 |
|---|
HMIS Snapshot: 2021 Homelessness Prevalence Estimate in Austin/Travis County
Ending Community Homelessness Coalition (ECHO)
May 21, 2021
About ECHO
The Ending Community Homelessness Coalition (ECHO) is the Austin/Travis County Continuum of Care lead agency tasked with planning and coordinating community-wide strategies to end homelessness in the Austin and Travis County geographic region. We work in collaboration with community nonprofits and government agencies to coordinate services and housing resources for people experiencing homelessness (PEH) in our community. ECHO manages the local Homeless Management Information System (HMIS) database and uses research and evidence-based practices to advocate for the resources to bring the local Homelessness Response System to scale and meet our community’s goal of making homelessness rare, brief, and non-recurring.
Acknowledgments
ECHO would like to acknowledge our incredible partner organizations, Austin and Travis County civic leaders, and the community at large for your support in the COVID-19 emergency response. Thank you for your continued commitment to ending homelessness in Austin/Travis County. This has been a very different year for our Homelessness Response System, but thanks to the hard work of our dedicated coalition, we have been able to continuously serve people experiencing homelessness while prioritizing the health and safety of all.
Overview
All Continuums of Care (CoCs) are required by federal statute to count people experiencing unsheltered and sheltered homelessness in their communities biennially. These Point in Time (PIT) Counts are point prevalence estimates of unsheltered and sheltered homelessness and are typically conducted on one day in the last ten days of January.1 The Austin/Travis County CoC has chosen to conduct the PIT Count annually. In 2020, 886 staff and volunteers conducted the count on January 25th. The Housing Inventory Count (HIC) is an annual inventory of the Homelessness Response System that provides a snapshot—based on data from the local Homeless Management Information System (HMIS) and completed on the same day as the PIT Count—of the number of units and beds available to people experiencing homelessness. Because service providers also provide data on how many people are using those units and beds, the HIC also provides a point prevalence estimate of the number of people experiencing sheltered homelessness. The PIT Count and HIC are both required by the United States Department of Housing and Urban Development (HUD) and are conducted by CoCs across the United States. The goal of these counts is to understand the number and demographic characteristics of people experiencing homelessness so that funding and services may be targeted appropriately. The data from this annual count contribute to both local and national efforts to end homelessness.2
In November of 2020, HUD’s Office of Special Needs Assistance Programs (SNAPS) announced that communities were required to follow reporting requirements to submit data for the sheltered count portion and allowed flexibility for communities to change their methodology and/or cancel the unsheltered portion of the PIT Count. CoCs were instructed on how to make the exception request to be approved by HUD, which included the need to document local emergency operations enacted by local government health entities and evidence of why the exception to cancel or change the methodology was necessary.
Due to COVID-19, ECHO, with approval from the CoC’s governing body, the Membership Council, decided that conducting an in-person unsheltered count in late January 2021 would put staff, volunteers, and people experiencing homelessness at risk of COVID-19 contraction (see more in the Background section below). Therefore, in November 2020, ECHO submitted a request to HUD, per the department’s guidelines, to not conduct an unsheltered PIT Count in 2021 and instead use HMIS data to estimate the extent of unsheltered homelessness in our community (see more in the Methodology section below).3 HUD approved the exception request in late November 2020.
Although ECHO and partner organizations did not conduct an unsheltered PIT Count in January 2021, the findings in this report, based on an alternative method using data from HMIS, provides our community with an informed estimate of the scope of unsheltered homelessness seeking permanent housing services in Austin/Travis County.
The methodology used in 2021 differs from the PIT (in-person) Counts conducted in the past. This year’s overall, demographic, and sub-population counts are all based on estimates using HMIS. HMIS includes information from people who have engaged with the Homelessness Response System at some point in time as opposed to individuals identified by volunteers as experiencing unsheltered homelessness during our traditional PIT Counts. There are advantages and disadvantages to the two methods, but because they are very different from each other, we do not attempt to compare this year’s counts with past years’ PIT Counts and urge readers to also not make those comparisons.4 We discuss the limitations of the HMIS method in the Limitations section below.
Background
COVID-19 and Homelessness
Austin/Travis County requested an exception from the standard methodology for the 2021 unsheltered count due to concerns about the ability to conduct a thorough count while ensuring the safety of our unsheltered neighbors, system staff members, and the volunteers necessary to thoroughly survey the area. In November 2020, two months prior to the PIT Count time period in January 2021, ECHO cited several reasons in a request to HUD for approval of the use of an alternative methodology to the count. This request came following guidance from HUD regarding the potential for exceptions from the unsheltered count in light of COVID-19, and Austin/Travis County is one of many communities across the country that were approved to use an alternative methodology.
In November 2020, at the time that ECHO submitted the exception request, Texas had more COVID-19 cases than any other state in the United States, and Austin/Travis County cases were on the rise and projected to continue to rise.5 Travis County had recently entered Stage 4 Risk-Based Guidelines, which indicated that individuals at high risk of complications due to COVID-19 should avoid gatherings of more than 2 people, and low risk individuals should avoid gatherings of more than 10 people.6 An early assessment of our local population found that more than half of people experiencing homelessness in our Coordinated Entry system7 met the Centers for Disease Control and Prevention (CDC) COVID-19 high risk criteria, and researchers across the country argued that people experiencing homelessness were at higher risk of COVID-19 infection than the general population.8
If the PIT Count was to be conducted as normal, the capacity for recruiting and training surveyors would be limited due to local regulations, and fewer surveyors would have been able to conduct the count. This would have likely led to an undercount of unsheltered homelessness. The unsheltered count would have needed to rely on Homelessness Response System agency staff, who are essential workers responsible for providing resources to people experiencing homelessness during COVID-19. Homelessness Response System staffing was already strained due to COVID-19 exposure, quarantining, and illness. To have pulled staff away from the emergency response to conduct an in-person count would have contributed more risk to agency staff and placed more strain on the system.
In the event that someone tested positive for COVID-19 following the unsheltered count, the lack of contact tracing capacity in Texas would have limited the ability to prevent further spread of the virus.9 The risk of potentially creating hot spots by conducting an in-person count was too high, especially considering the vulnerability of people experiencing homelessness to complications due to COVID-19.
Finally, to support people living unsheltered during COVID-19, many volunteers and organizations donated tents to encourage social distancing in encampments. The increase in the number of people living in tents was a good public health measure but would have likely contributed to an undercount in an in-person PIT Count—because PIT surveyors are only allowed to count and record survey responses if they can confirm someone is living in a tent. For all these reasons, the Austin/Travis County CoC determined that conducting an in-person unsheltered count would not be responsible, feasible, or accurate.
