Affordable Housing Development Intelligence

A tax-credit deal lives or dies on one comparison: the rent the income limits let you charge, against the rent the market already charges. This report runs that comparison for 1,551 counties on FY2026 HUD income limits, adds the income-qualified demand behind it from Census microdata, and ends with a one-page schedule for the 100 markets where the numbers work best.

20.9M

Renter households at or below 60% AMI

47.7% of all U.S. renters

526

Counties where a 60% unit beats market rent

of 1,551 studied

−$80

Median county rent advantage, 2BR

FMR minus the maximum credit rent

50.9%

Renters paying 30%+ of income on rent

25.7% pay more than half

Seven findings organize what follows. First, and most consequential: in 1,025 of 1,551 counties the open market already clears at or below the 60% AMI ceiling, so the credit buys no rent advantage there at all. The median county sits −$80 a month on that comparison. Second, the advantage concentrates precisely where land and construction cost most: the coastal and large-metro counties where a 60% unit undercuts the market by hundreds of dollars a month. Third, the qualified pool is enormous and poorly served: 20.9M renter households sit at or below 60% AMI, and 10.5M sit at or below 30%, the band almost no unsubsidized credit deal can reach.

Fourth, 50.9% of American renters are cost-burdened and 25.7% severely so, which is the demand case in one number. Fifth, supply is not following that need: the supply-gap section names counties with tens of thousands of renter households and effectively no multifamily permitting since 2022. Sixth, rent trajectory matters as much as rent level, because the ceiling moves with the area's income limits, and the counties where Fair Market Rents are climbing fastest are where a credit rent underwritten today ages best. Seventh, the markets that score highest on all six pillars are rarely the ones with the highest rents; they are mid-size metros with deep renter populations, real burden, and almost no new supply.

Every figure in this report descends from one calculation, so it is worth stating in full and in public. HUD publishes an area median income and, from it, income limits for household sizes one through eight. A tax-credit unit's maximum rent is 30% of the income limit for an imputed household size of 1.5 persons per bedroom, divided by twelve. The 60% band that most 9% and 4% deals underwrite to is defined as 120% of HUD's published 50% very-low-income limit, because HUD publishes no 60% table.

Worked through San Benito County, CA, which carries the San Benito County, CA HUD Metro FMR Area schedule and posts the largest rent advantage in the country. Its FY2026 area median income is $121,800. The 50% limit for a three-person household is $66,150, so the 60% limit is $79,380. A two-bedroom unit imputes to three persons, so its maximum gross rent is 30% of $79,380 divided by twelve, or $1,985 a month. The FY2026 Fair Market Rent for the same unit is $2,902, so a credit tenant there saves +$917 a month against the market. That difference, computed for every county and every bedroom count, is the report.

Note. Gross versus contract rent is the most common misreading of a table like this. The maximum rents above include an allowance for tenant-paid utilities, which each public housing authority sets locally and no federal file publishes; the rent a lease can actually charge is that figure minus the allowance, often $60 to $180 a month. Every rent in this report, credit and market alike, is gross, so the comparisons hold even though the levels are above what a lease shows.

Every figure is computed by a deterministic pipeline from published federal files, and the prose interpolates the same values the tables show; no number is typed by hand. The arithmetic, the universes and every known limitation:

