The Section 8 Rent Ceiling Index: 43 Years of HUD Fair Market Rent Data Analyzed

Fair Market Rent is one of the most quoted numbers in American residential property and one of the least examined. It sets the payment standards for the Housing Choice Voucher program, rent ceilings for HOME and Emergency Solutions Grants, and flat rents in public housing. Roughly 2.3 million households depend on it.

It is also published every year without much commentary on how it behaves over time. So we took HUD's full historical file, 4,764 FMR areas from 1983 to 2026, and looked.

This is a data piece. There is no advice in it and no conclusion about whether anybody should buy anything. It is a record of what the numbers do, published because the analysis did not appear to exist anywhere else. Our method is at the bottom and the code is available, so anyone can check or extend it.

What Fair Market Rent is, briefly

Fair Market Rent is HUD's estimate of the 40th percentile gross rent for a standard-quality unit in a given area, defined in 24 CFR 888.113. Gross rent includes tenant-paid utilities other than telephone, cable and internet.

HUD calculates it from American Community Survey data on rents paid by recent movers, adjusted for inflation using Consumer Price Index components, then projected forward with a trend factor. Two-bedroom units are the primary calculation because they are the most common and most reliably surveyed; other bedroom sizes are derived using bedroom ratios. FMRs are published annually and take effect on 1 October.

One point matters for reading everything below: Fair Market Rent is not a rent anybody pays. It is an input. Housing authorities set payment standards between 90% and 110% of it, and a separate rent reasonableness test constrains actual contract rents independently. The figures here describe the movement of a statistical ceiling, not of anyone's income.

The long run: 1983 to 2026

Median two-bedroom Fair Market Rent, 1983 to 2026

Plotted on fiscal years. HUD's file has no 1984 column, so the first segment spans two years and 1984 is left genuinely missing rather than interpolated. The five marked years are 1994, 1995, 2012, 2014 and 2020. Author's calculation on HUD's published FMR history file, as at August 2026.

Over the full period the median two-bedroom FMR rose 274%, a compound annual rate of 3.12%. For context, that is broadly in line with long-run consumer price inflation, which is unsurprising given that CPI components are an explicit input to the calculation.

The striking feature is not the trend but the variance around it. Annual changes range from a fall of 8.0% to a rise of 11.4%, and the proportion of areas moving against the national direction is frequently large.

The years the ceiling fell

The national median fell in five years out of forty-two: 1994, 1995, 2012, 2014 and 2020.

More useful than the median, though, is the breadth of decline, because a national median can rise while a third of areas fall. Ranked by the share of areas where the two-bedroom FMR declined:

Share of FMR areas where the two-bedroom ceiling fell, by year

There is no value for 1983, since there is no prior year to compare against, and the 1985 bar covers the two-year interval from 1983 because HUD's file has no 1984 column. Author's calculation on HUD's published FMR history file, as at August 2026.

Two things stand out. The 1994 and 1995 episode was more severe and more widespread than 2012, which is the year usually cited when FMR declines come up at all. And 2005, 2015 to 2019 and 2025 all saw a fifth to a third of areas decline while the national median rose, which means area-level movement is frequently invisible in the headline figure.

The 10% floor, and what happened before it

HUD's Housing Choice Voucher Guidebook states the rule plainly: "An FMR will never decrease by more than 10 percent from the previous year's FMR."

The data shows that limit binding from FY2019 onward. In every year from FY2019 to FY2026, the worst single-area decline is exactly 10.0% and no area falls further. Before FY2019 the picture is very different:

In 1994, close to a third of all FMR areas in the country saw the two-bedroom ceiling cut by more than 10%, and 553 areas were cut by more than a fifth. The largest single-area decline in the record is 48.9%, in 2012.

There is a second provision worth noting alongside the cap. HUD's guidebook requires a housing authority to conduct a fresh rent reasonableness determination when an FMR falls by ten percent or more. So an area reaching the floor does not simply see its ceiling reduced; it also triggers a mandatory review of existing contract rents.

FY2026 in detail

FY2026 is a strong year by historical standards. The median two-bedroom FMR rose 11.4%, which alongside 2024 at +11.0% and 2023 at +9.8% makes the last four years the strongest sustained period in the entire 43-year record.

Any claim that Section 8 rent ceilings are currently collapsing is not supported by the data.

The largest three-bedroom increases were Gates County, NC at +43.0%, Kalawao County, HI at +38.8% and Amherst Town-Northampton, MA at +37.4%. Seven metropolitan areas reached the 10.0% floor: Vallejo and Salinas in California, Lincoln County in North Carolina, Boise City and Gem County in Idaho, Waco in Texas, and New Orleans-Metairie in Louisiana.

