Racial Disparities and Homelessness in Western New York1 Executive Summary 2 Racial and Ethnicity Inequity in Poverty and Homelessness 3 Within the Homeless System 5 Vulnerability Among Different Races/Ethnicities 6 Implications of Equity Among Different Homeless Interventions 9 Equity in Successful Exits 10 Conclusion and Action Plan 12 September 2018 1 In this data analysis, Western New York is defined as Erie, Niagara, Genesee, Wyoming and Orleans County because these are areas the Homeless Alliance of WNY coordinate services with and administer the data for. Page 1 Executive Summary The Homeless Alliance of Western New York analyzed racial disparities among homelessness within Western New York and examined the homeless system’s equity serving different racial/ethnic groups in terms of receiving those services, prioritizing those services, and housing success rate. The ideal model for an unbiased homeless system would distribute assistance such that it is received in equal percentages across racial/ethnic groups as the percent of that racial group experiencing homelessness. One of the consequences of systemic racism is an overrepresentation of people of color among those who experience homelessness. Black people make up 11% of the general population of Western New York but they account for 25% of the people in poverty and 47% of the homeless population. When the number of people who experience homelessness are compared against the number of people who are in poverty, Black people living in poverty are 3 times more likely to experience homelessness compared to White people. These suggest that poverty rates alone do not explain the over-representation, but that systemic racism blunts the ability of people of color to recover from financial catastrophe to avoid homelessness relative to those who are White. We reviewed data entered in our Homelessness Management Information System pertaining to the race and ethnicity of clients who have experienced homelessness and those who have participated in the coordinated entry process and those who received housing vouchers assistance (through permanent supportive housing or rapid rehousing programs). Some programs, like emergency shelter, must serve all. However, in 2014, our system for distributing rapid-rehousing and permanent supportive housing vouchers changed from first-come, first serve to a coordinated entry process that prioritized those who had highest vulnerabilities on the VI-SPDAT triage tool. We found that higher rates of white people presented with more disabilities than other race and ethnic groups, and more scored in the higher categories of the vulnerability index. We found that implementing a need-based coordinated system sharply reduced the disproportionately white access to permanent housing that existed in 2013. We also found that the highest rate of successful exits from rapid rehousing are among Hispanic/Latinx clients and Black clients. The homelessness alleviation system is the safety-net of last resort for community members at their most dire financial circumstances. It is thus the system that pays the sharpest witness to the consequences of institutional racism and systemic barriers to financial resource acquisition. We will continue monitoring the demographics of those experiencing homelessness and those receiving services to ensure that our system is not perpetuating the same racism. Page 2 The purpose of this data analysis is to analyze racial disparities among homelessness within Western New York and examine the homeless system’s fairness across different racial/ethnic groups in terms of receiving services, prioritizing clients for services, and housing success rate. ​​Homeless system information is solely obtained from the Homeless Management Information System (HMIS), which stores client’s demographic information as well as program information, such as program start date, exit date and exit destination. 90% of the shelters, street outreach programs, transitional housing, rapid rehousing, and permanent supportive housing programs in the homeless system participate in this database. Racial and Ethnicity Inequity in Poverty and Homelessness Figure 1: Percentage of each race in WNY general population, those who experienced Poverty (United States Census Bureau - 2016 American Community Survey) and those experienced Homelessness (Homeless Management Information System - FY2017, October 1, 2016 – September 30, 2017). Because Hispanic/Latinx is considered an ethnicity, in this chart it is calculated separately from people's races so each section of the graph will be greater than 100%. Figure 1 shows that White people are the only group who experience poverty and homelessness at rates less than their proportion in the general population. They