. Comparing the performance to six other methods on three loan evaluation datasets, the proposed method shows superior interpretability and realizes similar predictive performance to the tree ensemble model. For practical loan evaluation, the proposed method provides decision-makers with an interpretable rule-based model, which could replace the opaque tree ensemble model in high-stakes decision-making. In addition, the proposed method could facilitate decision-makers in explaining the tree ensemble model by analyzing the important and valuable rules that are extracted from the original opaque model.

Abstraction Hierarchy Based Explainable Artificial Intelligence

This work explores the application of Cognitive Work Analysis (CWA) in the context of Explainable Artificial Intelligence (XAI). We built an AI system using a loan evaluation data set and applied an XAI technique to obtain data-driven explanations for predictions. Using an Abstraction Hierarchy (AH), we generated domain knowledge-based explanations to accompany data-driven explanations. An online experiment was conducted to test the usefulness of AH-based explanations. Participants read financial profiles of loan applicants, the AI system’s loan approval/rejection decisions, and explanations that justify the decisions. Presence or absence of AH-based explanations was manipulated, and participants’ perceptions of the explanation quality was measured. The results showed that providing AH-based explanations helped participants learn about the loan evaluation process and improved the perceived quality of explanations. We conclude that a CWA approach can increase understandability when explaining the decisions made by AI systems.

Pricing of Margin Call Stock Loan Based on the FMLS

In common stock loan, lenders face the risk that their loans will not be repaid if the stock price falls below loan, which limits the issuance and circulation of stock loans. The empirical test suggests that the log-return series of stock price in the US market reject the normal distribution and admit instead a subclass of the asymmetric distribution. In this paper, we investigate the model of the margin call stock loan problem under the assumption that the return of stock follows the finite moment log-stable process (FMLS). In this case, the pricing model of the margin call stock loan can be described by a space-fractional partial differential equation with a time-varying free boundary condition. We transform the free boundary problem to a linear complementarity problem, and the fully-implicit finite difference method that we used is unconditionally stable in both the integer and fractional order. The numerical experiments are carried out to demonstrate differences of the margin call stock loan model under the FMLS and the standard normal distribution. Last, we analyze the impact of key parameters in our model on the margin call stock loan evaluation and give some reasonable explanation.

Loan approval evaluation framework of public-private partnership project of battery storage power station under interval-valued intuitionistic fuzzy environment

The main objective of this paper is to propose a loan approval evaluation framework of public-private partnership project (PPP) of battery storage power station (BSPS) for the commercial bank to evaluate and select the optimal project to provide the loan and then promote the development of BSPS-PPP. In order to achieve this goal, this study proposed a loan approval evaluation index system which consist of 5 attributes, 14 criteria and 45 sub-criteria. The novelty of this loan evaluation index system is that it reflects the basic requirement of loan approval of the commercial bank based on the characteristics of BSPS and PPP project, the commercial banks can better prepare for possible risks of BSPS-PPP. The proposed decision model was established based on the interval-valued intuitionistic fuzzy value (IVIFV) and interval number. Its novelty is that ①The model not only satisfies the experts’ expression habit of uncertainty, but also can be handled the uncertainty in the performance evaluation value and the criteria’s weights properly; ②the proposed weight determining method based on the deviation maximization principle can not only maintain the will of decision makers as much as possible, but also decrease the uncertainty of weights. Finally, the effectiveness of loan approval evaluation framework was verified based on a case study of China. According to the application effect of the evaluation index system and the proposed model, 3 managerial implication are put forward.

Loan officer rotation and credit access: evidence from Madagascar

PurposeWith exclusive data from a commercial microfinance institution (MFI) in Madagascar, the purpose of this paper is to investigate if loan officer rotation (change of loan officer) has an effect on credit access (loan approval) in rural and in urban areas. The authors further analyze how the frequency of loan officer rotation affects credit access in rural and in urban areas.Design/methodology/approachThe authors apply propensity score matching to compare credit access between loan applicants who experienced loan officer rotation and loan applicants who experienced no loan officer rotation in rural and in urban areas.FindingsResults show that loan officer rotation has a positive and statistically significant effect on credit access. The authors observe further that loan officer rotation has a different effect on credit access in rural and in urban areas. Whilst rural loan applicants who experienced loan officer rotation are more likely to have credit access, urban loan applicants show no statistically significant effect of loan officer rotation on credit access. For the frequency effect on credit access, the authors observe that one loan officer rotation has a positive and statistically significant effect on credit access whereas results are mixed for two loan officer rotations.Research limitations/implicationsEven though the authors can show that loan officer rotation can improve credit access to loan applicants, especially in rural areas, the conditions in Madagascar are unique. Therefore, results need to be verified in other countries and institutional contexts.Practical implicationsFrom the perspective of MFI, the authors recommend that the management of MFI needs to provide better tools to loan officers to improve on the evaluation of agricultural loan products or standardize the assessment of agricultural loan products to improve on lending decisions. Further, if applicable, the authors recommend that MFI should consider using credit worthiness assessment procedures which rely less on loan officer’s judgment for loan evaluation, such as automated systems. From the perspective of loan applicants, the authors recommend that loan applicants should request for a change of loan officer if they experience successive loan applications rejection.Originality/valueTo the authors’ knowledge, this paper is the first to provide empirical evidence on the effect and frequency of loan officer rotation on credit access in Sub-Sahara Africa, and Madagascar, in particular.

