Facts about Public Housing

 

a.      Number of units — In 2007 there were nearly 1.2 million public housing units, of which slightly more than 1,150,000 were eligible for operating subsidies.[i]  Most of the difference is due to vacant units (only some of which are eligible for operating subsidies), including units in the process of demolition or disposition. 

 

b.      Occupied units — In 2007, about 1 million public housing units were occupied by families receiving assistance.[ii] 

 

c.      Loss of units — There has been a reduction of about 140,000 public housing units in the 12-year period 1995 -2007.[iii]      

 

HUD data indicate that it approved removal (via demolition or “disposition”) of more than 210,000 units from the public housing stock between 1995 and 2007, of which 40,000 units remained at the end of 2007.[iv]  (Units disposed of but not demolished may have been converted to vouchers or sold.)  At least 25,000 units have been replaced with new public housing units; at least another 22,000 replacement units have been funded.[v]

 

d.      Condition of units — Most public housing units were built more than 30 years ago[vi], a period which is often considered the “useful life” of residential units absent major renovation.  It is unclear what portion of the older stock has been comprehensively modernized.  (HUD is conducting a new study of public housing capital needs, which Congress directed it to release by July 2009.)  Less than 15 percent of the units are in substandard condition, according to available information from HUD’s inspection system.[vii]  Roughly half of the substandard units are severely distressed.[viii]

 

e.      Location of units — While 60% of public housing units are in central cities, 21% are in rural and non-metropolitan areas, with 19% in suburbs.[ix]  In the 1990s, family developments were much more likely to be in distressed neighborhoods than elderly/disabled developments.  The average public housing unit in a “good” neighborhood was part of a smaller elderly/disabled development in the suburbs; residents of such units were less likely to be minorities or female-headed households and had incomes roughly $2,000 higher than residents of other public housing.[x]  Only 7.5% of all public housing units were in census tracts with poverty rates below 10%, while more than a third were in heavily poor and/or minority census tracts.[xi]  About 1/6 of all public housing units were located in areas considered  “underclass” neighborhoods, based on high incidences of high school drop-out rates, male lack of attachment to the labor force, welfare receipt, and female heads of household.[xii]

 

No published analysis has updated these findings to account for the demolition of about 170,000 units since 1995 or the significant change in many urban neighborhoods after the 1990 census.[xiii] 

 

f.     Development Size — Most public housing developments are relatively small.  In 2000, the average project had 90 units and the median was 50 units.  Of nearly 13,000 total projects, fewer than 800 (or 6 percent) had 250 or more units.  Only 68 projects had more than 1,000 units; 59 of these were in New York City.  Only 34% of public housing units were part of “large” developments (with 250 or more units). In total, 25% of developments had 25 or fewer units; 53% had 50 or fewer units; and 77% had 100 or fewer units.[xiv] 

   

Public Housing Developments and their Unit Counts in 2000, by Development Size

 

1-10 units

11-25 units

26-50 units

51-100 units

101-249 units

250-500 units

501+ units

Number of projects

872

2,349

3,608

3,125

2,127

576

219

Cumulative number of projects

872

3,221

6,829

9,954

12,081

12,657

12,876

Percentage of projects

6.77%

18.24%

28.02%

24.27%

16.52%

4.47%

1.70%

Cumulative percentage of projects

6.77%

25.02%

53.04%

77.31%

93.83%

98.30%

100.00%

Total units

6,014

44,333

141,802

241,236

331,652

194,855

205,080

Cumulative number of units

6,014

50,347

192,149

433,385

765,037

959,892

1,164,972

Percentage of all units

0.52%

3.81%

12.17%

20.71%

28.47%

16.73%

17.60%

Cumulative percentage of units

0.52%

4.32%

16.49%

37.20%

65.67%

82.40%

100.00%

 

f.        Public Housing Agency Size — A total of 3,153 PHAs own public housing.  Slightly less than half of these agencies (1,495) administer both the public housing and Section 8 housing voucher programs.  The remainder (1,658) do not have voucher programs.  (Another 912 PHAs administer voucher programs but do not own public housing.) 