1. Income limits: HUD Income Limits FY2026 (growth vs FY2025), pulled from the HUD USER API. Year-over-year change compares the FY2025 and FY2026releases for the same county, which is how the rent ceiling actually moves; HUD applies caps and floors to that change, so it does not track local income growth one-for-one. HUD publishes area median income and the 30% (extremely low), 50% (very low) and 80% (low) limits for household sizes 1–8. Counties inherit their HUD area’s schedule, so several counties in one metro carry identical numbers by design; each profile names the area it belongs to.
2. The 60% bandis 120% of the published 50% limit. HUD publishes no 60% table; 1.2 × VLI is the statutory construction the sector underwrites to. Where a state or allocating agency uses income averaging, the applicable band differs by unit and this report’s 60% column is the reference point, not the rule.
3. Maximum rentsare 30% of the income limit for an imputed household size of 1.5 persons per bedroom (studios = 1 person), divided by twelve. Fractional imputed sizes interpolate linearly between the published limits, which is HUD’s own method.
4. All rents here are GROSS. They include tenant-paid utilities. The utility allowance is set locally by each public housing authority and appears in no federal file, so contract rent is lower, commonly by $60–180 a month. Market rents are gross too, so the comparisons hold.
5. The market comparator is HUD Fair Market Rents FY2026 (growth vs FY2023). Fair Market Rent is HUD’s 40th-percentile standard-quality rent for the area, which makes it a conservative proxy for asking rents in a strong submarket and a generous one in a weak submarket. The ACS median gross rent (ACS 2020–2024 five-year (renters, rents, burden, incomes)) is shown alongside where relevant and reflects rents already being paid, including long tenancies, so it runs below asking.
6. Income-qualified renter households (county and tract): HUD’s CHAS special tabulation, 2016-2020 ACS, on HUD Area Median Family Income (HAMFI) computed from 2016-2020 ACS income. CHAS is the custom Census cross-tabulation HUD commissions to count households its programmes serve, and it is the only published source giving households by HUD income band at county and tract grain. HUD publishes no newer release and no single-year version (small-area estimates need the five-year sample), so its bands are qualification against 2016-2020 thresholds, NOT against the FY2026limits used for the rents in this report. Its inputs here are vendored files downloaded from HUD’s open-data portal rather than an API pull.
7. The current-income basis: ACS PUMS 2024 one-year (income-qualified renter households), the most recent one-year income data published. Every renter household is compared against the limit for its own household size in its own state, with exact integer weights. Comparing that measurement to CHAS gives the drift factors reported in section 5 (nationally 0.96 at 30% and 0.93at 50%), which are applied to each county’s CHAS structure and today’s renter household count to produce the “current basis” columns. Those columns are SHIFT-SHARE ESTIMATES and are labelled as such wherever they appear; the CHAS column beside them is always the published count. The 60% band is computed from PUMS only, because CHAS does not publish it.
8. The affordability gap: CHAS “affordable and available” unit counts, divided by the households at or below the same band. A unit counts as available only if it is vacant or occupied by a household within the band, which is what distinguishes this from a simple count of cheap units. Units are matched by rent, not by subsidy status.
9. Tract concentration: the same CHAS tabulation at census-tract grain (2020 boundaries), summarised per county as the share of extremely-low-income renter households in its densest decile of tracts. This does not identify Qualified Census Tracts, which HUD designates annually on separate criteria not carried by any file here.
10. Renter households, burden, rents and incomes at county grain: ACS 2020–2024 five-year (renters, rents, burden, incomes). Renter households are occupied units minus owner-occupied. Median rent burden is the ACS median of gross rent as a share of household income. Five-year estimates trade currency for precision and carry sampling error.
11. Supply: Census Building Permits Survey, 2022–2025. Multifamily = units in structures of 5+ units. Permits are authorizations, not completions. The county permit series is complete from 2022, so every cumulative supply figure uses the 2022–2025 window.
12. Demand context: population and migration from Census PEP Vintage 2025; average wages from BLS QCEW 2025 Q4; sale prices from Redfin county medians, June 2026, county grain only.
13. The universe: counties with published HUD county income limits, 2025 population ≥ 25,000, at least 1,000 renter households, an FY2026 FMR, ACS rent and income, and a complete 2022–2025 permit series. That is 1,551 counties. The 100 profiled markets are the 100 highest-scoring counties with at least 10,000 renter households.
14. New England: CT, MA, ME, NH, RI and VT publish HUD income limits by town, not county, so their counties are absent from every county table here; state-grain sections cover all 51. Connecticut is excluded for a second reason as well: CHAS is published for its legacy counties, while the population and permit files use the planning regions that replaced them, and joining across that redefinition would fabricate markets.
15. The Feasibility Index: seven pillars (rent advantage ↑, renter share ↑, median rent burden ↑, multifamily permits per 1,000 renters ↓, affordable-and-available units per 100 ELI renters ↓, population growth ↑, FMR growth ↑), each z-scored across the qualifying counties, winsorized ±2.5σ, equally weighted, rescaled to 50±10. The two supply pillars are deliberately separate: one measures the flow of new construction, the other the standing affordable stock, and a market can be short on either alone. Equal weighting is an editorial choice; every pillar score is published so a reader who weights burden over advantage can re-rank the table.
16. What this report is not. It is a market-selection instrument, not a market study and not an underwriting model. It reads no state Qualified Allocation Plan, knows nothing about basis boosts, Difficult Development Areas or Qualified Census Tracts, carries no utility allowances, and does not price land, construction or credits. A credit application still needs the commissioned study; this report is for deciding which markets are worth commissioning one in.

Note. Ranks run 1 = best in the stated direction. A dash is a figure the source does not publish for that geography and is excluded from that ranking; it is never a zero. Assembled August 2026 against FY2026 limits, which HUD revises annually; each figure refers to its stated vintage, not the assembly date.