Between the Ceiling and the Floor

Definitions

Area median income. The figure every level is a percentage of, published by HUD each spring for every metropolitan area and every rural county. It is median family income, not median household income: the Census definition excludes people living alone and unrelated people sharing a home, both lower-income groups, so the figure sits above what a typical individual earns. That is the plain-language answer to why these percentages feel so high, and it falls hardest on one-person older households.

Affordability levels: 30, 50, 60 and 80 percent. A building is funded at a level, and that level sets two things at once: the most a household living there may earn, and the most the home may charge in rent. A lower level means cheaper rent and a lower income limit. Most new homes are built at the 60 percent level. Sixty percent is a ceiling, not a target. In tax credit housing the rent is set by the home, not by the tenant, so a household at 30 percent of area median income is fully eligible for a 60 percent building but pays the 60 percent rent. There is no 30-percent-of-your-income calculation unless a subsidy is attached to the home itself.

Rent-restricted. A home where the rent is fixed by a published table rather than by what the resident earns. The floor applies here. Where rent moves with income, through project-based rental assistance or a voucher, there is usually no minimum to prove. Writing "affordable housing" rather than "rent-restricted" is the mistake a housing partner will catch first.

The ceiling. The most a household may earn and still be eligible: the published income limit for that household size, in that area, at that level. Earn more and the household is not eligible. It does not change with the size of the home. Being over a ceiling does not mean a household is comfortable: at the 30 percent level the ceiling is far below what it costs to live on the open market.

The window. The band in between, where a household can apply: enough income to satisfy the property, and not so much that the household is ineligible. Clearing the floor is a ratio, not a verdict on affordability. A household can pass it and still be stretched once medicine, food and everything else is paid. Ceiling, floor and window are our words; HUD publishes an income limit, the IRS sets a maximum rent, and the floor has no official name.

The floor. The least a household must prove it earns. Most properties ask applicants to show income of two to three times the monthly rent. No national source records which multiple prevails. Each property sets its own, so it varies between buildings on the same street. Only two jurisdictions we could verify cap it: Portland at 2.5 times, and 2 times above a rent threshold, under Portland City Code 30.01.086; and Colorado at 2 times statewide, under Colorado Revised Statutes 38-12-904. Minneapolis requires landlords asking three times or more to consider alternative evidence. Most of the country has no cap, which means the floor is higher elsewhere, not lower.

Maximum rent. Not published separately. It is calculated under Internal Revenue Code section 42(g)(2) as 30 percent of the income limit for an assumed household size, over twelve months. The assumed size is one person for a studio and 1.5 per bedroom after, so a one bedroom uses the average of the one and two person limits and a two bedroom uses the three person limit. That averaging convention is universal in state agency rent tables. The figure is a gross rent: utilities sit inside the cap, and a utility allowance is subtracted where the resident pays their own.

Income averaging. Since 2018, a building may designate individual homes anywhere from 20 to 80 percent in ten-point steps, as long as the average is at or below 60 percent. So "is this building income averaging, and are any homes below 60 percent?" is a more useful question than "do I qualify?"

A Housing Choice Voucher. Your share is roughly 30 percent of your adjusted income, so there is no fixed rent to test against and the floor stops applying. It also works in ordinary market-rate housing, and tax credit properties must accept it. Washington State, Seattle, California and Portland require the voucher to be subtracted from the rent before any multiple is applied.

Household size. Everyone who will live in the home counts, including children. Unrelated people count too: 24 CFR 5.403 defines a family as "a group of persons residing together," with no relationship test. Census uses "family" to build the area median income figure; HUD uses it to mean whoever will live in the home. Area median income is a ruler, not a rulebook.

The room benchmark. Half the two-bedroom Fair Market Rent for the area, the pro-ration HUD uses for shared housing. It is a suggested starting rate, not a published one. In a shared home the home provider sets the rent, so treat this figure as the opening of a conversation.

Methodology

Income limits come from HUD's Multifamily Tax Subsidy Project tables, FY2026, effective May 1, 2026. That series governs tax credit properties, which is most of the housing this page describes. It is a different series from the Section 8 income limits: Section 8 limits can fall year over year where these cannot, and the 60 percent band does not exist in the Section 8 file at all. All four levels here come from the same series, so no two bars on the chart sit on a different basis.

Coverage. The 50 states and the District of Columbia: 2,623 HUD income limit areas, covering all 4,680 counties. Every county maps to exactly one area. Where a county sits inside a metropolitan area its limits are set market-wide, and the page says so, because that is itself the lesson: a reader in a metropolitan county learns their limits are not local, and a reader in a rural county learns theirs are their own. HUD also publishes limits for Puerto Rico, Guam, the U.S. Virgin Islands, American Samoa and the Northern Mariana Islands, and this page covered them until August 2026. They are out now because the rest of our research cannot follow: the state benefit averages, availability figures and spare-bedroom counts the companion pieces rely on are not published for the territories, so a reader who found their county here would find nothing anywhere else on the site. The brief you can download from this page is a national analysis rather than a lookup, and it still counts all 2,644 areas, territories included, which is why its coverage line reads differently from this one.

The income test multiple is a control, not a datum. The property sets it, and it varies between buildings on the same street, so publishing a local figure would mean publishing a number we could not stand behind. The reader sets it instead, and the real next step is to ask the property what multiple it uses.

Vintage. Eligibility figures print current and carry their effective date, because they state who is eligible today. This is a deliberate exception to the rule that governs the rest of this research family, where the Elder Index sets the vintage and HUD follows it. Within this page every figure sits in the same year, including the room benchmark, so the chart never compares two vintages. HUD publishes new limits each spring, normally April or May.

Nothing is estimated. Where a figure could not be retrieved, the area does not ship. The build fails loudly on a missing value rather than filling a gap.

Sources

This product uses HUD published data but is not endorsed or certified by HUD.