QUESTION
We conduct pre-funding, post-closing, and discretionary quality control reviews.
In an evaluation done in 2016, it was found that the quality control department had many auditing issues. The people who did the assessment are no longer with the company. Supposedly, the problems were corrected.
However, now Fannie has come in and criticized us for our quality control sampling methodologies. We were cited for defective discretionary audits, in addition to other adverse results.
So, now we’re writing you for some guidance.
What are the sampling methodologies for pre-funding and post-closing quality control in terms of discretionary audits?
ANSWER
Your question focuses on sampling strategies, so my response will concentrate on them. If you haven’t done an annual review of the rationale for selecting pre-funding and
post-closing loans, it comes as no surprise that Fannie became concerned.
Perhaps there was no attempt to evaluate trigger points, production volume
changes, market conditions, and responsive procedures. Maybe your quality
control department or your quality control vendor is insufficiently resourced
to mitigate risk.
Notwithstanding an annual review, you should be checking your quality control and loss mitigation risk factors at least quarterly. Reporting the QC findings to management should be ongoing.
One of the mini-audits in our Compliance Tune-up series is the QC Tune-up. It is popular because it determines if you adequately meet GSE and investor guidelines, a mandate that virtually all mortgage loan originating entities must implement, excluding certain self-generated portfolio loan products.
Click the link for information about the QC Tune-up.
Let’s be clear about Fannie’s expectations with regard to selection sampling. There are two methodologies: (1) Random Sampling and (2) Discretionary Sampling. The random sample provides a high-level view of loan quality. The discretionary sample is a subset of loans within the total population that focuses on specific characteristics, such as high-risk loans, adverse findings, and so forth.
Pre-Funding Selection
For pre-funding, you might be interested to know that Fannie does not have a minimum required sample size for pre-funding QC, nor does it require random selections for loans sampled in the pre-funding QC review. However, Fannie does require that the pre-funding sample size be relevant to the loan production volume.
The pre-funding selection sampling is discretionary, and there are two types of discretionary methodologies: Full and Component. Whether Full or Component, you will want to incorporate them into your Clear to Close procedures.
The full file selection is usually conducted on only a part of the loan population. In practice, there are multiple risk factors in a population of loans, so the full file selection leads to increasing certainty on the loan quality. Before our firm conducts a pre-funding audit, many clients first have selected specific characteristics, such as unique underwriting guidelines and documentation requirements or newly hired loan officers, processors, underwriters, or new third-party originators. Some clients focus on loans with multiple layers of credit risk, such as high loan-to-value ratios, low credit scores, and high debt-to-income ratios (DTI).
The component file selection focuses on a particular element(s) of the loan file itself that has the potential for elevated risk or loan characteristics identified with defects in a post-closing QC review. For instance, we have clients who target loans with a higher risk profile, focusing on likely areas that impact eligibility (i.e., higher DTIs, where, say, an undisclosed liability could affect eligibility).
Sometimes, a client will select on the component file basis to validate succinct procedures, such as loans that evinced process failures identified in the post-closing random sample (i.e., Form 4506-C execution rates). Another validation may be on internal exception policies and procedures to confirm all requirements are consistently followed. We’ve had clients select to validate the accuracy of loan quality tools used by their organization (i.e., fraud monitoring tools or undisclosed debt monitoring).
Many clients select loans on the component file basis that contain top trending defects identified in their post-closing random sample (i.e., gift funds or excessive interested party contributions). The goal is to test internal controls around each defect and the effectiveness of the corrective action plans and remediation efforts.
Post-Closing Selection
Ideally, pre-funding and post-closing are intrinsically interfaced. But post-closing QC takes us to the entirety of the loan transaction, including closing and legal documents that are not available during the pre-funding reviews. Most lenders choose random 10% sampling, though some go onto the statistical sampling method for the post-closing QC selection process. Loan quality results are reflective of the sampling method chosen.
Lenders generally use the random 10% sampling method with an annual production of 3,500 or fewer loans. The statistical sampling method is generally advantageous for lenders with an annual production of more than 3,500 loans.
Keeping it simple, lenders prefer the 10% sample selection because there is no need to manage a statistical calculation process. Also, it does not require a periodic evaluation to ensure the sample size is valid. However, the statistical sample selection produces statistically valid results that can be used to extrapolate loan quality conclusions. In that regard, for lenders with a consistent defect rate, it produces a predictable monthly sample that does not vary due to large swings in production volume.
Whether 10% random or statistical, a full file review must be completed on all loans selected for the post-closing QC process, with reverifications on all the data relied upon to qualify the borrower. In addition, all selections require a comprehensive collateral risk assessment of the appraisal used to support the value of the subject property.
Discretionary Audits
Because Fannie cited you for “defective discretionary audits,” I will provide a few extra guidelines for discretionary reviews. Keep in mind that discretionary selections allow you to optimize the reviews and target certain loan features (i.e., high-risk loan characteristics) identified in the pre-funding and post-closing random selections. This is why I mentioned the importance of regularly reviewing the selection criteria to ensure you effectively manage risk and your QC resources. Always keep the risk factors current!
To recapitulate, discretionary reviews may be full file reviews or targeted component reviews, which should allow you to increase the overall number of reviews or the ability to evaluate the risks.
Full file reviews require reverification of all components, whereas targeted reviews allow for reverification of only those elements being audited. In addition, targeted or component reviews are an effective way to narrow in on a particular risk element, product, or process within the loan origination process without completing a full file review.
Breaking this down further, there are two types of sampling methods used, one type for full file reviews and the other for component file reviews.
You would want to use the full file sampling method when selecting loans to review new hires, new products, or newly implemented processes. In this context, use discretionary reviews to ensure at least one loan from each third-party originator is pulled for a review annually.
Selections should be based on those risk attributes that are identified as top trending defects from pre-funding and investor review results. The goal is to analyze the root causes when developing action plans to prevent future defects from occurring or test the effectiveness of an implemented corrective action.
The component file review provides an opportunity to sample loans with known risks. For instance, the component or targeted review allows, among other things, for the ability to ensure the borrower was employed at the time of closing, the income used to qualify was accurately calculated, the assets were adequately documented, and property eligibility and validation of data were supportive of the appraised value. If there are defects in the foregoing criteria, the component file review can identify a rationale for needing a full file selection.
Click the links for information about quality control audits and the QC Tune-up.
Jonathan Foxx, Ph.D., MBA
Chairman & Managing Director
Lenders Compliance Group