Response to COVID-19: ProLodges and Expansion of Rapid Re-Housing
To prevent the spread of COVID-19, shelters in the Homelessness Response System temporarily reduced capacity, in alignment with the facility layout considerations of CDC guidance for homeless service providers. Because of this reduced shelter capacity, and perhaps in addition to individual-level concerns about COVID-19 transmission risk in congregate shelter environments, people experiencing homelessness in Austin/Travis County were more likely to be in unsheltered living situations than in shelters in January of 2021. The City of Austin, in response to COVID-19, established Protective Lodges (ProLodges)—temporary, non-congregate crisis facilities—that have compensated in part for the reduction in existing crisis bed capacity in the Homelessness Response System.
Five hotels were being used as ProLodges during the COVID-19 pandemic. The ProLodge project launched immediately in response to the pandemic, in late March 2020, with units dedicated to people experiencing homelessness in need of a space to self-quarantine for unconfirmed exposure and/or who meet CDC high-risk criteria for COVID-19 complications.10 People staying in the ProLodges receive on-site services, including health/medical services, case management, substance use resources, and dedicated housing exit strategy support. ProLodge stays are temporary, and the goal is to move people into permanent housing when they leave from the ProLodges.11
Dedicated funds toward the COVID-19 crisis response have also been used to expand Rapid Re-Housing capacity. Rapid Re-Housing is a permanent housing project type that provides short to medium-term rental assistance, case management, and housing location services to permanently house participants in Austin/Travis County. The majority of households permanently housed in our system each year are housed through Rapid Re-Housing programs. The Austin/Travis County CoC continues to center permanent housing solutions in its Homelessness Response System strategies.
Although it is impossible to discern the causal impact the COVID-19 response had on prevention of transmission among people experiencing homelessness, we know that these efforts allowed people experiencing homelessness to socially distance, isolate, and shelter in place, as well as receive services, such as COVID-19 testing. The system was able to efficiently adapt to serve countless people in crisis with new and expanded services while keeping staff and clients safe.
Increased Visibility of Unsheltered Homelessness in Austin/Travis County
The Austin/Travis County Homelessness Response System functions nearly at full capacity, serving as many clients as possible with limited resources. Many economic and social conditions and systemic inequities have led to a higher rate of inflow to homelessness than can be compensated for with housing resources allocated toward outflow from homelessness. With current resources, it is not possible to house every person experiencing homelessness in Austin/Travis County. With existing shelter capacity, it is also not possible to provide shelter for all. The CDC directed communities to connect people experiencing unsheltered homelessness with basic needs, including tents, phone chargers, and bottled water, to enable them to abide by COVID-19 safety guidance, including social distancing and quarantining. Thus, as a part of the COVID-19 crisis response, a tent and camping supplies distribution system has been essential. People experiencing unsheltered homelessness are better able to abide by COVID-19 safety guidance, including social distancing and quarantining, when tents are available. Tent and camping donations were collected and distributed in an effort to prevent transmission throughout unsheltered communities.
One of the consequences of making tents and camping supplies more accessible to people experiencing unsheltered homelessness is an increase in the total area occupied by encampments. There are simply more tents per unsheltered group of people than there were before, and these tents, especially when distanced from one another in a given area, increase the visibility of unsheltered homelessness. CDC COVID-19 guidance also recommends against displacing or disrupting encampment locations if individual housing options are not available. As the Homelessness Response System continues to function at capacity, encampments have been allowed to remain where they are during the pandemic. This could also increase the visibility of encampments, as people experiencing unsheltered homelessness can more easily accumulate belongings over time when not being required to frequently change locations.
Researchers that have compared encampment trends in the United States have found no direct relationship between the emergence of sizable encampments and a general increase in the overall population of people experiencing homelessness.12 An increase in the area occupied by encampments in Austin/Travis County may suggest changes to the scale of unsheltered homelessness, but this relationship is not direct and there are many extraneous factors involved. The visibility and scale of unsheltered homelessness must be considered as separate matters.
Another element of the COVID-19 response that could have impacted the visibility of encampments is the stand up of food distribution to encampments. In an effort to ensure that people experiencing unsheltered homelessness had the ability to shelter in place to prevent the spread of COVID-19, the City of Austin established the Eating Apart Together (EAT) initiative, a grocery and meal distribution project that delivered approximately 1 million meals to unsheltered communities since the onset of the pandemic. This initiative has allowed unsheltered people to stay put in encampments while maintaining the ability to meet food needs. The activity of distribution at encampments also may have drawn more attention to them and made them more visible.
Methodology
To estimate the number of people experiencing unsheltered homelessness who are seeking services, we used the by-name list (BNL) extracted from HMIS on January 28th, 2021. The BNL is a list of all heads of household who have entered the Homelessness Response System, completed a Coordinated Assessment (CA), and are either on a prioritization or program referral list. The prioritization list is a list of heads of household who completed a CA and have expressed a need for a permanent housing (PH) program intervention—either Rapid Re-Housing (RRH) or Permanent Supportive Housing (PSH)—but have not been connected to a specific program yet. Some people on this list are waiting for RRH, some for PSH, and others for either. The program referral list is a list of heads of household whose referral has been accepted by a PH program but who have not moved into housing yet. This process is locally known as the Coordinated Entry System.
Coordinated Entry is a consolidated housing application process whereby anyone experiencing homelessness in Austin/Travis County can apply for all possible housing programs at the same time and get the same answer about all immediate and future program openings.
Core Driving Principles are:
- Housing First – Housing is a basic human right, not a privilege or a reward for good behavior.
- Client Choice – Clients are the experts on their own experiences and needs, and most people experiencing homelessness today do NOT want to be homeless.
- Human Dignity – All people deserve to be treated with respect, empathy, and personal autonomy, regardless of past or current behavior.
- Equity – Systematic privilege is real and pervasive, and we must take affirmative steps in systems design and implementation to achieve equitable outcomes for groups who are systematically oppressed.
The BNL is updated daily by ECHO and service provider staff. People come off the list when they have moved into housing, a staff member knows they have found permanent housing on their own, or ECHO and service providers have been able to document to a reasonable level of certainty that a client is no longer experiencing homelessness in Austin/Travis County. We cannot know for sure whether everyone on this list is still actively looking for housing or—if they are marked as “unsheltered”—whether they are currently experiencing unsheltered homelessness. That is because they could have resolved their homelessness and a case manager does not know about that yet. The other most common reasons for a client exiting the list for reasons other than housing are leaving the area, long-term medical institutionalization, long-term incarceration, or death.