 

The overwhelming majority of PHAs are very small.  Nearly 88% of the PHAs that own public housing have 500 or fewer units.  These agencies manage 29.4% of the public housing units nationwide.  Almost half of agencies with any public housing have fewer than 100 public housing units.  Public-housing-only agencies are more likely to have small stocks of public housing than agencies that administer both public housing and vouchers.  More than half (58.42%) of all agencies that own public housing have fewer than 250 units overall (including vouchers where applicable).  On the other hand, the 34 agencies with the largest stocks of public housing, more than 5,000 units each, own 31.4% of the public housing units.[xv]

   

Characteristics of all agencies administering public housing in 2007, by their number of public housing units

Size (Number of Agency PH Units)

# of agencies

% of agencies with PH units

# of PH units

% of all PH units

# of agencies also having vouchers

% of agencies with both PH units and vouchers

# of vouchers

% of all vouchers administered by agencies with PH units

<100

1,493

47.35%

72,564

6.03%

409

27.36%

107,597

6.93%

100-500

1279

40.57%

281,401

23.37%

741

49.57%

370,122

23.85%

501-1000

202

6.41%

137,903

11.45%

173

11.57%

233,171

15.02%

1001-1500

69

2.19%

83,235

6.91%

66

4.41%

122,125

7.87%

1501-3000

86

2.72%

218,021

18.10%

84

5.62%

339,672

21.88%

5001-7501

12

0.38%

71,350

5.93%

11

0.74%

119,016

7.67%

7501-15000

8

0.25%

79,995

6.64%

8

0.54%

97,457

6.28%

15000+

4

0.13%

259,667

21.56%

3

0.20%

163,211

10.51%

TOTAL

3,153

 

1,204,136

 

1,495

 

 1,552,371[xvi]

 

 

  1. le with disabilities.[xxxiv]  Among families with children, those who receive most of their income from work stay in public housing slightly longer than those dependent on welfare.[xxxv] 

 




[i] HUD sources provide similar but not identical figures for the total number of public housing units in 2007 and the number of units eligible for subsidy.  HUD’s PIC system showed 1,194,958 units in November 2007.  HUD’s 2007 Operating Fund Annual Report indicates a total unit figure for fiscal year 2007 of 1,191,110.  (It is unlikely that the number of units increased after the end of the fiscal year.)   According to the Operating Fund Report, 1,151,819 public housing units were eligible for operating subsidies in 2007, including 84,534 public housing units owned by agencies in the Moving to Work (MTW) demonstration.  HUD’s 2007 Performance and Accountability Report indicates a slightly higher figure of 1,155,377 “units assisted.”  

 

[ii]  HUD’s 2007 Operating Fund Report indicates that there were 999,371 units owned by PHAs not in the MTW demonstration that were occupied by eligible families.  Less than 2,000 additional units were occupied by families not eligible for subsidy (PHA employees or police), and about another 3,000 were approved for special uses (e.g., day care centers).  A large portion of the 84,534 public housing units funded under MTW agreements likely still exist and are occupied by eligible families.  About 70,000 non-MTW units were unoccupied due to vacancies (including disasters) or were in the process of modernization.  Most of these units likely could be made available to eligible families in the near term.  In May 2008, HUD’s Resident Characteristics Report (RCR) data indicated 978,879 households living in public housing.  The RCR data do not include families living in public housing at many MTW agencies.

 

[iii] In 1995, there were about 1,329,000 public housing units.  (This analysis reduces HUD’s figure of 1,393,915 public housing units in 1995 by the approximately 65,000 “Indian” units that were shifted to funding under the Native American housing block grant in 1998.)  In 2007, HUD reports 1,191,110.  See note 1.

 

[iv] The HUD data, accessed May 27, 2008, are available at http://www.hud.gov/offices/pih/systems/pic/sac/demolition_disposition_report2.xls.  HUD reports these data by the date of the request for demolition/disposition, rather than by the date of approval, but requests to remove fewer than 7,000 units were submitted prior to 1995. 

