For the day one comes back
Why lenders send questionnaires back.
A returned questionnaire is rarely a wrong building. It is usually a right building described in a way underwriting cannot accept: a percentage computed on the wrong base, a count given in dollars, an answer with no source behind it. Here are the ten that come back most, what the lender is actually asking in each, and the answer that ends it.
The ten
Each one lists what the lender is asking, the answer that gets it returned, and what settles it. All of it applies to the Full Review regime that took effect for applications dated on or after August 3, 2026. What changed that day.
-
The reserve percentage, computed on the wrong base
The question: does the budget allocate at least 10% of assessment income to replacement reserves, rising to 15% for applications dated on or after January 4, 2027.
What comes back: the percentage the association already believes, carried over from a budget footnote. Underwriting recomputes it, and the recomputed figure governs.
What settles it: the division written out, with pass-through and incidental income taken out of the denominator first. Bulk cable, utilities billed straight through, master association pass-throughs and special assessment income all come out. A building can pass on the stated figure and fail on the computed one, and the two straddling the floor is the single most common return we see.
-
A reserve study doing work it cannot do
The question: if the budget does not meet the floor, is there a reserve study that supports the lower allocation.
What comes back: a study completed more than three years before the application date, offered as the substitute, or a budget that funds less than the study's own highest recommended figure.
What settles it: the completion date stated plainly, and the funding compared against the highest recommendation rather than a baseline method, which is no longer accepted for applications on or after August 3, 2026.
-
Delinquency answered in dollars
The question: how many units are 60 or more days past due on regular assessments, against a limit of 15% of units.
What comes back: a dollar figure, an aging report on a 30 or 90 day basis, or a ledger that still carries balances for units that have since transferred.
What settles it: a count of units, on a 60 day basis, reconciled before it is reported. Special assessment delinquency is tested separately and is not blended into the same count.
-
Occupancy that was estimated
The question: how many units are owner occupied, second homes, or investor owned.
What comes back: a confident figure with nothing behind it. Occupancy concentration limits were retired for established projects in March 2026, so this answer often feels harmless. It is not: it is the classic misrepresentation trap, and it is the one that has produced litigated losses for managers.
What settles it: the source and the as-of date, or a plain statement that the association does not maintain occupancy records. Both are acceptable. A guess is not.
-
Single-entity ownership counted too narrowly
The question: does any one person or entity own more than 20% of the units in a project of 21 units or more.
What comes back: a count that leaves out leased units, developer retained units, or units held by the association itself and rented out.
What settles it: holdings aggregated the way the guide aggregates them, itemized by holder, with the leased portion shown. Note also that Freddie Mac uses 25% at that size, so the answer can pass one agency and fail the other.
-
A special assessment whose purpose contradicts the repair answer
The question: the purpose, approval date, original amount, amount remaining and expected payoff of every current and planned special assessment, and separately whether there are critical repairs.
What comes back: a purpose written in engineering language, spalling, structural, remediation, sitting beside a No on critical repairs. The lender pulls the minutes, finds the mismatch, and the file is worse than if the repair had simply been disclosed.
What settles it: disclose the condition, attach the engineer's report rather than characterizing it, and let the two answers agree.
-
Litigation answered yes or no
The question: is the association a party to any litigation, and if so, what kind.
What comes back: a bare yes, which reads as disqualifying when it usually is not, or a bare no, which is worse when a matter surfaces later. Arbitration and mediation count too.
What settles it: disclose every matter, then classify each into the carve-out it falls in, with the facts the carve-out needs: the proceeding type, whether the insurer is defending, and the limits against the demand where that matters. A collection action the association brought is not a project defect, and saying so in the right words is the difference between a cleared file and a dead one.
-
Insurance answered with adjectives
The question: coverage amounts, deductibles, endorsements and loss settlement basis.
What comes back: "adequate", "standard", or a net effective deductible with the layers hidden. For applications dated on or after July 1, 2026 the per-unit deductible is capped at $50,000, and the per-occurrence deductible may not exceed 5% of coverage.
What settles it: figures transcribed from the declarations page with the peril, the basis and how each applies, including buy-back layers shown separately rather than netted.
-
Answering to the wrong vintage of the rules
The question: which rules apply at all.
What comes back: answers written to whichever version the manager learned, when four separate dates now route the file: March 18, 2026 for the items effective immediately, July 1 for the deductible cap, August 3 for Full Review and the reserve study method, and January 4, 2027 for the 15% floor.
What settles it: the lender's application date, stated at the top, with every answer given to the regime that date selects.
-
Material answers with no source
The question: all of them.
What comes back: a form filled in correctly with nothing behind it, which survives until a lender asks one follow-up question and the manager has to reconstruct the answer weeks later.
What settles it: every material answer carrying the document it came from and that document's date. It costs nothing at the time and it is the whole defense afterwards.
Where these come from
They are the failure patterns encoded in the checks we run on every package, drawn from the published Fannie Mae and Freddie Mac requirements and from what lenders actually return. Six of them are pure arithmetic, which means you can test a building against them yourself in about two minutes.
This page is general information about published agency requirements, not legal advice, not statutory interpretation, and not a lender decision. Your lender applies its own review and its own overlays.
Test one of your buildings
The readiness check runs six of these on your own figures, in your browser, with no signup and nothing sent anywhere. If you would rather see the whole thing done, the first package for any new client is free.