represent 84% of the general population, but only about 10% of them experience poverty (Figure 2). Every other group, who together total only 16% of the general population, experiences poverty at more than three times that rate. These rates are already disproportionate, but when we further analyzed the rates that people in poverty also experience homelessness, the results were still inconsistent. Black and Hispanic/Latinx people experiencing poverty also experienced homelessness at triple and double the rates respectively of White people (Figure 3). These suggest that poverty rates alone do not explain the over-representation of people of color in the homeless system. Page 3 Figure 2: Percentage of people experiencing poverty in each group based on the total population of that group Figure 3: Percentage of people experiencing homelessness in each group based on the number of people experiencing poverty in that group Based on our 2017 Homelessness Summary Brief, 29.15% of people experienced homelessness due to eviction, either by primary tenant or by landlord, and 10.55% of the them due to job loss in Erie2. According to the Buffalo City School District Graduation Rate Data 4 Year Outcome as of June, 2018, the White students’ high school graduation rate is 74% but the Black and Hispanic/Latinx students’ graduation rates are only 62% and 51% respectively.3 T​ his, combined with the fact that unemployment 2 https://wnyhomeless.org/app/uploads/2017-Erie-Homelessness-Brief-Final-Version.pdf 3 Buffalo City School District Graduation Rate Data 4 Year Outcome as of June. Accessed September 10, 2018 https://data.nysed.gov/gradrate.php?year=2017&instid=800000052968 Page 4 rates are higher for those with less than a high school degree (12.8% compared to the 4.9% of those who graduate from high school4), does not make it surprising that the unemployment rate for African American and Hispanic/Latinx workers in the city in recent years has averaged over 13.6% and 10.2% respectively while White workers have averaged a 4.9% unemployment rate5. For those who are working, people of color are not making as much as their White counterpart. ​Partnership of Public Good’s 2016 report points out that “In Buffalo-Niagara, whites have a median income of over $55,000. By contrast, for blacks the median income is under $25,000 and for Hispanics it is under $27,000.” With increasing rent, affording housing becomes difficult for low-income households. 42% of Native American households, 37% of Hispanic/Latinx households, 30% of black households, and 29% of Asian or Pacific Island households pay more than half their income on housing every month compare to only 15% of the White households.6 According to our 2017 Homelessness Summary Brief, a significant portion of clients enter the homeless system after being released from an institution like jail or prison (9.87% of clients in Erie County7, 18.98% of clients in Niagara County8). This likely affects the Black and Hispanic/Latinx communities at a much higher rate. In Partnership for the Public Good’s 2013 report, ​Alarming Disparities: The Disproportionate Number of African American and Hispanic People in Erie County Criminal Justice System 9, while African Americans made up less than 14% of the population in Erie County, they accounted for 43% of arrests. The report found this to be a result of residential segregation among other factors. When Black and Hispanic/Latinx communities are involved in the criminal justice system disproportionately relative to other racial and ethnic groups, the fabric of poor, segregated communities is further damaged and more families are pushed into poverty. Homelessness reflects the failure of our social systems to serve people of all racial and ethnic groups equally in housing, education, employment, wealth accumulation, and justice. Within the Homeless System In order to examine equity in the homeless service delivery system, we decided to review the vulnerability of all clients who experienced homelessness compared to those who were enrolled, review the enrollment rate in two types of permanent housing interventions (Permanent supportive housing and Rapid rehousing), and review the success rates among different interventions. In this data analysis, we compare only FY2013 data to the most recent data from August 1, 2017 to July 31, 2018 due to a significant change to our homeless system. New York State is a right-to-shelter state, so the standard of intervention for emergency shelters is that all people are served. In FY2014, we began 4 Partnership for the Public Good, “Employment Data for Buffalo-Niagara” (Partnership for the Public Good, 2018). Available from Partnership for the Public Good, www.ppgbuffalo.org. 5 2012-2016 American community Survey 5 year estimates for Erie County 6 U.S. Housing and Urban Development Department, Affirmatively Furthering Fair Housing Map Tool, Table 10 – Demographics of Households with Severe Housing Cost Burden (July 2016), https://egis.hud.gov/affht/#. 