Regional and state divergence

The national figure conceals a very wide regional spread.

More than a third of Southern areas saw the two-bedroom ceiling fall in a year when the Northeast rose nearly 12% with almost no declines.

At state level the concentration is sharper still. States where the largest share of areas saw the three-bedroom ceiling fall:

Mississippi is the clear outlier. Nine out of ten of its FMR areas saw the three-bedroom ceiling fall, at a median decline of 9.82%, which is close to the maximum HUD permits.

At the other end, Maine, Rhode Island and Vermont recorded no declining areas at all, with Maine's median three-bedroom FMR rising 23.64%.

Unit size matters

Declines are consistently more common on larger units, and the gradient is monotonic: 17.0% of areas fell on efficiencies, 14.9% on one-beds, 15.5% on two-beds, 18.7% on three-beds and 20.3% on four-beds.

We do not have a confident explanation. Larger-unit FMRs are derived from two-bedroom figures using bedroom ratios estimated from American Community Survey data, and those ratios are updated annually, so it is plausible that ratio revisions introduce additional variance at the larger sizes. That is a hypothesis, not a finding.

Which areas are most and least stable

Taking the 2,615 areas with at least 20 years of continuous three-bedroom data, we counted how many years each area's ceiling declined.

The median area saw a decline in 5 of the last 25 years. Only 4 areas out of 2,615 never recorded a decline. Ten areas declined in 10 or more years.

The most stable areas with a FY2026 three-bedroom FMR above $1,000 are overwhelmingly rural counties in Iowa and Illinois, several with zero or one decline in 25 years. The least stable include Denali Borough and North Slope Borough in Alaska, Rich County in Utah, and Eastern Worcester County in Massachusetts, each with 11 declines.

For reference, some larger metropolitan areas:

Note that Kansas City combines a low decline count with high volatility and the strongest growth of the group, which is a reminder that the two measures capture different things.

What we looked for and did not find

Negative results are worth reporting.

We tested whether the three-bedroom premium over two-bedroom is being compressed over time, which would systematically disadvantage larger units independently of market conditions. It is not. The national median three-bed to two-bed ratio was 1.2920 in FY2001 and 1.2935 in FY2026, a change of +0.12%. The four-bed to two-bed ratio moved from 1.4737 to 1.4537, a decline of 1.36%, which is small and not clearly directional.

We also checked whether unusually large annual increases are systematically followed by declines. The record is not clean enough to support that. The 2013 rise of +7.4% was followed by a fall, and 2019's +4.7% was followed by a fall, but the 1994 and 2012 declines were both preceded by modest rather than large increases. We would not draw a cyclical conclusion from 42 observations with this much noise.

Limitations

The dataset has gaps and quirks that anyone reusing this analysis should know about.

There is no 1984 column in HUD's file, so the 1983 to 1985 comparison spans two years. FMR area definitions change over time as the Office of Management and Budget revises metropolitan statistical area boundaries, which means an area compared across 25 years is not always geographically identical. HUD revised the FY2026 figures in April 2026, and we have used the revised file. FY2026 values arrive in HUD's file as text with currency formatting while earlier years are numeric, which is an easy source of error.

Most importantly, this analysis describes movement in a published statistic, not movement in anybody's rent or income. Payment standards are set locally within a range, actual contract rents are constrained separately by rent reasonableness, and changes reach existing tenancies only at recertification. Nothing here should be read as a measure of landlord receipts or tenant costs.

Method and reproducibility

Source data is HUD's complete FMR history file, FMR_All_1983_2026.csv, published on HUD's Fair Market Rents dataset page and downloadable without registration. It contains 4,764 FMR areas and 310 columns covering all bedroom sizes from 1983 to 2026.

All figures above are calculated on the full population of areas rather than a sample. Where we report medians they are across areas, unweighted by population or housing stock, which means a rural county counts the same as a large metropolitan area. That is a deliberate choice for measuring the behaviour of the FMR system itself, but it means these figures are not estimates of what a typical voucher household experiences.

Area-level comparisons are matched on HUD's area name and state FIPS code. We would flag one trap for anyone repeating this: matching areas on name alone produces errors, because several distinct FMR areas share a city name. "Cleveland" matches areas in Arkansas, North Carolina, Ohio and Tennessee.

We intend to update this analysis each autumn when HUD publishes the following fiscal year. If you want the code or have found an error in ours, we would rather hear about it than not.

Readers looking for what any of this means for a rental property investment rather than for the statistic itself will find that discussed separately in is Section 8 a good investment. We have deliberately kept the two apart.