A CBR Decision Support System for Loan Evaluation

A CBR Decision Support System for Loan Evaluation

Loan evaluation in P2P lending based on Random Forest optimized by genetic algorithm with profit score

Loan evaluation is an effective method for credit risk assessment in peer-to-peer (P2P) lending and significantly affects lender investment decisions as well as his/her profits. Besides traditional methods of loan evaluation, machine learning has gained increased attention and has achieved better performance for P2P lending, especially regarding the Random Forest approach. However, the loan evaluation model based on Random Forest aims to improve the overall accuracy, which cannot guarantee that the lender profit is maximized when the overall accuracy is maximized because the profits of each loan are different. To further improve the loan evaluation effect and lender profits, Random Forest optimized using a genetic algorithm with profit score (RFoGAPS) is proposed. First, considering the actual and potential returns and losses, a new profit score is proposed and taken as the optimization objective. Second, the genetic algorithm is used to optimize the combination of decision trees in Random Forest. Then, the dataset of Lending Club is used to evaluate the proposed method. Experimental results show that the RFoGAPS can obtain higher profits for lenders compared with actual profit and traditional methods. Some suggestions are proposed based on experimental results to facilitate the healthy development of P2P lending.

Reintermediation in Fintech: Evidence from Online Lending

The peer-to-peer loan market was designed to bring together borrowers and lenders without banks as middlemen. Yet over time P2P lending platforms have evolved into new intermediaries, performing essentially all tasks related to loan evaluation. By contrast, lenders are overwhelmingly passive and automatically fund almost all loans on offer. The dominant role of lending platforms without skin in the game makes the market vulnerable to moral hazard, checked by the threat of institutional investors' withdrawal. Our findings suggest that in markets without private information reintermediation may arise naturally as the platform's expertise in data analysis crowds out that of investors.

Stock loan valuation based on the Finite Moment Log-Stable process

The empirical test suggests that the log-return series of stock price in US market reject the normal distribution and admit instead a subclass of the asymmetric distribution. In this paper, we investigate the stock loan problem under the assumption that the return of stock follows the finite moment log-stable process (FMLS). In this case, the pricing model of stock loan can be described by a space-fractional partial differential equation with time-varying free boundary condition. Firstly, a penalty term is introduced to change the original problem to be defined on a fixed domain, and then a fully-implicit difference scheme has been developed. Secondly, based on the fully-implicit scheme, we prove that the stock loan value generated by the penalty method cannot fall below the value obtained when the stock loan is exercised early. Thirdly, the numerical experiments are carried out to demonstrate differences of stock loan model under the FMLS and the standard normal distribution. Optimal redemption strategy of stock loan has been achieved. Furthermore the impact of key parameters in our model on the stock loan evaluation are analyzed, and some reasonable explanation are given.

Cost-sensitive boosted tree for loan evaluation in peer-to-peer lending

Abstract Loan evaluation is an effective tool for credit risk assessment in the peer-to-peer (P2P) lending market. However, most of the traditional loan evaluation models assume a balanced misclassification cost, which is far from business reality and leaves opportunities for investigating the cost-sensitive modelling approaches. Moreover, few existing studies consider the special characteristics of P2P lending. In this paper, we propose a cost-sensitive boosted tree loan evaluation model by incorporating cost-sensitive learning and extreme gradient boosting (XGBoost) to enhance the capability of discriminating potential default borrowers. Therefore, a portfolio allocation model that converts the portfolio optimization problem into an integer linear programming is proposed as a decision support system for unprofessional lenders. Unlike existing studies, which mainly evaluate the model through accuracy-based metrics, this study measures the expected profitability based on the annualized rate of return (ARR) and proposes a new metric (ARR curve) reflecting the ARR over different rejection rates. Two real-world P2P lending datasets are examined, and explanatory data analysis shows that explanatory variables differ in predicting the probability of default and profitability. Experimental results reveal that the area under the ARR curve is not a considerable metric in evaluating models from a profit perspective. Conventional cost-insensitive models may even lead to negative returns, whereas the proposed loan evaluation and portfolio allocation model are the best performing methods over the two datasets in terms of profitability.

Credit Scoring and the Quality of Business Credit During the Crisis

The Great Recession produced a sharp worsening of credit quality for banks and firms alike. In this context, this paper investigates whether in Italy during the period 2008–2011 credit quality was higher for the banks that had adopted credit scoring techniques before the crisis and actually incorporated them into their loan evaluation process. To this end, I use a unique dataset built by merging information about banks organization variables drawn from the Regional Bank Lending Survey, the main banks’ balance sheet indicators and data on non‐performing loans drawn from the Italian Central Credit Register and from banks’ supervisory reports to the Bank of Italy.The main finding is that, after controlling for other supply factors, during the financial crisis the banks that relied more heavily on credit scoring models had higher‐quality loans to non‐financial companies. This result is consistent with the thesis that scoring models mitigated the negative impact of the crisis by enhancing banks’ screening ability. However, the positive effect is significant only for loans granted to large firms; for smaller firms, the absence of a significant effect could be explained by the greater importance traditionally attached to other means of assessing their creditworthiness, mainly soft information.