Note: Blank means data not available

Occupancy

Occupancy rates are only collected for Public Housing. We assume plausible levels of occupancy in other programs for purposes of calculating the completeness of reporting.

Completeness

Overall reporting is 73% of occupied units (so we have 27% missing data). Public Housing, Certificates+Vouchers, Section 8 New, and Section 236 have the highest reporting, from 73–80%. The smaller programs range from 48% to 65% complete. In Tax Credits, no reporting is required.

On another aspect of completeness, we believe this count of Tax Credit units placed in service is fairly complete. There are estimates that over 600,000 Tax Credit units have been allocated, but for this report the appropriate figure is the number of unit s that have been placed in service, given above.

Some of the households counted above in Section 236 projects have incomes high enough to pay full rent, without subsidy. These households are not reported to HUD and their characteristics are not included in this report.

Household Sizes

Sizes range from 1.6 people per unit in Section 8 New projects, to 3.7 people per unit in Indian Housing. This variation is related to age: programs with smaller household sizes have more elderly, as shown on the next page.

Total People Served

We can multiply housing units times the average people per unit to find the total number of people served by the programs: about 11 million.

Average rents range very narrowly from $160 per month in Moderate Rehabilitation projects to $197 per month in Certificates+Vouchers. Rents in Indian Housing do not mirror their relatively higher incomes, since most Indian Housing is Mutual Ownership, with a distinct calculation for household cost. The calculations are shown on the data collection form (50058).

Incomes range from $7,200 per year in Moderate Rehabilitation to $16,000 per year in Indian Housing. If we divide by average household size, income per person ranges from $3,100 in Moderate Rehabilitation and $3,200 in Certificates+Vouchers, to $5,300 in Section 8 New. The lowest averages of income per person are in the programs with most welfare recipients. Welfare here includes only Aid to Families with Dependent Children and General Assistance.

Income mix is expressed as a percent of the average income. The larger the percent, the more income mix there is.(1) These figures reflect the spread of incomes across the country, including high and low income areas. The mix is usually narrower in any one project or neighborhood.(2) The national income mix is relatively wider in Public and Indian Housing and in the 'Other Subsidy' projects. These programs serve a wide range of incomes across the country. The mix is narrowest in Section 8 New, where most households are elderly, living on Social Security or Supplemental Security Income (SSI); these incomes vary little between or within projects. Some Section 8, Section 236, and 'Other Subsidy' projects have unsubsidized households, who are not counted in this report, and extend the income mix. The surroundings of Certificates+Vouchers likewise include unsubsidized higher income households. Figures include different household sizes, without adjustment, so some mix is due to larger households typically having (and needing) more income than smaller households.

Age of Head

Section 8 New has by far the highest concentration of elderly at 60%. Public Housing and Section 236 are next at 33–35%, and the other programs are 16–23%. It may be noted that 5% of subsidized households (about 230,000) have a head or spouse age 85 or older, which gerontologists term the 'very old.'

There are large numbers of people with disabilities in some programs. Overall a fifth of the non-elderly have disabilities. Also, a fifth to two-fifths of the elderly have disabilities, but this is not necessarily surprising; many of us develop disabiliti es as we age. The definition includes mental and physical disabilities; for a summary of the definition, see the last page of form 50058.

Minorities

Overall, 56% of households are minorities (38% Black, 14% Hispanic, 2% Asian, 3% Native American). Indian Housing, of course, is almost entirely minority (for a recent description of Indian Housing see Kingsley et al 1996). The lowest minority rate, 36%, is in Section 8 New projects.

The Average Difference column measures separation between minority and white subsidized households. The larger the number, the more separately subsidized whites & minorities live from each other. The scale is from 0 to 50. It is a broad indicator, and more detailed analysis with individual racial and ethnic groups can be done (Goering et al 1995).

For each Public Housing project, 'Average Difference' is the difference between: (a) % minority for the project and (b) overall % minority for Public Housing at the agency. We average the figures from each project to have a total for each agency, state or other area.(3) Averages for states and the United States exclude agencies with under 5% or over 95% minority, or only one project, since differences are not meaningful in such agencies.

In Public Housing, households are assigned to projects by the housing agency.(4) Thus, racial differences would not be expected in projects, except by chance, or by households of one race moving out of or refusing to move into particular projects . The 13% national figure is a mix of random variation in some agencies, current practices, and past segregation patterns.(5) It may be useful in reviewing a particular agency to notice whether its figure is above or below this national average, but no particular figure proves discrimination. Particular practices at individual agencies need to be studied before any conclusions can be drawn.

