Chapter 11: Income Documentation: W-2, Hourly, Overtime, Bonus, Commission, and the Rules for Counting Income

"The borrower tells you what they earn. The underwriter tells you what they may count. Your entire job in this chapter is knowing, on day one, how far apart those two numbers are." — constructed; the working rule of this chapter

Overview

There is a sentence that ends more transactions than any credit score in this business, and it is usually said cheerfully, on a first call, by a person telling the absolute truth: "I make about ninety-eight thousand."

They do. Their tax return says so. Their employer would confirm it. And the number an underwriter will put in the denominator of every ratio in Chapter 4 might be \$94,000, or \$84,000, or — if the bonus that made up the difference started nine months ago — \$70,000. Nobody lied. Nobody made an error. The borrower answered the question they were asked, which was what do you earn, and the underwriter is answering a different question entirely, which is what can I prove will still be arriving in three years.

That gap is this chapter. It is the single largest source of the difference between a pre-approval letter and a closing, and it is almost entirely knowable on day one — if you ask the right four questions in the first twenty minutes instead of discovering the answers on day forty.

Here is the good news, and it is better news than most new loan officers expect. Income documentation looks like a memorization problem: two years for overtime, six months of receipt for child support, an award letter for Social Security, a lease and a factor for rents, a different rule for part-time than for a second job, and a stack of exceptions on top. It is not a memorization problem. Every rule in this chapter is one of three questions applied to a specific kind of money. Learn the three questions and you can derive the rule for an income type you have never seen, which you will need to do, because borrowers earn money in ways no guideline anticipated and the guidelines themselves are revised continuously.

We will build the Linden Street income file from zero to \$10,500.00 a month, one documented component at a time — a nurse's base rate and her shift differential, a sales representative's salary and four years of commissions — and at the end of the chapter every dollar of it will be tied to a specific document and a specific rule. Then we will look at the Fulton Avenue file, where the same rules produce a number \$583 a month below what the borrower's accountant told them, and understand precisely why the underwriter is allowed to do that.

In this chapter, you will learn to:

• State the three questions underwriting asks about any income, and derive rules from them
• Calculate qualifying income for salaried, hourly, overtime, bonus, shift-differential, and commission borrowers, correctly, in dollars
• Apply the 24-month average and explain why rising and declining trends are treated asymmetrically
• Reconcile a paystub's year-to-date figure against W-2s and a Verification of Employment
• Handle rental, retirement, Social Security, disability, alimony, and child support income
• Recognize employment gaps, job changes, and offer letters as stability questions, not paperwork
• Order and read the verification documents — VOE, Form 4506-C, tax transcripts — and know what each one actually proves

Learning Paths

🎓 Exam — §11.1, §11.3, and §11.7. The SAFE test loves the averaging rules and the non-taxable gross-up concept. Know the three questions cold; most stems are one of them in costume. 🏠 New LO — §11.2, §11.3, §11.4, and §11.9. These four sections are the difference between a pre-approval that survives and one that does not. §11.9 is where your first-year files will die. 🤝 Partner — §11.1 and §11.10. If you understand why verified income is not the same as earned income, you will stop being surprised by a lender's conditions and start warning your clients before they write the offer. 📊 Operations — §11.10 and the day-40 discovery. Verification timing is a workflow problem before it is a guideline problem.

11.1 The three questions underwriting asks about income

Before any rule, the frame.

An underwriter reviewing income is not auditing the past. That is the mistake nearly every new loan officer makes, and it makes the whole rulebook feel arbitrary. The underwriter is forecasting. They are being asked to certify, to an investor several steps removed who will hold this loan for years, that a specific household will still be able to make a \$3,033.72 payment long after everyone in this transaction has forgotten the address. Every documentation rule in this chapter is a tool for making that forecast defensible.

So the underwriter asks three questions about every dollar. In order.

Question 1: Is it stable? Does this money have a history? Not "did they receive it," but "have they received it long enough and regularly enough that a pattern exists?" A pattern is what you extrapolate from. A single receipt is an anecdote. This is the question that produces every history-length rule in the chapter — two years of overtime, two years of commission, six months of child support received, two years in the same seasonal line of work.

Question 2: Is it likely to continue? Forward, not backward. The common formulation in this business is a three-year forward expectation: the income should reasonably be expected to continue for at least the next three years. For wage income from a current employer, a lender generally does not document three years forward — it documents that the borrower is employed, that no termination is known, and that the employer expects the pay component to continue. For income that ends on a date — child support, an alimony order, a note receivable, a temporary disability award — the continuance test bites hard and directly, because the ending date is printed on a document. Verify the specific continuance documentation your program requires; it varies by income type and it is revised.

Question 3: Can it be documented? By a third party the underwriter will believe. This is the narrowest of the three and it disqualifies the most income in practice. Cash tips a restaurant server genuinely earns and genuinely spends are real income and are frequently not qualifying income, because no third-party record exists. Money a borrower's parents send every month is real and is almost never countable. The rule is not a judgment about the borrower's honesty; it is a statement about what an investor's file review can verify years later.

THE THREE QUESTIONS — the whole chapter in one diagram

                          A DOLLAR THE BORROWER RECEIVES
                                      │
                    ┌─────────────────┴──────────────────┐
                    │  1. IS IT STABLE?                  │
                    │     Does it have a HISTORY?        │── no ──► NOT qualifying
                    │     How far back must I look?      │          income
                    └─────────────────┬──────────────────┘
                                     yes
                    ┌─────────────────┴──────────────────┐
                    │  2. IS IT LIKELY TO CONTINUE?      │
                    │     Forward. Commonly ~3 years.    │── no ──► NOT qualifying
                    │     Who says so, and on what form? │          income
                    └─────────────────┬──────────────────┘
                                     yes
                    ┌─────────────────┴──────────────────┐
                    │  3. CAN IT BE DOCUMENTED?          │
                    │     By a third party the           │── no ──► NOT qualifying
                    │     underwriter will believe.      │          income
                    └─────────────────┬──────────────────┘
                                     yes
                                      ▼
                    ┌────────────────────────────────────┐
                    │  QUALIFYING INCOME                 │
                    │  The number that goes in the       │
                    │  denominator of every ratio in     │
                    │  Chapter 4.                        │
                    └────────────────────────────────────┘

Two definitions, and then we can go to work.

Qualifying income is the monthly income an underwriter will actually use to compute ratios: the portion of a household's earnings that passes all three questions. It is a number produced by the lender, not reported by the borrower.

Stable monthly income is the same idea stated from the stability side — income with a demonstrated history and a reasonable expectation of continuance, expressed as a monthly figure. The two terms are used almost interchangeably on the desk. Where they differ in emphasis, "stable monthly income" points at questions 1 and 2, and "qualifying income" points at all three plus whatever the specific program permits.

Now look at how much of the rest of the chapter is already derivable. Here is the whole rulebook, compressed. Every row is the three questions applied to one kind of money.

Every "commonly" in that table is doing real work. History lengths, continuance windows, and acceptable documentation vary by agency, program, income type, and automated underwriting recommendation, and they are revised. Verify against the current Fannie Mae Selling Guide, Freddie Mac Seller/Servicer Guide, HUD Handbook 4000.1, or the VA and USDA handbooks — and against your own employer's overlays — before you rely on any of them for a specific file.

📞 On the Phone

This is the twenty-minute conversation, and the four questions in it are the highest-value questions you will ask all year.

The question everyone asks: "What's your annual income?"

What you actually ask, in this order:

1. "Walk me through how you get paid. Is there a base, and then something on top of the base?"
2. "For anything on top — overtime, bonus, commission, shift differential — how long have you been getting it? Was last year bigger or smaller than the year before?"
3. "Has anything changed at work in the last two years, or is anything about to change? A new job, a promotion, a schedule change, going from salary to commission, a reduction in hours?"
4. "Is there any money coming in that isn't from an employer? Rent, support, retirement, a disability benefit, a side business?"