We analyze the BNL in two steps. First, we estimate the average length that a client on the by-name list experienced unsheltered homelessness. To do so, we examine the length of homelessness for all people experiencing unsheltered homelessness on the by-name list who had sought any service from the Homelessness Response System (HRS) in the 12 months prior to the 2021 PIT Count date (January 28th).13 We then narrowed in on clients whose living situation prior to seeking services from the HRS was unsheltered homelessness and, for each client, estimated that person’s length of unsheltered homelessness. We then calculated the average length of unsheltered homelessness for those clients and estimate that this is the average amount of time that people experience unsheltered homelessness in Austin/Travis county.14 Table 1 in the appendix shows the average length that people experienced unsheltered homelessness in the last three years.
After estimating the average length of unsheltered homelessness, we use the BNL and estimate the number of people on that list who both 1) had “touched the system” (e.g. taken an assessment, stayed in a shelter, visited a day resource center) within the average length of stay calculated in step 1 and, 2) were marked as “unsheltered” on the BNL. We then took into account household size of all of those head of household clients on the by-name list to estimate the total number of people experiencing unsheltered homelessness on January 28th, 2021.
Results
The results in the following sections focus on four primary questions. All of the data on unsheltered homelessness come from HMIS and should not be compared to PIT Count numbers from previous years.
What was the point prevalence of homelessness in Austin/Travis County in January 2021?
What are the main subpopulation and demographic breakdowns of people experiencing homelessness at that same time?
How do those breakdowns differ by whether someone was living unsheltered, in a shelter, or in a ProLodge?
How do subpopulation and demographic breakdowns in 2021 compare to 2020, using the same HMIS/BNL method we use in this year’s report?
Notes
When collecting information in the HMIS database, households self report information they are comfortable disclosing. HRS staff recognize the trauma associated with homelessness and prioritize the safety concerns of those experiencing homelessness, so some data elements are not collected. Therefore, the total numbers for the different tables and figures may differ. We note these totals and discrepancies throughout the report. We report counts and percentages to provide both precision and a basis for comparison. For percentages reported, we use the number of people who self reported each characteristic as the denominator (i.e., people whose, say, race or gender we do not have information on are not counted when calculating the percentage).
In tables where we compare characteristics by living situation or between years (tables 4 through 11), we test for statistical significance using a chi-squared test, a common measure used to assess whether two categorical variables statistically significantly differ. By most standards, a p-value above 0.05 means that the difference (e.g., between 2020 and 2021 for a given variable) is not statistically significant. For space considerations, we do not show p-values in tables 4 through 8 but indicate whether the differences are statistically significant in the notes below the table. We show p-values in tables 9 through 11 (comparisons between 2020 and 2021).
Estimated Prevalence of Homelessness in 2021 Using Alternative HMIS Method
Table 1 and figure 1 below show the point prevalence of homelessness among people seeking services, based on HMIS data, in Austin/Travis County. The results point to three main findings about the prevalence of homelessness in the Homelessness Response System (HRS) using HMIS data since 2019. First, the total number of people experiencing homelessness and seeking services from the HRS increased by 5.6% from January 2019 to January 2020 but decreased slightly (-1.1%) from January 2020 to January 2021. Second, in the same time period, unsheltered homelessness increased by 21.9% between 2019 and 2020 but dipped by 1.1% from 2020 to 2021. And third, sheltered homelessness has seen two consecutive years of decline, falling about 20% from 2019 to 2020 and another 24% from 2020 to 2021. Despite the decline in sheltered homelessness in the past year, 209 people were living in the ProLodges—the temporary, non-congregate crisis facilities that have been used for crisis lodging as part of the City of Austin’s emergency response to COVID-19. This means that the number of people seeking temporary lodging, either through congregate shelters or the non-congregate ProLodges, in January 2021 was 922, a 1.1% decrease compared to the number of people seeking temporary lodging in 2020.
- The HMIS and the in-person unsheltered PIT Count methods are not interchangeable and cannot be compared to estimate year to year changes.
- While the visibility of homelessness may have increased, estimates of the scale of unsheltered homelessness have not.
- We anticipate a possible delayed rise in measures of the scale of homelessness when the eviction moratorium and unemployment benefits end, as rents continue rising, and as people begin to feel safer accessing services.
Table 1. Point Prevalence of Homelessness in Austin/Travis County Homelessness Response System, 2019-2021
| 2019 | 2020 | % Change | 2021 | % Change | |
|---|---|---|---|---|---|
| Crisis Lodging | 1169 | 932 | −20.3% | 922 | −1.1% |
| Shelter | 1169 | 932 | −20.3% | 713 | −23.5% |
| ProLodges | NA | NA | NA | 209 | NA |
| Unsheltered | 1855 | 2262 | 21.9% | 2238 | −1.1% |
| Total | 3024 | 3194 | 5.6% | 3160 | −1.1% |
Sources: Sheltered and unsheltered counts come from HMIS, and the ProLodges count comes from the City of Austin’s Emergency Operations Center.
Date of 2021 Count: January 28, 2021
Sources: Sheltered and unsheltered counts come from HMIS, and the ProLodges count comes from the City of Austin’s Emergency Operations Center.
Sources: Homeless Management Information System (HMIS), the City of Austin, and the Texas Demographic Center
The left-hand panel of figure 2 above replicates the numbers from table 1, and the middle panel of figure 2 shows the population of Travis County, which has increased by about 3.4% from 2019 to 2021. Taking into account population, the right panel in figure 2 shows the number of people experiencing homelessness per 1,000,000 residents. That number has been relatively stable since 2019, with a three-year average of 2415 or 0.24% of the Travis County population.
Subpopulation and Demographic Characteristics of Homelessness in 2021 Using Alternative HMIS Method
Figure 3 and tables 2 and 3 show the main subpopulation and demographic breakdowns of all people experiencing homelessness in the Homelessness Response System in January 2021. The numbers show that a large majority (69%) of people experiencing homelessness live with at least one disabling condition and half are chronically homeless.15 Forty-one percent of people experiencing homelessness reported being survivors of domestic violence.
Table 2 shows the gender and age breakdown of people experiencing homelessness. Notably, 21% of people are under the age of 18 and approximately 0.6% of the population identify as transgender or gender non-conforming. Table 3 shows race and ethnicity breakdowns. Because of the way that some service providers submit their data to ECHO, we could not disaggregate race and ethnicity (for example, to get an estimated number of non-Hispanic White people in the system). The numbers do show that large proportions of the population are Black (37%) and Hispanic or Latino (32%). Most notably, Black people are over-represented among people experiencing homelessness, as compared to the population of Austin/Travis County.
