The role of credit scores in consumer lending today

RMA Journal, The, Oct, 2003 by Elizabeth Mays

The following article is Chapter 1 of Credit Scoring for Risk Managers: The Handbook for Lenders. Elizabeth Mays' book is to be published early in 2004 by Thomson South-Western, a division of Thomson Learning.

We've seen a revolution in the financial services industry in the last ten to fifteen years as lenders have embraced automated decision making and model technologies to speed loan decisions and manage credit risk. Although credit scoring has been used by lenders since the 1950s, in the last decade it's become pervasive throughout the consumer lending arena and has expanded to products like residential mortgages and small business loans.

In an article on credit scoring for small business lending, Loretta J. Mester (1) describes a 1997 Federal Reserve survey in which 70% of large banks indicated they use credit scoring in their small-business lending. Credit scoring has become the standard method used to evaluate the credit quality of residential mortgages since the introduction of automated decisioning technologies by the Government Sponsored Enterprises (GSEs), Freddie Mac and Fannie Mae in the mid 1990s.

Mester notes several benefits from credit scoring. First, it provides great efficiencies and time savings in the loan approval process. She cites a study indicating the traditional approval process for small business loans can take as much as twelve and one half hours over as long as two weeks. This can be reduced to under an hour with the use of credit scoring.

A second benefit of credit scoring is reduced subjectivity in the loan approval process. With traditional underwriting, the standards to which applications are held can vary depending on who the decision maker is. Human judgment is affected by past experiences. With credit scoring, lenders can ensure they are applying the same standards to all applicants regardless of race, gender, or other applicant characteristics.

David J. Hand (2) notes another benefit of credit scoring over traditional underwriting which is that score models permit the loan decision to take into account more factors than could a human making the decision judgmentally.

U.S. Federal Reserve Chairman Alan Greenspan stated in an October 2002 speech to the American Banker's Association "credit scoring technologies have sharply reduced the cost of credit evaluation and improved the consistency, speed, and accuracy of credit decisions."

The benefits of credit scoring don't just apply to the loan acquisition process but to credit scores used for account management as well. That is, using credit scores for loan collection and modification decisions, line management, and loss recovery strategies call speed these decisions, remove bias from them, and help lenders make the right decisions.

In the same speech, Chairman Greenspan further noted, when discussing credit scoring technologies "their use also has expanded well beyond their original purpose of assessing credit risk. Today they are used for assessing the risk-adjusted profitability of account relationships, for establishing the initial and ongoing credit limits available to borrowers, and for assisting in a range of activities in loan servicing, including fraud detection, delinquency intervention, and loss mitigation. These diverse applications have played a major role in promoting the efficiency and expanding the scope of our credit-delivery systems and allowing lenders to broaden the populations they are willing and able to serve profitably."

In the next section we discuss in more detail how lenders are using scores today. Following that is a short discussion on the latest methods and techniques for building and using credit scores.

Use of Scores in Loan Acquisition

Today many lenders (certainly the vast majority of large lenders) have their own proprietary custom acquisition scores. A custom score is one which is built for a specific product using a lender's own data. Often, a custom score will contain characteristics based on data from the loan application such as monthly income or debt to income ratio while generic scores are based only on credit bureau data.

Generic scores are built with a wide array of consumer credit data and typically don't focus on a specific loan product or specific set of borrowers. The most well known example is the FICO score. (3) The FICO score is based on models built by a leading scorecard developer, Fair Isaac Company and is designed to rank the likelihood that an applicant will go 90 days delinquent on any consumer credit loan or account within the next two years.

Lenders who use custom scores may still use generic credit scores to make quick decisions about the highest credit quality customers. In this case, the generic score is reviewed, frequently in combination with other data provided by the applicant, to get a preliminary risk assessment. If the borrower appears to pose very low risk, the loan may be funded with little or no further review. If the generic score leaves any question that the applicant presents little risk, more information is obtained so that the custom score can be calculated. As an example of this, many indirect auto lenders will commit to fund a borrower's loan on the basis of a high FICO score as long as certain other lending parameters are not exceeded. For applicants with lower FICO scores additional information is obtained and the custom score is calculated before a decision is rendered.