What goes wrong when you skip them: a borrower says "ninety-eight thousand," you write \$8,166.67 on the worksheet, and the file is priced, quoted, and under contract before anyone learns that \$28,000 of it is a bonus paid for the first time last February. The number was never \$8,166.67. It was \$7,833.33 — and you found out on day forty instead of day one.

Question 3 is the one nobody asks and the one that saves files. Ask it again at day thirty.

Notice something about the four questions: none of them requires you to know a guideline. They are just the three underwriting questions in a borrower's vocabulary. That is the point of the frame. You are not quizzing them; you are collecting the facts that determine which rules apply.

11.2 Salaried and hourly

Start with the easy case, because the easy case contains an arithmetic error that costs real files.

Salaried income

A salaried borrower has an annual salary that does not depend on hours worked. The rule is the simplest in the chapter: use the current base salary, divided by twelve. No averaging. If they got a raise last month, you use the new number, not last year's W-2. The W-2 is there to corroborate employment and to catch a story that does not hang together — not to set the figure.

The trap is the pay frequency. Employers pay on four common cycles, and only one of them lets you multiply by a familiar number:

Bi-weekly is not twice a month. There are 26 bi-weekly periods in a year and 24 semi-monthly ones, and the difference is two full paychecks. A loan officer who multiplies a bi-weekly gross by two has understated the borrower's income by 2/26 — 7.69% — and will do it consistently, on every bi-weekly file, until somebody catches it.

Watch what that does on a real file. Borrower 1 on Linden Street is a registered nurse paid bi-weekly. Her regular gross is \$2,640.00 per period.

• Correct: \$2,640.00 × 26 ÷ 12 = **\$5,720.00** per month.
• The error: \$2,640.00 × 2 = **\$5,280.00** per month.
• Understated by \$440.00 a month.

Household qualifying income would fall from \$10,500.00 to \$10,060.00, and the back-end ratio on this file — total obligations \$4,479.72 (Chapter 4) — would move from 42.66% to \$4,479.72 ÷ \$10,060.00 = 44.53%. Nearly two full points of debt-to-income, invented out of a multiplication.

Hourly income

Hourly income has no salary to divide, so you build it: hourly rate × hours per week × 52 ÷ 12. For a full-time schedule of forty hours, that reduces to the familiar shortcut rate × 2,080 ÷ 12, because 40 × 52 = 2,080.

For Borrower 1: \$33.00 × 2,080 ÷ 12 = \$68,640 ÷ 12 = \$5,720.00 per month.

The critical word in that formula is contracted. You do not get to assume forty hours. The Verification of Employment states the hours, and hospitals in particular run schedules that are not forty — three twelve-hour shifts is thirty-six. If this borrower's VOE had said 36 guaranteed hours rather than 40:

$$\$33.00 \times 36 \times 52 \div 12 = \$61{,}776 \div 12 = \$5{,}148.00$$

\$572.00 a month less**, from one box on one form. Household income would be \$9,928.00 and the back-end ratio would be \$4,479.72 ÷ \$9,928.00 = 45.12%**. On the Linden Street file the VOE confirms forty hours per week, eighty per pay period, and that is why \$5,720.00 stands.

When hours genuinely vary — a per-diem nurse, a warehouse worker on a fluctuating schedule, a retail associate — you have left "hourly base" and entered variable income, and §11.3's averaging applies to the hours, not just to a premium on top of them.