7 https://wnyhomeless.org/app/uploads/2017-Erie-Homelessness-Brief-Final-Version.pdf 8 https://wnyhomeless.org/app/uploads/2017-Niagara-Homelessness-Brief-Final-Version.pdf 9 ALARMING DISPARITIES: The Disproportionate Number of African American and Hispanic People in Erie County Criminal Justice System. Open Buffalo Report presented by Erin Carman Page 5 implementing a coordinated entry system so that all homeless-specific housing programs serve clients based on vulnerabilities. This replaced admissions based on a first-come, first-serve basis or program-based criteria and requirements such as a client maintaining sobriety, having income, not having criminal history, and so forth. We have been making progress since FY2014. In the summer of 2017 we felt we significantly achieved our goal of adopting the coordinated entry system across the entire homeless service system. As a result, we chose to present our system’s data before coordinated entry started and the most recent data, and we plan on monitoring this data moving forward. Although Hispanic/Latinx is considered an ethnicity by the United States Census, not a race, we find it is important to separate this group to gain a better understanding of how they may face social and economic complications compared to their counterparts of other races. For this reason, we decided to separate the Hispanic/Latinx population by pulling all the people who report their ethnicity as Hispanic/Latinx, regardless their race, in the following charts. The categories are not mutually exclusive as a consequence. Also, as shown in Figure 1, our homeless system has a very low representation of Asian, Hawaiian, Native American, and multi-racial populations. Each individual in these groups will have a significant impact on the data so we chose to group them together in order to minimize an individual’s impact. Vulnerability Among Different Races/Ethnicities In order to be prioritized by permanent housing assistance, we consider the following criteria: 1. Vulnerability 2. Disability 3. History of homelessness We will not address homeless history in this analysis because it could be documented outside of HMIS, our database for recording information of clients experiencing homelessness. In our community, we adopted the Vulnerability Index-Service Prioritization Decision Assistance Prescreen Tool (VI-SPDAT)10. This tool is internationally adopted and designed to triage clients by asking a series of questions on their housing history, risks, disability, socialization, daily functioning, and wellness and giving them a vulnerability index score based on their responses. The score, ranging from 0 to 26, determines whether a client has high, moderate, or low vulnerability. In general, clients who score in the high vulnerability range (8+) are recommended to receive permanent supportive housing, clients who score in the moderate vulnerability range (4-7) are recommended to receive rapid rehousing, and clients who score in the low vulnerability range (0-3) should be able to self-resolve with mainstream resources. In a community with limited resources like housing, not all clients in the high vulnerability range may be served in a permanent supportive housing program; instead, they may be served by the rapid rehousing programs to ensure those with the greatest needs are being served somewhere. This, in turn, affects how many people can be served in rapid rehousing since those programs are also affected by limited resources. Ultimately, by prioritizing people with the highest levels of vulnerability, we don’t serve as many people with lower scores like 4-5. 10 http://www.orgcode.com/products Page 6 Clients with VI-SPDAT (Aug 2017 - July 2018) Black White Other Race Hispanic/Latinx Clients Served 2236 1587 368 746 Clients Assessed 577 (26%) 382 (24%) 63 (17%) 169 (23%) Table 1: Number of total clients served compared to the number of each race/ethnic who were assessed by the coordinated entry tool Table 1 above shows the number of people who were assessed with the VI-SPDAT. Not all clients are assessed. Those who do not engage with a shelter’s or outreach team’s case manager or who only stay in a shelter for a short period of time (e.g. less than 5 days) would not receive an assessment. However, once they are assessed, this value, combined with the individual’s homeless history and disability status, is used to decide a client’s priority for housing assistance. Of all people served in the last year, nearly a quarter of all black, white, and