Income Limits, Fair Market Rents, Area Median Incomes, and Poverty Levels

Download the income limits, fair market rents, area median incomes, and poverty levels. This is a self-executing file (*.exe, 960 KB) that when extracted, contains:

• AHS97Inc Read Me.txt - the documentation file.
• AHS97IncLim.ASC - the data in fixed-column ASCII format (7.6 MB after extraction).
• ahs-medinc-fmr-ipov.sas - the SAS program used to generate the data.

READ ME

This file contains estimates of Fair Market Rents, income limits, area median incomes, and poverty levels for the occupied housing units in the public use microdata of the American Housing Survey national sample for 1997. These estimates were prepared by staff members and contractors at the U.S. Department of Housing and Urban Development using their standard methods. The data in this file can be matched with the AHS 1997 national file by using the CONTROL variable.

Note that these data were calculated by HUD using the AHS public use microdata, plus HUD fair market rents, HUD income limits, HUD estimates of area median incomes, and other public data, such as poverty levels and the Consumer Price Index. All of these data are publicly available. No confidential data were used.

Fair market rents can be downloaded from: http://www.huduser.org/datasets/fmr.html

Income limits can be downloaded from: http://www.huduser.org/datasets/il.html

NOTES ON DATA:

1. Fair market rent is in dollars per month. All income measures are dollars per year.

2. The poverty level is calculated given household size, number of adults, elderly, and year.

3. The fair market rent, income limit, and median income calculations were performed by using code translated into SAS from HUD's BOAT package, a FORTRAN-based program. Below are some notes from the BOAT documentation that help explain the average, minimum, and maximum estimates: "...Data include LMED median income, L50 very low income limit, L80 low income limit, FMR fair market rent. (LMED is estimated from all family sizes, without adjustment; the others are adjusted for household size or bedrooms, as appropriate.)"

"Locations defined in the National AHS (and this routine) include parts of some MSAs (using 1983 geography). Other areas are subdivided by climate, metro/nonmetro, and census region. Each area, even a 1983 metro area, may include more than one FMR and Income Limit area now. Therefore, we looked at each county (MCD in New England) separately, found its Location on AHS, and its FMRs and Limits now. Then for each Location defined in AHS we had one or more sets of FMRs and Limits. We took an average, weighted by population, but also took the lowest and highest. These are available in FMR (average), FMRA (minimum), and FMRB (maximum)."

"Thus, if you want to be sure you are studying households with rents below the FMR for their area you would use ZSMHC < FMRA (OR TO GET HOUSEHOLDS ABOVE THEIR FMR: ZSMHC > FMRB). Similar A and B variables exist for LMED, L50, and L80. (Note: in processing we had to generate FMRs and Limits for 16 of Virginia's independent cities which are not specifically published, e.g. Falls Church was given the Washington metro figures.)"

4. Users of the AHS data in SAS datasets may use the input statement below as part of a DATA step to read the data into SAS:

input	@1	control	$12.
	@14	ipov	8.0
	@23	fmr	8.0
	@32	lmed	8.0
	@41	l50	8.0
	@50	l80	8.0
	@59	fmra	8.0
	@68	lmeda	8.0
	@77	l50a	8.0
	@86	l80a	8.0
	@95	fmrb	8.0
	@104	lmedb	8.0
	@113	l50b	8.0
	@122	l80b	8.0
	;

Note that the first line of the file contains the variable names. Thus, if you run the above input statement as part of a SAS program, you will get error messages refering to the first record. You may safely ignore these messages, as the file will be created properly.

5. The ahs-medinc-fmr-ipov.sas program is included for reference only, so that you can see how the data was generated. It contains references to directories on local computers at HUD and will not run "as is" on your installation. It also reads a data file, dt16-t16.fmr, which is not included in this archive.

6. HUD is indebted to Paul E. Burke for writing the FORTRAN code to calculate these data when he was employed at HUD. Ming Chow and Anh T. Nguyen, contractors for HUD, translated the code into SAS and checked the results. Any complaints or questions should be directed to David A. Vandenbroucke, using the contact information included in the zip archive. You can also join the AHS eList to discuss any issues relating to the data.

REVISION HISTORY

09/14/1999 9:48 AM
This release fixes computational problems that set some records to all zeroes.

6/3/99 1:27 PM
This release fixes the fact that the original file used only 10 digits for CONTROL instead of 12. It is also based on the new release of the microdata.

5/26/99 10:12 AM
Original file released.