Fair Lending – One Loan and One Gap Is All It Takes

Complying with fair lending laws has quantitative and qualitative aspects both of which must be covered in thorough and well-written procedures. One case study of an FDIC (Federal Deposit Insurance Corporation) audit demonstrates just how critical developing and upgrading procedures for both elements are to a company.

It only takes one loan for a problem

For the year being reviewed, the company reported over 43,000 loans on the HMDA LAR (Home Mortgage Disclosure Act loan application register), and one loan became the focal point in a fair lending review. The DTI (debt-to-income ratio) was 71%, the borrower had a few recent late payments on revolving and installment accounts, and the subject property needed some repair. How did a loan get all the way to the underwriter with that DTI? Simple—the tax transcript came in late and revealed the borrower had neglected to disclose four rental properties causing substantial losses on his tax returns. The loan was denied due to insufficient income, unacceptable property, and delinquent credit. Most underwriters would have come to a similar conclusion. The FDIC auditor saw this loan differently.

Each reason for denial was considered separately and not in the context of the whole picture. The loan was a 95% LTV (loan-to-value ratio) refinance of the borrower’s current home. All of the exterior wood siding was rotted along the ground and needed replacing. The borrower did not have the cash or equity to complete the repairs; yet, the auditor inquired as to why we, as the lender, didn’t offer a HELOC (home equity line of credit) for the repairs. My job was to answer calmly with our HELOC product requirements and underwriting guides. The auditor then proceeded to look through hundreds of approved loans to see if we had provided HELOCs for repairs on other refinance loans. This was not our company practice so there was not a matching approved loan.

The credit became the next discussion point. Why did we deny a loan with only a few delinquent credit accounts when the borrower had other good credit? This was clearly a layer of risk the underwriter was understandably concerned about but not a strong stand-alone reason for denial. The auditor found several approved loans with similar credit scenarios. Each of the approved loans had DTI’s within the guides and a mix of other stable factors such as lower LTV, better property condition, and adequate cash reserves. The auditor set the issue aside as an unresolved discrepancy until all other factors could be investigated. This raised the tension level of the audit as the auditor was not convinced all things were equal.

The income calculation became the pivotal point that would tip the scale to a positive or negative conclusion. Although the net rental income calculation may seem straightforward to the average mortgage professional, the auditor needed to see it was applied fairly and consistently. The income calculations for all four rental properties were based on the net income shown on Schedule E plus depreciation and divided by the applicable number of months. The auditor wanted to know why other factors weren’t considered such as equity in the property or utilities paid by the tenants, or one-time repairs and the possibility the borrower could raise the rent. Our answer was logical and aligned with GSE (government-sponsored enterprise) requirements. This didn’t stop the loan review from expanding to check if our underwriters had been consistent in their analysis and calculation of rental income. All totaled, the auditor reviewed over 400 approved loans to see if any had been approved under similar conditions as the one denied loan.

The quality of the review counts

Each loan became a story problem to present to the auditor. (No one tells you ahead that your life will be one story problem after the next when you get into the mortgage business!) The defense of the 400 plus loans started with good documentation, organized and easy to find. The team had done their job making sure tax returns and leases were fully documented. Good comments by originators and processors in the notes screen left the right trail of information to know the borrower was not ignored and was given accurate information at application and throughout processing. When a document was misplaced, the support team worked hard to get their hands on it fast. The training was paying off at this point.

There were dozens of loans with rental income scenarios and, of course, reviewed by different underwriters. Each loan included full details of the income approach and notes when unique factors were considered. In no case was there evidence of management waiving documentation or a company requirement. The underwriters did not take an aggressive average of income to get a DTI in line. Factors not assessed by automated underwriting were addressed. The manual income analysis held up to the tough standards of a fair lending review because of clear boundaries set within policies and procedures.

The manual steps in the loan process are an obvious target for any auditor. There are over 24 specific areas such as fraud, public record items, credit disputes, rental properties, source of funds, and others that cannot be fully assessed by automation. There are sophisticated analytical tools to identify pricing consistency at the time a loan is locked and other technology to compare HMDA data, such as overall denial percentages at the company versus regional demographics and branch locations. Pricing inconsistencies, for example, are facts and averages that can even be monitored almost in real time. However, the manual assessments are still an important part of underwriting and an area of fair lending exposure that lenders see in hindsight through post-closing QC results.

There is often a gap in the process because post-closing QC (quality control) typically doesn’t have the data needed to determine if an error, such as an incorrect income calculation, actually is a fair lending concern as well. The quality control management system must include an additional track for the manual factors impacting fair lending, such as underwriting errors. The track should also merge and improve on the pre-closing “second look” underwriting outcomes. Very often this step doesn’t have specific data, checklists, or worksheets associated with it to aid in explaining the thought process months later.

Second looks get another look

Regarding our one targeted loan with the 71% DTI, that did receive a second look by a senior underwriter, as did all denied loans. The note by the underwriter was very brief since it seemed like an open and shut case. This didn’t help our defense or the perception of our second look process. After all, the whole premise of a second look is that the borrower actually gets a second chance. Fortunately, the one brief note was ‘FHA max DTI 43%. ’ This at least demonstrated another product was considered. Every second look process should include notes about alternative products that were considered, evidence of re-run through AUS (automated underwriting system), if possible, and reasons why the change will not work. In this case, the DTI far exceeded the allowable limit, even in automated underwriting. This closed the door for the auditor to challenge calculating income differently based on FHA guidelines.

As the auditor’s time on-site approached the four-month mark, the tone began to change. The auditor understood the collateral was in fair condition and the income was calculated correctly. We agreed that the denial codes should generally not have included ‘unacceptable credit’ when the credit alone is not a reason for denial. However, in this case the credit did represent a layer of risk that was not acceptable for such a high DTI.

Fair Lending requires a fair interview

The final discussion was the adequacy of the application interview. Had the originator asked enough questions to identify there were rental properties? Was the borrower given a fair and complete application along with adequate counseling about qualification and income requirements?

This originator was one of the good ones, always consistent and detail oriented. We had no reason to believe the ‘other real estate owned’ on the application was avoided or omitted. The borrower was provided a list of documents to bring to application and showed up without the tax returns. Reminder emails for the tax returns were responded to very late, and, hence, the tax transcript was the exposing document. As the audit concluded, there was not a matching approved loan even when each denial code was audited independent of the others. The bank earned an outstanding CRA rating.

Use the audit to improve your processes

Surviving an audit is a good opportunity to close gaps where new light appeared. The few we found were:
  • Complete written documentation for the denied loan second look process;
  • Develop a fair lending review track for underwriter errors cited in pre- and post-closing QC;
  • Monitor denied loan percentages for protected classes, each month and add loans to the fair lending review track on a discretionary basis;
  • Improve the procedures for escalating an underwriting exception; and
  • Provide additional training for loan originators covering the importance of asking every question on the loan application and receiving clear responses from the borrower.

Companies often rely on employees to bring a unique scenario or problem to the supervisor for discussion. Generally, the thought process and discussions aren’t documented in the loan. This results in documentation gaps that must be filled by the compliance or QC officer facing the auditor. This is where well-written procedures come into play and help deliver on fair lending practices.