The Waiting List Is the Policy

The gap

Need vs supply

Every household in this bar is a renter earning under 30% of what a typical family in their area earns. A third of them have a subsidy. The other two thirds are not on a shorter list or a slower track - there is simply no unit for them.

Coverage: both figures are complete counts, not samples - the HUD occupancy and income shares are 100% reported in the all-programs row, and CHAS covers all 3,221 counties. The two files have different vintages (HUD 2025, CHAS 2017-2021 ACS) and slightly different income tests (HUD's share of local median income vs CHAS's HAMFI), so read 33.3% as close to right, not to the decimal.

How 5.1 million units become 3.5 million poor households housed

Units under contract 5,132,918 what HUD pays for
Occupied 4,471,285 12.9% sit empty
Of those, extremely low income 3,532,315 79% of tenants
Households who qualify 10,620,313 CHAS 2017-2021

The wait, where it is measured

Months, by state

HUD records how long each household waited before it moved in. It records nothing at all about the households still waiting - they enter the file on the day they are housed, and not before. So this is the wait experienced by those admitted, and the true wait is longer than anything shown here. Even so, it varies by a factor of 5.7: Maryland at 57 months, Wyoming at 10.

Coverage: 55 of 55 states and territories publish a wait figure (100%). Nationally the field is far thinner: it is suppressed or missing in 41.1% of all county rows in the file. HUD marks those cells -1 (missing), -4 (suppressed, fewer than 11 families) or -5 (non-reporting). They are excluded here, never averaged in - a naive mean of the raw column comes out negative.

Table: every state, in full

Months waited before move-in among admitted households, units under contract, and share of extremely-low-income need met, by state.
State	Wait (mo, those admitted)	Units	Occupied	ELI need met
Maryland	57	102,353	91,459	38.5%
New Hampshire	43	22,563	20,225	43.2%
Delaware	40	13,186	11,060	33.4%
Northern Mariana Islands	40	492	465	no data
New Jersey	39	169,445	153,600	36.6%
New York	39	588,668	516,539	40.3%
Connecticut	35	84,206	75,399	39.8%
Florida	33	205,176	180,144	25.6%
Oregon	33	57,559	50,217	31.9%
Maine	32	27,425	23,207	43.9%
Massachusetts	31	195,871	182,701	47.6%
California	28	507,349	449,875	27.2%
Hawaii	28	23,624	19,295	39.9%
Michigan	28	146,787	128,123	34%
Alaska	27	7,935	6,870	28.8%
Missouri	27	93,924	75,826	32.3%
Rhode Island	27	38,494	35,246	56.2%
Arizona	26	41,985	37,420	17.8%
Nevada	26	24,578	21,856	19.1%
Guam	25	3,474	3,314	no data
U.S. Virgin Islands	25	6,191	4,489	no data
Utah	25	20,331	17,485	25.8%
Virginia	25	107,467	92,877	31.3%
Louisiana	24	94,457	81,263	37.9%
Ohio	24	226,712	198,038	39.8%
Mississippi	23	56,893	49,336	41.3%
Minnesota	22	92,586	81,765	40.3%
Iowa	21	41,324	34,504	30%
North Carolina	21	128,527	104,659	26.7%
Tennessee	21	106,452	91,928	36.1%
Wisconsin	21	78,710	68,005	30%
Georgia	20	135,704	113,062	29.1%
Idaho	20	12,942	11,165	25.2%
Montana	20	14,113	11,432	28.8%
Pennsylvania	20	224,531	193,029	35.7%
Indiana	19	89,420	76,942	31.1%
New Mexico	19	27,103	21,143	27.3%
Texas	19	288,627	248,565	24.5%
Alabama	18	92,548	80,491	39%
Illinois	18	231,185	202,644	36.6%
Colorado	17	63,715	56,780	30.2%
South Carolina	17	63,728	55,836	30.2%
Vermont	16	13,563	11,196	46.1%
Arkansas	15	52,339	42,818	35.7%
Kansas	15	35,419	28,447	26.6%
Nebraska	15	28,297	23,272	31.8%
Washington	15	96,312	86,484	33.9%
District of Columbia	14	34,553	27,697	45.9%
Oklahoma	13	54,103	45,878	32.3%
South Dakota	13	14,057	11,083	36.4%
Kentucky	12	83,998	73,448	39.3%
North Dakota	12	14,311	10,528	31.2%
Puerto Rico	12	107,040	96,144	51.2%
West Virginia	10	34,674	30,734	42.1%
Wyoming	10	5,809	5,205	25.4%

Guam, Northern Mariana Islands, Puerto Rico, U.S. Virgin Islands appear in the table but not on the map: the Albers-USA projection has no geometry for them. CHAS publishes no county file for them either, so their ELI need is shown as no data.

Who gets in

Share of households, by program

Nationally, 43% of subsidized households are headed by someone 62 or older and 24% include a disabled member. But the average hides the design: two of these programs admit almost no one else. 202/PRAC is elderly housing by statute and 811/PRAC is disability housing by statute. Vouchers are the only door families with children go through in any number.