 

[v] The HOPE VI Quarterly report for January – March 2006 (the most recent available) shows 25,401 ACC replacement units actually constructed, out of 47,415 planned through 3/31/06.

 

[vi] In 2000, 57 percent of public housing units were in developments more than 30 years old; 38 percent were in developments 15-30 years old, and only 5 percent of units were in developments built less than 15 years prior.  Stockard, James G. Jr., Gregory A. Byrne, Kevin Day, Gretchen A. Maneval, Lora A. Nielsen, and Katherine James, Public Housing Operating Cost Study.  Cambridge, MA: Harvard University Graduate School of Design (2003).  Available at: http://www.gsd.harvard.edu/research/research_centers/phocs/documents/Final%206.12.pdf.  Since 2000, some of the oldest units have been demolished, while others have aged into the “more than 30 years old” category.

[vii] HUD’s 2007 Performance and Accountability Report indicates that 85.7 percent of public housing units met or exceeded HUD’s physical condition standards.  HUD, FY 2007 Performance and Accountability Report, p. 182.  If HUD used as the “universe” for this figure the 1.15 million units eligible for subsidy, this would indicate that about 165,000 units are in substandard physical condition.  This figure may be high from the perspective of major repairs needed, as units can be judged substandard based on findings that would cost little or nothing to repair, such as lack of smoke alarms or debris on the grounds.  G. Thomas Kingsley et al., Lessons from HOPE VI for the Future of Public Housing, Urban Institute (2004). On the other hand, HUD’s physical condition standards do not measure conformity with modern standards of adequate living or storage space, and do not assess neighborhood or safety factors. 

 

[viii]  After a systematic review of data from HUD’s Management Information Systems, researchers at the Urban Institute estimated that between 47,000 and 82,000 severely distressed units remained in 2007, in addition to about 10,000 units already approved for demolition but not yet demolished.  Turner, Margery Austin, Mark Woolley, G. Thomas Kingsley, Susan J. Popkin, Diane Levy, and Elizabeth Cove, Estimating the Public Costs and Benefits of HOPE VI Investments: Methodological Report. Washington, DC: Urban Institute (2007).  Available at: http://www.urban.org/UploadedPDF/411497_cost_benefits_hope_VI.pdf.  In addition to physical condition (based on adjusted HUD inspection scores to eliminate minor defects), the Urban Institute analysis considered the poverty rate of the census tract — above 30 percent for the entire range — and the percentage of families in the developments relying primarily on welfare income (30 percent or more for the lower estimate and 25 percent or more for the higher estimate).  It appears that the data on income of project residents is from 1998.  If current income data were used and similar thresholds applied, the analysis might yield fewer “severely distressed” projects, given the significant decline in welfare receipt.  See n. __ below.  It is unclear whether the census tract poverty rates were from 1990 or 2000; use of earlier rates also may bias the results. In 2005, GAO estimated that 2% of elderly/disabled developments showed signs of severe distress, compared to 12% of family developments and 15% of mixed-use developments.  U.S. Government Accountability Office, Distressed Conditions in Developments for the Elderly and Persons with Disabilities and Strategies Used for Improvement, GAO-06-163 (2005), available at:  http://www.gao.gov/new.items/d06163.pdf.  Overall, the Urban Institute estimate represents approximately 4 to 7 percent of the stock, while the GAO estimates represent 9 percent.  The discrepancy could reflect additional distressed buildings having been demolished since January 2005.  (Both the GAO and Urban Institute reports use HUD’s Real Estate Assessment Center (REAC) data.) 

[ix] Econsult Corporation, Assessing the Economic Benefits of Public Housing: Final Report, 2007.  Available at: http://www.clpha.org/uploads/final_report.pdf.  It is not clear what data were used to determine these percentages.