📄 Read the File

```text FIGURE 11.1 — "What a paystub actually proves" [the Linden Street file] THE DOCUMENT Bi-weekly earnings statement, Borrower 1, regional hospital system. Pay period 08/17-08/30; pay date 09/05; period 18 of 26. Delivered on day 5 with the completed application, along with the prior period. THE CONTEXT A $385,000 purchase, 5% down, conventional, two borrowers. Nothing has been verified by a third party yet. This is the first document in the file that contains a number an underwriter will use. WHAT IT SHOWS EARNINGS RATE HOURS CURRENT YTD Regular 33.00 80.00 2,640.00 47,520.00 Shift diff 3.00 62.00 186.00 3,030.00 Overtime 49.50 0.00 0.00 1,980.00 --------------------------------------------------------- GROSS 2,826.00 52,530.00

               Three facts an underwriter takes from this. (a) The base rate is
               $33.00 and the schedule is 80 hours per period -> $5,720.00/month.
               (b) The employer breaks shift differential and overtime out as
               separate line items, which is what makes them separately averageable.
               (c) YTD gross of $52,530.00 across 18 of 26 periods annualizes to
               $75,876.67, or $6,323.06/month -- within $23.06 of the $6,300.00 we
               intend to qualify her at.

WHAT IT DOESN'T It does not prove she is still employed on the day of closing; a paystub is evidence about a period that has ended. It does not say whether the shift differential is a permanent assignment or a temporary coverage arrangement -- 62 differential hours in this period against 1,010 year-to-date is a pattern, not a promise. It does not show last year or the year before, so it cannot establish the two-year history that overtime and differential require. And it says nothing at all about Borrower 2. THE DECISION Order the written Verification of Employment today (day 5 is early; the requests go out day 7) and specifically request the base / OT / bonus / commission breakout and the employer's statement on continuance. Collect two years of W-2s. Do not put $6,323.06 in the worksheet -- put $6,300.00, and be able to say why. THE LESSON A paystub establishes the rate and the pattern. It cannot establish the history, and it cannot establish the future. Every variable income component needs two more documents behind it. ```

Constructed. Pay statement formats vary enormously by employer and payroll provider; the line-item structure above is the one you hope for and frequently will not get.

🧮 Run the Numbers

Annualizing a year-to-date figure — and the two ways to do it, only one of which is right here.

Borrower 1's paystub shows YTD gross \$52,530.00 through the pay period ending August 30, which is period 18 of 26.

Method A — by pay periods elapsed (the correct method for a bi-weekly earner):

$$\frac{\$52{,}530.00}{18} = \$2{,}918.33 \text{ per period} \quad\rightarrow\quad \$2{,}918.33 \times 26 = \$75{,}876.67 \text{ per year}$$

$$\$75{,}876.67 \div 12 = \$6{,}323.06 \text{ per month}$$

Method B — by calendar months elapsed (August 30 ≈ eight months in):

$$\frac{\$52{,}530.00}{8} = \$6{,}566.25 \text{ per month} \quad\rightarrow\quad \$78{,}795.00 \text{ per year}$$

The two methods differ by \$243.19 a month, which is not a rounding difference — it is a \$2,918.33 difference in annual income. The reason: eighteen bi-weekly periods is 18 ÷ 26 = 69.23% of the year, while August 30 is only 8 ÷ 12 = 66.67% of the year. Method B divides by too small a fraction and overstates.

The interpretation, which matters more than the arithmetic. Method A gives \$6,323.06. We are qualifying her at \$6,300.00** — \$23.06 lower, a difference of 0.37%. That is the reconciliation working. The base component (\$5,720.00) is identical under both approaches because base is taken at the current rate; the entire \$23.06 sits in the variable component, where the year-to-date is running slightly ahead of the 24-month average of \$580.00. Running ahead is fine. Running behind is a condition.** Annualize the YTD on every file and compare it to your worksheet before you submit — when the two disagree by more than a percent or two, you have found something, and it is better to find it on day five.

(Illustrative pay-period calendar; annualization conventions vary by lender and by income type.)

The habit this section is really teaching is reconciliation. You will have three documents saying three slightly different things: a paystub with a rate and a YTD, W-2s with prior-year totals, and a VOE with the employer's own numbers. They will not match exactly, and they are not supposed to — a W-2 reports taxable wages after pre-tax deductions and can include items you may not count, while a VOE reports what the employer's payroll system thinks. Your job is not to make them equal. Your job is to be able to explain, in one sentence per component, why they differ. An underwriter who cannot get that sentence from the file writes a condition asking for it.

11.3 Overtime, bonus, and shift differential: the averaging rules

Now the part of the job that separates loan officers.

Variable income is any compensation whose amount is not fixed by the employment arrangement: overtime, bonus, shift differential, incentive pay, tips, piece rates, and commission (which is important enough to get §11.4 to itself). It is real money. Borrowers live on it. And it is treated completely differently from base pay for a reason that follows directly from question 1: base pay has a rate; variable pay only has a history.

The averaging rule

Apply the three questions and the rule falls out.

Question 1 — is it stable? Variable income commonly requires a two-year history. Not two receipts; two years. Some programs and some automated underwriting recommendations will accept a shorter history — often stated as at least twelve months — with additional documentation, and the treatment differs across agencies and across income types. Verify.

Question 2 — will it continue? This is the employer's answer, not yours and not the borrower's. The written Verification of Employment has a box for it. If the employer says the overtime is not likely to continue, the income is gone regardless of how long the history is. Read that box.

Question 3 — can it be documented? Two years of W-2s establish the totals; the VOE breaks the totals into components; the year-to-date paystub shows the current year is not falling apart. Three documents, three jobs.

And the arithmetic: sum the variable component over 24 months and divide by 24. That is the 24-month average. It is a monthly figure, so it drops straight into the worksheet next to base.

$$\text{24-month average} = \frac{\text{(prior-prior year variable)} + \text{(prior year variable)}}{24}$$

For Borrower 1 on the Linden Street file, the VOE and the two W-2s break out shift differential and overtime as follows:

$$\frac{\$13{,}920}{24} = \$580.00 \text{ per month}$$

Her qualifying income is therefore \$5,720.00 base + \$580.00 variable = \$6,300.00, and every dollar of it now has a document behind it.

Why averaging, and why 24 months

Because the underwriter is forecasting. A single year of overtime can be a hiring freeze, a colleague on medical leave, a contract that will not repeat. Two years is the shortest window in which a pattern is distinguishable from an event — and even then only barely, which is why question 2 exists as a separate test. The average is not an estimate of what the borrower will earn next year. It is a deliberately conservative statement of what a forecaster can defend.

Notice what the rule does not do. It does not reward a good year. It does not credit a raise in overtime rates. It does not consider that the borrower has already earned more year-to-date than they did in the whole of the prior year. All of those are true facts about the borrower and none of them changes the number, and this is the conversation you will have over and over.

The trend test, and the asymmetry that runs the whole chapter

Here is the rule that matters most, and it is the one worth reading twice.

If the variable income is rising, the 24-month average is usable. The older, smaller year drags the average below the recent year, and the underwriter accepts the lower number.

If the variable income is declining, the average is not usable, and the lower, most-recent figure governs. Averaging a declining stream would produce a number the borrower is no longer earning, and an underwriter who forecast with it would be forecasting a past that has already stopped happening.

Stated as one sentence: both branches of the rule select the more conservative number. That is not a coincidence and it is not two rules. It is one rule — forecast defensively — applied to two shapes of data.

THE TREND TEST — one rule, two shapes                     [constructed teaching example]

  RISING                                    DECLINING
  yr-2  ████████████        $19,800         yr-2  ████████████████   $109,500
  yr-1  ██████████████      $23,400         yr-1  ███████████████    $107,000

  24-mo average  = $1,800.00/mo             24-mo average  = $9,020.83/mo
  most recent yr = $1,950.00/mo             most recent yr = $8,916.67/mo
                   ^^^^^^^^^                                 ^^^^^^^^^
  USE $1,800.00 -- the average, which       USE $8,916.67 -- the recent year,
  is the LOWER figure.                      which is the LOWER figure.

  Left: the Linden Street commission.  Right: the Fulton Avenue file.
  Same principle. The underwriter takes the lower number both times.

The right-hand column is the Fulton Avenue file — a self-employed contractor running a six-employee residential HVAC business as an S-corporation. Two years of analysis produce \$109,500** and **\$107,000, a 2.3% year-over-year decline. The 24-month average is \$216,500 ÷ 24 = **\$9,020.83 a month; the most recent year alone is \$107,000 ÷ 12 = \$8,916.67 a month**; and because the trend declined, **the underwriter uses \$8,916.67.**

Two things about that file belong here and the rest does not.

First, the number the borrower expected. Their accountant told them they "make about \$9,500 a month." The underwriter's figure is \$583.33 a month lower, and the borrower will experience that as the lender calling them a liar. It is not. It is the same forecast-defensively rule you just applied to a nurse's overtime, running on a larger number.

Second, the decline is small. A 2.3% drop is nothing — a slow winter, one truck replaced, a customer who paid in January instead of December. The rule does not have a materiality threshold you can argue your way around. Down is down, and down means the recent year governs. If you take one operational habit from this section, take this one: when a borrower's variable income declined year over year, price and qualify the file on the lower figure from the very first conversation. You can always deliver good news later. You cannot un-deliver a pre-approval.

The mechanics of how those two Fulton Avenue numbers were produced — the tax returns, the add-backs, the cash-flow analysis form — belong to Chapter 32, which is where self-employed income is taught properly. What matters here is that the trend test does not care where the numbers came from. It is the same test.

🎓 NMLS Exam Watch

Variable-income averaging is heavily tested, and the exam's favorite trick is to give you a declining stream and see whether you average it anyway. Read the two years before you reach for the calculator.

The stem usually looks like: "A borrower received \$12,000 in bonus income two years ago and \$9,000 last year. What monthly bonus income may be used to qualify?" The arithmetic trap answer is \$21,000 ÷ 24 = \$875. The correct reasoning is that the income declined, so the most recent year governs: \$9,000 ÷ 12 = **\$750**.

Two related distinctions candidates miss under time pressure:

• Base pay is not averaged. A raise counts immediately; a salaried borrower's prior-year W-2 does not drag the current salary down. Averaging applies to variable income.
• History and continuance are separate tests. A four-year overtime history with an employer who states the overtime is ending yields zero qualifying overtime. Length of history cannot cure a failed continuance test.

Remember also that specific history lengths are program-dependent and change. The exam tests the principle; your file is governed by the current guide.

Bonus income, specifically

