Self-reported impacts of the COVID-19 pandemic for people experiencing homelessness in Sacramento, California

ABSTRACT

The COVID-19 pandemic has likely substantially compounded hardships for people experiencing homelessness. In addition to their already heightened health risks, shelter-in-place orders and recommended physical distancing have constrained available services. Though people experiencing homelessness have surely also been impacted economically, the extent of these impacts remains unclear. This study documents self-reported disease and economic impacts of the coronavirus pandemic for people experiencing homelessness in Sacramento, California. The study analyzes survey data from 198 people experiencing homelessness, collected in collaboration with a homelessness services agency in October 2020. The article contextualizes these data with comparisons to Sacramento’s point-in-time survey of homelessness and a sample of low-income housed Californians. The results suggest relatively limited exposure to COVID-19 among people experiencing homelessness in Sacramento. Income and employment losses were more common, but still less pronounced for people experiencing homelessness than for low-income housed Californians. However, these lower economic losses mainly reflect enduring deprivation prior to the pandemic. People experiencing homelessness also received stimulus funds in the spring of 2020 at much lower rates than low-income housed Californians. Overall, the study adds to an emerging empirical literature on the diverse impacts of the COVID-19 pandemic for people experiencing homelessness.

KEYWORDS:

• Homelessness
• poverty
• stimulus payments
• COVID-19
• service provision

Introduction

The COVID-19 pandemic may have profoundly compounded pre-existing hardships for people experiencing homelessness. Among the U.S. population overall, both the health and economic impacts of the COVID-19 pandemic have been much greater for already disadvantaged people. Low-income people and People of Color have disproportionately contracted and died from the disease, and their economic hardships disproportionately increased (Mendez-Smith & Klee, 2020). People experiencing homelessness bear multiple health and economic burdens that make them especially vulnerable to these negative impacts (Culhane et al., 2020; Lima et al., 2020; Perri et al., 2020).

This study documents the self-reported health and economic effects of the coronavirus pandemic for people experiencing homelessness in Sacramento, California. The study analyzes survey data from 198 people experiencing homelessness, collected October 26–28, 2020, in collaboration with Loaves & Fishes, a homelessness services agency. This survey (hereafter “L&F Survey”) measured both perceived disease impacts (i.e. exposure to the coronavirus, access to testing, congregate shelter avoidance) and economic impacts (i.e. job loss, income loss, receipt of the stimulus payments in spring 2020). These events could be especially detrimental for people experiencing homelessness given their constrained resources, perpetuating their experience of homelessness (Paat et al., 2019).

The article contextualizes these patterns with multiple additional data sources. First, the article compares rates of homelessness and COVID-19 cases and deaths in Sacramento to other California cities, the state of California, and the nation as a whole. Second, the article compares the L&F Survey sample’s characteristics to estimates from Sacramento’s 2019 point-in-time survey of homelessness, which represents Sacramento’s population experiencing homelessness (Baiocchi et al., 2019). Third, the article presents concurrent estimates of economic resources and pandemic-related losses for California’s low-income housed population using the Household Pulse Survey (Census Bureau, 2020). These comparisons suggest the results from the L&F Survey may apply to most people experiencing homelessness in Sacramento and highlight notable differences from low-income housed Californians. The study’s results contribute to a growing body of empirical work documenting the pandemic’s complex consequences for people experiencing homelessness.

Background & context

Unequal impacts of the pandemic

Many advocates and experts have prominently highlighted high vulnerability to the coronavirus (or SARS-CoV-2, the virus that causes the disease COVID-19) for people experiencing homeless (Culhane et al., 2020; Lima et al., 2020; Perri et al., 2020). People experiencing homelessness have limited access to sanitation and may not be able to physically distance in congregate shelters or crowded encampments. Compared to the housed population, a greater fraction of people experiencing homelessness are older and have disabilities or chronic diseases that make them especially vulnerable to COVID-19. As a result, simulations early in the pandemic predicted potentially catastrophic rates of COVID-19 infections and deaths among people experiencing homelessness (Culhane et al., 2020).

