Case Study 11.2 — When Verification Broke: Employment in 2020, and a File That Died on Day 40

Type: real, public industry event, examined through a clearly labeled composite file · Sources: Tier 1 (agencies, documented market events), Tier 2 (industry practice), Tier 3 (the composite file and all its figures)

The first case study showed a rule being written. This one shows the rule's limit — the failure that perfect documentation discipline does not prevent, and what a loan officer's discipline changes anyway.

Background: the month the labor market moved faster than the file

In March 2020, the COVID-19 pandemic produced an abrupt, very large contraction in United States employment. Restaurants, hotels, retail stores, gyms, theaters, salons, and travel businesses closed or curtailed operations within days. Furloughs and layoffs arrived faster than any underwriting process was built to absorb.

For mortgage origination the disruption hit at an exact structural point: the gap between when income is verified and when the loan funds.

Recall §11.10's sequence. A written Verification of Employment is obtained early — on the Linden Street file, requested on day 7. The loan funds much later — day 51. The verbal VOE exists precisely to cover that gap, confirming days before funding that the borrower is still employed. In ordinary times the verbal VOE is a formality that occasionally catches a job change. In the spring of 2020 it became the most consequential document in the file, because the gap it covers had turned into the whole story.

Two documented responses followed.

The agencies issued temporary flexibilities. Fannie Mae and Freddie Mac published a series of temporary selling-guide flexibilities beginning in March 2020, addressing employment verification when employers were closed or unreachable, appraisal and property inspection alternatives, and additional documentation for self-employed borrowers. These were extended repeatedly and later retired. Note the shape of the relief: it changed how verification could be performed, not whether it had to be.

Lenders tightened, sometimes far beyond agency requirements. Common overlays included verbal verifications performed the day of closing rather than days before, borrower attestations of continued employment signed at the closing table, additional documentation for self-employed borrowers such as recent business account statements and a year-to-date profit and loss statement together with an attestation that the business was operating, higher minimum credit scores, higher reserve requirements, and the temporary suspension of some products entirely. Overlays varied enormously by lender, which is precisely the guideline-versus-overlay distinction Chapter 14 takes apart.

The composite file

This file is a constructed composite, assembled from documented industry patterns of that period. It is not a real borrower, the figures are illustrative, and no real transaction is being described.

A restaurant general manager, employed by the same regional group for six years, buying a first home.

A completely ordinary file. Six years with one employer answers question 1 emphatically. The bonus had a clean two-year history and it rose year over year, so §11.3's trend test permitted the 24-month average. The written VOE was in the file on day 12 with the base/bonus breakout the underwriter needed. Conditional approval issued on day 27. Nine conditions, seven cleared by day 34.

Day 40. Eight days from closing. The verbal VOE.

The employer's HR line confirms it: the borrower has been furloughed, indefinitely, effective the prior Friday. Not terminated — furloughed, with a stated intention to recall when the dining rooms reopen, on a date nobody could name.

Apply the three questions to a furlough and the answer is immediate and brutal.

• Question 1, stability: six years of documented history. Passes.
• Question 3, documentation: flawless. Passes.
• Question 2, continuance: the employer cannot state that the income will continue, because the employer does not know.

One failed question, and qualifying income falls from \$6,200.00 to nothing that can be counted. The unemployment compensation the borrower began receiving is not countable outside a documented seasonal pattern (§11.5). There is no restructure. There is no compensating factor. There is no ratio to improve, because there is no denominator.

The file was denied. Sunk costs at the point of denial: appraisal \$650.00, credit report \$85.00**, home inspection **\$450.00 — \$1,185.00 the borrower does not get back.

What the loan officer's discipline actually changed

Nothing about §11.10's habits would have saved this transaction. Nobody could have asked a better question at application. This is the honest limit of the chapter, and it deserves to be stated plainly: verification proves the past and the present. It cannot prove the future, and it is not supposed to.