Elderly (62+)
Disabled
With children

The three shares are independent, not parts of a whole: a household can be both elderly and disabled, so they are grouped rather than stacked and do not sum to 100%. "With children" is HUD's no-children share, inverted. Composition is published for every program shown here, with no suppression. Not charted: S236/BMIR and 811 Project Rental Assistance (PRA) - legacy programs now holding 0 occupied units between them, where a share of almost nothing would read as loudly as a share of millions.

Household composition by HUD program.
Program	Units	Households	Elderly 62+	Disabled	With children	Extremely low income
Housing Choice Vouchers	2,797,793	2,334,114	35%	26%	36%	79%
Project Based Section 8	1,321,244	1,208,945	54%	19%	26%	82%
Public Housing	848,351	772,585	38%	23%	34%	74%
202/PRAC	120,345	116,132	100%	8%	0%	79%
811/PRAC	33,372	30,501	34%	91%	2%	87%
Mod Rehab	8,723	9,009	36%	34%	16%	90%
S236/BMIR	3,090	0	52%	8%	26%	37%
811 Project Rental Assistance (PRA)	0	0	10%	77%	12%	97%

Who pays for it

Per unit-month, 2025

A subsidized household pays $452 a month and HUD pays $1,135. The tenant covers 28.5% of the cost of their own home - not because the rent is cheap, but because the statute caps their share at roughly 30% of an adjusted income that averages $18,483 a year, or 21% of what a typical family in the same area earns.

All HUD programs $452 $1,135 tenant 28%
Housing Choice Vouchers $493 $1,255 tenant 28%
Project Based Section 8 $389 $1,150 tenant 25%
Public Housing $439 $856 tenant 34%
202/PRAC $401 $571 tenant 41%
811/PRAC $354 $588 tenant 38%
Mod Rehab $298 $812 tenant 27%

Tenant pays
HUD pays

Multiplied out - $1,135 a month across 4,471,285 occupied units - HUD's side of this split runs to roughly $60.9 billion a year. That is an implied figure, not an appropriation: it is the per-unit subsidy in the file times twelve times the occupied units, and it excludes administrative cost, capital funds, and the two programs below. Read it as the order of magnitude of the transfer, not as a budget line. Not charted: S236/BMIR and 811 Project Rental Assistance (PRA) - HUD suppresses the subsidy side of their split, and half a split is not a split.

Seventeen years of standing still

2009-2025

The number of subsidized homes has barely moved since 2009: 4,970,225 then, 5,132,918 now, a change of 3.3% across seventeen years in which the country added more than twenty million people. The number of people living in them fell 7.1%, because the households moving in are smaller and older than the ones they replaced. Meanwhile the subsidy HUD pays per unit rose 84%. The program did not grow. It got more expensive to hold in place.

Units under contract, and people living in them

Units under contract
People served

What a unit costs each month, and who pays it

Tenant pays
HUD pays

Dollars are nominal - not inflation-adjusted. The Picture file publishes no deflator and we do not import one, so the rise in both lines is part real and part inflation. The ratio between them is the honest read: the tenant's share of the cost of their own home has fallen from 32% to 28%, because their incomes did not keep pace with the rent HUD must cover. Occupancy is not charted: HUD did not publish an occupied-units column before 2025, and a back-filled number is not a measured one.

Table: every year, 2009-2025

HUD subsidized housing by year, 2009 to 2025.
Year	Units	People	Tenant pays	HUD pays	Wait (mo, those admitted)	Geography
2009	4,970,225	9,536,969	$286	$617	18	pre-2020 census
2010	5,095,126	9,859,194	$288	$631	18	pre-2020 census
2011	5,147,935	10,099,094	$289	$631	20	pre-2020 census
2012	5,168,778	10,027,278	$297	$640	20	pre-2020 census
2013	5,180,467	10,072,890	$304	$637	18	pre-2020 census
2014	5,031,773	9,834,571	$321	$666	26	pre-2020 census
2015	5,038,578	9,853,342	$328	$680	26	pre-2020 census
2016	5,015,895	9,785,085	$332	$697	26	pre-2020 census
2017	5,018,939	9,653,388	$337	$693	27	pre-2020 census
2018	5,035,824	9,535,360	$346	$743	26	pre-2020 census
2019	5,034,685	9,439,919	$357	$765	26	pre-2020 census
2020	5,076,615	9,338,358	$355	$810	27	pre-2020 census
2021	5,098,041	9,170,091	$364	$839	27	pre-2020 census
2022	5,124,368	9,027,284	$386	$899	25	2020 census
2023	5,129,585	9,045,714	$416	$989	25	2020 census
2024	5,149,303	9,039,779	$433	$1067	27	2020 census
2025	5,132,918	8,860,364	$452	$1135	25	2020 census

Where you live

3,126 counties

The national picture is an average of 3,126 very different local ones. Pick a state to see every county in it: how many subsidized homes there are, how long the people who got one waited, and what share of the local need that covers.