Widespread transmission of the coronavirus is especially likely in congregate living places, including shelters for people experiencing homelessness (Culhane et al., 2020). Shelter outbreaks were of especially high concern early in the pandemic, given the resource constraints many shelters face (Newman & Donley, 2017). By April 2020, coronavirus clusters appeared in shelters in Boston, San Francisco, and Seattle (Mosites et al., 2020), with many more reported across the country (Finnigan, 2020). In early May 2020, Continuums of Care (local coordinating agencies for homelessness services) across the country still reported shortages in access to personal protective equipment for shelter staff, coronavirus testing, and capacity to quarantine or isolate symptomatic or positively tested people (Rice et al., 2020).

Expanded public health responses by local governments, health care providers, and homelessness services organizations may have effectively prevented worst-case predictions in many places. Early and widespread coronavirus testing, isolating those testing positive for the virus, helped constrain coronavirus clusters in congregate shelters in cities like Dallas, San Diego, and Seattle (Benavides & Nukpezah, 2020; Marquez et al., 2020; Tobolowsky et al., 2020). Lower transmission rates for the coronavirus in outdoor spaces, like tent encampments, may have reduced risk for many people experiencing homelessness. Among the mostly unsheltered people experiencing homelessness in Los Angeles County, the test positivity rate is lower than among the overall population (Weber, 2020). A study of young people experiencing homelessness in the Los Angeles area also found relatively high access to coronavirus precautions, like sanitation and physical distancing (Tucker et al., 2020).

In addition to unequal exposure to COVID-19, economic hardships like job loss, housing loss, and food insecurity have grown disproportionately among lower-income people during the pandemic (Mendez-Smith & Klee, 2020). Despite lower-than-feared disease impacts, news reports have documented how the pandemic has severely compounded hardship for many people experiencing homelessness (Berry-Jester & Hart, 2020). Some shelters closed due to virus outbreaks and many others reduced capacity to implement CDC-recommended physical distancing (Finnigan, 2020; Mosites et al., 2020). Access to many behavioral health services similarly became more limited (Tucker et al., 2020). Anecdotally, increased applications for services and demands for affordable housing among previously middle-income people may further hamper access to services or housing for those experiencing homelessness (Berry-Jester & Hart, 2020).

Sacramento’s rates of homelessness and COVID-19 in comparison

Sacramento, California, has notable similarities and differences from other areas of the United States in terms of both homelessness and the extent of the COVID-19 pandemic. Table 1 compares Sacramento County to the state of California, the nation, and Los Angeles and San Francisco Counties. Sacramento’s rate of homelessness in 2019 was similar to California’s overall rate, and both were more than twice the national rate. In comparison, Los Angeles and San Francisco Counties were outliers with extremely high rates of homelessness. Though less commonly studied than Los Angeles or San Francisco, Sacramento’s contexts of homelessness arguably better represent those in other cities around the country.

Table 1. Comparison of homelessness rates in 2019 and COVID-19 case/death rates as of October 25, 2020.

	Sacramento County	Los Angeles County	San Francisco County	California	United States
Homelessness Rate in 2019	35.9	58.7	91.1	38.3	17.3
Cumulative COVID-19 Case Rate	162.8	298.6	137.3	229.2	260.7
Cumulative COVID-19 Death Rate	3.1	7.0	1.6	4.4	6.8

Notes: All rates are per 10,000 people. Numbers of people experiencing homelessness from 2019 point-in-time counts. Total populations from U.S. Census Bureau. Cumulative COVID-19 cases and deaths as of October 25, 2020, from https://usafacts.org/visualizations/coronavirus-covid-19-spread-map

Rates of COVID-19 cases and deaths have been relatively low in Sacramento County. As of October 25, 2020 (immediately prior to survey data collection), the cumulative rate of COVID-19 cases and deaths in Sacramento County was lower than in the state overall and much lower than in the nation.

