HUD Minimum Vacancy Rates

This HUD required vacancy floor, however, only applies to one particular NOI underwriting. HUD sizes a loan against two separate net operating incomes. The DSCR NOI, which the underwriter builds to test debt service coverage, is bound by the minimum. The Market NOI, which the appraiser builds and which drives appraised value and the loan-to-value test, is not — the appraiser sets whatever vacancy and bad debt factor the property and its market support. A sponsor comparing HUD’s 7% against Fannie’s 5% and concluding HUD is the more conservative execution is comparing one of HUD’s two numbers against Fannie’s only one. Our paper on DSCR NOI versus appraised value NOI works through what the difference between the two NOIs means on a deal where LTV or DSCR rates determine the loan amount.

The floors apply across the new construction and substantial rehabilitation programs — 220, 221(d)(4), 231, and 241(a) — and to 223(f) acquisitions and refinances. Section 232 healthcare properties are underwritten under a separate set of standards and are outside the scope of this paper.

The 5% tier is the one sponsors most often get wrong, because both conditions have to be met. A LIHTC property that clears the set-aside test but whose restricted rents sit at or near market rents does not qualify for 5% — meaning it will underwrite at 7%, the same as a market-rate deal. The rent advantage is part of earning the lower vacancy rate, not the set-aside by itself.

The 223(a)(7) streamlined refinance does not use these tiers at all. HUD underwrites a 223(a)(7) at the property’s actual occupancy. There is no new appraisal on a 223(a)(7) and no programmatic DSCR vacancy floor applied to it, which means a well-occupied property refinancing under a 223(a)(7) is not carrying the 7% haircut that the same property would carry under 223(f).

Again, these are minimums, not fixed assumptions. HUD’s guidance is explicit that underwritten vacancy and collection loss may be set above the minimums where the appraised or actual loss rates at the property support a higher figure. A property running 9% economic vacancy will most likely underwrite to a 9%. The floor only does work on properties performing better than it.

Physical vacancy and bad debt are two components of one number. HUD’s line item is the minimum vacancy and collection loss rate, and bad debt is the collection loss. A property running 4% physical vacancy with 1% bad debt is at 5% economic loss, which sits under the 7% floor, so the floor governs and the property underwrites at 7% for the DSCR loan amount.

A common clarification is that concessions are not in that number. HUD requires that concessions come off separately, on top of the underwritten vacancy. Take that same property and add 4% in concessions, and the deduction from Gross Residential Rent Potential is 7% plus 4%, not 7%. On $1,000,000 of Rent Potential the sponsor expecting their 5% economic loss to govern is looking at $950,000 of effective gross income. The underwritten figure is $890,000. At a 6.00% rate on 35-year amortization, that $60,000 gap is roughly $735,000 of loan proceeds.

How concessions get underwritten on top of the programmatic vacancy rate is a judgement to the underwriter and HUD. One of the most important aspects in underwriting concessions is whether and when they burn off. A lease-up incentive that ran for two quarters and has stopped is a different underwriting question from one month free offered on every renewal in a soft submarket. The first can support carrying a reduced concession figure forward or none at all; the second might get underwritten at what the property is actually giving away, because it doesn’t look like the concessions are going to stop. Often these two situations are analyzed via the burn-off schedule, and how much of it the underwriter is willing to credit depends on how well the pattern is documented.

Which is why a concession schedule is worth asking for even though it isn’t on the formal checklist — what was offered, to how many units, and over what period. A lot of lenders work to shorten the document request, and there is a case for that, but a shorter list is only better if what came off it wasn’t going to help you. The argument for a reduced concession figure has to be built from the leasing history rather than asserted, and the schedule is what builds it. Our 223(f) document checklist covers what HUD actually requires.

As stated earlier, HUD’s minimum vacancy only governs the DSCR NOI, which is the figure the underwriter builds to size the loan against the debt service coverage test. It does not govern the appraiser’s NOI — which determines the loan to value (LTV) test mortgage amount. Our papers on how HUD sizes a 223(f) mortgage and how HUD sizes a 221(d)(4) mortgage walk through the full set of sizing tests.

The appraiser establishes a vacancy and bad debt factor from the property’s own history, the rent comparables, and long-term occupancy expectations in the market. HUD’s direction to the appraiser is that the occupancy estimate be based on the property’s actual occupancy without regard to HUD’s DSCR vacancy floors, and that the appraised vacancy factor reflect long-term occupancy rates expected to continue. This means that a market-rate property can be appraised at 5% while the same property underwrites at 7% for the DSCR NOI.

That gap is critical when the LTV test is determining the loan amount. The appraiser’s vacancy feeds the Market NOI, the Market NOI divided by the market cap rate produces appraised value, and the LTV ceiling is a percentage of that value. If the deal is debt-service-constrained, a lower appraisal vacancy moves the appraisal but does nothing to the loan amount.