Figure 4 shows some of the major disparities in the Homelessness Response System. In addition to showing the large over-representation of Black people among people experiencing homelessness (i.e., Black people make up 37% of people experiencing homelessness, but only 8% of the Travis County population), figure 4 shows that people experiencing homelessness are much more likely to have a disability than the rest of the population. Veterans are also slightly over-represented among people experiencing homelessness.
Notes: 1) Demographic characteristics are not mutually exclusive. 2) “Family with Children” includes all people living in households with a person under the age of 18. 3) “Living in Youth-Headed Household” includes all people living in a household headed by someone under the age of 25.
Table 2. Gender and Age Breakdown, 2021: Percent of People Experiencing Homelessness
| Characteristic | N = 3151 |
|---|---|
| Gender | |
| Male | 1708 (56%) |
| Female | 1317 (43%) |
| Transgender | 15 (0.5%) |
| Gender non-conforming | 2 (<0.1%) |
| Unknown | 109 |
| Age Breakdown | |
| Under 18 | 669 (21%) |
| Between 18 and 24 | 251 (8.0%) |
| Over 24 | 2223 (71%) |
| Unknown | 8 |
Table 3. Race and Ethnicity Breakdown, 2021: Percent of People Experiencing Homelessness
| Characteristic | N = 3151 |
|---|---|
| Race | |
| American Indian or Alaska Native | 29 (1.0%) |
| Asian | 15 (0.5%) |
| Black | 1079 (37%) |
| Native Hawaiian or Other Pacific Islander | 5 (0.2%) |
| Two or more races | 200 (6.8%) |
| White | 1624 (55%) |
| Unknown | 199 |
| Ethnicity | |
| Hispanic or Latino | 968 (32%) |
| Non-Hispanic/Non-Latino | 2060 (68%) |
| Unknown | 123 |
Note: Race and ethnicity categories are not mutually exclusive.
Sources: Austin/Travis HMIS; American Community Survey.16
Notes: 1) Race and ethnicity categories are not mutually exclusive. 2) Race categories included in plot are ones with a substantive difference between people experiencing homelessness and all people living in Travis County.
Subpopulation and Demographic Characteristics of Homelessness in 2021 by Living Situation
Tables 4 through 8 show whether people who live in a shelter, unsheltered, or in one of the five ProLodges differ on the subpopulation and demographic characteristics presented above. The biggest differences in table 4 indicate that people living in the ProLodges are more likely to have a disability, people living in a shelter are more likely to be part of a family or a veteran, and people living unsheltered are more likely to be chronically homeless and domestic violence survivors.
The biggest differences related to gender, race, and ethnicity suggest that people living in the ProLodges are more likely to be male, shelters have a slightly higher representation among Black people, and White people are slightly more likely to be unsheltered. Further, Hispanics or Latinos are more likely to live in shelters than unsheltered or in one of the ProLodges.
Table 8 shows the impact of Living Situation on vulnerability to various circumstances related to homelessness. A majority of people in the ProLodges, sheltered and unsheltered people experiencing homelessness have not been in stable housing for at least one year. People experiencing sheltered homelessness are more likely than people living unsheltered or in a ProLodge to have been in stable housing within the past year, and people in the ProLodges are the least likely. People in shelters are more likely to report currently being able to meet their basic needs than are unsheltered people or people in the ProLodges. Trauma was reported as a common cause or contributing factor that led to homelessness across all groups, but especially so among ProLodge guests. People in the ProLodges are also more likely to have been hospitalized and/or to have had an emergency room visit in the last 6 months. Unsheltered people are most likely to report having recently been attacked or beaten up.17
Table 4. Subpopulation Differences in 2021 by Living Situation: Percent of People Experiencing Homelessness
| Characteristic | ProLodge, N = 200 | Sheltered, N = 713 | Unsheltered, N = 2238 |
|---|---|---|---|
| Living in Youth-Headed Household | 2 (1.0%) | 78 (11%) | 260 (12%) |
| Living with a Disability | 181 (90%) | 258 (51%) | 1589 (71%) |
| Chronically Homeless | 83 (42%) | 153 (21%) | 1371 (61%) |
| Domestic Violence Survivor | 76 (39%) | 222 (31%) | 999 (45%) |
| Family with Children | 0 (0%) | 350 (49%) | 708 (32%) |
| Veteran | 13 (6.6%) | 115 (16%) | 148 (7.0%) |
Notes: 1) Cells highlighted in beige show most likely living situation for each subpopulation. 2) The correlation between living situation and subpopulation, using a chi-squared test, is statistically significant. 3) “Family with Children” includes all people living in households with a person under the age of 18. 4) “Living in Youth-Headed Household” includes all people living in a household headed by someone under the age of 25. For definitions, see table 3 in the Appendix.
Table 5. Gender Differences in 2021 by Living Situation: Percent of People Experiencing Homelessness
| Characteristic | ProLodge, N = 200 | Sheltered, N = 713 | Unsheltered, N = 2238 |
|---|---|---|---|
| Gender | |||
| Male | 136 (69%) | 394 (55%) | 1178 (55%) |
| Female | 61 (31%) | 312 (44%) | 944 (44%) |
| Transgender | 1 (0.5%) | 4 (0.6%) | 10 (0.5%) |
| Gender non-conforming | 0 (0%) | 0 (0%) | 2 (<0.1%) |
Notes: 1) Cells highlighted in beige show most likely living situation for each gender. 2) The correlation between living situation and gender, using a chi-squared test, is statistically significant.
Table 6. Race Differences in 2021 by Living Situation: Percent of People Experiencing Homelessness
| Characteristic | ProLodge, N = 200 | Sheltered, N = 713 | Unsheltered, N = 2238 |
|---|---|---|---|
| Race | |||
| American Indian or Alaska Native | 4 (2.1%) | 3 (0.4%) | 22 (1.1%) |
| Asian | 0 (0%) | 2 (0.3%) | 13 (0.6%) |
| Black | 71 (37%) | 272 (39%) | 736 (36%) |
| Native Hawaiian or Other Pacific Islander | 1 (0.5%) | 1 (0.1%) | 3 (0.1%) |
| Two or more races | 15 (7.8%) | 61 (8.7%) | 124 (6.0%) |
| White | 102 (53%) | 360 (52%) | 1162 (56%) |
Notes: 1) Race categories include people who identify as both Hispanics and non-Hispanics. 2) Cells highlighted in beige show most likely living situation for each race. 3) The correlation between living situation and race, using a chi-squared test, is statistically significant.