[x] A “good” tract is defined as one with a median income above the mean of the median income of all tracts, a proportion of adults over 25 who are college graduates at least one standard deviation over the mean for all tracts, and a proportion of adults in professional/managerial occupations at least one standard deviation over the mean.  Newman, Sandra J. and Ann B. Schnare, “’…And a Suitable Living Environment’: The Failure of Housing Programs to Deliver on Neighborhood Quality,” Housing Policy Debate 8(4): 703-741 (1997).

 

[xi] More precisely, 36.5% of public housing units were in tracts with more than 40% poverty, 50% were in majority-minority neighborhoods, 37.6% with more than 80% minority residents.  Newman and Schnare, n. 10.

 

[xii] Newman, Sandra J. and Ann B. Schnare, “’…And a Suitable Living Environment’: The Failure of Housing Programs to Deliver on Neighborhood Quality,” Housing Policy Debate 8(4): 703-741 (1997); Goering, John, Ali Kamely and Todd Richardson, The Location and Racial Composition of Public Housing in the United States: An Analysis of the Racial Occupancy and Location of Public Housing Developments.  Washington, DC: US Department of Housing and Urban Development (1994).    Because public housing developments can include hundreds of units in dozens of buildings, it is tempting to attribute the census tract’s demographics to the characteristics of the public housing residents themselves.  However, Newman and Schnare note that public housing actually represents a very small percentage of the housing in a census tract.  In only 3 percent of cases do public housing residents represent more than 25 percent of total tract households.

[xiii] Jargowsky, Paul A.  (2003).  Stunning Progress, Hidden Problems: The Dramatic Decline of Concentrated Poverty in the 1990s.  Brookings Institution Center on Urban and Metropolitan Policy.  Available at: http://www.brookings.edu/~/media/Files/rc/reports/2003/05demographics_jargowsky/jargowskypoverty.pdf.

 

[xiv] U.S. Department of Housing and Urban Development, Picture of Subsidized Households – 2000.  Available at: http://www.huduser.org/picture2000/index.html.  These are the latest data that HUD has made public.  It is likely that the current profile is somewhat smaller, as larger projects as well as a large number of family high-rises have been demolished in the intervening years.  The figures in the text exclude 839 developments labeled by HUD as “scattered site,” with a total of 45,046 units, frequently single-family homes.

[xv] CBPP analysis of HUD RCR data.  (The data on authorized vouchers are as of January 2007; the public housing data are as of 3/31/07.)  The data do not indicate cases in which several small agencies consolidate their resources and share a single administrative framework.  Generally, mixed agencies have similar numbers of public housing units and vouchers.  There are, however, many exceptions to this rule.  For 274 agencies, representing 18.33% of all mixed agencies, the number of vouchers is more than 4 times the number of public housing units.  Large proportions of these agencies are in western states such as California, Colorado, Oregon, and Texas.  There are 62 mixed agencies with a much larger stock of public housing than vouchers (ratio greater than 4:1), representing 4.15% of all mixed agencies.  These agencies tend to have a significantly smaller total stock than the agencies with many more vouchers than public housing units.  They cluster in the Mid-Atlantic region (New York, New Jersey), but are spread more widely across the country than the voucher-heavy mixed agencies. 

 

[xvi] This is the total number of vouchers administered by agencies that have public housing units.  An additional 612,438 vouchers are administered by voucher-only agencies, for a total of 2,164,809 vouchers overall as of January 2007.

 

[xvii] As explained further below, HUD’s March 2008 RCR count of 2,119,000 individuals fails to include all agencies.  (The May data show only 2,000 more residents than the March data we analyzed in detail.)  If the 1,084,000 public housing units eligible for operating subsidy in 2007 (see note i) that were either occupied or owned by a PHA in the MTW demonstration all housed the average size household of 2.2 persons, public housing would have 2,384,590 residents.

 

[xviii] These percentages are based on HUD’s RCR data accessed May 7, 2008 (for the period from January 1, 2007 through April 30, 2008).  They are virtually the same as the data reported by GAO in 2005.  See the next note.

 

[xix] U.S. Government Accountability Office, Distressed Conditions in Developments for the Elderly and Persons with Disabilities and Strategies Used for Improvement, GAO-06-163 (2005). Available at: http://www.gao.gov/new.items/d06163.pdf.