Bonus deserves a note because its timing confuses people. A bonus paid once a year in February means that as of January, the borrower has received the same number of bonuses they had a year ago — and in a February-to-February file, a borrower can look like they have "two years of bonus" on the strength of two payments fourteen months apart. Whether that satisfies the history test is a guideline question with a real answer, and the answer varies. Ask it explicitly rather than assuming, and put the payment dates in the file.

Also: read the bonus plan, or at least ask about it. A discretionary bonus that the employer states may not recur is a different animal from a contractual incentive tied to a formula, and the VOE's continuance box is where that difference becomes visible.

Shift differential

Shift differential — the premium an employer pays for nights, weekends, or holidays — is variable income and averages like overtime. It has one practical wrinkle worth knowing: some employers do not break it out on the paystub at all, folding it into the regular rate, and some pay it at a different rate for different shifts. If the differential is not a separate line item, you cannot separately average it, and an underwriter may either fold it into an effective hourly rate or refuse it.

The Linden Street VOE breaks it out, which is why \$580.00 could be computed. When it does not, ask the employer for a payroll detail report. That request costs one email on day seven and saves a condition on day thirty.

11.4 Commission income

Commission income is compensation paid as a share of sales or production. It is the purest form of variable income — a salesperson's commission is a direct function of an outcome nobody controls — and it is governed by the same two-year history and 24-month average you just learned, with three additional wrinkles that matter on real files.

Borrower 2, and why the average is the number

Borrower 2 on Linden Street is an outside sales representative, four years with the same company, paid a W-2 base plus commission. The base is straightforward: \$28,800 a year ÷ 12 = \$2,400.00 a month, current rate, no averaging, exactly like §11.2.

The commissions are the work.

🧮 Run the Numbers

The 24-month commission average, and what the conservative rule costs this household.

From the VOE's commission breakout, corroborated by two years of W-2s:

	Two years ago	Last year	24-month total
Commission	\$19,800 | \$23,400	\$43,200

The 24-month average:

$$\frac{\$43{,}200}{24} = \$1{,}800.00 \text{ per month}$$

The trend: \$23,400 − \$19,800 = \$3,600, an increase of \$3,600 ÷ \$19,800 = 18.18%. Rising. Therefore the average is usable, and Borrower 2's qualifying income is \$2,400.00 + \$1,800.00 = \$4,200.00 per month.

Now price what the rule costs them. Last year alone, this borrower earned \$23,400 ÷ 12 = \$1,950.00 a month** in commission. The rule uses \$1,800.00. The household is qualifying on \$150.00 a month less** than the most recent year actually produced.

What does \$150 do here? Total obligations on this file are \$4,479.72 (PITI + MI of \$3,033.72 plus \$1,446.00 of other monthly debt — Chapter 4).

Income used	Back-end DTI
\$10,500.00 (the 24-month average) | \$4,479.72 ÷ \$10,500.00 = 42.66%
\$10,650.00 (last year's commission) | \$4,479.72 ÷ \$10,650.00 = 42.06%

Sixty basis points of debt-to-income, permanently unavailable to this borrower because the conservative rule takes the lower of the two figures. On this file it does not matter — 42.66% clears. On a file sitting at 45%, it is the whole transaction.

The interpretation. A rising commission stream is good news that you do not get to spend. Say that to the borrower on the first call, in those words, and you will never have the argument.

That is the asymmetry from §11.3, now with a price tag on it. Rising: you use the average, which is lower than the recent year. Declining: you use the recent year, which is lower than the average. There is no configuration of the data in which the borrower gets the higher number.

Wrinkle one: the share-of-income threshold

Many programs treat commission differently when it makes up a large share of a borrower's compensation — a threshold commonly stated at 25% of the borrower's income. Above that line, programs have historically required additional documentation, which at various times has included personal tax returns and an analysis of unreimbursed business expenses reported by the borrower. The treatment of this threshold has changed, differs among agencies, and differs again with the automated underwriting recommendation. Verify the current requirement before you promise a borrower which documents they will or will not need.

Compute it for Borrower 2 anyway, because the arithmetic is instructive:

$$\frac{\$1{,}800.00}{\$4{,}200.00} = 42.86\%$$

Commission is nearly forty-three percent of this borrower's income. Note the denominator: the test is generally applied to the individual borrower's own income, not to household income. Against the household's \$10,500.00 the same commission is only \$1,800.00 ÷ \$10,500.00 = 17.14%, which would look comfortably under any threshold and would be the wrong calculation. Getting the denominator right is the difference between anticipating a condition and being surprised by one.

Wrinkle two: draws, and what a draw actually is

Some commission plans pay a draw — an advance against future commissions, recovered from later earnings. A draw is not base salary even when it arrives on the same schedule and looks identical on a paystub. A recoverable draw is, economically, a loan the employer makes against income the borrower has not yet earned, and treating it as base pay would count the same dollar twice: once when advanced and once when earned.

Ask the question directly: "Is your draw recoverable — do they take it back out of commissions you earn later?" If the answer is yes, the qualifying income is the commission, averaged, not the draw. If the borrower does not know, the VOE and the compensation plan will say.

Wrinkle three: W-2 commission versus 1099 commission

This is a boundary, and it is worth naming plainly because it changes which chapter governs the file.

A commissioned salesperson who receives a W-2 is an employee. Everything in this section applies.

A commissioned salesperson who receives a 1099 is, for mortgage purposes, generally treated as self-employed, and the entire analysis changes: different documentation, different history requirements, different arithmetic, and a business-expense analysis that can reduce the qualifying figure substantially below the gross receipts. That is Chapter 32's material, in full. Your job in the first twenty minutes is only to find out which one you have, and the question is one sentence: "At the end of the year, does your employer send you a W-2 or a 1099?"

Borrower 2 receives a W-2. That single fact is why this file's income analysis fits in one worksheet and Fulton Avenue's does not.

11.5 Part-time, second jobs, and seasonal work

Three related situations, one shared question: is this a pattern or is it a stretch?

Part-time and second-job income

A second job is exactly the kind of income that separates the three questions from each other. Question 3 is trivially satisfied — there are paystubs and a W-2. Question 2 is usually fine — the borrower still holds the job. Question 1 is the whole fight, because a second job undertaken in order to qualify for the mortgage is precisely the income an underwriter must not count.

The common requirement is a two-year uninterrupted history of the part-time or secondary employment, with the income then averaged over that period. Uninterrupted is doing work in that sentence: a second job held for two years across three different employers may be acceptable where a second job held for fourteen months is not, because the first shows a sustained pattern of working two jobs and the second shows a recent change in circumstances. As always, the specific requirement varies by program and by whether the automated underwriting findings ask for more or less. Verify.

The conversation that follows is a hard one and you should have it early:

"You've been driving for the delivery service eight months and it's about six hundred a month. I have to tell you now that we very likely cannot count it — most programs want two years on a second job before they'll use it. That doesn't mean it's not helping you; it means it isn't in the ratio. Let's build the file on what we can document and treat that six hundred as your cushion."

That is not a rejection of the borrower's effort. It is the difference between a pre-approval that holds and one that collapses in underwriting, and borrowers consistently respond better to it than loan officers expect.

Seasonal work

Seasonal employment — construction in a northern climate, agricultural work, resort and tourism staffing, school-year positions — is income that predictably stops and predictably resumes. Applying the three questions:

• Stability commonly requires a two-year history in the same seasonal line of work, which may span more than one employer.
• Continuance is evidenced by the employer's stated intent to rehire for the coming season. Without it, the history is a record of jobs that ended.
• Documentation is W-2s, the VOE, and — where the borrower receives it — a documented history of seasonal unemployment compensation, which some programs allow to be included when it is a consistent and documented part of a seasonal pattern. That allowance is program-specific and has conditions. Verify it; never assume it.

The arithmetic is the ordinary one: total the seasonal earnings over the documented period and divide by the number of months, including the off-season months. A worker who earns \$30,200 in seven months of the year has monthly qualifying income of \$30,200 ÷ 12, not \$30,200 ÷ 7. The underwriter is funding twelve mortgage payments, not seven.

The per-diem and PRN case

Worth a paragraph because Borrower 1's profession is full of it. Health care systems staff heavily with per-diem or "PRN" shifts — hours picked up voluntarily, at a premium, with no guaranteed schedule. Economically it behaves like overtime and it is treated like variable income: two years of history, averaged, with the employer's statement on continuance. What makes it dangerous is that it can be a very large share of a nurse's actual take-home, so a borrower who picks up two extra shifts a week may be living on income that the file will count at a two-year average or, if the pattern is only ten months old, not at all.

Ask, at application: "Are any of those hours shifts you volunteer for, or are they all your regular schedule?"

11.6 Rental income

Rental income is where borrower expectations and underwriting arithmetic diverge most violently, and the reason is that a borrower thinks in rent collected while an underwriter thinks in net rental income.

The structure

There are two ways rental income gets documented, and which applies depends on how long the borrower has owned the property:

• With a tax-return history, the property's reported rents and expenses establish the income. The reported net is adjusted for non-cash items — depreciation being the obvious one — because depreciation reduced taxable income without reducing cash. The full mechanics of reading the rental schedules on a tax return sit with the tax-return chapters; what you need here is the shape.
• Without a tax-return history — a property purchased recently, or the subject property itself if it is being purchased with a lease in place — a lease agreement and often an appraiser's comparable rent schedule establish the gross rent, and the lender applies a vacancy and maintenance factor. That factor is commonly 25%, meaning 75% of the gross rent counts. The factor, and whether it applies at all in a given scenario, varies by agency and program. Verify.

Then the property's own housing expense is subtracted. Rental income is a net concept:

NET RENTAL INCOME -- the arithmetic borrowers do not expect
                                                     [constructed teaching example]

  gross monthly lease rent                                     $1,800.00
  less vacancy/maintenance factor (illustrative 25%)            - 450.00
  ────────────────────────────────────────────────────────────────────────
  adjusted gross rents                                         $1,350.00
  less the rental property's own PITIA
      (principal, interest, taxes, insurance, association)    -1,520.00
  ────────────────────────────────────────────────────────────────────────
  NET RENTAL INCOME                                            -$170.00

  A NEGATIVE result is not "zero income."
  It is a $170.00 MONTHLY LIABILITY added to the borrower's debts.