Table 7. Ethnicity Differences in 2021 by Living Situation: Percent of People Experiencing Homelessness
| Characteristic | ProLodge, N = 200 | Sheltered, N = 713 | Unsheltered, N = 2238 |
|---|---|---|---|
| Ethnicity | |||
| Hispanic or Latino | 40 (20%) | 257 (37%) | 671 (32%) |
| Non-Hispanic/Non-Latino | 156 (80%) | 447 (63%) | 1457 (68%) |
Notes: 1) Ethnicity categories include people of all races. 2) Cells highlighted in beige show most likely living situation for each ethnicity. 2) The correlation between living situation and ethnicity, using a chi-squared test, is statistically significant.
Table 8. Vulnerability Differences in 2021 by Living Situation: Percent of People Experiencing Homelessness
| Characteristic | ProLodge, N = 200 | Sheltered, N = 504 | Unsheltered, N = 2238 | p-value |
|---|---|---|---|---|
| Last Time in Permanent Housing | <0.001 | |||
| Currently in stable housing | 0% | 0% | 0.4% | |
| Less than 1 year | 23% | 40% | 28% | |
| One year or more | 77% | 60% | 71% | |
| Basic Needs Currently Met | 61% | 66% | 42% | <0.001 |
| Homelessness Caused by Trauma | 74% | 62% | 71% | 0.004 |
| Hospitalized in Last 6 Months | 48% | 32% | 32% | <0.001 |
| ER Visit in Last 6 Months | 67% | 51% | 62% | <0.001 |
| Recently Attacked or Beaten Up | 54% | 34% | 57% | <0.001 |
Notes: 1) Not all people seeking services in the Homelessness Response System take a CA. For full text of the questions asked, see table 4 in the Appendix. 2) Cells highlighted in beige show most likely living situation for how people answer each of the questions.
Source: Data in this table come from questions asked in the Coordinated Assessment (CA), which is mostly based on the Vulnerability Index - Service Prioritization Decision Assistance Tool (VI-SPDAT).
Comparisons Between 2020 and 2021 Using Alternative HMIS Method
Tables 9 through 11 show differences in subpopulation and demographic characteristics between January 2020 and January 2021. The tables indicate that there was no significant change in subpopulation or demographic composition of the population experiencing homelessness and seeking permanent housing resources. Any changes in the percent of the population within each subpopulation characteristic or demographic group between 2020 and 2021 can be attributed to variation in the data and not to any substantive increase or decrease in subpopulation or demographic composition.
Table 9. Subpopulation Differences between 2020 and 2021: Percent of People Experiencing Homelessness
| Characteristic | 2020, N = 3194 | 2021, N = 3151 | p-value |
|---|---|---|---|
| Living with a Disability | 71% | 69% | 0.087 |
| Chronically Homeless | 51% | 51% | 0.96 |
| Domestic Violence Survivor | 42% | 41% | 0.76 |
| Family with Children | 35% | 34% | 0.36 |
| Living in Youth-Headed Household | 9.7% | 11% | 0.18 |
| Veteran | 8.8% | 9.1% | 0.72 |
Notes: 1) Demographic characteristics are not mutually exclusive. 2) “Family with Children” includes all people living in households with a person under the age of 18. 3) “Living in Youth-Headed Household” includes all people living in a household headed by someone under the age of 25. For definitions, see table 3 in the Appendix.
Table 10. Gender and Age Breakdown Differences between 2020 and 2021: Percent of People Experiencing Homelessness
| Characteristic | 2020, N = 3194 | 2021, N = 3151 | p-value |
|---|---|---|---|
| Gender | 0.58 | ||
| Male | 56% | 56% | |
| Female | 43% | 43% | |
| Transgender | 0.5% | 0.5% | |
| Gender non-conforming | 0.2% | <0.1% | |
| Age Breakdown | 0.40 | ||
| Under 18 | 22% | 21% | |
| Between 18 and 24 | 7.3% | 8.0% | |
| Over 24 | 70% | 71% |
Table 11. Race and Ethnicity Breakdown Differences between 2020 and 2021: Percent of People Experiencing Homelessness
| Characteristic | 2020, N = 3194 | 2021, N = 3151 | p-value |
|---|---|---|---|
| Race | 0.21 | ||
| American Indian or Alaska Native | 1.3% | 1.0% | |
| Asian | 0.4% | 0.5% | |
| Black | 39% | 37% | |
| Native Hawaiian or Other Pacific Islander | 0.2% | 0.2% | |
| Two or more races | 6.0% | 6.8% | |
| White | 53% | 55% | |
| Ethnicity | 0.17 | ||
| Hispanic or Latino | 30% | 32% | |
| Non-Hispanic/Non-Latino | 70% | 68% |
Note: Race and ethnicity categories are not mutually exclusive.
System Capacity Over Time
Despite the challenges that COVID-19 has posed for many communities and the strain it has placed on the Homelessness Response System here in Austin/Travis County, the two figures below show that service providers are still connecting people with permanent housing.18 Figure 5 shows the number of shelter and permanent housing beds in the system since 2011. The figure indicates that, while shelter bed capacity has seen a consistent decline since 2014, the number of permanent housing beds has gone up considerably in the last ten years. In the last year alone, the number of permanent housing beds in the system increased from 2066 in 2020 to 2571 in 2021, a 24% increase. Figure 6 below shows the number of people permanently housed by the Homelessness Response System. The numbers show people who were experiencing homelessness and either 1) exited a non-Permanent Housing program (e.g., emergency shelter, street outreach) to a permanent housing destination (e.g., living in a rental unit on their own or living with family permanently), and/or 2) moved into a Rapid Re-Housing or Permanent Supportive Housing program.
This approach focuses on quickly and effectively connecting individuals and families experiencing homelessness with permanent housing and removing barriers to entry such as sobriety and treatment of service participation requirements.
Research Shows that Housing First participants:
- Access housing faster and are more likely to remain housed.
- Are less likely to use emergency services including hospitals, jails, and emergency shelters.
- Can utilize up to $23,000 less per person per year in public resources than when in a shelter program.
See Byrne et al. (2016), Gubits et al. (2015), Gulcur et al. (2003), Stefancic and Tsemberis (2007), Tsemberis and Eisenberg (2000), Woodhall-Melnik and Dunn (2016).
Note: Capacity figures are snapshots of the number of beds in the system in late January of each year.