 

[xx] McNickle, Larry, “Service Coordinators: An Essential Link for Seniors to Age in Place,” Journal of Housing & Community Development 64(4): 40-47 (2007).  (The source of the cited figure is not provided.)

 

[xxi] U.S. Government Accountability Office, Distressed Conditions in Developments for the Elderly and Persons with Disabilities and Strategies Used for Improvement, GAO-06-163 (2005). Available at: http://www.gao.gov/new.items/d06163.pdf.

 

[xxii] In the HOPE VI Panel Study, which addressed only “family” developments, 13 percent of the households were headed by a person age 62 or older.  These seniors had been both in their units and in their public housing communities longer than their younger counterparts.  Although older residents in the HOPE VI Panel Study are disproportionately minority women, there are more male-headed households than in the younger age groups.  Smith, Robin E. and Kadija Ferryman.  Saying Goodbye: Relocating Senior Citizens in the HOPE VI Panel Study.  Washington, DC: Urban Institute (2006).  Available at: http://www.urban.org/UploadedPDF/311279_Roof_10.pdf.

 

[xxiii] According to HUD’s Resident Characteristics Report (RCR) for the period from January 1, 2007 to April 30, 2008, 46% of public housing households are African-American, 51% are white, 2% are Asian, and 1% are American Indian.  Of those households, 23% are Hispanic/Latino.  (The RCR website does not cross-reference “race” and “Hispanic origin.”)  Using 2004 data, HUD reported that 44% of public housing households were non-Hispanic blacks, while 32% were non-Hispanic whites, 21% were Hispanic, 2% were Asian and 1% were Native American.  HUD, “Sixth Annual Report to Congress on Public Housing and Rental Assistance Programs,” October, 2005 (known as the “QHWRA Report”).  Both sets of figures are based on data reported by PHAs to the Multifamily Tenant Characteristics System (MTCS).  The absence of data from some agencies with large numbers of public housing units may somewhat understate the percentage of minority households.  (E.g., no data are reported for Oakland (CA), the District of Columbia, Atlanta, Baltimore or Philadelphia, among the 16 agencies that were part of the MTW demonstration during this period and have no resident data on RCR.)  Moreover, these figures include only the small number of Gulf Coast residents who had returned to public housing following Hurricane Katrina.

 

[xxiv] Goering, John, Ali Kamely and Todd Richardson,  “The Location and Racial Composition of Public Housing in the United States: An Analysis of the Racial Occupancy and Location of Public Housing Developments,”  Washington, DC: US Department of Housing and Urban Development (1994).  

[xxv] RCR data for the period from January 1, 2007 to April 30, 2008 show an average public housing household annual income of $12,717. Eighteen percent of households have less than $5,000 of annual income, while 16 percent have incomes of more than $20,000.  In relation to local incomes, 63 percent have incomes less than 30% of the area median income (AMI) level, adjusted for family size; 16 percent have incomes between 30-50% AMI, 7 percent have incomes between 50-80% AMI, and 2% have incomes above 80% AMI.  No income data was available for 12% of households.  The HOPE VI Panel Study found that while the elderly residents also typically were extremely poor, they had somewhat higher incomes on average than their younger neighbors, which is likely because SSI and Social Security provide a better safety net than what is now available for families.  Smith, Robin E. and Kadija Ferryman, Saying Goodbye: Relocating Senior Citizens in the HOPE VI Panel Study, Washington, DC: Urban Institute (2006).  Available at: http://www.urban.org/UploadedPDF/311279_Roof_10.pdf.

 

[xxvi] In 2004, 19 percent of families with children in public housing reported welfare as their largest source of income, compared with 35 percent in 1997.   Paralleling this change, in 2004 half of families with children reported some earned income, and for 47 percent work was their largest source of income, while in 1997 only 36 percent relied primarily on earned income.  HUD’s Sixth Annual QHWRA Report (2005).  As noted above, 15 percent of families with children are headed by a person who is elderly or has disabilities.  It appears that the HUD data in this note include these households in the “families with children” category.

