Fearing FNMA's New Loan Quality Red Flags? by Brad Kelso

by 20 May 2010

Simple Insights on two of FNMA’s Loan Quality Initiatives Can Save Repurchase Risk

Two of FNMA’s recently communicated -- Loan Quality Initiatives (Announcement SEL -2010-01 March 2, 2010) validating borrowers’ SSN/ or Tax Identifications and Confirming Undisclosed Liabilities -- seem daunting but only require a little more understanding for originators to navigate without concern. The key is in understanding what FNMA is after, how existing tools can help, and then knowing how to escalate should existing data or results prove insufficient. 
Bottom line, it’s not as tough as it sounds.
First, let’s review FNMA’s SSN/Tax ID validation request.
Simplifying the new LQI requirements, the originator must essentially check four elements of a borrower’s SSN:
1) its basic format
2) whether it has been legitimately issued
3) whether it is a deceased person or
4) whether the issue date is inconsistent with age   
The good news is that Social Security number use in connection with credit reporting is so pervasive that all three bureaus offer reputable and accurate sets of fraud alerts to assure that your borrower is properly connected to the SSN provided.  To be specific, in combination, the three bureaus offer 31 fraud alerts triggered off the borrower’s SSN, and most are specific to the issues raised by FNMA.
Job one then, is simply to make sure the originator has at least one of these fraud alert product sets -Transunion’s High Risk Fraud (formerly HAWK), Experian’s Fraud Shield or Equifax’s Safescan - included on the credit reports.  Two sets of alerts (from separate bureaus) are preferred and are an inexpensive best practice to assure that identity risk is minimized.
Important use facts (and limitations) for bureau alerts and SSN validation:
  1. The bureaus’ fraud alerts are only as good as the number of creditors contributing to that borrower’s report. The fewer the number of creditors on the file, the fewer the data points you will have to validate a borrower’s identity attributes definitively.  “Thin files” (those having 5 or fewer creditors reporting) will require more care, review and possible escalation.
  2. The creditors’ reported raw data is rarely ‘verified’ as being true and accurate, meaning, that, for example,  Macys’ best efforts to report data is unlikely to include a verification of the SSN it is reporting. Therefore, the validity derived from a credit report is essentially driven by consensus and corroboration from amongst all the reported creditors:
a.       For example, the trace data section of the credit report may indicate… “56 times the SSN 012-34-5678 was reported as connected to John Doe” and that “2 times the SSN was 013-34-5678.”
In this case, a 97% consensus is pretty strong validation especially given the absolute number (58) of reported SSN connections to Mr. Doe. Note too, in this case it’s pretty obvious that the alternative SSN’s reported were simple input errors – one digit off – a fairly common issue.
  1. Good News: There is absolutely no debate or second guessing for TWO of the bureau alerts!
a.       ‘SSN not issued’ (All bureaus have definitive checks against the SSA’s master file of SSN issued numbers)
b.      ‘SSN is associated with a deceased individual’ (same)
  1. What about ‘SSN Issuance Date inconsistent (prior to) reported DOB’This alert will require some secondary assessment by the creditor.
a.       The bureau is able to compare the SSN issuance date with the reported DOB of the borrower and issues a warning of any dates that look “odd” BUT, here again the quality of the reported DOB’s coming from the creditor is historically weak and as with SSN itself, not typically verified by the creditor. Consequently, this message is a good warning that effectively requires escalation.
b.      Fortunately, the SSA (via SSA Form 89) offers a definitive means of verifying a Borrowers’ DOB in connection with the SSN. This form requires signature by the borrower and acts just as an IRS Form 4506-T does for income tax filings, in granting authorization to check the authoritative records. Once armed with the definitive DOB in combination, this obligation under FNMA’s LQI is easily mitigated or identified with authority.
  1. ITIN validation
a.       The bureau alerts do not today address ITIN, therefore, validating these must be done through a 4506-T request of at least the prior year tax forms or W2’s associated with the borrower.
Secondly, let’s review FNMA’s basics regarding Undisclosed Liabilities.
Here the requirements have been presented more vaguely and yet by subsequent discussion and research can be boiled down to two straightforward requirements:
 1) Determine that any and ALL liabilities that could tangibly affect the future repayment of a new mortgage are disclosed; and
2) That all previously disclosed credit inquiries are researched to assure they have not become new tangible liabilities prior to close.
Practically, the best way to assure both situations are mitigated begins with ordering a new credit report not greater than 10 days prior to the close of the loan. This practice would most likely manifest either of the two issues; a new credit relationship whose repayment terms tangibly impact the underwriting or a change in indebtedness (increased balances) with existing creditors.
Important use facts (and limitations) for meeting undisclosed liabilities:
  1. FNMA’s threshold for re-underwriting a DU case file expressly refers to changes in debt payment minimums that exceed 2% of the qualifying income. Payment changes below this percentage should be disclosed, but should not require re-underwriting the casefile.
  2. FNMA has expressed that a new credit report must be within 10 days of the close of the loan. Credit reports older than that will not adequately meet the timing.
  3. Commonly, most credit reports offer a balance summary section that will allow you to quickly compare a change in overall debt balances BUT, as FNMA’s focus is on minimum payments, the total debt payment threshold will have to be recalculated. The rationale being that creditors’ payment terms vary especially on revolving debt. Similarly, large increases in balances on home equity lines may not necessarily increase payments as dramatically as revolving debt.
  4. At this time, potential changes in FICO scores are not included in the LQI however, and potential ‘newly’ returned middle scores below 620 would likely be a red flag as it falls below the floor for conforming lending.
  5. For following credit inquiries shown in the initial credit report, the timing of a second credit report may not necessarily have allowed enough reporting time to pass (bureau reporting is done in 30-45 day cycles) to have captured a new indebtedness. Consequently, originators seeking complete insulation from repurchase risk should be prepared to follow-up by phone with any of these inquiries based on additional discussion at the time of application and review with the applicant.
  6. The applicant should be thoroughly screened prior to signing the 1003 application for the possibility of undisclosed liabilities that are not being reported through the credit bureaus. The reality is that such indebtedness is quite common and the credit report cannot be relied upon as the sole source for identifying this issue. In fact credit reports only reflect those creditors who subscribe and many creditors are too small. Examples of common unreported liabilities include: local/small trade retailers (jewelry, grocery), Rent-to-Own operations, private automobile financing, or medical repayment agreements. All of these must be included by active inquiry and unfortunately are not easily identified or verified by automated means.
Don’t be intimidated by the new LQI Initiatives. Understanding the LQI requirements for SSN validation and undisclosed liabilities will allow you to address them fairly simply with tools that are readily available to you.
Brad Kelso is the Executive VP, Sales and Marketing at Informative Research, with a cumulative 22 years in financial services. Prior to joining Informative Research, Brad led Countrywide’s credit fraud initiatives and system development efforts with credits as a national expert and speaker on Authorized User Score Fraud. He is the primary architect of two products related to identity fraud for the mortgage industry.  Brad can be reached at (800) 473-4633 ext. 150 or e-mail bradk@informativeresearch.com.



Is TILA-RESPA a good or bad thing long term?