Since 1974 the U.S. Department of Housing and Urban Development (HUD) has helped low-income households obtain better rental housing and reduce the share of their income that goes toward rent through a program that relies on the private rental market. As of 1997, 1.4 million households held Section 8 certificates or vouchers, which allow them to rent eligible units in the private market and receive rental subsidies from the Federal Government. A key parameter in operating the certificate and voucher programs is the Fair Market Rent (FMR). This article explains what FMRs are, how they function in these programs, and how HUD calculates them.

The Role of FMRs

FMRs play different roles in the certificate and voucher programs. In both programs, FMRs set limits. In the certificate program, FMRs set limits on what units can be rented; in the voucher program, FMRs set limits on the subsidy provided to the household. Certificate program households cannot rent units with gross rents exceeding the FMR; the recipients receive a subsidy equal to the difference between the gross rent and 30 percent of their incomes.¹ Voucher program households receive a subsidy equal to the difference between the FMR and 30 percent of their monthly incomes. Participants in the voucher program can choose units to live in with gross rents higher than the FMR, but they must pay the full cost of the difference between the gross rent and the FMR, plus 30 percent of their income.

FMRs function primarily to control costs. Research has shown that program recipients act rationally and choose units with gross rents close to the FMRs, that is, the best units available under program rules.² On average, higher FMRs would mean higher costs per household served and, within a fixed budget, fewer households would be assisted. FMRs also help ensure reasonableness in the Section 8 assisted housing program. The taxpaying public would probably be unwilling to support an assisted housing program that put low-income families in housing units that were substantially better than typical rental units.

While budget realities and sensitivities to public acceptance exert pressures to set FMRs at low levels, other concerns create countervailing incentives to raise FMRs. For the Section 8 program to work properly, certificate and voucher holders must have an adequate supply of decent, safe, and sanitary rental units to choose from. Higher quality units command higher rents, so FMRs must be sufficiently high to provide acceptable choices for participants. In addition, the certificate and voucher programs were designed to allow assisted households to choose among different neighborhoods. The FMRs must also be high enough to provide acceptable choices among neighborhoods.

The FMR Standard

Since Congress established the Section 8 program in 1974, there have been three definitions of FMRs. The current definition, which became effective in 1995, contains several elements:

The FMR is the 40th percentile of gross rents for typical, non-substandard rental units occupied by recent movers in a local housing market.³

40th percentile: The 40th percentile is that point in a distribution of numbers at which 40 percent of the numbers are less than or equal to it and 60 percent of the numbers are greater than or equal to it. In the set of numbers {$395, $458, $486, $517, $675}, $458 would be the 40th percentile.⁴ The 40th percentile is similar in concept to a median; the median is the 50th percentile.

Gross rents: Gross rent is the sum of the rent paid to the owner plus any utility costs incurred by the tenant. Utilities include electricity, gas, water and sewer, and trash removal services but not telephone service. If the owner pays for all utilities, then gross rent equals the rent paid to the owner.

Typical, non-substandard rental units: In developing the FMR, the following units are excluded: public housing units, rental units built in the last 2 years, rental units considered substandard in quality, seasonal rentals, and rental units on 10 or more acres. The definition excludes public housing units to prevent subsidized rents from skewing the distribution. Similarly, rental units built in the last 2 years are excluded to eliminate units at the higher end of the market from the distribution. Since Section 8 seeks to improve the quality of housing occupied by lower-income families, substandard units are eliminated from the distribution. Finally, seasonal units and units on large plots are not part of the market intended for Section 8 recipients.

Occupied by recent movers: The definition of FMRs recognizes that owners often charge new tenants more than current tenants with leases. For this reason, the definition limits the distribution used to calculate the FMR to units occupied by recent movers. HUD has found that recent mover units typically rent for approximately 6 percent more than other units. However, this relationship can vary widely. In markets where demand is bidding up the price of rental housing, the spread between recent mover rents and other rents can be much higher. HUD has also observed markets where recent movers pay less than current tenants with leases because the demand for rental units has failed to keep up with supply.