Follow that through on a hypothetical qualifying file: a borrower with \$7,000.00 of other monthly income, a proposed \$2,100.00 PITI on the new home, and \$600.00 of other monthly debts.

The borrower walked into this conversation believing the rental added \$1,800.00 a month to their income. It subtracted \$170.00 — a swing of \$1,970.00 in the borrower's mental model — and it moved the ratio nearly two and a half points in the wrong direction.

Three practical points

A departing residence is a rental question. When a borrower is buying a new home and keeping the old one as a rental, the old property's full housing expense is a liability unless documented rental income offsets it, and programs impose specific requirements — an executed lease, sometimes evidence of a security deposit and its deposit into the borrower's account, sometimes a landlord-experience test. Requirements vary; verify before you tell a borrower they can "just rent it out."

Landlord experience can matter. Some program and product rules distinguish borrowers with a documented history of managing rental property from first-time landlords. Ask.

Boarder or accessory-unit income exists in some programs and not others, typically under specific first-time-buyer or affordable products with documentation requirements of their own. If a borrower mentions a room rented to a friend, do not count it and do not dismiss it — find out whether the program you are placing them in has a path.

🔍 Check Your Understanding

1. A borrower is paid \$3,100.00 bi-weekly. What is their gross monthly income, and what is the number a careless loan officer writes down instead?
2. Overtime was \$8,400 two years ago and \$7,100 last year. What monthly overtime income qualifies, and why is it not \$645.83?
3. A borrower's second job produces \$700 a month and began eleven months ago. Which of the three questions does it fail?

(1: \$3,100 × 26 ÷ 12 = \$6,716.67; the careless answer is \$6,200.00, a \$516.67 error. 2: declining, so the most recent year governs — \$7,100 ÷ 12 = \$591.67; \$645.83 is the 24-month average, which a declining stream does not earn. 3: question 1 — stability. It is documentable and it is likely to continue; it simply has no two-year history.)

11.7 Retirement, Social Security, disability, and non-taxable income gross-ups

Income that does not come from an employer follows exactly the same three questions, but each one is answered by a different document.

Retirement and pension income

For a defined pension or annuity, the award letter or the plan's benefit statement establishes both the amount and the continuance — including, critically, whether the benefit ends on a date or steps down. A pension that terminates in twenty-six months fails a three-year continuance test just as squarely as a child-support order that ends in twenty-six months. Read the letter for an end date before you read it for an amount.

For distributions from a retirement account — an IRA, a 401(k) — the questions are different because the money is an asset being spent, not a benefit being paid. Programs generally require evidence that the distributions are regular, that they are expected to continue, and that the account holds enough to sustain them across the required continuance window. An account paying \$2,000 a month with \$40,000 left in it does not support a three-year expectation; it supports twenty months. That arithmetic — balance divided by distribution — is the test in one line, and specific requirements about how the balance is evaluated vary by program. Verify.

Social Security and disability

Social Security retirement, survivor, and disability benefits, Supplemental Security Income, and VA disability compensation are documented by the award letter (and typically corroborated by the 1099 the agency issues and by the deposits landing in the bank account). The continuance question turns on whether the specific benefit has a defined expiration — some do, particularly benefits paid on behalf of a dependent child, which end at a stated age.

A note on how to ask. A loan officer may not inquire into the nature or severity of a disability. What you may and must document is the income: its amount, its documentation, and whether it has a stated end date. Chapter 25 covers the fair-lending framework around this properly, and it is not a technicality — it is one of the clearest lines in the rulebook.

The gross-up, and why you must never quote the percentage from memory

Here is the principle, and it is genuinely elegant.

Two borrowers each receive \$2,400.00 a month. One receives it as wages and pays income tax on it. The other receives it as a non-taxable benefit and pays no income tax on it. Their gross incomes are identical and their spendable incomes are not — the non-taxable borrower has more money available for a mortgage payment out of the same gross figure. A ratio computed on gross income therefore understates the non-taxable borrower's real capacity, and the qualifying comparison between them is unfair.

The correction is the gross-up: qualifying income for verified non-taxable income is increased by a percentage, so that it can be compared on the same footing as taxable income.

$$\text{grossed-up income} = \text{non-taxable income} \times (1 + p)$$

The percentage $p$ varies by program and it has changed. Different agencies have used different figures; some tie the adjustment to the borrower's actual tax situation as evidenced by returns; some cap it. Do not carry a number in your head, do not quote one to a borrower, and do not put one in a pre-approval letter without confirming the current figure for the specific program you are placing the loan in. This is the changing-numbers discipline from Chapter 1's framing applied to a place where being twelve months out of date costs a borrower a house.

Three further points that are stable even though the percentage is not:

1. Only verified non-taxable income may be grossed up. You must document that the income is in fact not taxable — an award letter, the absence of the income on tax returns, or program-specific evidence. "Social Security is non-taxable" is not universally true; a portion of benefits can be taxable depending on total income.
2. Never gross up taxable income. A pension that is fully taxable gets no adjustment, even though it sits on the same award letter as something that might.
3. The gross-up affects the ratio, not the borrower's bank account. The borrower does not receive more money. Say so plainly, because "we grossed up your income" sounds to a layperson like an accounting trick, and it is not one — it is a fairness correction.

🧮 Run the Numbers

What the gross-up does, and why the percentage is worth confirming.

A constructed borrower, not Linden Street. Monthly income: \$2,400.00 in verified non-taxable Social Security plus \$1,600.00 in a fully taxable pension. Total housing expense plus other monthly debts: \$1,850.00.

Without any gross-up:

$$\text{income} = \$2{,}400.00 + \$1{,}600.00 = \$4{,}000.00 \qquad \text{DTI} = \frac{\$1{,}850.00}{\$4{,}000.00} = 46.25\%$$

With an illustrative 15% gross-up — this figure is used solely to show the arithmetic; confirm the current percentage for your program:

$$\$2{,}400.00 \times 1.15 = \$2{,}760.00 \quad\rightarrow\quad \text{income} = \$4{,}360.00 \qquad \text{DTI} = \frac{\$1{,}850.00}{\$4{,}360.00} = 42.43\%$$

With an illustrative 25% gross-up — again, illustrative only:

$$\$2{,}400.00 \times 1.25 = \$3{,}000.00 \quad\rightarrow\quad \text{income} = \$4{,}600.00 \qquad \text{DTI} = \frac{\$1{,}850.00}{\$4{,}600.00} = 40.22\%$$

The interpretation. The same borrower, the same benefits, the same house: 46.25%, 42.43%, or 40.22% depending entirely on a percentage set by the program. Nothing about the household changed. This is why the number is not a detail — for a file sitting near a threshold, the gross-up percentage is the approval — and it is exactly why quoting it from memory is malpractice. Look it up, on this file, today. Then note in the file where you looked it up.

11.8 Alimony, child support, and other contractual income

Income that arrives because a document says it must is the cleanest possible application of the three questions, because the document answers question 2 by itself.

Alimony, maintenance, and child support received

Question 1 — stability. Programs commonly require documented receipt, frequently stated as around six months, established by bank statements, cancelled checks, or a state disbursement unit's payment record. A court order alone proves an obligation exists, not that anyone is paying it.

Question 2 — continuance. The order or decree states an end date, or states the conditions that produce one — most commonly a child reaching a specified age. Count the months. A three-year continuance expectation means roughly 36 months remaining, so support ending in 26 months commonly will not count, while support ending in 47 months commonly will.

Note the shape of this and how it differs from Chapter 4's treatment of a debt with few payments left. Both rules look at how many payments remain, and they point in opposite directions: a debt about to end may be excludable, while income about to end must be excluded. Both are the same underlying instinct — the underwriter cares about the future, not the past — pointed at opposite sides of the ratio.

Question 3 — documentation. The executed order, decree, or separation agreement plus evidence of receipt. Specific requirements, including how consistency of receipt is judged and what a lender does with irregular payments, vary by program. Verify.

Alimony and child support paid

The mirror image, and a place where a real structural choice exists. Support the borrower pays is an obligation. Some programs permit alimony paid to be treated as a reduction to gross monthly income rather than as a monthly liability — which changes both the numerator and the denominator of the ratio and can produce a materially different result from treating it as a debt. Whether that option is available, and to which kinds of support it applies, is program-specific and has changed. Do not assume it; check, and check which treatment produces the better outcome for the file, because it is not always obvious.

Other contractual and miscellaneous income

The same three questions dispose of most of what remains:

• Trust income — the trust document establishes the amount and, crucially, the duration; distributions must generally be documented as received and expected to continue.
• Notes receivable — the note establishes the payment and the remaining term; the continuance test is the maturity date, and proof of receipt is required.
• Royalties — documented by tax schedules and a contract, with the continuance question turning on the term of the underlying agreement.
• Foster care income — allowed by some programs with specific documentation and history requirements. Verify before promising.
• Automobile or housing allowances — countable in some programs where a history exists; frequently the associated expense must also be considered.
• Military entitlements — base pay plus documented allowances and special pays, evidenced by the Leave and Earnings Statement, with continuance depending on the entitlement. Chapter 17 covers VA lending; this is an earned benefit and the documentation is straightforward once you know what the LES shows.

And the recurring list of income that is real and generally does not count: cash income with no third-party record; unreimbursed reimbursements; one-time capital gains; a raise the borrower has been promised but not received; gift money from family that arrives monthly; income from a job the borrower has not started (except under the narrow offer-letter path in §11.9); and unemployment compensation outside a documented seasonal pattern.

The professional habit here is not to memorize the list. It is to run any unfamiliar income through the three questions out loud in front of the borrower. You will be right most of the time, you will know exactly which question to research when you are unsure, and — most valuably — the borrower will understand why the answer is what it is instead of experiencing it as an arbitrary refusal.

11.9 Employment gaps, job changes, and offer letters

Everything so far has assumed a borrower who is employed today and was employed two years ago by the same people. Most borrowers are not that borrower.

Employment gaps

An employment gap is a period during the documented history in which the borrower was not employed. It is a stability question, and the underwriter's concern is not moral: a gap raises the probability of another gap, and a gap during the loan term is a missed payment.

