Mortgage Application Fraud Detection

BUSINESS BENEFIT SERIES: Mortgage Fraud

Industry: Mortgage Origination

Problem: Risk managers are unable to thoroughly review all loans for applicant fraud. Manual reviews are not an effective means for identifying trends in prior loan history or behavior patterns for the applicant or other parties of the transaction.

Solution: Through Vector ^™ , score all loan applications with FraudMark ^™    to automatically identify potential fraud in the pre-funding stage, and send only the risky loans for review.

Mortgage Application Fraud Detection

Typical Business Process:

To reduce risk associated with fraudulent loan applications, lenders review a random sample of applications. Manual review involves verifying loan applicant information through supplemental third-party reports to validate the borrower name, social security number, stated income, employment history and more. Random samples represent a portion of the lender's total loan volume but do not include all potentially fraudulent applications, leaving the lender at risk of underwriting fraudulent loans. Risk managers examine a single loan application and applicant at a time, and do not have access to critical behavioral trend information that can help them quickly identify fraud.

Alternative Business Process:

Score each loan application using FraudMark ^TM , available through Vector ^TM   , to automatically identify the highest risk loans. Based on identified patterns, FraudMark will provide a score of 1 - 999, with 999 being the riskiest along with up to 15 risk indicators, such as disproportionate age bracket-to-income ratios, inconsistent employment characteristics, and credit risks that differ from the typical applicant. Risk indicators improve the effectiveness of quality control reviews by pinpointing specific areas of concern that require further research. FraudMark also allows you to instantly re-score a loan application at any point during the loan origination process as additional data becomes available.

Using pattern recognition technology, FraudMark tracks behavior traits for applicants, brokers and account executives, as well as historical patterns of fraudulent and non-fraudulent loan applications. Leveraging a mortgage database that contains over 150 million mortgage loan attributes, FraudMark uses characteristics of past loans and parties involved in the transaction to detect deviations from normal behavior, predicting the degree of risk that the application contains fraud that could result in a financial loss.

Expected Results:

Decrease fraud losses and reduce data verification costs by automatically identifying loans that have the highest associated fraud. Rather than relying on a random sampling, all loans can be run through FraudMark in the pre-funding stage, allowing risk managers to focus their efforts on only the riskiest loans. Because Fraud- Mark provides specific risk indicators, managers know which parts of an application require deeper review, thereby improving productivity and further increasing work flow efficiency.