A Picture of Subsidized Households Detailed Explanation of the Variables

1. U.S. totals
   
2. Total of all housing agencies of a particular size. Actual size is shown in name and at end of record. Only covers programs run by housing agencies*. Agency is classified based on total units in all programs.
   
3. State total, also used for DC and outlying areas
   
4. Housing agency total*
   
5. No longer used
   
6. Project record (not for Certificates+Vouchers, Moderate Rehabilitation) A 'project' is a building or group of buildings funded at about the same time.
   
7. Census tract summary
   

   A Census tract is an area of about 1,500 homes, chosen by local communities in cooperation with the Census Bureau, as an area that is somewhat homogeneous socio-economically. Maps were published by the 1990 Census, showing tract boundaries
   

* Note that programs 6-9, (S-L) are not included in housing agency summaries, since they are independently managed.

For summary level 6, projects: name is in the first 12 characters of this item, a brief address is in the second 12 characters and city is in the last 8. When a project has more than one address, a central one was picked, as discussed below under "Latitude."

For any project whose agency does not have the same code as the beginning of the project number, we put a note in the name (e.g. RQ001001 is noted as being: @RQ005). This is in positions 1-6 of the name, or in 19-24 if street address was blank.

For summary level 7 in program 4 (C), Census tracts in Certificates+Vouchers this item shows: county and tract, in the format: "Cert+Vch:county123 1234.67tract"

For summary level 7 in program 1 (A), Census tract summaries across all programs, this item shows: population, renter households, and percent in various programs, in the format: "12345pop 2345ren 12P 12C 12H 12L". The abbreviations are as above, except H is a total of programs 6, 7 and 8 (S, T and F).

In summary levels 1-3, this code has a very brief label, like AL for Alabama. For other summary levels, the meaning varies by program:

In program 1 (A: i.e., for tract summaries that include all programs present in the tract), this code shows: state, tilde (~), HUD units as % of housing units in the tract in 1990, and number of housing units in the tract in 1990, like "AL~12% 1234 "

In programs 2-5 (B, P, C, K) this code shows: state abbreviation, then blank and agency#, then blank and project# if any, like: "AK 001 001" In program 2 or B, an M after the project# (in the 11th character) means Mutual Ownership. Portability households are coded with the agencythey are billed to (whence they moved).

In program 6 or S this code shows: state, 2-digit HUD office, and project number, like "AK01R123456"

In programs 7-8, (T, F) this code shows: state, underline, 3-digit HUD office and project number, like "AK_11144666" For this group, if state was unknown, we generated it from the HUD field office code. There may be some errors where offices cross state lines.

In program 9 or L this code shows: state, curly bracket ({), then state-defined project ID, if any, like "AK{87-123"

This count is of 6/1/97 (same as tenant data) for Public and Indian Housing, Certifictes, Vouchers and Moderate Rehabilitation. In other programs it is not as recent as the tenant data; it is 9/30/95, except Tax Credits, it is 12/94.

If an agency overall has more Certificates+Vouchers than it reports on, we assume each tract has extra Certificates+Vouchers in the same proportion (but we assume no more than twice as many in a tract as the number reported).

The number of units was missing and had to be estimated for a few projects. In Tax Credits, when we show 35 units we had no data (and used 35 since it is the national average); projects which truly had 35 or 36 units are shown as 36. In Ofc. of Housing (programs 6-8 or S-F), 66 units mean we had no data (and used 66, the national average); 67 means actually 66 or 67. In Public & Indian Housing, zero units are shown for 95 projects. It is believed most are not active project numbers.

Elsewhere no data are actually available, and numbers are assumed, purely for calculating reporting rates: In program 6 (S) occupancy rates are stipulated at 98%, and in programs 7-8 (T, F) at 96% (Wallace 1993, page 2-27, though his figures are based on rent revenue, and occupancy is less). In other programs we stipulate 99%, since almost all units are assumed occupied.

This includes rent the household pays to the agency or landlord, and an estimate of any separately metered utilities they pay (and also ownership costs in Indian ownership programs).

Based on households with rent in the range $1-$2,498; other rents are considered errors. Virtually all zero rents are missing data, since the minimum rent (10% of income) applied to positive income (see note on missing data in income) gives positive gross rent. However if gross rent is lower than estimated utilities, rent to the agency or landlord may be zero.

This is item 14 on form HUD-50058 (highest of: 10% of gross income, 30% of adjusted income, or welfare rent), except in vouchers, where it is item 22k (total family contribution), and Indian Mutual, where it is item 17h (family cost).

For Public & Indian housing, this only includes 1996 operating and modernization spending, since the construction costs have already been paid for, and opportunity cost is not available. It omits Drug Elimination Grants, which are the only other major source of current spending.

Operating subsidy to all agencies, and modernization funding to large agencies (generally over 250 units), is distributed every year by formula so we use the actual funding.