Hispanic/Latinx people were assessed and given a VI-SPDAT score; the percent of people of other races was lower at 17%. This would indicate that most people are being assessed at an equal rate. Figure 4: Percentages of the clients who were assessed and scored at levels needed to be considered for permanent housing programs. This includes data from the VI-SPDAT, VI-F-SPDAT, and TAY-SPDAT Figure 4 looks only at the clients who were assessed in coordinated entry and​ ​scored 4-7 or 8 or above on the vulnerability index scale. ​ A​ lthough a higher percentage of Black clients are being assessed (Table 1), compared to other racial and ethnic groups this chart shows that a smaller percentage of Black clients are scoring in the high vulnerability range and instead are more often scoring in the moderate and low ranges. This chart also shows that white people are scoring at a higher rate than other groups in the range of high vulnerability. Page 7 Figure 5: Percentage of people in each racial/ethnic group who identified having one, two, or three disabling conditions (alcohol/drug use, mental health, or physical health) Clients must have a long-term disability to be eligible for permanent supportive housing. Although rapid rehousing does not require a disability, our system prioritizes those who are most in need. A person with a disability is more likely to be served. Figure 5 shows White people who have experienced homelessness have a higher rate of having a disability (53%) as well as having multiple disabilities (29%). Clients with multiple disabling conditions, for example substance abuse and mental illness, when compared with either substance abuse or mental illness alone, demonstrated increased service utilization, housing instability and homelessness, and violent or criminal behavior11. They are more likely to need permanent supportive housing instead of Rapid rehousing or mainstream resources. Both the disability status chart (Figure 5) and the vulnerability index score data (Figure 4) indicate that White people who become homeless appear to be more vulnerable. People of color are more likely to become homeless, but they have lower vulnerability index scores, and smaller percentages of disabilities. We suspect this is because situational causes of homelessness, such as job loss, an eviction, or illness, may be more likely to result in homelessness for people of color as they experience a structural disadvantage in society when it comes acquiring fiscal resources through work, getting access to apartments, and having an informal safety net to jump into. By having barriers to economic participation, they become more vulnerable if anything goes wrong. Racial discrimination in other systems thus prevents them from obtaining resources and maintaining stable housing compared to white people. Prevention and diversion strategies could be used to assist people in order to reduce homelessness. Further studies could be conducted regarding other factors that could contribute to the score, e.g. barriers of client engagement, homeless history, and prior living situations. 11 Timko, C., & Sempel, J. M. (2004). Short-term outcomes of matching dual diagnosis patients’ symptom severity to treatment intensity. Journal of Substance Abuse Treatment, 26(3), 209-218. Page 8 Implications of Equity Among Different Homeless Interventions The ideal model for an unbiased homeless system would distribute assistance such that it is received in equal percentages across racial/ethnic groups as the percent of that racial group experiencing homelessness. ​This presumes that all racial groups experience vulnerability and other eligibility factors at the same rates. For example, as seen in Figure 1, the total ​number of people in each race experiencing homelessness is different. Although there might be more Black people being served compared to Hispanic/Latinx, if the same proportion of each race is being served in a housing intervention, this system would, in theory, suggest that we present a fair opportunity to both groups. Figure 6: Percentage of each racial/ethnic group who entered a permanent supportive housing program compared to all clients who were homeless during the two reporting periods New Admissions in PSH FY 2013 Aug 17 - July 18 Other Hispanic/ Other Hispanic/ Black White Black White Race Latinx Race Latinx Clients who were homeless 1683 1008 500 603 2236 1587 368 746 during reporting period 73 76 10 16 68 62 10 18 Clients Entering PSH (4.3%) (7.5%) (2.0%) (2.7%) (3.0%) (3.9%) (2.7%) (2.4%) Table 2: Number of people of each racial/ethnic group who entered a permanent supportive housing program compared to all clients who were homeless during the two reporting periods Permanent supportive housing (PSH) is a long-term housing