Constructing a Loan Loss Model for Consumer Loans

The Idaho First National Bank formed a research team to develop multiple regression models to help predict loan losses. The team was successful in developing a useful model for consumer loans, which can be used to forecast loan losses 6 months into the future. The model has a good statistical curve...

Full description

Saved in:
Bibliographic Details
Published in:Journal of retail banking services : JRBS 1987-10, Vol.9 (3), p.64
Main Authors: Hogan, James D, Frankle, Alan, Merz, C Mike
Format: Article
Language:English
Subjects:
Agricultural Loans
Banking industry
Case studies
Computer based modeling
Consumer credit
Corporate Credit Services
Loan loss allowances
Loan losses
Loans
Mortgages
Problem Loans
Regression analysis
Statistical methods
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The Idaho First National Bank formed a research team to develop multiple regression models to help predict loan losses. The team was successful in developing a useful model for consumer loans, which can be used to forecast loan losses 6 months into the future. The model has a good statistical curve fit and is intuitively appealing. The model uses only 4 independent variables: 1. loan balance, 2. nonaccruals, 3. initial unemployment insurance claims, and 4. nonagricultural employment. Other banks also may find this model helpful, or they may wish to develop their own models. Those wishing to develop a model should consider the 4-step process used by Idaho First National: 1. Select a project team with skills in statistics, an understanding of the bank's lending process, and knowledge of the customer base. 2. Select the variables and gather the data. 3. Construct and check the model. 4. Record the results in a formal report.
ISSN:0195-2064