In a local housing market: To achieve cost control and workability objectives, FMRs should be calculated in the context of the housing market in which certificate and voucher recipients shop. HUD generally uses Federal Office of Management and Budget (OMB) metropolitan area definitions to define local housing markets in urban contexts. Since OMB defines metropolitan areas on the basis of commuting patterns, they represent the housing market available to households working within the area. As commuting patterns have broadened, HUD has had to construct narrower local housing markets in a few metropolitan areas. For example, when OMB redefined the Washington, D.C., metropolitan area after the 1990 census, it added two counties in West Virginia and some distant, primarily rural counties in Virginia. HUD subsequently deleted these counties from the Washington, D.C., fair market rent area. In nonmetropolitan areas, HUD uses the county as the local housing market. Overall, HUD estimates FMRs for 354 metropolitan areas, 2,366 counties in nonmetropolitan areas, and 16 counties dropped by HUD from the OMB definitions of 6 metropolitan areas.

FMRs are set for rental units of different bedroom sizes, and Section 8 rules based on household size and the age and sex of children determine what size unit a household can choose.

How HUD Estimates FMRs

Failure to distinguish between the definition of FMRs and how HUD estimates them has been the source of considerable confusion. For example, the definition does not specify who is a recent mover; depending on the available data, HUD will count households who moved in within the past 15 to 22 months as recent movers. More fundamentally, the policy tradeoff between lower costs per family served, public acceptability of the housing provided, and the ability to provide an adequate range of choice among units and neighborhoods occurs at the definition level, not at the measurement level. Once the definition is chosen, HUD strives to provide the most current and accurate measurement of the definition in each FMR area. If in a particular market the FMR definition results in high per-unit costs or severe difficulties on the part of certificate and voucher recipients in finding suitable units, HUD has only limited ability to respond to these problems if it has measured the FMR correctly.⁵ (See discussion of exceptions below.)

HUD faces three problems in setting FMRs annually for 2,736 areas. First, the Department must obtain sufficient data on gross rents to calculate a 40th percentile. Second, it has to adjust the data for the various elements in the definition, such as recent movers and non-substandard housing. Third, if the data are not current, HUD has to adjust the calculated 40th percentiles to reflect current rental prices and then project them into the fiscal year for which it is estimating FMRs.

By law, HUD follows a formalized process in setting FMRs. In late April or early May of each year, HUD publishes in the Federal Register proposed FMRs for every FMR area for the forthcoming fiscal year (October 1 through September 30). The Federal Register notice invites comments from the public on the appropriateness of the proposed FMRs. Comments are generally due around July 1.⁶ After HUD reviews the comments, the Department publishes final FMRs in late September to take effect on October 1.

HUD uses four types of data for setting FMRs. From the U.S. Census Bureau, HUD has obtained a special extract of the 1990 census with data on gross rents for every metropolitan area and every nonmetropolitan county in the Nation. From its American Housing Survey (AHS), HUD has data on gross rents for 47 metropolitan areas for 1 or more years since 1990. HUD also gathers data every year on gross rents for 50 to 60 metropolitan areas or counties using Random Digit Dialing (RDD) telephone surveys. Finally, from the public comment process, HUD has information provided by local housing authorities and others for specific FMR areas.

Of these four sources, the census extract contains the largest samples and provides consistent data for all FMR areas. However, the census extract is the least current of the data sources. The AHS provides consistent data 1 to 7 years prior to the fiscal year for which FMRs are being estimated. Moreover, sample sizes among recent movers can be small in AHS samples for some areas. RDD surveys provide data on gross rents for January through July of the year in which the FMRs become effective on October 1. Publicly provided data are generally current but are idiosyncratic and often unusable.⁷

In all cases HUD uses these data to estimate the FMR for two-bedroom units only. Nationally, two-bedroom units account for 43 percent of the rental stock. Being the most common unit, they are the easiest units for which to obtain data. HUD estimates FMRs for efficiencies, one-bedroom units, three-bedroom units, and units of other sizes using the two-bedroom estimate as a base.⁸

Table 1 explains how HUD adjusts the various data sources to fit all the elements in the FMR definition. In general, HUD's own data sources -- the 1990 census extract, the AHS, and RDDs -- allow direct adaptation to attain the FMR definition. With data obtained through public comments, HUD frequently has to make adjustments based on 1990 census data or AHS data.

Table 1. How HUD Matches the FMR Definition With Different Data Sources

The final step in estimating two-bedroom FMR is to update and project the measured FMR from the date of measurement to the middle of the forthcoming fiscal year, that is, to April 1 of the following year. HUD uses a combination of sources and techniques to update and project. The approach used depends on the particular FMR area and the date of the data used to establish the 40th percentile.