Table 12. Shelter and Permanent Housing Bed Capacity by Program Type, 2019-2021
| Project Type | 2019 | 2020 | 2021 |
|---|---|---|---|
| Rapid Re-Housing | 727 | 905 | 1409 |
| Permanent Supportive Housing | 1083 | 1161 | 1162 |
| Emergency Shelter | 867 | 757 | 708 |
| Transitional Housing | 317 | 211 | 207 |
| Safe Haven | 15 | 61 | 15 |
Note: Capacity figures are snapshots of the number of beds in the system in late January of each year.
Source: Austin/Travis County HMIS
Permanent Housing Includes:
Permanent Supportive Housing (PSH): An evidence-based practice and permanent housing intervention for persons with a disability that includes ongoing rental subsidy and offers intensive support services throughout program participation.
Rapid Re-Housing (RRH): A short-term permanent housing intervention that includes housing locations services, rental assistance not to exceed 24 months, and housing-focused case management.
Minimal Housing Assistance (MHA): Support services designed to assist individuals with low housing barriers or persons likely to self-resolve their homelessness.
Discussion
Possible Explanations of the Point Prevalence in 2021
Many contextual factors could play a role in the lack of substantial change in the total number of people experiencing homelessness and seeking services from the HRS between January 2020 to January 2021.19 Although a greater number of people were permanently housed in 2020 than in any year prior, the estimated count of people experiencing homelessness in Austin/Travis County only slightly decreased. An increased inflow into homelessness in the community could explain the lack of change in the scale of homelessness despite increased outflow from homelessness to permanent housing.
The impact of the COVID-19 pandemic on economic factors and policy provides important context for these findings. Low wage workers in Travis County experienced a greater decrease in employment than in any other major urban county in the state, declining considerably at the onset of the pandemic, and hovering below the low wage workers’ employment rate in other major urban counties in Texas. This may be attributed to the entertainment and service industry-centered local economy in the Austin area. Figure 7 below shows these low-wage employment changes since January 2020. However, it is possible that unemployment insurance could have alleviated the negative impact of this trend on low-income households in the community, at least in part.
Notes: Low-wage workers are those whose annual income is less than $27,000 a year.
Source: Opportunity Insights Economic Tracker20
Other important factors that may have alleviated the negative impact of economic decline on housing stability are the stabilization of rent prices and the eviction moratoriums put in place at the federal and local levels. Figure 8 shows that the steadily increasing cost of rent in Austin plateaued in 2020, with a slight dip in the average cost of rent. This may have prevented episodes of homelessness for people at risk of homelessness in the past year. Further, the eviction moratorium almost certainly prevented homelessness. The economic impact of the COVID-19 pandemic may have had a larger negative impact on housing stability if not for such policies.21 Figure 9 shows that, from March 2020 to early 2021, evictions went down substantially in Austin compared to past years.
Source: Zillow22
Sources: Eviction Lab via BASTA Austin, Open Austin, and January Advisors.23
Looking Ahead
While there have been some policies and resources put in place since the beginning of COVID-19 that have likely kept the number of people experiencing homelessness from going up, there are considerable challenges in the coming years. Researchers have estimated that, as a result of the economic effects of COVID-19, chronic homelessness in the United States may increase by about 49 percent and that an extra 603,000 working-age adults in the United States may experience some kind of homelessness by 2023.24
It is likely, therefore, that the economic effects of the past year have not yet been felt for many people. Additionally, the national eviction moratorium will end on June 30, 2021, and the local eviction moratorium is slated to end on August 1, 2021. It is very likely that people at risk of homelessness have accumulated rental debt since eviction moratoriums went into effect, and it is possible that the delayed impact of the pandemic will have a greater effect on homelessness once eviction moratoriums end. These factors could impose significant strain on the capacity of the Homelessness Response System in the near future.
Additionally, Austin continues to see major strains on its housing market. The supply of housing, like much of the country, has not kept up with demand, driving sale prices up. And while rental prices plateaued in 2020, they have begun to climb in early 2021.25 This will put many low-income earners, many of whom are already cost-burdened by housing, at increased risk of experiencing homelessness. Figure 10 shows that Austin’s extremely low-income earners are some of the most cost burdened in the state of Texas.
Further, the rise in sales prices in the last year have made purchasing a home less feasible for many potential buyers, leading them to rent instead, and, in turn, making the rental market more competitive. This impacts all renters, including and, perhaps especially, those at risk of homelessness or those who were previously homeless. Housing vouchers as a form of housing assistance theoretically make low-income renters competitive rental applicants, but property owners can legally choose to not accept vouchers for these applicants. Owners have no obligation to accept vouchers, and even if they do accept them, they may still require applicants to earn as much as three times the cost of rent.26 Rising rent prices will exacerbate conditions for low-income renters, increasing the rent burden on top of a competitive and exclusionary market.
Acquiring and maintaining stable housing for formerly homeless persons is also complicated by the accessibility of fully digital systems. This issue was exacerbated by the COVID-19 pandemic, in which conversations with property owners, property tours, and paperwork were often fully digital, but issues with digital accessibility were already present, particularly with payment delivery. There is a common disconnect between rental assistance programs and property payment systems, in which rental assistance is provided in the form of checks, which are unusable by clients who are expected to make payments through online portals and/or apps. This has in many cases led to the accumulation of rental debt, particularly since the onset of COVID-19.
Notes: 1) Extremely low-income renters households are those who make less than 30% of area median income. 2) Severe housing cost burden is defined as paying more than 50% of income on housing and utilities.
Source: The National Low Income Housing Coalition.
Limitations
The alternative method we use this year to estimate unsheltered homelessness relies on HMIS data and differs considerably from a typical point in time (PIT) count. Both methods have their advantages and disadvantages. While researchers have noted the limitations of conducting a PIT Count, one main advantage the PIT Count has is that surveyors (volunteers and service provider staff) try to count individuals experiencing unsheltered homelessness, regardless of whether they have sought services from the Homelessness Response System.27
In this report, using the alternative method to estimate unsheltered homelessness, we have counted people experiencing unsheltered homelessness who are seeking permanent housing services. This count can perhaps better be defined as a snapshot of what is happening in the Homelessness Response System. The subpopulation and demographic statistics drawn from the method used in this report are not necessarily representative of the entire population experiencing unsheltered homelessness in Austin/Travis County.
The alternative method we used has a few limitations. It is unknown exactly how many people on the by-name list (BNL) are experiencing homelessness on any given day. It is possible that there are people who have self-resolved their homelessness who have remained on the BNL and are included in the count even though they are no longer experiencing homelessness. Thus, the scale of unsheltered homelessness could have been overestimated. However, because not all service providers participate in Coordinated Entry for client prioritization and referral, we presume that there are people served outside the system who are not captured using the alternative method for estimating unsheltered homelessness used in this report.
