Behavioural Economics

Behavioural Scoring Model for Managing Fraud in a Two-Sided Credence Goods Market: An Empirical Analysis

Abstract:

Fraud prevention in the banking industry is of utmost importance, particularly in loan transactions where borrowers may provide incorrect information. This article explores the use of a behavioural scoring model to mitigate fraudulent practices such as predatory lending and window-dressing credit data in a two-sided credence goods market. By combining game theoretical modelling and econometric analysis, the study proposes effective strategies for designing loans and financial products. Empirical insights from borrower reactions, lender behaviour, and optimal loan contract provisions provide valuable guidance for credit evaluation and management. The regression analysis highlights the impact of specific variables on the model's accuracy and predictive power. The study emphasizes the need to incorporate behavioural biases to prevent fraud and enhance market efficiency.

Introduction:

Managing fraud in loan transactions is a critical challenge for the banking industry. This article addresses the need for a behavioural scoring model in a two-sided credence goods market to prevent predatory lending and window-dressing credit data. By combining game theoretical modelling and econometric analysis, the study aims to propose effective strategies for designing loans and financial products.

Theoretical Framework:

2.1 Two-Sided Credence Goods Market: The study establishes the conceptual framework of a two-sided credence goods market, highlighting the information asymmetry between borrowers and banks. Banks are considered expert sellers of financial products, possessing superior knowledge compared to borrowers. The study explores the dynamics and incentives of both borrowers and lenders using game theoretical modelling.

2.2 Game Theoretical Modelling: Game theoretical modelling is employed to understand the motivations and behaviours of borrowers and lenders in the market. The study examines equilibrium solutions for predatory lending, fair play, and window-dressing credit data. A signalling game is utilized to analyse conditions under which banks offer fair terms or engage in predatory strategies.

Empirical Insights:

3.1 Borrower Reactions: Survey data analysis provides insights into borrowers’ reactions when they receive unfavourable deals compared to market conditions. Borrowers demonstrate a tendency to switch banks, renegotiate for better contracts, or make advance payments despite unfavourable terms. These responses indicate borrower retaliation and a desire to improve their situation.

3.2 Lender Behaviour: Empirical data also sheds light on lender behaviour, specifically the responses of branch managers. Branch managers show a willingness to offer better deals when credit is accurately stated or understated. However, there is variation in branch managers' responses when credit is overstated, with some opting for bad deals or rejecting borrowers. These findings illustrate the complexities of lender decision-making in the credence goods market.

Regression Analysis:

The study employs logistic regression models to analyse the impact of various variables on the model's accuracy. Four models are considered, including data before and after the Covid lockdowns. The regression analysis provides insights into the significance of explanatory variables and identifies optimal combinations for predicting borrower behaviour and managing fraud.

Optimal Loan Contract Provisions:

The study formulates objective functions for borrowers and lenders, considering microeconomic factors such as loan principal, interest rate, equity, and collateral value. The Nash equilibrium of the financing contract is determined based on these variables. Striking a balance between borrower and lender objectives is crucial for designing optimal loan contracts.

Conclusion:

The research presents a comprehensive analysis of a two-sided credence goods market and proposes a behavioural scoring model to prevent fraud and enhance market efficiency. Empirical insights from borrower reactions, lender behaviour, and regression analysis highlight the complexities and challenges involved in credit evaluation and management. The study contributes to the literature by incorporating behavioural biases and providing guidance for designing loans and financial products in a credence goods market. Future research can further explore the effectiveness and implications of the proposed model in real-world banking loan transactions.

Policy Recommendations:

Increased information and transparency in the banking industry are recommended to mitigate the impact of behavioural biases and moral hazard issues. Creating an information clearinghouse supported by the government, consumer groups, or major industry players can enhance market efficiency and prevent fraudulent practices. Designing rules for the information clearinghouse should consider geographical and cultural differences to effectively manage behavioural biases and prevent moral hazards.

Limitations of the Study:

The study acknowledges limitations in game theoretical modelling, particularly assumptions regarding utility functions and risk aversion. The econometric part of the study is subject to parametric assumptions. Future research can address these limitations and further refine the behavioural scoring model.

Keywords: Behavioural scoring model, fraud prevention, two-sided credence goods market, predatory lending, window-dressing credit data, game-theoretical modelling, econometric analysis, borrower reactions, lender behaviour, optimal loan contract provisions, information clearinghouse.