For Certificates+Vouchers, which do not have projects, 'Average Difference' is based on the location of subsidized households in Census tracts: For each Census tract we take the difference between: (a) % minority among Certificate+Voucher holders in the tract and (b) overall % minority among Certificate+Voucher holders at the agency. For averages we exclude tracts with 10 or fewer households reported, since tracts with so few subsidized households have little chance of matching the agency average. Averages for states and the United States also exclude agencies with under 5% or over 95% minority, or only one tract, since differences are not meaningful in such agencies.

This average difference is not the difference between subsidized households and the whole tract, but that can easily be calculated, since the tract's % minority is shown near the end of the record (and summarized in the next table). The difference shown above is the difference between subsidized households in the tract and subsidized households in the agency.

The average difference happens to be 13 points in both Public Housing and Certificates+Vouchers, but this does not mean the programs operate in the same manner. As the next table shows, Certificate+Voucher holders live in neighborhoods with many fewer minorities than Public Housing tenants: 35% minority on average, compared to 63%. The measure on the table above is different: Minority and white Certificate+Voucher holders live in different neighborhoods from each other to some degree. This difference reflects that households often lived in racially disparate neighborhoods before they applied to the programs. Some stayed in place. Others moved, but not far (Léger and Kennedy 1990). Thus, the minority and white subsidized households do not end up equally distributed in the same neighborhoods as each other. We do know from the next table that Certificate+Voucher holders end up in more integrated neighborhoods than Public Housing tenants. But there is still a difference in the locations of white and minority Certificate+Voucher holders.

The minority difference is limited by the percent minority in the agency: for example some large agencies are nearly all minority and can only have small differences. An agency that is 5% or 95% minority can have an average difference as high as 9.5, while an agency that is 20% or 80% minority can have an average difference as high as 32. A 50% minority agency can have an average difference as high as 50.

One could use numerous other statistics to summarize racial differences among projects or Census tracts.

There is some variety among the programs in social items. Couples with children are not common, except in Indian Housing, since they tend to have more income than elderly and single parents. The first two columns above can be added to count all households with children. These figures are based on age, so they include foster children, nephews, grandchildren; anyone under 18. Female heads are common, since they include most households with no spouse present, as well as most elderly. The extent of over-housing may limit access for large households. It results partly from households that are not relocated to smaller units when their children grow up and leave.

Financial Characteristics

Assets over $5,000 are taken into account in setting rent. They are not common except in Section 8 New, with its many elderly households. Utility allowances are given when individual metering and billing are used, so they show the extent of individual metering. The size of the allowance is estimated by each Housing Agency to reflect typical bills. Households may actually spend more or less.

Length of Stay and Wait

For some programs we have length of time since the household moved in. It averages 6 years, with little variation. The average total stay of households, until they move out or die, is twice this span, or 12 years.(6) Time on waiting lists varies more than length of stay, from 10 months in Public Housing, to 25 months in Certificates+Vouchers. This is the average waiting time for people moving in, who tend to be in the higher priority categories ('preferences'), which give them the shortest waits. The average wait for people remaining on the waiting list may be longer.

Neighborhoods

To describe the neighborhoods of these households, we show three measures of the Census tracts surrounding them. A Census tract is an area averaging 1,500 homes, chosen by local communities in cooperation with the Census Bureau, as an area that is somewhat homogeneous socioeconomically. In low density areas Census tracts cover large land areas, so they are broader than the common idea of a neighborhood. The figures on the Census tract include the subsidized households there, so large projects can dominate or overwhelm the neighborhood figures. Tax Credit projects return to the discussion here, since we do have their addresses, and can describe their neighborhoods, even though they do not have to report on their occupants.

On average in the neighborhoods of most of the subsidy programs, 19–28% of households have incomes below the poverty limit, except in neighborhoods of Public Housing, which average 38%. There is more range in neighborhood minority rates, from 34–40% minority for Section 8 New, Certificates+Vouchers, Tax Credits, and Section 236, to 63% minority in the neighborhoods of Public Housing projects.

To provide a different view of these neighborhoods, we show the percent of households in the neighborhood that own single family homes. The extremes are a low of 25% for Public Housing neighborhoods, and highs of 40–43% for Tax Credit and Certificate+Voucher neighborhoods.

These two tables count the subsidized households by some characteristics of their tracts. Note that the total units are lower than in other tables, since we lack some addresses and cannot put them into tracts.

The first table confirms that much Public Housing is in tracts with high poverty rates, as mentioned on the previous page. Other programs tend to be in less poor tracts.

Similarly Public Housing tends to be in minority tracts. 'Other Subsidies' are in minority tracts almost as much, but the remaining programs tend to be in tracts with fewer minorities.

Spatial Concentration

The data file has a summary record for each Census tract. There are about 61,000 Census tracts in the country, with a 1990 population of 250 million. Subsidized housing is in about 45,000 Census tracts, with a 1990 population of 200 million. Many of these tracts have no subsidized projects, and only a few Certificates+Vouchers, which households can take to any unit with a moderate rent. Even Certificates+Vouchers are fairly concentrated, because of households' preferences, and a lack of transportation and child care to permit a wide search in a reasonable time. The reports show some of this concentration: they show the tracts with the most subsidized units. The data file allows one to compare all tracts, with and without various kinds of subsidies.