For smaller agencies modernization funding is competitive, and they do not all receive it each year, so the actual funding in any one year would not reflect average long term costs. As an approximation we found the average modernization per unit, and multiplied this average by the units in each agency to estimate the long term expected value of modernization spending. A higher fraction of Indian agencies receive modernization than of other agencies, so we did the average separately for Indian and non-Indian agencies.

At each agency, for purposes of this summary, total spending is averaged per occupied unit per month, without differentiating among projects. In fact some projects receive more or less than average each year.

For Section 8, total spending is calculated for each unit (from the rent calculation form) and averaged, so averages should vary correctly in each tract and agency. In Certificates+Vouchers and Moderate Rehabilitation our estimate also includes the 8% administrative fee.

In other programs spending is not yet available, so figures exclude section 236 and other subsidies. Some spending shows up in Section 236 and Other Subsidies, but this is just Section 8 (Loan Management), averaged over the households that receive it.

This is basically total income before adjustments, expected for the next 12 months, but it does exclude some types of income which HUD regulations do not count at all, such as earnings of minors, and scholarships.

Averages below $1,000 are shown as .00-.99 thousand dollars. Averages below $10,000 are shown as 1.0-9.9 thousand dollars. Higher averages are shown as 10. to 99. thousand dollars. (All software that we are familiar with can handle this floating decimal point; normally declare the maximum width: 3 and maximum decimal positions: 2.)

Note that remaining households have majority of income from another source, such as pensions, Social Security, SSI, unemployment benefits or child support, or have a mixture, with no type giving a majority of income

This is the coefficient of variation, which is the standard deviation of income, as a percent of the average income. The standard deviation measures how far each household's income is from the average income. Incomes that are far above or be low average have large effects, since the differences are squared before being averaged.

Local median incomes are estimated regularly by HUD, because they are used in setting income limits for housing subsidies. (The most common income limit is 50% of local median income, adjusted for household size. There are some state and national floors and ceilings, which restrict the amount of variation from area to area.) Here we compare with adjusted median income (i.e. twice the 50% limit), so for example if local median income is $30,000 (and the 50% limit is $15,000), then a subsidized household with $10,000 income has 33% of local median income.

Every year, HUD estimates the median income of families in each metropolitan area and each non-metropolitan county, in order to establish income limits. That estimate is based on both subsidized and unsubsidized families. 'Families,' used in establishing median incomes, include only people living together who are related by blood, marriage or adoption; people living alone are not used for establishing median incomes, but of course they are eligible for subsidy, and are included in all the statistics on subsidized housing.

When we say 'adjusted for household size' we mean the following: We make no adjustment in households with 4 people. We reduce the denominator 10% for each person under 4 in a household (3 people: 90% of the 4-person denominator; 2 people: 80%; 1 person: 70%). We increase the denominator 8% for each person over 4 in the household (5 people: 108% of the 4-person denominator; 6 people: 116%; etc.).

For Public Housing projects, this is the difference between: % minority for the project, and overall % minority for Public Housing at the agency. Difference is shown for each project (in absolute value). Average of these differences i s shown for each agency.

For Certificates+Vouchers (which do not have projects), difference is based on Census tracts with more than 10 households reported: average difference between: % minority among Certificates+Vouchers in the tract, and overall % minority among Certif icates+Vouchers at the agency.

Summaries of this item exclude agencies that are under 5% or over 95% minority, and agencies with only one project or tract, since there cannot be much difference there.

We treat units with no bedrooms reported and more than 2 people as missing data, and omit them from numerator & denominator of bedroom percents.

Households are reported annually, on a flow basis, so for good recent data the average should be about 6 months. Larger averages mean recent data have not been received, or have not been accepted, because of errors.

For projects with only one address, we give its latitude and longitude. For projects with more than one address, we pick the address that is closest to the median latitude and median longitude of the whole project, and use this address' latitude, longitude, zip code, county, tract, etc. Other files have more address information, but no data on occupants (see Subsidized Housing Projects' Geographic Codes, page 17)

Negative means West longitude, which covers most of the United States. Zero is in Britain; negative numbers go to -180 in the Aleutians; positive numbers go to +180, also in the Aleutians.

Distances in miles between points a and b can be calculated (where f=.0174532925 and converts from degrees to radians):

IF abs(longa-longb) < .001 THEN

distance = abs(lata-latb) * 69.055

ELSE

distance = 3956.56 * arcos{ [sin(lata*f) * sin(latb*f)]
+ [cos(lata*f) * cos(latb*f) * cos(abs(longa-longb)*f)] }

END IF

This, like other items, is averaged for all projects, weighted by number of units in each project

0 means not in any metropolitan area (blank or -1 means status unknown)

8888 means more than one metro area applies.

MSA and PMSA (Primary Metropolitan Statistical Area) codes form one consistent series; the distinction is minor: PMSAs are pieces of even larger "Consolidated" areas (CMSAs), while MSAs are free-standing.

These are statistical areas established by the Office of Management and Budget. HUD has subdivided some for setting Fair Market Rents and Income Limits, and HUD's areas are NOT reflected here.