Application scoring is the use of a statistical scorecard to rate a new application for credit at the moment a lender decides whether to approve it. The scorecard combines details from the application form, such as income, employment and housing status, with data from a credit bureau, and turns them into a single number that estimates the chance of the applicant falling seriously behind. The lender sets a cut-off score and may vary the limit, price or deposit by score band as part of its credit underwriting.
It differs from behaviour scoring, which rates existing accounts using their own repayment and spending history, and from generic bureau scores such as the FICO Score, which many lenders feed into their own custom application scorecards as one input. Models are built from past applicants' outcomes, increasingly with machine learning. In the United States, a declined applicant must receive an adverse action notice giving the main reasons, under the Equal Credit Opportunity Act.
Ref 1050000778