The scale of unsheltered homelessness could have been underestimated for a few reasons. There are a few reasons people could have been excluded from the count. The accessibility of Coordinated Entry during the COVID-19 pandemic has been limited by access to technology, both web and phone, which suggests the possibility of the exclusion of people who lack access to the technology required to participate. This is furthered by the possibility that COVID-19 safety concerns about staying in congregate shelter may have negatively impacted service engagement overall. It is also possible that Homelessness Response System staff’s focus on the COVID-19 response shifted focus away from conducting Coordinated Assessments.
Despite these limitations, the alternative method we used has several advantages. This method provides a more accurate snapshot of the scope of homelessness among people in the Homelessness Response System and, therefore, gives service providers and civic leaders a clearer picture of the resources and policies needed to house people experiencing unsheltered and sheltered homelessness. The HMIS method also allows us to provide more accurate and comprehensive subpopulation and demographic breakdowns than what is typically available from the in-person PIT Count. And, finally, because the HMIS method relies on a database that is updated by service providers on a regular basis, it allows us to report on the scope of homelessness more frequently than we would be able to if relying exclusively on an in-person count.
Community and Homelessness Response System Efforts to End Homelessness
Community partners, city leaders, and ECHO continue to work on efforts to end homelessness in Austin/Travis County. Here, we provide highlights of some of these ongoing and future efforts.
- The City of Austin purchased two hotels to convert into bridge and permanent housing for people experiencing homelessness in 2020, adding two more thus far in 2021. Each of these properties adds dozens of new units to the Homelessness Response System.
- In 2020, SAFE provided Rapid Re-Housing services to 170 youth and adult survivors of violence and abuse, a 48% increase over 2019.
- In 2020, LifeWorks assisted more than 283 youth exit homelessness. Since the 2018 concerted effort to end youth homelessness, LifeWorks and partners have housed more than 800 youth.
- In 2020, Family Eldercare permanently housed 125 older adults and distributed over $1.5 million to support housing stability and prevent homelessness for vulnerable populations. Family Eldercare also assisted 1,058 individuals with benefits enrollment for a total value of $2,030,553 over the last year.
- In 2020, The Other Ones Foundation housed 25 people, removed 467,810 pounds of trash from green spaces in Austin, and paid out $430,805 in earned income to people experiencing homelessness.
- The Salvation Army began using a 5-year grant from the Bezos Day One Fund to provide Housing Assistance, Child-Centric Services, COVID 19, and emergency shelter operations needs for families experiencing homelessness at the Salvation Army Rathgeber Center and the Austin Shelter for Women and Children. During the last calendar year, over 560 children were sheltered. The official slogan of the $2.5 million effort to enhance our community response to family homelessness is “No Child Sleeps Outside!”
- In 2020, Integral Care completed 163 intakes or assessments for mental health services through the 3,000 Oak Springs Clinic for individuals at the ProLodges and housed a total of 142 people.
- Integral Care analyzed data from Terrace at Oak Springs, a Permanent Supportive Housing development that uses a Housing First model, comparing the six months after move-in to the six months before move-in for 50 people and found over a 70% reduction in arrests, over a 60% reduction in medical inpatient admissions, and an almost 90% reduction in medical inpatient days. This led to public costs savings of more than $336,000 over the course of six months.
- Front Steps reduced the number of people staying nightly in shelter to proactively combat the transmission of COVID-19 and connected additional clients to one of the five ProLodges, where, with Front Steps’ support, the daily average of people served reached 305. These COVID-19 precautions kept people sheltered and safe, and reduced the number of positive cases among people experiencing homelessness to below national averages.
- The Eating Apart Together (EAT) Initiative, a City of Austin-led partnership to deliver shelf-stable and prepared meals directly to encampments across Travis County during the pandemic, delivered more than 500,000 meals in 2020 and will reach 1 million meals delivered in May of 2021.
- COVID-19 Protective Lodges (ProLodges) connected 164 people to permanent housing and 13 to temporary housing as of January 28, 2021. That number has since increased to more than 200.
- ECHO is helping coordinate $15.5M of COVID-19 Emergency Solutions Grant (ESG-CV) funds to serve at least 488 households between 2020 and 2022.
- ECHO and the Austin Justice Coalition launched the How to House campaign in December 2020 to form new partnerships with property owners and managers and advance anti-racist housing policies and practices. ECHO has facilitated 88 initial move-ins since the start of the campaign.
- The Summit to address unsheltered homelessness was launched in March of 2021. Partners from the Summit have committed to housing an additional 3,000 people in the next 3 years.
- New federal housing vouchers are slated to support people experiencing homelessness, including 242 Emergency Housing Vouchers from the American Rescue Plan Act, 750 Housing Choice Vouchers over the next 5 years, and 50 Mainstream Vouchers for non-elderly disabled households.
- Caritas will launch the Espero Austin at Rutland, adding 172 units by December 2022.
- The At Home Initiative, a multi-partner effort, is a collaborative project being funded by the St. David’s Foundation, Seton, Ascension, Travis County, the City of Austin, the Department of Housing and Urban Development, and the Department of Justice and will provide housing and supportive services for 250 individuals in Permanent Supportive Housing.
- Foundation Communities opened Waterloo Terrace (132 affordable housing units) in December 2020.*
- Mobile Loaves and Fishes will add 1,400 units in the next 10 years.*
* These organizations do not participate fully in the Homeless Management Information System and therefore these unit numbers would not be reflected in the count of “people housed” in the Homelessness Response System.
HUD Racial Equity Demonstration Project
Austin/Travis County was one of eight communities chosen for the first cohort of the HUD Racial Equity Demonstration Project, which ran from fall 2020 to spring 2021. Project core team member and CoC Equity Task Group co-chair Quiana Fisher notes, “The HUD Racial Equity Demonstration Project required us to take a critical look at our Homelessness Response System. This project has allowed us to look at the gaps in our system, recognize what communities we have excluded from the table, and solicit partnership and feedback.”
On next steps for the project, Fisher comments, “I have recognized that the answers to eliminating disparate impact and ending homelessness are inside communities most impacted and not necessarily within our formal system. Through partnership and active engagement with those with lived experience, we can build trust, find solutions, and implement systems change because those closest to the problem are also closest to the solutions. We must become accountable to the people most impacted by our work and in our community that is disproportionally Black people experiencing homelessness.”
