Skip to main content

Homelessness: HUD Should Help Communities Better Leverage Data to Estimate Homelessness

GAO-22-104445 Published: Nov 22, 2021. Publicly Released: Nov 22, 2021.
Jump To:

Fast Facts

The Department of Housing and Urban Development takes a "point-in-time" count to gauge how many people are experiencing homelessness on a single night. HUD relies on local communities to provide the count.

During the pandemic, HUD supplemented in-person enumerations by allowing local communities to use data collected by public and nonprofit agencies on people using their services—an approach that likely improved accuracy of point-in-time counts.

We recommended that HUD provide communities more information about how best to use such data to improve point-in-time counts moving forward.

A person sleeping on a bench

Skip to Highlights


What GAO Found

The Point-in-Time (PIT) count is a nationwide count of people experiencing homelessness on a single night, conducted by Continuums of Care (CoC)—local planning bodies that coordinate homelessness services. The Department of Housing and Urban Development (HUD) allows CoCs to use different methods to estimate homeless populations—including a census (complete count), sampling, or a combination of these. For counting unsheltered individuals (those on the street or in other uninhabitable places), HUD requires CoCs to use in-person methods—for example, by having enumerators visually locate and attempt to ask questions of these individuals on the night of the count. HUD permits CoCs to also use administrative data—that is, records collected by public and nonprofit agencies on people who use their services. However, HUD does not provide CoCs with examples of how to extract and use administrative data for the unsheltered count. By doing so, HUD could help improve the quality and consistency of CoCs' estimates and position CoCs to provide better estimates, particularly if in-person counts are again disrupted, as they were in 2021, during the COVID-19 pandemic.

The U.S. PIT count is similar to Canada's and England's approaches in that they are nationally administered and localities can choose among various approved methods to conduct in-person local counts. The Netherlands and Australia use more centralized methods and statistical analyses to develop estimates. For example, Australia produces an estimate using data from the general census of the population.

Little comprehensive data exist on PIT count costs, but a GAO survey of 41 CoCs provided information on funding sources and key resources required from their most recent unsheltered PIT count prior to 2021:

  • Of the 41 CoCs, 31 used HUD funds, 19 used state or local funds, and 10 used private donations (often in combination with government funds).
  • All 41 CoCs reported using volunteers to complete their PIT counts, with large cities using the most volunteer hours.
  • Respondents reported an average of 4.8 work hours (paid staff and volunteers) for every person counted in their PIT count of unsheltered individuals. The most common PIT count costs were for incentives for volunteers and meals.

Examples of Homeless Encampments in Oakland, California, in 2021

Examples of Homeless Encampments in Oakland, California, in 2021

Why GAO Did This Study

HUD's PIT count is a key tool for estimating the size of the U.S. homeless population. However, developing an accurate understanding of the extent of homelessness is challenging due to the hidden nature of the population. Further, some members of Congress and others have raised questions about the reliability of HUD's estimates.

GAO was asked to review the PIT count and alternative methods for estimating the size of homeless populations. This report (1) examines communities' approaches for counting people experiencing homelessness and HUD's guidance for using these approaches, (2) describes approaches used by selected foreign countries to estimate their homeless populations, and (3) describes what is known about funding sources and resources expended by selected communities in conducing the PIT count.

GAO conducted a literature review to identify methods to estimate homelessness and selected four countries for case study based on a literature review and recommendations from researchers. GAO also surveyed a nongeneralizable sample of 60 CoCs and received responses from 41 of them about PIT count costs and funding sources, reviewed agency guidance and documents, and interviewed U.S. and foreign government officials.


GAO recommends that HUD provide CoCs additional information about how they can use administrative data to improve the accuracy of their unsheltered PIT count. HUD agreed with our recommendation.

Recommendations for Executive Action

Agency Affected Recommendation Status
Department of Housing and Urban Development HUD's Deputy Assistant Secretary for Special Needs should provide additional tools and information about how CoCs can use Homeless Management Information System and other administrative data to help improve the accuracy of their unsheltered Point-in-Time count. (Recommendation 1)
Open – Partially Addressed
As of February 2024, HUD has begun to take steps to address this recommendation. For example, in preparation for the 2024 unsheltered PIT count, HUD officials provided CoCs informal guidance on how to use administrative data to improve the accuracy of their PIT counts. HUD officials noted that a larger initiative is underway, pending budgetary resources, to develop more formal guidance. To fully implement this recommendation, HUD should develop more formal guidance or information and distribute it widely to all CoCs.

Full Report

Office of Public Affairs


CensusCommunitiesData collectionGrant programsHomelessnessHousingpandemicsUrban developmentGeographyCensus takers