COVID-19 cases among people experiencing homelessness in Sacramento County have been especially low. Of the 1,315 coronavirus tests administered across several sites in that time, only nine have been positive (Sacramento County, 2020). The county, in collaboration with local service providers and the state government (i.e. California’s Project Roomkey), has been renting hotel rooms and establishing other quarantine/isolation spaces for people experiencing homelessness who tested positive for the coronavirus or were in high-risk groups. Between April 8 and September 9, 2020, 1,073 people received temporary quarantine/isolation housing (Sacramento County, 2020) relative to an estimated 10,000 people who experienced homelessness in Sacramento County at any point in 2019 (Baiocchi et al., 2019).

In contrast to the relatively low case rates, the pandemic’s economic impacts on Sacramento County may have been more pronounced. California was the first state to implement a shelter-in-place order on March 19, 2020. Sacramento County’s unemployment official rate spiked to 14.5% in April 2020, similar to the official national unemployment rate. However, Sacramento’s official unemployment rate (9.8%) remained higher than the national rate in September 2020 (U.S. Bureau of Labor Statistics, 2020).

Study aims

This study examines self-reported disease and economic impacts of the COVID-19 pandemic for people experiencing homelessness in Sacramento, California. Given the patterns described above, one could expect relatively few self-reported disease impacts (i.e. exposure to the disease) but relatively strong economic impacts of the pandemic (i.e. lost income or work). The study also compares these self-reported economic impacts to a sample of low-income housed Californians. Finally, I also compare access to economic relief, specifically the federal stimulus checks in spring 2020 from the CARES Act, for people experiencing homelessness and low-income housed Californians.

Data & methods

L&F survey

The L&F Survey, this study’s primary data, was collected through Loaves & Fishes, a large homelessness service provider in Sacramento, California. From October 26–28, 2020, the author and a small team of volunteers (four to nine at a time, 25 total) conducted face-to-face surveys with adults (observed age range: 21–79) receiving free take-away breakfasts and lunches. The survey was administered outside; all survey takers and participants wore facemasks. Loaves & Fishes is located north of Sacramento’s downtown area, close to a temporary shelter, and within walking distance from several large tent encampments. The organization provides many services in addition to meals, including sanitation services, clothing, medical services, and a school program for children experiencing homelessness. During the survey collection periods, Loaves & Fishes served 1,363 breakfasts and lunches over five meal services. Based on survey questions about the type and frequency of Loaves & Fishes service usage, the number of unique individuals receiving these meals likely exceeded 700.¹

Respondents typically completed the survey in 10–15 min and received a $5 gift card for their time. The survey asked three question batteries: evaluations of Loaves & Fishes services; respondent background information, including demographics, economic resources, and experiences of homelessness; and self-reported impacts of the COVID-19 pandemic.

The team collected 236 completed surveys (only three respondents started but did not finish the survey). The analytic sample includes 198 respondents who self-reported currently experiencing homelessness. Item non-response was generally low (less than 8%). In addition to point estimates, Table A1 in the online appendix presents standard errors, observation numbers, and item non-response rates for all key variables.

Multiple selection processes likely influenced the composition of the L&F Survey sample. First, the sample only included people using the service provider’s breakfast and lunch distribution by definition. People experiencing homelessness who systematically do not use Loaves & Fishes or meal services in general were not represented. Second, people receiving meals in combination with other services, like the county’s COVID-19 temporary shelter program, were also likely not represented. Third, the survey only represented people still experiencing homelessness late in the pandemic. It misses people who experienced homelessness earlier in the pandemic but transitioned out of homelessness.

The first part of the analysis compares the demographic characteristics and conditions related to homelessness in the L&F Survey to data on the broader population of people experiencing homelessness in Sacramento and low-income housed Californians (described below). Key demographic variables include: current age in years; gender (male = 1, other = 0); race/ethnicity (non-Latina/o White, non-Latina/o Black, Latina/o of any race, other); and family status (unmarried and without children = 1; other = 0).

