FICO Scores
FICO® scores were developed by Fair Isaac & Company, Inc. for each of the credit repositories. The scores are: (Equifax) Beacon®, (Experian formerly TRW) Experian/FICO and (TransUnion) Empirica®. They are simply repository scores meaning they only consider the information contained in a person's credit file; they do not consider a persons income, savings or amount of a down payment for a mortgage.
The scores were designed to assess risk. They are useful in directing applications to specific loan programs and to set levels of underwriting, i.e. streamline, traditional or second review. The scores are objective, consistent, accurate and fast.
Many people in the mortgage business are skeptical about the accuracy of FICO scores. Scoring has only been an integral part of the mortgage process in the past few years; however, the scores have been in use since the 1950's by retail merchants, credit card companies, insurance companies and banks for consumer lending. The data from large scoring projects emphasizes the accuracy, the predictive quality of the scores. Large portfolios have been scored for mortgage servicing and investment groups, and again, they demonstrate that FICO scores work.
The scores were developed from each repository's database using actual loan performance. A sample of over 750,000 consumers per repository was used. The repositories have each made great strides to increase the accuracy of their respective database through computer technology and internal monitoring. There is a new standard reporting format for credit grantors to use when sending electronic information to the repositories; this is the critical first step to providing accurate data.
The scores use a multiple scorecard design. Each repository uses 10 individual scorecards, and the models at each repository are the same. This increases accuracy and optimizes the predictive variables for each subpopulation. (For example, a borrower with two 30-day late payments will be scored against a population with some minor delinquencies.) This feature may cause a borrower with delinquencies to score in the same range as a borrower without delinquencies. Scorecards are reviewed and updated every twenty-four months.
The actual scoring process is proprietary, and the algorithms are copyrighted. We can share the predictive variables, the portion of the credit file considered and the weight as provided by Fair Isaac. They are:
- Previous credit performance (35%)
- Trade line information specific to payment history
- Current level of indebtedness (30%)
- Current balance compared to the high credit
- Time credit has been in use (15%)
- Types of credit available (15%)
- Installment loans, revolving accounts, debit accounts
- Pursuit of new credit (less than 5%)
;)
;)