The appraiser is not required to use 7%, but is not prohibited from using it either. HUD permits the appraiser to use the programmatic minimums to develop market value where the appraiser considers it appropriate, and the latitude runs in both directions — the appraiser can conclude a factor below 7% where the property and the market support it, and can conclude one above 7% where they do not. What we see in practice on market-rate properties in healthy markets is 5%, but sometimes they could go lower.

When the appraiser does conclude above 7%, the DSCR underwriting usually moves with it. The floor is a minimum, not a fixed assumption, so a property whose appraised loss rate supports 9% is probably being underwritten at 9% on both sides. The separation between the two numbers only exists on properties performing better than the floor.

Commercial space is underwritten on its own occupancy limits, separate from the residential floors. On a 223(f), underwritten commercial physical occupancy is the lesser of 90%, the property’s actual commercial occupancy, or what the market indicates. On new construction the ceiling is the lesser of 80% or the market. The appraiser has to note the actual market-derived commercial vacancy in the report, and the appraised value of the commercial component has to assume the lesser of what the market shows or the program’s limit. A property with meaningful retail is running two vacancy analyses, and the commercial one is the more conservative of the two.

Both HUD and Fannie Mae impose a minimum loss deduction that applies whether or not the property has actually lost that much income. The two are built differently, and on a well-run market-rate property Fannie’s is the lower number — on the DSCR mortgage test. Everything in this section compares HUD’s DSCR NOI against Fannie’s single underwritten cash flow. The appraised value side of a HUD deal is not bound by the 7% at all, which is what the second of the two differences below turns on.

HUD’s is a flat floor by property type: 7% market rate, 5% for LIHTC properties with a 10% rent advantage, 3% for properties with HAP contracts on 90% or more of units. Fannie’s is a greater-of test. Physical vacancy, concessions, and bad debt together must equal the greater of 5% of gross potential rent, or the difference between gross potential rent and the property’s annualized trailing three-month net rental collections.

Vacancy is one input, and it is worth putting next to the others. HUD’s coverage requirement on a market-rate deal is 1.15 against a 35-year amortization, reset by Mortgagee Letter 2025-03. Fannie’s is set by Form 4660 rather than by program and tiers by market, product, and sponsor, but 1.25 against a 30-year amortization is normal on a conventional deal. Those two differences run the other way, and they run harder than the vacancy does. HUD’s 7% vacancy produces less net operating income than Fannie’s 5%, and the HUD loan still sizes larger on the coverage test, because 1.15 against a 35-year amortization more than offsets two points of vacancy. How much larger depends on where the two executions are priced and on the property’s expense load. Our paper on how a DSCR-constrained mortgage works walks through the mechanics.

Two structural differences follow from the way the floors themselves are built.

Concessions sit inside Fannie’s floor and outside HUD’s

Fannie’s minimum covers vacancy, concessions, and bad debt as one combined deduction, so a property giving away 4% in concessions has already consumed most of its 5% floor. HUD deducts concessions separately, on top of the 7%. The property that gave away 4% takes 7% plus 4%.

On paper that favors Fannie. In practice it rarely does, because the two conditions don’t normally co-occur. A property offering heavy concessions is usually offering them because they are experiencing high vacancy — a stabilized asset holding sub-5% vacancy and bad debt is not typically a property that needs to give away a month of free rent. And once concessions are running heavy enough to matter, the property’s trailing three-month collections have fallen behind its gross potential rent, which means Fannie’s floor is no longer 5%. The greater-of test has switched over to the collections gap, and that gap is capturing the concessions anyway. Fannie’s structural advantage on this line is mostly theoretical.

Fannie’s floor reaches the value conclusion; HUD’s does not

Fannie’s deduction sits inside the underwritten net cash flow, and that single cash flow drives both the coverage test and the value used for sizing. HUD’s 7% binds only the DSCR NOI. The appraiser builds the Market NOI on a market-supported factor, and on a 223(f) where LTV is the binding test, that is the number setting the loan. A property that appraises at 5% and underwrites at 7% for coverage is getting the benefit of the lower rate on the test that controls — an outcome Fannie’s structure does not produce, because there is only one cash flow.

Fannie Mae terms are typical of high leverage transactions. Some strong borrowers and/or markets may be different.

The affordable tiers converge more than the market-rate ones. HUD’s 3% for HAP properties is a flat rate based on the Section 8 contract. Fannie permits an affordable property to underwrite below 5%, to no less than 3%, but conditions it — the property has to sit in a qualifying market, restricted rents generally have to run at least 10% below comparable market rents, and the economic vacancy has to be supported by current data plus three years of history showing no less than the actual rate. Same floor, reached by demonstration rather than by category. HUD would also make us underwrite above 3% if the project wasn’t performing at that economic vacancy historically.

One more difference worth knowing on a property whose collections are slipping. Fannie’s greater-of test rises automatically when trailing collections fall behind gross potential rent, so a property with a collections problem sees its minimum deduction climb on its own. HUD’s floor does not move that way, however, HUD can question whether the current rent roll represents sustainable income if they see collections falling and can underwrite the Rent Potential lower, and the MAP Guide separately permits a vacancy rate above the minimum where the property’s history supports one.