The Bureau of Labor Statistics publishes annual rent price index data from the consumer price index (CPI) for 95 FMR areas. HUD uses the CPI data to update FMRs for these areas and then projects the FMRs forward using approximately a 3-percent annual rate of increase. In estimating proposed FMRs for fiscal year 2000, HUD will have available CPI data for these metropolitan areas through December 1998. HUD will then inflate the end-of-1998 FMRs 15 months forward to April 1, 2000, using the assumed 3-percent annual rate.

For other areas that have FMRs based on data from 1993 or later, HUD uses special RDD surveys to update to the end of 1998 and then projects forward to April 1, 2000, again using the assumed 3-percent annual rate. Prior to 1993 HUD used CPI rental price indices, which measure rent changes for 4 large regions: the Northeast, the Midwest, the South, and the West. Experience during the 1980s with such broad-based indices was not satisfactory, so beginning in 1993, HUD funded RDD surveys to measure rent changes for 20 smaller areas. The 20 areas are derived from the metropolitan and nonmetropolitan parts of each of the 10 Federal regions with one modification: The metropolitan part of each Federal region omits all the metropolitan areas for which HUD has CPI data.⁹

For areas without local CPIs that have FMRs based on data from 1990, 1991, or 1992, HUD uses the regional CPI rental indices to update these areas to the end of 1993 and then uses RDD surveys to finish the updating to the end of 1998. Again HUD projects these FMRs forward to April 1, 2000, using the assumed 3-percent annual rate.

Example

Table 2 shows how HUD has computed the two-bedroom FMR for Nashville throughout the decade. The starting point was the 1990 census. Because the FMR definition used the 45th percentile prior to 1995, HUD used the extract from the 1990 census to calculate a 45th percentile of $455 as of April 1990, the date of the census. HUD then adjusted this estimate upward by 1 percent to account for the requirement in the definition to exclude substandard units. (This 1-percent adjustment used both national AHS data and census data for the Nashville area.) HUD then used CPI regional data to update the adjusted census estimate to the end of 1993. Then HUD used its RDD survey for the metropolitan portion of the Southeast region to obtain an end-of-1994 estimate of $521.¹⁰ When the FMR definition was changed to the 40th percentile in 1995, the end-of-1994 estimate decreased by 2.1 percent to $509. An RDD survey for the metropolitan portion of the Southeast region raised this estimate to $525 at the end of 1995.

Table 2. Example of FMR Calculation, Nashville Tennessee

In 1996 HUD conducted an RDD survey to update its baseline for Nashville. The survey found that the 40th percentile had increased to $578. The new baseline replaced all previous estimates and became the starting point for the 1996 and subsequent FMRs. The current (fiscal year 1999) estimate of $626 was obtained by updating the $578 to the end of 1997 using the RDD survey for the metropolitan portion of the Southeast region and then projecting forward to April 1, 1999, using the assumed 3-percent annual rate.

State Minimums, Exceptions, and Web Site

Sparsely populated, nonmetropolitan counties present special difficulties in setting accurate FMRs. These counties typically have small Section 8 programs, making it cost-inefficient for HUD to conduct RDD surveys to update their FMRs.¹¹ Because the nonmetropolitan regional RDDs may miss small, hot real estate markets, updating from the 1990 census can introduce errors over time. Generally, the public comment process will identify these areas. Unfortunately, it is these small, nonmetropolitan areas that have the most difficulty providing the information HUD needs to change an FMR. For these reasons, HUD introduced State minimums in 1996. For fiscal year 1996 HUD set a minimum FMR for each State that is the lesser of the average nonmetropolitan FMR for the State or $450. Since 1996 HUD has used State minimums, updating the $450 limit for overall inflation in rents. In fiscal year 1999, 1,721 of the 2,382 county FMRs were State minimums.

By statute, HUD can increase the FMR by up to 20 percent for a portion of an FMR area. This provision allows HUD to respond to situations where an areawide 40th percentile FMR does not provide the desired range of choice among units or neighborhoods. There are three limitations on exceptions:

• The exception area can contain no more than 50 percent of the population of the overall FMR area.

• The exception FMR must equal the 40th percentile rent within the exception area.

• The exception must respond to problems observed in operating the Section 8 program in the exception area or be part of an active effort to provide housing opportunities outside poverty areas.¹²

A complete list of FMRs can be obtained over the Internet.