Incorporating multilevel macroeconomic variables into credit scoring for online consumer lending

Xia, Yufei; Li, Yinguo; He, Lingyun; Xu, Yixin; Meng, Yiqun

doi:10.1016/j.elerap.2021.101095

Cited by 21 publications

(23 citation statements)

References 80 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Nevertheless, some scholars suggest that the application of hard information in credit scoring places too much weight on the borrowers' negative credit history, thereby limiting and adversely affecting their repayment behavior (Giannetti & Jentzsch, 2013). In addition, the delay in updating borrowers' hard information may lead to information asymmetry between borrowers and lenders, which reduces the effectiveness of credit risk evaluation (Xia et al, 2021). Therefore, the development of hard information used in credit scoring is facing challenges (Wang et al, 2020).…”

Section: Credit Scoringmentioning

confidence: 99%

“…However, identifying borrowers' credit quality using these signals has some drawbacks. For example, the updating of borrowers' credit history lags (Xia et al, 2021) and some individuals (such as young people) lack sufficient credit records (Wang et al, 2020). It is due to these issues that lenders are less efficient to evaluate borrowers' credit risk based on such signals.…”

Section: Signaling Theorymentioning

confidence: 99%

“…As digital credit services and their products have become increasingly prevalent, lenders are able to evaluate borrowers' credit risk from a more multidimensional perspective. A consensus exists that borrowers' soft information can alleviate the information asymmetry in the evaluation of their credit risk (Bachmann et al, 2011;Liang & He, 2020), since lenders can make full use of the additional information (Hu et al, 2019;Xia et al, 2021). In this case, borrowers' soft information is the signal they send to lenders.…”

Section: Signaling Theorymentioning

confidence: 99%

“…For instance, borrowers' digital footprints on the Internet, such as their browsing histories of web pages (Berg et al, 2020;Orlova, 2021), mobile phone usage histories (Björkegren & Grissen, 2020;Ma et al, 2018;Pedro et al, 2015), and the e-mail usage histories (Djeundje et al, 2021), are collected and used to determine their credit risk. The credit risk evaluation incorporating borrowers' soft information breaks the limitations and broadens the application boundaries of credit scoring compared to conventional credit scoring based on hard information (Berg et al, 2020;Xia et al, 2021). However, further discussion is required regarding the reliability and accuracy of soft information used in credit risk evaluation.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

How Can We Learn from Borrowers’ Online Behaviors? The Signal Effect of Borrowers’ Platform Involvement on Their Credit Risk

Tang¹,

Feng²,

Zhu³

et al. 2022

Preprint

View full text Add to dashboard Cite

A growing number of borrowers are applying for digital credit through Internet platforms due to the integration of digital credit services the Internet. However, further empirical evidence is needed to explore how a borrower’s platform behaviors affect its credit risk. As such, our study uses signaling theory as the theoretical foundation to explore the overall effects of a borrower's platform involvement intensity on its credit risk based on a large consumer credit application dataset. The main finding shows the increase in a borrower’s involvement intensity reduces its likelihood of defaulting. We attribute it to the platform's belief that borrowers with high involvement intensity have the higher value to the platform. In addition, we examine how a borrower's involvement intensity is moderated by several factors, such as the stability of its platform involvement intensity and its credit history. Due to the importance of digital credit services in microfinance, we have provided useful implications for achieving win-win outcomes in the credit market for the stakeholders.

show abstract

Section: Credit Scoringmentioning

confidence: 99%

Section: Signaling Theorymentioning

confidence: 99%

Section: Signaling Theorymentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

How Can We Learn from Borrowers’ Online Behaviors? The Signal Effect of Borrowers’ Platform Involvement on Their Credit Risk

Tang¹,

Feng²,

Zhu³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Se por um lado a estimação da PD a nível individual auxilia as instituições financeiras na decisão de se conceder ou não um crédito (Xia et al, 2021;Zhou et al, 2021) e de se precificar estas operações (Brezigar-Masten et al, 2021;Dinh & Kleimeier, 2007;Featherstone et al, 2006;Luo et al, 2016), por outro a estimação da PD a nível de portfólio contribui para decisões gerenciais de alocação de capital (Bohn & Stein, 2009;Bülbül et al, 2019;Fraisse & Laporte, 2022) e é determinante para o cálculo da reserva de capital regulatório (Moreira, 2010;Thomas, 2009).…”

Section: Introductionunclassified

De Basileia para o campo: estimando a estrutura de dependência de >i<default>/i< em portfólios de crédito rural por meio de cópulas

Rensi¹

View full text Add to dashboard Cite

show abstract

Deep learning meets decision trees: An application of a heterogeneous deep forest approach in credit scoring for online consumer lending

Xia

Guo

et al. 2022

Journal of Forecasting

View full text Add to dashboard Cite

Online consumer lending has recently been growing rapidly, but it faces high credit risk. For this problem, developing powerful credit scoring models has become an effective solution and can be achieved from three aspects: modeling approach, data source, and evaluation measure. This paper proposes a novel model that departs from those in previous studies in threefold. First, a heterogeneous deep forest model that combines deep learning architecture and tree‐based ensemble classifiers is proposed as the modeling approach. Second, a Bayesian‐based macroeconomic variable optimization method is developed to determine the macroeconomic variables and the corresponding lag term, and the selected macroeconomic variables are used as supplementary data source for modeling. Lastly, a series of capital charge error measures is proposed to evaluate credit scoring models from a regulatory perspective. The proposal is evaluated on multiple large datasets under performance measures on predictive accuracy, profitability, and capital charge errors. Frequentist and Bayesian nonparametric significance tests are used to examine the statistical significance of heterogeneous deep forest and benchmarks. Three main conclusions can be reached from the comparison. First, heterogeneous deep forest significantly outperforms the industry benchmarks over all the evaluation measures. Second, the predictive performance is enhanced after incorporating the selected macroeconomic variables and the corresponding lag, and the result remains robust under cross‐validation and forward‐chaining validation. Third, the capital charge errors reflect the model performance from a regulatory perspective and thus lead to different rankings from those when evaluating predictive accuracy and profitability.

show abstract

Incorporating multilevel macroeconomic variables into credit scoring for online consumer lending

Cited by 21 publications

References 80 publications

How Can We Learn from Borrowers’ Online Behaviors? The Signal Effect of Borrowers’ Platform Involvement on Their Credit Risk

How Can We Learn from Borrowers’ Online Behaviors? The Signal Effect of Borrowers’ Platform Involvement on Their Credit Risk

De Basileia para o campo: estimando a estrutura de dependência de >i<default>/i< em portfólios de crédito rural por meio de cópulas

Deep learning meets decision trees: An application of a heterogeneous deep forest approach in credit scoring for online consumer lending

Contact Info

Product

Resources

About