Scoring was once perceived as a complex decision tool, with little incremental predictive power. Nevertheless with the statistical validation of the predictive power of scorecards and with the ease of interpretation of score ranges, scoring models have gained wide acceptance. To evaluate loan applicants and manage existing accounts, financial organisations are increasingly using credit scoring models.
These models are effective portfolio and risk management tools if they are employed appropriately. The use of such models has increased tremendously as a tool for underwriting and administering all forms of retail credit, including credit cards, direct and indirect installment loans, residential mortgages, home equity loans, and small business credit.
To add to this growth, small and medium-sized banks have joined the large banks and other organisations in taking advantage of this tool. Non-traditional uses for scores are also being created. For instance, some scoring models have been developed on the commercial side for the purpose of ascertaining the probability of fraud or theft.
However, credit scoring is a complex process and has become a serious issue in the lending community. While the use of credit scoring models has grown rapidly, the fair-lending issue has also undergone transformation. At one time (1990s) discrimination was not perceived as a considerable issue in the credit market, primarily due to associating discrimination with the obvious, intentional bias. Furthermore, in the same period, regulatory bodies have expanded their attention to encompass both consumer and business credit. Today, an imbalance in observed denial or high cost lending rates between groups can signal regulatory and enforcement agencies that a deeper probe is indicated. At an industry conference late in 2005, a representative from the Department of Justice said that they would investigate a half a percent difference in average APR spread between a protected class and non-protected class, even if the overall spread was below the established thresholds for higher priced loans set by the Federal Reserve Board. Consequences associated with non-compliance are severe, including large fines and class-action lawsuits.
Credit scoring models can assist bankers with credit decisions and can help enhance compliance with fair lending laws and regulations. This executive summary examines the current deployment of credit scoring models, the challenges encountered and the future trends in relation to Fair Lending regulations.
Contents
Management Summary
Introduction
Current Use of Models
Pitfalls with Current Models
Concerns with Application of Credit Scoring Models
Performance
Internal Audit Procedures
Regulatory Compliance
Future Trends
Conclusions
Download
PDF Ebook Credit Scoring Vulnerability to Fair Lending Risk