Longest waits for those admitted

Counties with at least 500 units, so a single admission cannot top the list.

Least of the local need met

Share of extremely-low-income renter households with a subsidy. Same 500-unit floor.

Wait figures exist for 2605 of the 3126 counties HUD reports (83.3%); the rest are suppressed (fewer than 11 families), missing, or non-reporting, and are shown as "no data" rather than as a zero. A further 14,327 units across 56 rows could not be placed in any county by HUD and appear nowhere on the map or in these lists.

Browse by state

Every figure on this page is computed from two real federal bulk files. Nothing here is estimated, modelled, or stood in for. The dashboard is badged Full for that reason, and the code that derives each number is in the repository beside it.

HUD Picture of Subsidized Households, snapshot dated 31 December 2025, plus the annual files back to 2009 for the trend. This is HUD's own administrative census of every unit it has under contract: 5.13 million of them, reported by the housing authorities and owners who run them.
HUD CHAS, built on the 2017-2021 American Community Survey. This supplies the one thing the Picture file does not contain: a denominator. The Picture file can tell you who got in. It cannot tell you how many wanted to.

The wait is not the waiting list

This is the most important caveat on the page. HUD's months_waiting measures the time a household waited before it moved in. A household still on a list has no row in this file - it enters HUD's records on the day it is housed, and not one day earlier. The figure is therefore survivor-biased: it describes the people who made it through, and it almost certainly understates how long the queue really is. Every label on this site says "the wait experienced by those admitted." None of them say "the waiting list," because this data cannot support that sentence. Many authorities have closed their lists entirely; those households are not late in the queue, they are not in it.

Suppressed and missing cells

HUD does not leave a cell blank when it has no value - it writes a negative code. Per the 2025 data dictionary: -1 missing, -4 suppressed (fewer than 11 reported families, for privacy), -5 non-reporting (under half the families reported). These are not edge cases. In the county file, 41.1% of all rows carry one of them for the wait field: 5,936 are missing, 3,100 are suppressed (fewer than 11 reported families), 531 are non-reporting (under 50% reporting rate).

Averaged naively, those codes pull the result negative. Every number here filters them out before any arithmetic, and every chart that touches a suppressed field prints the share of rows that survived. For the county wait that is 83.3% (2,605 of 3,126 counties); for the states, 100%; for the eight programs, only 37.5% - HUD publishes a wait for vouchers, public housing and Mod Rehab, and for nothing else. Suppressed places are drawn in a neutral gray that is not a step on the color ramp, and labelled "no data" in every table, so an absence can never be misread as a short wait.

How the 33% is calculated

CHAS counts 10,620,313 renter households at or under 30% of area median income - the eligible pool. HUD reports 4,471,285 occupied subsidized units, of which 79% are that poor, giving 3,532,315 extremely-low-income households actually housed. 3,532,315 / 10,620,313 = 33.3%. The remaining 7,087,998 receive nothing.

We compare poor-and-housed against poor-and-eligible, rather than all 4,471,285 assisted households against a poverty-only denominator. The looser version credits the 21% of tenants who earn above the line toward need they do not represent, and flatters the program by several million households. Two honest caveats remain: the files have different vintages (2025 against a 2017-2021 survey), and their income tests are near-neighbours rather than twins - HUD's is a share of local median income, CHAS's is a share of HUD Area Median Family Income. Read 33.3% as close to right, not exact to the decimal.

Units are not households

5,132,918 units are under contract; 4,471,285 are occupied. The 12.9% difference is real vacancy - units being turned over, repaired, or held off-line. This page says "units" when it means units and "households" when it means households, and never trades one for the other to make a number look better.

Counting each program exactly once

The Picture file lists sub-programs (tenant-based and project-based vouchers, Moving-to-Work and not) as rows inside each program's total, not beside it. Summing the file naively double-counts more than three million voucher units. Every figure here reads only the totals, and the build asserts it: the eight program totals sum to 5,132,918 units against the all-programs row's 5,132,918 - an exact match, checked on every run, and the build fails if it ever stops matching. On occupied units the two differ by 1, which is HUD's own rounding, not ours.

What you are not seeing

The households still waiting - the entire subject of the page - are invisible in this data by construction, which is why the denominator has to be borrowed from a survey. Nor does this file cover the Low-Income Housing Tax Credit, the largest affordable-housing program in the country, which is Treasury's and not HUD's. 14,327 units across 56 rows could not be placed in a county by HUD and are excluded from the map and the county pages, though they remain in the national total. Dollar figures are nominal, never inflation-adjusted. And the annual series has a seam: 2022 is published under both the old and the 2020-census geographies, so we take 2009-2021 from the former and 2022-2025 from the latter, counting each year once.

Generated 2026-07-13 02:52 UTC

Source: HUD Picture of Subsidized Households (2025 snapshot, 2009-2025 trend) joined to HUD CHAS 2017-2021 ACS for the eligible-household denominator

↑ Top kvlak / limn - all dashboards →