The common practice is that a gap of roughly thirty days or more requires a written letter of explanation, and an extended gap — often discussed around the six-month mark — may require the borrower to have been back at work for a period before the income is considered stable. Both the trigger and the cure vary by program and by the automated underwriting recommendation. Verify.

What makes a gap survivable is almost always the reason, and the reasons underwriters see constantly are ordinary human ones: school, a medical event, parental leave, caring for a parent, relocation with a spouse, a plant closure. Get the letter of explanation early, get it specific, and where a document supports it — a school transcript, a discharge summary is not needed but an enrollment record is, a spouse's relocation orders — attach it. A vague letter invites a second condition; a specific letter with a document ends the conversation.

Job changes

Changing employers is normal and is usually not a problem. The questions to ask, in order:

• Same line of work? A move within the same field and skill set generally preserves the stability history. A move into an unrelated field generally does not, and may restart it.
• Same compensation structure? This is the one that surprises people. A borrower who moves from a \$95,000 salary to a \$50,000 base plus commission at a higher total expected compensation has just converted \$45,000 of documented base into variable income with no history at the new employer. Their qualifying income may have dropped by \$3,750 a month while their actual earnings went up.
• Probationary period? Some employers impose one; some programs care.
• Increase or decrease? An increase is fine. A decrease is the beginning of a stability conversation.

Offer letters and income that has not started

Some programs permit qualifying on income from employment the borrower has not yet begun, under conditions that are narrow and specific. The conditions typically include a non-contingent written offer, a start date within a defined window relative to closing, and sufficient verified reserves to cover payments until the income begins — and often an additional verification once employment starts. Availability and terms vary by agency, program, and lender overlay, and they change. Do not tell a borrower this is possible until you have confirmed it for the specific program.

The right posture on the phone is: "There is a path for this on some programs. Send me the offer letter today and give me until tomorrow morning, and I'll tell you whether it works on the program we're using — and if it doesn't, what would." That is honest, it is fast, and it does not promise anything.

The habit that prevents all of it

Ask at application, and ask again around day thirty:

"Between now and closing — is there any chance you change jobs, get promoted into a different pay structure, switch from salary to commission, reduce your hours, take unpaid leave, or start a business? Even if it's good news. Especially if it's good news."

Write the answer in the file. Every one of the disasters in §11.10 was preventable by that question, asked twice.

11.10 Verification: VOE, 4506-C, transcripts, and the day-40 discovery

The borrower's documents establish what happened. Verification establishes that a third party agrees.

The Verification of Employment

A Verification of Employment (VOE) is a lender's request to an employer to confirm employment and compensation. It comes in two forms and they do different jobs.

A written VOE — the industry's standard form is Fannie Mae Form 1005, Request for Verification of Employment, for which Freddie Mac publishes an equivalent — is sent to the employer, usually to human resources or payroll, and returned completed and signed. It confirms dates of employment, position, current base pay and pay frequency, the prior two years' earnings broken out by base, overtime, bonus, and commission, the probability of continued employment, and any pending change. That breakout is the document on which §11.3 and §11.4 depend; without it you are averaging from W-2 totals that mix components together.

A verbal VOE (a VVOE) is a documented telephone or database confirmation that the borrower is still employed, obtained shortly before closing. Its entire purpose is question 2 in real time: the written VOE proved employment as of day ten, and the loan funds on day fifty-one. Timing requirements are program-specific and commonly stated as within a small number of business days of the note date. Verify the current requirement.

Increasingly both are satisfied through third-party employment verification databases, which return payroll data directly from participating employers. They are fast and they are excellent when the employer participates. When the employer does not, you are back to a form, a fax number, and an HR department that processes verifications on Thursdays — which is a calendar problem, not a guideline problem, and it is why the Linden Street VOE requests went out on day 7 rather than at submission.

THE VERIFICATION STACK -- what each document actually proves
                                                     [constructed teaching example]

  DOCUMENT                 PROVES                          DOES NOT PROVE
  ─────────────────────────────────────────────────────────────────────────────
  Paystub                  current rate; YTD pattern       history; continuance;
                                                           employment tomorrow
  W-2 (2 years)            prior-year totals; employer     the split between base
                           relationship                    and variable components
  Written VOE (1005)       dates, position, the            that they are still
                           BASE/OT/BONUS/COMM breakout,    employed at closing
                           employer's continuance view
  Verbal VOE               still employed, today           amounts
  4506-C + transcript      what was actually filed         current-year income
                           with the IRS
  Bank statements          money arrived                   where it came from (Ch.12)
  ─────────────────────────────────────────────────────────────────────────────
  No single document does the job. The file is the argument; each document is
  one premise in it.

Form 4506-C and tax transcripts

Form 4506-C is the IRS form by which a taxpayer authorizes a lender to obtain transcripts of their tax filings through the IRS's Income Verification Express Service. The borrower signs it; the lender submits it; the IRS returns a tax transcript — a summary of what the IRS actually has on file.

Why bother, when the borrower already handed you W-2s? Because a W-2 is a piece of paper the borrower gave you, and a transcript is what the IRS has. The transcript is an independent check against altered or fabricated documents, and it is one of the most effective fraud controls in the process. Chapter 27 covers income misrepresentation properly.

Two transcript types matter here:

• A tax return transcript shows line items from the return as filed.
• A wage and income transcript shows the information returns filed about the borrower — W-2s, 1099s, and similar — which is the one that corroborates a W-2 directly and which also reveals income sources the borrower did not mention.

The IRS has revised these forms and the transcript delivery process more than once; confirm the current form, the current authorization language, and your lender's specific requirement.

⚖️ Compliance Check

Three overlapping obligations sit on top of income verification.

Ability-to-Repay. Under the Truth in Lending Act as amended by the Dodd-Frank Act and implemented in Regulation Z, a creditor making a covered residential mortgage loan must make a reasonable, good-faith determination that the borrower has a reasonable ability to repay — and must verify income and assets using reasonably reliable third-party records. This is the legal foundation under everything in this chapter. It is why "the borrower told me" is not a documentation standard and why stated-income lending for covered transactions effectively ended. Case Study 1 works through it.

Authorization and privacy. A VOE, a 4506-C, and a third-party database inquiry all involve the borrower's private financial information. The Gramm-Leach-Bliley Act governs how you safeguard it; the borrower's signed authorization governs whether you may seek it at all. Do not call an employer before you have authorization, and do not email unencrypted paystubs.

Equal Credit Opportunity Act / Regulation B. How you ask about income is regulated. You may not inquire into the nature or severity of a disability. Income from public assistance may not be discounted because of its source. A borrower need not disclose alimony, child support, or separate maintenance unless they want it considered — and you must tell them so.

Requirements change and state law varies. Verify current rules with your compliance department and your regulator before you rely on any summary, including this one.

The day-40 discovery

Which brings us to the failure this section exists to prevent.

⚠️ Where Deals Die

The verbal VOE on day forty.

The mechanism, and it is always the same. The file was approved on documented income. The borrower then did something completely reasonable and told nobody, because nobody told them it mattered. The verbal VOE — the one obtained days before closing to satisfy question 2 in real time — is where the lender finds out.

A constructed file, not Linden Street. A borrower qualified at \$7,400.00 of gross monthly income with \$3,050.00** of total obligations: a back-end ratio of **\$3,050.00 ÷ \$7,400.00 = 41.22%, approved with conditions on day twenty-six. On day 40, the verbal VOE comes back from the employer: terminated, effective Friday — the borrower resigned. They accepted a better job. They start in two weeks. They were excited and they did not think to mention it, because nobody had asked them since the application.

The new position pays a \$78,000 base with a bonus plan. The bonus has no history and cannot be counted. Documentable income falls to \$78,000 ÷ 12 = **\$6,500.00**, and the ratio becomes:

$$\frac{\$3{,}050.00}{\$6{,}500.00} = 46.92\%$$

The file is now a different file. The approval's income condition is void, the findings must be re-run, and the new employment cannot be verified until the borrower has started and, on most programs, produced a paystub. That is weeks the contract does not have. The lock will need an extension — on a \$300,000 loan, a fifteen-day extension priced at an illustrative 0.125 point is \$375.00 somebody now has to pay — and the closing date in the contract will be missed, which is a negotiation with a seller who has movers scheduled.

What the disciplined loan officer does instead. Three things, none of which cost anything:

1. Ask at application whether any employment change is contemplated, and write the answer in the file.
2. Tell the borrower explicitly — in the same breath as the do-not-open-new-credit conversation — that a job change, a promotion into a commission structure, a schedule reduction, or unpaid leave before closing can end the loan, and that the correct move is to call you before accepting anything, not after.
3. Ask again around day thirty, before the verbal VOE, while there is still time to restructure.

Borrowers do not conceal these things. They simply do not know that a promotion can be a problem. The only person in the transaction whose job it is to tell them is you.

The wider lesson generalizes past employment. Verification is not a formality performed on facts you already know; it is the moment the file's assumptions meet reality, and it happens near the end by design. Everything you can pull forward — the written VOE on day seven, the 4506-C signed at application, the year-to-date reconciliation done on day five instead of at submission — converts a day-forty catastrophe into a day-five conversation.

🗂️ The Loan File

Chapter 11 contribution: the income worksheet — \$10,500.00, built from four documented components.

This is the chapter where the denominator of every ratio in this file gets built. Here it is, one component at a time, each with the document that supports it and the rule that governs it.

Foot it: \$5,720.00 + \$580.00 + \$2,400.00 + \$1,800.00 = \$10,500.00. Borrower 1 contributes \$6,300.00 and Borrower 2 contributes \$4,200.00.

📄 Read the File