 

[xxvii] Riccio, James A.,. Mobilizing Public Housing Communities for Work: Origins and Early Accomplishments of the Jobs-Plus Demonstration, MDRC (1999).  Available at:  http://www.mdrc.org/publications/75/full.pdfMartinez, John M., The Employment Experiences of Public Housing Residents: Findings from the Jobs-Plus Baseline Survey,  MDRC (2002).  Available at: http://www.mdrc.org/publications/25/full.pdf; Levy, Diane K. and Mark Woolley,  Relocation is Not Enough: Employment Barriers amond HOPE VI Families, Washington, DC: Urban Institute (2007). Available at: http://www.urban.org/UploadedPDF/311489_HOPEVI_Employment.pdf

 

[xxviii] Popkin, Susan J., Mary K. Cunningham, and Martha Burt, “Public Housing Transformation and the Hard-to-House,” Housing Policy Debate 16(1): 1-24 (2005). 

[xxix] GAO, n. xix above.

 

[xxx] Smith, Robin E. and Kadija Ferryman, Saying Goodbye: Relocating Senior Citizens in the HOPE VI Panel Study,  Washington, DC: Urban Institute (2006).  Available at: http://www.urban.org/UploadedPDF/311279_Roof_10.pdf.

 

[xxxi] Length-of-stay analyses are methodologically complicated.  Using somewhat different methods to analyze HUD data, one study found a median length of stay in 2000 of 4.69 years, while another found a median of 3.97 years for the period 1995 – 2002.  Lubell, Jeffrey M., Mark Shroder, and Barry Steffen, "Work Participation and Length of Stay in HUD-Assisted Housing," Cityscape 6(2):207-223 (2003); Thompson, Dianne T., “Evaluating Length of Stay in Assisted Housing Programs: A Methodological Note,” Cityscape 9(1):217-238 (2007).  Thompson also estimated a median “survival” time in the public housing program of 8.3 years.  This methodology aims to account for the likely lengths of stay of households still residing in assisted housing at the time of the analysis.

[xxxii] The Thompson analysis found a mean length of stay in public housing of 7.46 years; the analysis by Lubell et al. found a mean of 8.5 years.  Thompson also analyzed length of stay for families that shifted from the public housing to the voucher program (or vice versa) during the 8 year period she studied.  Including this latter group is particularly important in light of the substantial number of public housing units demolished during this time period, as noted above; families living in these units often shifted to the voucher program.  These “mixed program” households had a mean length of stay of 6.33 years, and a median of 5.25 years.

 

[xxxiii] Thompson, p. 224.  The analysis by Lubell et al. highlights the segment of long stayers who raised children in public housing and remained after their children were grown.

 

[xxxiv] Lubell et al. found that families with children with a non-elderly, non-disabled head had a median length of stay of 3.17 years and an average (mean) length of stay of 5.59 years, while Thompson found a slightly longer median length of stay of 3.86 years and a shorter mean length of stay of 2.12 years for such households.  For the elderly, Lubell et al. found a median length of stay of 8.44 years, and for households headed by a person with disabilities, 4.05 years.  (Thompson does not present data separately for these groups.)

 

[xxxv] Lubell et al. found that the median length of stay for working families with children (receiving more than half of their income from earnings) is 3.59 years, exceeding the median of is 2.73 years for families with children for whom more than half of their income comes from welfare.  For families with zero reported income, length of stay was shorter than all other groups at every percentile. 

 

[xxxvi] About 60 percent of housing vouchers are used by families with children, unlike in the public housing program where families with children make up about 40 percent of households.  Lubell et al. found that families with children with a non-elderly, non-disabled head lived in public housing for about 1.6 years more on average than they received voucher assistance; Thompson found less of a difference.  For all households, Thompson found a mean length of stay of 7.46 years in public housing compared with 4.24 years in the voucher program; Lubell et al. found a mean of 8.5 years in public housing compared with 4.75 years in the voucher program.