intervention that provides clients who have a long-term disability and an extended homeless history with a housing subsidy and supportive services. As shown in Figure 6, FY2013 had significantly more White people enrolled in PSH compared to the other racial and ethnic groups. However, the most recent data indicates that we enroll clients at a much more even rate across all groups, although White people experiencing homelessness are still being served at a greater rate than others. This can be explained using the vulnerability and disability data Page 9 from earlier which showed that White people who experienced homelessness have higher rates of higher vulnerability and disability (Figures 4 and 5). Figure 7: Percentage of each racial/ethnic group who entered a rapid rehousing program compared to all clients who were homeless during the two reporting periods New Admissions in RRH FY 2013 Aug 17 - July 18 Other Hispanic/ Black White Black White Other Hispanic/ Race Latinx Race Latinx Clients who were homeless 1683 1008 500 603 2236 1587 368 746 during reporting period 51 49 7 21 207 162 32 53 Clients Entering RRH (3.0%) (4.9%) (1.4%) (3.5%) (9.3%) (10.2%) (8.7%) (7.1%) Table 3: Number of people of each racial/ethnic group who entered a rapid rehousing program compared to all clients who were homeless during the two reporting periods Rapid rehousing is a housing intervention that assists homeless households for a short period of time. Its focus is to help clients search for affordable housing and maintain stabilized housing. Figure 7​ ​compares the percentage of each racial/ethnic group served in Rapid rehousing and shows a slightly less biased system, indicating that the two periods were both fairly equitable among all races and ethnicities. When compared to the vulnerability and disability data, Hispanic/Latinx might be undeserved in the most recent period. Between FY2013 and our most recent data period, we have significantly increased this our rapid rehousing program which has led to more than a 300% increase in the total number of people served. This means that although there was a 32% increase in Black people experiencing homelessness compared to five years ago, they are now 3 times more likely to receive rapid rehousing. Equity in Successful Exits In addition to looking at how equitably each race is being served with the implementation of coordinated entry, we also looked at the rates of successful exits from our programs. In this case, we Page 10 considered all exits to permanent destinations (e.g. remain in their own apartment or move to subsidized housing) as successful. Because of how small the number of exits are that occur from permanent supportive housing programs, we chose not to include that data in our analysis. We chose to compare successful exit rates for each race/ethnicity from Rapid rehousing and from all of our services combined. Figure 8: Percentage of each racial/ethnic group who successfully exited to a permanent destination from Rapid rehousing based on the number of clients who exited Rapid rehousing within the reporting period Successful Exits from RRH FY 2013 Aug 17 - July 18 Other Hispanic/ Other Hispanic/ Black White Black WhiteRace Latinx Race Latinx Clients Exiting within the 18 17 1 3 88 82 13 24 reporting period Clients Exiting Successfully 10 5 0 1 72 63 10 20 from RRH (55.6%) (29.4%) (0.0%) (33.3%) (81.8%) (76.8%) (76.9%) (83.3%) Table 4: Numbers of each racial/ethnic group who successfully exited to a permanent destination from Rapid rehousing based on the number of clients who exited Rapid rehousing within the reporting period Figure 8 and Table 4 compare the rates of successful exits from Rapid rehousing programs to a permanent destination based on the total number of exits from Rapid rehousing programs.​ ​Overall, the success rate in the most recent reporting period is a lot higher compared to FY2013. It’s difficult to compare success rates when the total number of people exiting is very small, like other race and Hispanic/Latinx clients in FY2013 (Table 4). Unlike data in the enrollment charts, the percentage of white people exiting successfully is smaller than the other groups in the most recent reporting period. It is worth noting however that the differences between all group’s success rates in this reporting period are very small. Page 11 Figure 9: Percentage of each racial/ethnic group who successfully exited to a permanent destination based on the number of clients who exited any homeless program within the reporting period Successful Exits from Emergency Shelter, Transitional Housing, Street Outreach, Permanent Supportive Housing, and Rapid Rehousing FY 2013 Aug 17 - July 18 Other Hispanic/ Other Hispanic/ Black White Black White Race Latinx Race Latinx