Appendix
Table 1. Estimated Average Length of Unsheltered Homelessness on the By-Name List, 2019-2021
| Year | Number of Days |
|---|---|
| 2019 | 138 |
| 2020 | 158 |
| 2021 | 180 |
Table 2. Point in Time Count vs. HMIS Comparison: Unsheltered Count Only
| Year | PIT | HMIS Method | HMIS/PIT Ratio |
|---|---|---|---|
| 2019 | 1086 | 1855 | 1.71 |
| 2020 | 1575 | 2262 | 1.44 |
| 2021 | NA | 2238 | NA |
Table 3. Definitions
| Measure | Definition |
|---|---|
| Living with a Disability | A person living with a disability experiences one or more of a physical, mental, or emotional impairment, including an impairment caused by alcohol or drug abuse, post-traumatic stress disorder, or brain injury that is expected to be long-continuing or of indefinite duration and substantially impedes the ability to live independently. Other disabling conditions include developmental disabilities and HIV or AIDS. |
| Chronically Homeless | A person experiencing chronic homelessness is living with a disability and has been living in a place not meant for human habitation, a safe haven, or and emergency shelter for at least 12 consecutive months, or on at least four separate occasions in the last 3 years where the combined occasions total a length of time of at least 12 months. |
| Family with Children | Includes the total number of people in households with a person under the age of 18. |
| Living in Youth-Headed Household | Includes all people living in a household headed by someone under the age of 25. |
Table 4. Questions Asked in the Coordinated Assessment Presented in Table 8
| Coordinated Entry Assessment Question |
|---|
| How long has it been since you lived in permanent stable housing? |
| Are you currently able to take care of basic needs like bathing, changing clothes, using a restroom, getting food and clean water and other things like that? |
| Has your current period of homelessness been caused by an experience of emotional, physical, psychological, sexual, or other type of abuse, or by any other trauma you have experienced? |
| Been hospitalized as an inpatient? |
| Received health care at an emergency department/room? |
| Have you been attacked or beaten up since you’ve become homeless? |
References
For more information about the Department of Housing and Urban Development’s guidance on conducting a PIT, see Department of Housing and Urban Development (2014).↩︎
For more information about PIT and HIC and for national-level data, see Department of Housing and Urban Development (n.d.).↩︎
See Department of Housing and Urban Development (2020b) for details about HUD’s exception criteria.↩︎
HUD has acknowledged that PIT Count data submitted by communities across the nation will likely differ due to COVID-19, and communities will not be penalized for changing methodologies. Moreover, the inability to use data to compare local or national progress is to be expected and may have additional impacts in the strategies used by HUD during funding competitions. Communities are encouraged to collect and report data to the best of their ability while prioritizing their attention and efforts toward supporting and protecting those most at-risk and most impacted by COVID-19.↩︎
January 2021 saw the highest number of new COVID-19 cases in Travis County. See Austin Public Health (n.d.) for Travis County’s COVID-19 dashboard.↩︎
See Austin Public Health (2020) for Austin Public Health’s COVID-19 guidelines.↩︎
Coordinated Entry provides a single entry point for people experiencing homelessness to access vital community resources.↩︎
See Baggett et al. (2020); Iwundu, Santa Maria, and Hernandez (2020); Lima et al. (2020); Perri, Dosani, and Hwang (2020)↩︎
Researchers at Johns Hopkins University and National Public Radio estimated, in October 2020, that Texas had less than a quarter of the number of contact tracers that the state needed at the time. See Simmons-Duffin (2020) for more.↩︎
People in the ProLodges were screened for COVID-19 risk and vulnerability and were admitted to the ProLodges based on their risk status.↩︎
As of May 7th, there were 4 ProLodges with a total capacity of 290 units. For more information about the ProLodges, see Sanders (2021) and City of Austin (2021).↩︎
See Herring and Lutz (2015).↩︎
To estimate this, we use the Length of Stay data element in HMIS. Because that data element is a categorical variable, we created a continuous variable that was equal to the minimum number of days in the Length of Stay value (for example, for someone who had Value = “One week or more, but less than one month,” the value of the new variable would be 7). We used the minimum number of days, rather than a midpoint, for two reasons. First, the range for each value in the data element is large, and second, one of the values (“One year or longer”) has no end value. For these two reasons, using a midpoint would have likely overestimated the length of stay for clients.↩︎
For details on the two HMIS data elements (Prior Living Situation and Length of Stay) we used in this estimation process, see Department of Housing and Urban Development (2020a).↩︎
We use HUD definitions for disability and chronicity. A person living with a disability experiences one or more of a physical, mental, or emotional impairment, including an impairment caused by alcohol or drug abuse, post-traumatic stress disorder, or brain injury that is expected to be long-continuing or of indefinite duration and substantially impedes the ability to live independently. Other disabling conditions include developmental disabilities and HIV or AIDS. A person experiencing chronic homelessness is living with a disability and has been living in a place not meant for human habitation, a safe haven, or emergency shelter for at least 12 consecutive months, or on at least four separate occasions in the last 3 years where the combined occasions total a length of time of at least 12 months.↩︎
See Census Bureau (n.d.).↩︎
For a list of the Coordinated Assessment questions we analyzed for table 8, see table 4 in the Appendix.↩︎
For more on Housing First, see the text box and the following sources: Byrne et al. (2016), Gubits et al. (2015), Gulcur et al. (2003), Stefancic and Tsemberis (2007), Tsemberis and Eisenberg (2000), Woodhall-Melnik and Dunn (2016).↩︎
For research on local, community indicators associated with homelessness, see Fargo et al. (2013); Glynn and Fox (2019); Hanratty (2017); Department of Housing and Urban Development Office of Policy Development and Research (2019); Elliott and Krivo (1991).↩︎
See Opportunity Insights (n.d.).↩︎
See Collinson and Reed (2018) for research on the effects of evictions on homelessness.↩︎
See Zillow Research (n.d.).↩︎
See Eviction Lab (n.d.).↩︎
See Flaming et al. (2021).↩︎
See McGlinchy (2021).↩︎
For more about how Texas law prevents landlords from being punished for discriminating against families with vouchers, see Walters and Satija (2018).↩︎
For an analysis of the limitations of the PIT Count, see Schneider, Brisson, and Burnes (2018).↩︎
Please note:
Our alternative methodology for estimating how many people are experiencing unsheltered homelessness this year uses data from HMIS. Estimates of unsheltered homelessness based on the alternative methodology used in 2021 are not interchangeable with in-person PIT Count estimates and therefore cannot be compared to past PIT Count estimates.