What discipline changed was everything downstream of the discovery.

The \$4,000.00 earnest money survived. The financing contingency was still in force because the loan officer had refused, three weeks earlier, to sign off on an early contingency release the buyer's agent had asked for to "make the file look stronger." That refusal cost a slightly awkward phone call in week three and saved four thousand dollars in week six.

The borrower learned it from their loan officer, by telephone, within an hour — not from a denial letter, and not from a closing that failed to happen with a moving truck already loaded.

The seller was told the same day, which preserved a relationship that mattered eleven months later when the borrower was recalled to work, re-documented with a fresh two-year history intact, and closed on a different house.

None of that is in a guideline. All of it is the job.

The mirror image: the income that went up and still did not count

The same period produced the opposite failure, and it is worth a paragraph because it tests whether you have actually internalized §11.3.

A hospital nurse — the same profession as Linden Street's Borrower 1 — began receiving crisis staffing premiums and incentive shift pay of roughly \$1,400.00 a month in the spring of 2020. Their take-home rose sharply. Their qualifying income did not move by a single dollar.

Why? Question 1: no history — the premium was months old, not years. Question 2: the employer stated explicitly that the incentive was temporary and tied to conditions expected to end. Two failed questions. The borrower experienced this as absurd, and said so, in the words loan officers heard constantly that year: "I am making more money than I have ever made in my life and you are telling me I qualify for less house."

Both are true statements, and they are not in conflict. The borrower was describing the present. The underwriter was describing a forecast. Those are different questions, and the entire chapter is the difference between them.

The lesson

Three, in order of usefulness.

1. Verification is a snapshot with a timestamp, and the timestamp matters as much as the content. Every verification in the file has an age, and the age of the oldest load-bearing document is a risk you are carrying whether you have named it or not.

2. The rule that seems harsh in one direction is the same rule that seems harsh in the other. The nurse's incentive pay and the manager's furlough are the same continuance test, and a loan officer who argues against it in one case has no principled objection in the other.

3. What you control is not the outcome; it is what the borrower knows and when, and which of their protections are still intact when the news arrives. The contingency, the phone call within the hour, and the relationship that let them close eleven months later were all decisions made before day 40, by someone who had considered that day 40 might go badly.

Discussion questions

1. The agencies' 2020 flexibilities changed how verification could be performed but never whether it had to be. Why is that distinction the whole story? What would have had to be true for the relief to go further, and what would it have cost?

2. The composite borrower had six years with one employer and a rising bonus — a file that scored well on every stability measure available. Does that mean the stability measures failed? Argue that they did, then argue that they measured exactly what they claim to measure and nothing more.

3. The loan officer declined to support an early release of the financing contingency three weeks before the furlough. That refusal produced an awkward conversation and saved \$4,000. Describe the equivalent decision in a file you can imagine handling this month, and say honestly whether you would make it.

4. Compare the furloughed manager and the nurse with incentive pay. Write the single sentence that correctly explains both outcomes to a borrower, and check that it does not sound like an excuse in either direction.

5. This case study concerns a genuinely unforeseeable event. Case Study 11.1 concerned a foreseeable and largely self-inflicted one. What does a rulebook owe to each? Should ability-to-repay analysis attempt to price the risk of a pandemic-scale shock, and what would happen to first-time buyers if it did?

Tier 1: the COVID-19 pandemic and the associated 2020 employment contraction as documented public events; Fannie Mae and Freddie Mac as institutions; the temporary selling-guide flexibilities issued beginning in March 2020 and later retired. Tier 2: descriptions of lender overlays and verification practices during the period, which varied substantially by lender and are not asserted as uniform. Tier 3: the composite file and every figure in it — income, ratios, costs, dates, and the nurse's incentive pay — all constructed for teaching. No statistics, enforcement figures, loss rates, or named companies are asserted. Verify all current verification requirements with your compliance department.