```text FIGURE 11.2 — "The income worksheet" [the Linden Street file] THE DOCUMENT Lender income calculation worksheet, both borrowers, prepared day 10 after the written VOEs returned (requested day 7). One page. It is the underwriter's first stop and the document a post-closing audit will re-perform line by line. THE CONTEXT A $385,000 purchase, 5% down, conventional, 45-day contract. Credit was pulled day 1 and the representative score is 706. The AUS returned Approve/Eligible on day 6 using income the borrowers stated; this worksheet is what turns that into a supportable file. WHAT IT SHOWS BORROWER 1 -- registered nurse, W-2, 3 yrs base $33.00/hr x 2,080 / 12 $5,720.00 shift diff + OT 24-mo avg ($13,920 / 24) $580.00 ----------- $6,300.00 BORROWER 2 -- outside sales, W-2 base + commission, 4 yrs base $28,800 / 12 $2,400.00 commission 24-mo avg ($43,200 / 24) $1,800.00 ----------- $4,200.00 =========================================================== TOTAL QUALIFYING MONTHLY INCOME $10,500.00

               Both variable components trend UP (differential/OT $6,720 ->
               $7,200; commission $19,800 -> $23,400, +18.18%), so the 24-month
               average is the usable figure in both cases.

WHAT IT DOESN'T It does not show that Borrower 2 could have qualified on $1,950.00 of commission using last year alone -- $150.00/month the rule does not allow. It does not prove either borrower is employed on the day of closing; that is the verbal VOE's job, days before funding. It does not address the $4,900 commission deposit sitting in the bank statements, which is an ASSET-sourcing question (Chapter 12), not an income question -- the commission itself is already inside the 24-month average and adds nothing further to income. And it is silent on affordability: $10,500.00 is what may be counted, not what is comfortable. THE DECISION Lock the worksheet. Quote, disclose, and pre-approve at $10,500.00 -- not at the $10,650.00 last year's commission would suggest, and not at the $6,323.06 Borrower 1's year-to-date annualizes toward. Tell both borrowers today, in plain language, which pieces of their pay the file counts and which it does not, and why. THE LESSON A worksheet is not arithmetic; it is an argument, and every line needs a document behind it and a rule above it. If you cannot name both for a line, that line is not income yet. ```

