The Declining-Income Rule for Self-Employed Borrowers

What the declining-income rule actually says

Self-employed income is qualified from your federal tax returns, and the direction of the trend matters as much as the amount. The framework most lenders follow comes from the Fannie Mae Selling Guide, section B3-3.2 (self-employment income), and it splits into two branches:

• Stable or increasing income → the underwriter averages your last two years: (year 1 + year 2) ÷ 24 = qualifying monthly income. You never get to use only the newer, higher year.
• Declining income → the average is off the table. The underwriter uses the lower, most recent year only — and must also be satisfied the decline has stopped. If the income "is not stable," meaning the drop appears to be continuing, the guide says it may not be usable at all, which in practice means a denial or a request to come back later.

The asymmetry surprises people. Rising income gets diluted by the weaker prior year; falling income gets no credit for the stronger prior year. Underwriting is deliberately conservative in both directions: the question is never "what did you earn at your peak?" but "what can we count on going forward?"

Two more things ride along with the lower-year figure. First, deeper scrutiny: a declining trend routinely triggers requests for a year-to-date profit-and-loss statement, business bank statements, and a written explanation. Second, discretion: the Selling Guide leaves "has the income stabilized?" to underwriter judgment, so two lenders can read the same file differently. That's also why individual lenders layer their own overlays on top (more on thresholds below).

Worked example: the same two returns, reversed

Take two borrowers with identical tax returns in opposite order. Both earned $96,000 in one year and $82,000 in the other (net Schedule C income, after add-backs are already applied).

Same $178,000 of two-year income, but Borrower B qualifies with $583 less per month — about 8% less income, purely from the order of the years. At a 6.5% 30-year rate, the $164/mo difference in housing budget is roughly $26,000 of loan amount if it all went to principal and interest (each $1 of monthly P&I carries about $158 of loan at that rate). In a real approval, part of the gap is absorbed by property tax and insurance scaling with price, so the purchase-price difference is somewhat smaller — but it is real money either way.

And note what Borrower A does not get: the full $96,000. Her qualifying income is $7,417/mo, not $8,000/mo. The average always pulls the rising borrower down toward the older year.

What counts as "declining" — magnitude and trend

The Selling Guide does not publish a bright-line percentage, so it helps to think in three bands. These thresholds are how lenders commonly behave as of July 2026, not published rules — treat them as directional:

Any decline: the average is gone

Even a small dip — $96,000 to $94,000 — technically makes the recent year the ceiling. Some underwriters will still average a trivial decline (1–3%) with a justification note, but you should plan on the lower year being your number the moment the trend points down.

Roughly 10–20% down: expect scrutiny

Many lenders' internal overlays flag declines in this range for a mandatory second look: a year-to-date P&L (sometimes CPA-prepared), 2–3 months of business bank statements, and a letter of explanation. The file can absolutely still close — on the lower year — if the documents show the bleeding has stopped.

20%+ down, or two declining years in a row: usability is in question

A steep drop or a continuing downtrend attacks the premise that the income is stable at all. Here the underwriter isn't asking "which year do I use?" but "can I use this income?" Without strong evidence of stabilization — a YTD P&L annualizing at or above the recent year — expect a suspense (more conditions) or a denial. This is the band where waiting a year or switching programs (see bank-statement loans) becomes the practical conversation.

One nuance: why income fell matters. A documented one-time event — a medical leave, the loss of one large client since replaced, a deliberate business relocation — reads very differently from a slow erosion with no story. Underwriters are allowed to use judgment, and a clean explanation gives them something to hang an approval on.

Five ways to mitigate a decline

1. Wait one filing year. The most reliable fix. If this year is back on track, next year's return makes the recent year the higher year — the decline disappears from the two-year window entirely. Expensive in time, free in everything else.
2. Bring a strong year-to-date P&L. A YTD profit-and-loss statement (CPA-prepared carries more weight) that annualizes at or above your recent year is the single best document for proving stabilization. Back it with matching business bank statements — underwriters check that deposits support the P&L.
3. Write the explanation letter properly. One page: what caused the drop, why it was one-time, what has already recovered, with dates and numbers. Vague letters ("business was slow") hurt more than they help.
4. Lower the ask. A larger down payment cuts the loan and the DTI the lower income has to support; paying off a car loan or card does the same from the debt side. If your file works comfortably at the lower-year income, the trend matters much less. Our guide to 2026 DTI limits shows exactly how much room each cap gives you.
5. Consider a different documentation path. If tax returns simply understate a healthy business, bank-statement programs qualify you on 12–24 months of deposits instead — at a higher rate and larger down payment. Compare honestly before switching.

How our calculator applies the rule

The self-employed affordability calculator implements the branch exactly as underwriters do. It first adds your add-backs to each year, then compares:

• If the recent year (with add-backs) is at or above the prior year: monthly income = (year 1 + year 2 + 2 × add-backs) ÷ 24
• If the recent year is lower: monthly income = (year 1 + add-backs) ÷ 12

A badge on the results panel tells you which branch fired — "2-year average" or "declining income: recent year only" — so you can see the rule working, and toggle your own numbers to measure exactly what a decline costs you. What it can't model is the judgment layer: stabilization evidence, overlays, and outright usability calls stay with a human underwriter.