Clients Exiting 1559 906 468 571 2327 1625 383 796 Clients Exiting Successfully 382 273 45 99 609 454 103 177 from homeless system (24.5%) (30.1%) (9.6%) (17.3%) (26.2%) (27.9%) (26.9%) (22.2%) Table 5: Numbers of each racial/ethnic group who successfully exited to a permanent destination based on the number of clients who exited any homeless program within the reporting period The data comparing all exits from the two reporting periods also shows that racial and ethnic groups are exiting successfully at more similar rates than they did in previous years (Figure 9, Table 5). Out of all clients who exited from any program in the most recent reporting period, White clients have the greatest rate of successful exits (27.9%). However, all other groups have improved compared to FY13, especially other races which saw the greatest increase in successful exits (9.6% to 26.9%) and went from having the smallest rate of success to the second highest. Despite Hispanic/Latinx clients having the most success in Rapid rehousing programs (Figure 7), this data shows that we can do more to ensure their success in all other programs. Conclusion and Action Plan ● Any effort to end homelessness must address the range of issues that have resulted from racial inequity. ​The homeless alleviation system cannot achieve that goal alone. We need to end all systems of racial bias and together we can provide equal opportunities to all residences. ● HMIS data shows that 35% of the clients only stayed in shelter for less than 7 days. For those who stayed and got assessed, the majority of clients have a low to moderate vulnerability score. Since these clients have fewer destabilizing conditions working against them, they would benefit from a​ coordinated entry system that expanded its preventive services like linkage of Page 12 employment and other mainstream resources, mediation between landlord or primary tenant, and the search for flexible funding to meet people’s immediate needs, especially Black people who have higher chances falling from poverty into homelessness. ● With implementation of coordinated entry, there are some improvements on reducing racial inequity in all homeless programs, especially in permanent housing.​ ​ ​But overall we believe the current homeless system are providing equitable services to those who experienced homeless. ​​The Homeless Alliance and the Coordinated Entry Lead agencies will continue monitor the process by conducting client and provider surveys and analyzing data to ensure all clients have equal access to the housing and services that they need. ● Based on American Community Survey12, 67% of the social workers are White and this is significantly different from the population we serve (only 32% are white). Further study could be conducted related to racial composition of senior/executive staff and front line staff, engagement with clients, involving trust, culture differences, and engagement style. ​We need to create greater racial/ethnic representation on the decision-making bodies of the homeless system including executive staff and frontline staff. ● Asian, Native American, Hawaiian and Hispanic/Latinx are slightly underserved in the housing interventions that we offer. We may need to further examine the data, e.g. look into homeless history . Additionally, it’s important that we ensure different language services are offered and all clients understand their rights to receive services. ● With housing intervention, people of color could achieve higher outcomes in maintaining stable housing.​​ We will continue to advocate for increased assistance for all people who experience homelessness and ensure that this experience will be brief and nonrecurring. ● Erie County passed new legislation to ensure fair housing in April 2018. The new legislation expands upon protections based on race, color, national origin, religion, sex, disability and the presence of children conferred by the national legislation, by barring discrimination in housing against individuals on the basis of their age, marital status, source of income, military status, sexual orientation, gender identity, immigration status or familial status. This helps clients who are homeless receive mainstream assistance or in a housing program to navigate housing. However, a recent report raised questions about the adequacy of the city’s enforcement on Buffalo’s Fair Housing Ordinance which prevents discrimination based upon source of income since adopted in 2006.13 W​ e need to continue to implement and enforce the fair housing protections with the local government. 12 Racial among social work:https://datausa.io/profile/soc/211020/#demographics 13 http://www.investigativepost.org/2018/07/05/buffalo-not-enforcing-its-fair-housing-law/ Page 13