Machine Learning-Driven Lending Decisions in Bank Consumer Finance

Machine Learning-Driven Lending Decisions in Bank Consumer Finance

Xiaoning Wang, Yi Tang, Anna Grazia Quaranta
DOI: 10.4018/IJISSCM.348337
Article PDF Download
Open access articles are freely available for download

Abstract

This paper investigates the bank lending decision process for internet consumer finance using machine learning. It focuses on microloans and compares Logistic Regression and GBDT models for credit risk assessment. Variables are filtered and recoded via Information Value and WoE methods to enhance discrimination between defaulting and performing users. Experimental results utilizing these models predict credit risk and optimize using AUC values. Additionally, it develops a fixed-effect regression model to explore how bank-specific factors affect systemic risk, revealing that larger banks reduce risk, while higher returns, non-performing loans, and equity volatility elevate it, with inconclusive effects from leverage ratio.
Article Preview
Top

Literature Review

Scholars have pointed out a distinction between Internet finance in a narrow sense and a broad sense, and the difference between the two lies in whether they include the Internet-based form of finance (Orlova, 2020). Bazarbash presented a similar view that Internet finance is a general term for financial activities in which financial service providers provide financial services in banking, insurance, and securities through an Internet platform, with Internet technology as the technical support (Bazarbash, 2019). Internet finance, in a narrow sense, refers to the establishment of a bridge between financial service providers and users through Internet technology, establishing a new financial operation model, while Internet finance, in a broad sense, includes not only the financial operation model but also the external environment such as Internet financial institutions and Internet regulation (Caird & Hallett, 2019). Specifically, some scholars’ understanding of Internet finance mainly lies in the narrow scope of Internet finance, which is considered to be an innovation in financial forms. Among them, scholars believe that the “new finance” of the Internet is not a kind of financial exuberance but a change in the financial sales and access channels, whether online or offline; the transaction is a financial contract. Internet finance is probably just a stage in the development of the financial industry, a transitional form of the financial industry in the process of reform (Kakderi et al., 2019). Tsarchopoulos et al believe that Internet finance enhances the efficiency and user experience of financial services through Internet technology, broadens the new boundaries of financial services, and thus further promotes the reform of the financial sector but does not change the main functions of finance in the six areas of payment, price discovery, financing, risk management, resource allocation, and dealing with information asymmetry (Tsarchopoulos et al., 2017). In other words, Internet finance does not derive new financial functions but only improves the efficiency and experience of financial services.

The earliest risk models were based on modern investment theory. To give the model an optimal loan allocation decision, Monte Carlo simulation was used to obtain key parameters such as yield and standard deviation for each year of the loan term, artificially introducing various constraints, such as risk limits, adjusting the strategy, international practice of laws and regulations, and business management, and finally establish return maximization as the objective function and calculate it (Komninos et al., 2019). Starting from the financial crisis in 2008, the International Accounting Standards Board (IASB) established an expected credit loss model for expected loss (EL) under the Basel framework and proposed new impairment accruals such as International Financial Reporting Standards (IFRS) 9. The relevant research argues that the implementation of the new credit risk regulatory model, involving profit and loss and credit risk, requires the collection of data calculation systems distributed by two major departments of finance and risk management and operationally requires consideration of whether the forecasts of different asset-related parameters can meet the adequacy and reliability of historical data (Rana et al., 2019). Accordingly, scholars believe that the new regulation will affect the lending strategies of state-owned banks and joint-stock banks, suggesting that commercial banks should re-examine their market exposures, strengthen the dynamic early warning of DVO1 for “current profit and loss changes in financial assets at fair value,” and encourage commercial banks to use interest rate derivatives for risk hedging (Trencher & Karvonen, 2020).

The banks should also encourage commercial banks to use interest rate derivatives to hedge their risks and require further strengthening of their credit rating capabilities for customers.

Complete Article List

Search this Journal:
Reset
Volume 17: 1 Issue (2024)
Volume 16: 1 Issue (2023)
Volume 15: 7 Issues (2022): 6 Released, 1 Forthcoming
Volume 14: 4 Issues (2021)
Volume 13: 4 Issues (2020)
Volume 12: 4 Issues (2019)
Volume 11: 4 Issues (2018)
Volume 10: 4 Issues (2017)
Volume 9: 4 Issues (2016)
Volume 8: 4 Issues (2015)
Volume 7: 4 Issues (2014)
Volume 6: 4 Issues (2013)
Volume 5: 4 Issues (2012)
Volume 4: 4 Issues (2011)
Volume 3: 4 Issues (2010)
Volume 2: 4 Issues (2009)
Volume 1: 4 Issues (2008)
View Complete Journal Contents Listing