We have tallied subsidized units by some of the characteristics of their tracts. Note that, as on the previous page, the total units are lower than in other tables, and some statistics differ, since we lack some addresses and cannot place them in tracts. Furthermore this table, like all tables in the report, is based on the public summary file, so it does not have data where few households were reported, and data were suppressed. The rate of concentration ranges widely as a percent of the homes in the tract, but the median subsidized unit is in a tract that is 12% subsidized.

Incomes and ages of subsidized households do not vary consistently based on the level of subsidy concentration in the tract. However the highest incomes are in the most concentrated tracts. These tend to be in large cities, where incomes of subsidized households tend to be higher than average. The oldest households are in tracts where 10–29% of households are subsidized (perhaps reflecting the presence of small elderly projects).

Minority and poverty rates vary more. Minorities tend to be in tracts with high concentrations of subsidized housing: there are high minority rates both in the subsidized housing and in the tract as a whole when subsidy concentration is high. Also the tracts with a lot of subsidized housing tend to have high poverty rates (as expected, since subsidized units are included in the poverty count for the tract, and raise the poverty rate).

The following table shows similar data, with the point of view changed, from the perspective of subsidized units, to the perspective of the total population of these Census tracts. Eighty percent of people in the United States live in Census tracts where there is some subsidized housing.

The body of the report shows the number of subsidized housing units and their characteristics by size of agency. The table above counts the number of agencies and projects. There are relatively few large agencies, but they have most of the units.

Where one agency has more than one numeric code, we count it only once, but other publications may count each code separately. Also we count agencies based on total units in all programs. If they were grouped by the number of units in any one program, there would be fewer large agencies and more small ones.

Most projects are small, under 50 units. But most units are in larger projects, those with 100 or more units. The smallest projects are common in all programs, while the very largest projects are primarily in Public Housing. This table omits the 95 Public and Indian projects in the data file with zero units, since they are usually not really separate projects.

Multiple Subsidies

Privately owned projects in some programs can have multiple subsidies. This table shows how we have categorized the subsidy records at HUD, matching on FHA and Section 8 project numbers to avoid double-counting. The number of subsidized units here is slightly less than in other tables, since active projects in the files with zero units are counted as zero here, but their size is estimated for other tables.

Most LMSA/RAPSUP units are in Section 236 and BMIR projects, which are subsidized anyway. The remainder are added to the subsidized count. A few can even be in Section 8 New projects, where Section 8 only subsidizes part of the project, and LMSA can be provided to help the other units if the project runs into difficulty.

There is still a limited amount of double-counting in the figures, since Vouchers may be used in Section 236 projects, and Section 8 may be used in Tax Credit projects, and we do not yet subtract out the overlap.

We present selected information here, without further comment, for the 12 largest agencies. This is meant as a convenient reference. Total units (in parentheses) include Moderate Rehabilitation, which is not shown separately on this page. Fuller informati on on these and other large agencies is in the body of the report.

Blanks above show missing data. HUD and many of these agencies are working to improve the data, so future releases will be able to fill in most of these blanks.

A Picture of Subsidized Households, order # forthcoming. HUD, 1996. Summary data on paper, internet & cdrom for the United States, each State, each housing agency, and for Public, Indian, Section 8, FHA and Tax Credit projects and their Census tracts.

Subsidized Housing Projects' Geographic Codes, form HUD–951, AVI–1004, gopher://huduser.gov:73/11/2/f/shp OR https://www.huduser.gov/data.html HUD, 1996. Latitude, longitude, names, Zip codes and Census geography for HUD projects.

Development and Analysis of the National Low-Income Housing Tax Credit Database, order # HUD–7306, AVI–60 https://www.huduser.gov/lihtc/ HUD, 1996. Analyzes data from 1992–94, but database goes back to 1987, when the program began.

Housing Problems and Needs of American Indians and Alaska Natives, order # HUD–7173, by G. Thomas Kingsley, Maris Mikelsons and Carla Herbig, Urban Institute. HUD, May 1996.

Assisted Housing Quality Control, order # HUD–7219, by Suzanne Loux, Mary Sistek and Frank Wann. HUD, 1996. Measures errors in the form 50058 and 50059 data used in this report (though not using income matching to detect unreported sources of incom e). The largest dollar errors come from not getting income verifications on earnings and pensions, and, more surprisingly, not using verified amounts even when they are obtained. Net errors in rent average $4 underpayment per month per household.

Location and Racial Composition of Public Housing in the United States, order # HUD–6557, by John Goering, Ali Kamely and Todd Richardson. HUD, March 1995.