Constructed. Worksheet formats vary by lender; the components, documents, and rules are the content that matters.

What this settles. The denominator. Every ratio in this file now rests on \$10,500.00: housing 28.89%, back-end 42.66% against total obligations of \$4,479.72 (Chapter 4). It also settles the shape of the conversation — both borrowers now know that roughly \$2,380 a month of their pay is variable income counted at a two-year average, and that a change at either job changes the loan.

What it does not settle. Whether the money is there — assets, the \$10,000 gift, and the \$4,900 commission deposit are Chapter 12's. Whether \$3,033.72 is affordable as opposed to approvable, which is a different question the ratios cannot answer. And whether either job survives the next forty-one days, which is why the verbal VOE exists.

Open questions carried forward:

• The commission trend. It rose 18.18% and we counted \$1,800.00. If it had fallen, this file would be qualifying at \$10,350.00 and the back-end ratio would be 43.28% — a materially different conversation. Revisited when the automated findings are read in Chapter 15.
• The variable share. \$2,380.00 of \$10,500.00 — 22.67% of household income — is variable and averaged. That is a compensating-factor and a risk-layering question. Chapters 14 and 15.
• The employment risk. Two employed borrowers, two verbal VOEs, fifty-one days. §11.10, and Chapter 19's condition discipline.

Your task. In Appendix C's workbook, rebuild the four-line worksheet from the source documents without looking at the total, then check that it foots to \$10,500.00. Then do the harder half: for each of the four lines, write the single sentence you would say to the borrower explaining what that line is and why it is that number. If a sentence takes more than about twenty seconds to say, you do not understand the rule yet.

Conclusion

Underwriting asks three questions about every dollar a borrower receives. Is it stable — does it have a history? Is it likely to continue — commonly a three-year forward expectation? Can it be documented by a third party the underwriter will believe? Every rule in this chapter is one of those three questions applied to a particular kind of money, which is why the rulebook is derivable rather than memorizable.

Base pay is taken at the current rate and never averaged, with pay frequency the arithmetic trap: bi-weekly is twenty-six periods, not twenty-four, and getting that wrong understates income by 7.69% every single time. Variable income — overtime, bonus, shift differential, commission — commonly needs a two-year history and is averaged over twenty-four months, and the trend test decides which figure governs. Rising: use the average. Declining: use the lower, most-recent figure. Both branches select the more conservative number, because the underwriter is forecasting rather than auditing, and a forecast built on income that has already started falling is not a forecast at all.

Everything else follows the same shape. Rental income is net, not gross, and a net loss is a liability. Non-taxable income may be grossed up so it can be compared fairly against taxable income — by a percentage that varies by program, has changed, and must be confirmed on every file rather than recalled. Support income needs documented receipt behind it and enough months remaining in front of it. A second job needs two years. And none of it is real until a third party says so, which is what the written VOE, the verbal VOE, Form 4506-C, and the tax transcript are for.

The operational core of the chapter is smaller than the rulebook: find out how the borrower gets paid, in detail, in the first twenty minutes — and ask, twice, whether anything is about to change. Every failure in §11.10 was a question nobody asked on day five, discovered on day forty, at a price.

The Linden Street file now has a denominator: \$10,500.00 a month, four components, each with a document and a rule. The borrowers are still shopping — an online lender has quoted them a lower rate — and that comparison is waiting at the end of the book.

Next: the money has to actually be there. Chapter 12 takes up assets — what counts as verified funds, the large-deposit rule that will require the borrowers to source that \$4,900 commission deposit, how a \$10,000 gift is documented so an underwriter can accept it, and what "reserves" means when the ratios are already tight.

Key Terms

Qualifying income — the monthly income an underwriter will actually use in the ratios: earnings that are stable, reasonably expected to continue, and documentable by third-party records. Produced by the lender, not reported by the borrower. (Ch.11)

Stable monthly income — income with a demonstrated history and a reasonable expectation of continuance, expressed as a monthly figure; the stability side of qualifying income. (Ch.11)

Verification of Employment (VOE) — a lender's third-party confirmation of employment and compensation. Written (industry standard Form 1005) confirms dates, position, pay, and the base/overtime/bonus/commission breakout; verbal confirms shortly before closing that the borrower is still employed. (Ch.11)

Base pay — the fixed component of compensation set by the employment arrangement: a salary or an hourly rate times contracted hours. Taken at the current rate and never averaged. (Ch.11)

Variable income — compensation whose amount is not fixed by the employment arrangement, including overtime, bonus, shift differential, incentive pay, and commission. Requires history and averaging. (Ch.11)

Overtime — premium pay for hours beyond the standard schedule; variable income, commonly requiring a two-year history and averaged over 24 months. (Ch.11)

Bonus — incentive compensation paid periodically or annually; variable income, commonly requiring a two-year history and averaged over 24 months. (Ch.11)

Commission income — compensation paid as a share of sales or production; variable income requiring history, averaging, and an employer statement of continuance. W-2 commission is employee income; 1099 commission is generally treated as self-employment (Ch.32). (Ch.11)

Shift differential — a premium paid for working nights, weekends, or holidays; variable income, averaged like overtime, and separately countable only when the employer breaks it out. (Ch.11)

24-month average — the standard method for converting variable income to a monthly figure: the sum of the component over twenty-four months divided by twenty-four. (Ch.11)

Declining income — a variable income stream lower in the most recent period than in the prior one; the 24-month average is not usable and the lower, most-recent figure governs. (Ch.11)

Employment gap — a period within the documented employment history in which the borrower was not employed; commonly requires a written explanation beyond roughly thirty days and can affect income stability when extended. (Ch.11)

Form 4506-C — the IRS form by which a borrower authorizes a lender to obtain tax transcripts through the Income Verification Express Service. (Ch.11)

Tax transcript — an IRS-produced summary of what was filed. A tax return transcript shows return line items; a wage and income transcript shows the W-2s, 1099s, and other information returns filed about the taxpayer. (Ch.11)

Year-to-date (YTD) — cumulative earnings from January 1 through the current pay period, shown on a paystub; annualized by pay periods elapsed and used to confirm that current-year income supports the qualifying figure. (Ch.11)

Gross-up — the upward adjustment applied to verified non-taxable income so it can be compared fairly against taxable income in a ratio. The percentage varies by program, has changed, and must be confirmed on every file. (Ch.11)

Spaced Review

1. (Ch. 4 + Ch. 11) A colleague multiplies a bi-weekly gross of \$2,640.00 by two and reports \$5,280.00 of monthly base income for Borrower 1. Compute the correct figure, state the dollar error, and compute what the mistake does to Linden Street's back-end ratio given total obligations of \$4,479.72.

2. (Ch. 9 + Ch. 11) The Uniform Residential Loan Application asks the borrower to report base, overtime, bonus, commission, and military entitlements as separate lines rather than as one total. Using §11.3, explain why the form is built that way and what you lose if a borrower lumps them together.

3. (Ch. 4 + Ch. 11) Chapter 4's treatment of an installment debt with ten or fewer payments remaining and this chapter's continuance test both count remaining months, and they point in opposite directions. State both rules in one sentence each and explain why the same instinct produces opposite treatments.

4. (Ch. 9 + Ch. 11) One of the six items that constitutes an application under the disclosure rules is income. Is that the borrower's stated income or the lender's verified qualifying income? Explain what your answer implies about how early the Loan Estimate clock can start relative to when you actually know what the file's income is.

5. (Ch. 11) Borrower 2's commission rose from \$19,800 to \$23,400. Recompute the household qualifying income and both ratios as they would stand if those two figures were reversed, and state in one sentence why the underwriter is permitted to use a different method for the two scenarios.