Family Data on Public & Indian Projects, August 1993, order # AVI–27, by Paul Burke. gopher://huduser.gov:73/11/2/f/famdata or https://www.huduser.gov/data.html HUD, November 1993. This is an earlier version of A Picture of Subsidized Households 19 96, prepared on a comparable basis, also with data on individual projects and agencies. It excludes Section 8, FHA and Tax Credits.

The Assessment of the HUD–Insured Multifamily Housing Stock, Final Report, v. 1, Current Status, order # HUD–6261, by James E. Wallace, Abt Associates, September 1993.

Characteristics of HUD–Assisted Renters and Their Units in 1989, order # HUD–5961, by Connie Casey. HUD, March 1992. This has data (from sample interviews in the American Housing Survey) on the housing and neighborhood quality of the units people l ive in.

"Characteristics of HUD-Assisted Tenants," Appendix C in HUD Annual Report 1990. HUD, 1992. A very brief summary of Characteristics of HUD-Assisted Renters and Their Units, based on the American Housing Survey.

"Report on Subsidized Tenants," Appendix C in HUD Annual Report 1989. HUD, 1991. Data for selected samples in various areas, from the same data base as this report, and from research samples.

"Report on Subsidized Tenants," Appendix C in HUD Annual Report 1988. HUD, 1990. An earlier and very brief version of Characteristics of HUD-Assisted Renters and Their Units, based on the 1987 American Housing Survey.

Recipient Housing in the Housing Voucher and Certificate Programs, order # HUD–5597, by Mireille Léger and Stephen Kennedy, Abt Associates. HUD, May 1990.

"Data on Individual Tenants," chapters 8.3 and 9.3 in Researcher's Guide to HUD Data order # HUD–3846, third edition. HUD, August 1984.

"Trends in Subsidized Housing 1974–1981", order # HUD–3034, by Paul Burke. HUD, March 1984.

"Analysis of the Data Quality of the Section 8 Tenant Characteristics Data Base," Office of Program Planning and Evaluation. HUD, San Francisco Regional Office May 1979.

Lower Income Housing Assistance Program, order # HUD–817. HUD, 1978.

"Occupants of Subsidized & FHA Housing," chapter 9 in Researcher's Guide to HUD Data order # HUD–223, second edition. HUD, July 1978.

"Suspended Subsidy Programs," chapter 4 in Housing in the Seventies. HUD, 1974.

Justification of the Estimates, volumes on Office of Housing, Office of Public & Indian Housing (formerly Office of Housing Management). HUD, Budget Office, annual.

Statistical Yearbook 1966 to 1979. HUD & Government Printing Office.

Footnotes

(1) This is the coefficient of variation, which is the standard deviation of income, as a percent of the average income. The standard deviation measures dispersion: how far each household's income is from the average income. Incomes that are far above or below average have large effects, since the differences are squared before being averaged. In a normal distribution two thirds of cases are within ±1 standard deviation of the average. For example in Public Housing, the standard deviation of household income is $6,500, which is 76% of the average household income, $8,500, as shown.

(2) To understand why the national figure is bigger than most local figures, consider an example of two equal-sized projects: one where everyone has $20,000 income, and another where everyone has $5,000. The mix is zero in each project, but the combined m ix is quite wide (60%). National figures are not averages of local figures, but measure the overall national mix. The following table does give average mix, from projects with enough reporting. Public and Indian Housing still have a wider mix of incomes i n their projects than most, but not by much. They serve a wide range of incomes across the country, but not so wide in each project.

(3) For example an agency where every project has the same % minority, has zero difference in each project, and zero average difference for the agency as a whole. An agency with 80 blacks in one project and 80 whites in the other has the maximum difference of 50 points. Such an agency is 50% minority, so difference in the black project is: 100% minus 50%, and in the white project is: 0% minus 50%. Difference is measured in absolute value, so the average difference is 50 points. Racial separation can also be measured in other ways.

(4) Tenant assignment in Public Housing is a complex subject. In general over the past 20 years, households were placed on an agency-wide waiting list in order of their federal and local 'preferences' and their date and time of application. ('Preference' means a priority category, like homelessness.) When a housing unit became available, it was offered to the household at the top of the waiting list, regardless of minority status (unless for example an effort was being made to overcome proven discrimination in the past). If they turned down the units offered (they typically got one or three chances), they dropped to the bottom of the waiting list.

(5) HUD enforces the federal fair housing acts, which prohibit discrimination in federally assisted and most private housing.

(6) Mathematically this doubling comes from the fact that data forms are submitted evenly throughout each person's stay. Therefore the average data form submitted is half-way through the average stay.

(7) Some Pittsburgh households were originally mis-coded into Philadelphia. This report has made efforts to identify most of them and report them correctly in Pittsburgh.