I recoded several of the variables measuring conditions related to homelessness for comparison with the point-in-time survey of homelessness. Length of residency in Sacramento was measured in years, months, weeks, and days, converted to a binary variable (at least one year = 1; less than one year = 0). The survey asked respondents where they slept last night, converted into a binary variable for being unsheltered (outside without a tent, outside with a tent, car/bus/vehicle = 1; hotel, friend’s/relative’s home, temporary shelter, transitional or permanent housing = 0). About halfway through data collection, the survey added a series of questions about difficult experiences or conditions. One of these questions asked respondents if they had ever been diagnosed with a physical disability that limits their work (yes = 1, no = 0). Another asked if respondents had been diagnosed with a serious mental health challenge, like major depression, bipolar disorder, or schizophrenia (yes = 1, no = 0).

The L&F Survey measured monthly income with two questions. Total income for October included the following categories: none, $1–$100; $101–$250; $251–$500; $501–$,1000; $1,001–$2,000; and over $2,000. The survey also asked respondents which income sources they had in October: full- or part-time work; unemployment benefits; TANF or CalWorks; General Assistance; SSI or SSA (Social Security); child support payments; SNAP/CalFresh; or other. The main income variables for comparison with the survey of low-income housed Californians are whether the respondent had earned income in October (any income from full- or part-time work = 1, no = 0) and whether the respondent received SNAP/CalFresh benefits (yes = 1, no = 0). Some income variables are not directly comparable with the comparison surveys, so I describe them in the text without presenting estimates in the tables.

The second part of the analysis estimates self-reported impacts of the COVID-19 pandemic. Four questions asked about potential exposure to COVID-19. The first question asked whether respondents believe they have been sick with COVID-19, and the second asked if they believe they have been exposed to someone who was sick with COVID-19. The third question asked if respondents have been tested for the coronavirus. The fourth asked respondents if they have avoided staying in a shelter because they were afraid of the coronavirus.

Three questions asked about economic impacts of the pandemic. The first question asked respondents to rate their income in October compared to their income in February 2020: much lower, somewhat lower, about the same, somewhat higher, much higher. The second asked respondents if they have been laid off or lost a job due to the pandemic. The third asked respondents if they received a stimulus check from the federal government in the spring of 2020.

Sacramento point-in-time survey of homelessness

I compare the demographics and conditions related homelessness in the L&F Survey to published estimates from the 2019 point-in-time (PIT) count and survey of homelessness in Sacramento County (Baiocchi et al., 2019). The PIT count was an approximate census of homelessness conducted the nights of January 30 and 31, 2019. Shelters counted residents on January 30 while volunteer teams canvased 168 sites across the county on both nights. The face-to-face survey of a random sample of people experiencing homelessness provided demographic information (with sampling weights applied). Like all PIT counts, the Sacramento count likely missed some people experiencing homelessness. PIT counts also miss people cycling into and out of homelessness. The number of people experiencing homeless at any point during the year may have been twice as high as the PIT count (Baiocchi et al., 2019).

The demographic and homelessness variables in the L&F Survey are coded as similarly as possible to the PIT survey for comparison. The race/ethnicity measures are a notable exception. The Sacramento PIT survey includes both Latinas/os and non-Latinas/os as White or Black, and Latinas/os include those of any race. The PIT survey’s question on work disabilities included physical and mental disabilities. The PIT survey included a small number of people under age 18.

Household pulse survey

The Census Bureau’s Household Pulse Survey (HPS), collected October 14–26 2020 (Census Bureau, 2020), provides a comparison sample of low-income housed Californians. The HPS is a short online survey conducted every one or two weeks by the Census Bureau to track the social and economic impacts of the pandemic. A very large, random sample of adults was invited to take the survey (based on addresses), but the overall response rate was only 8.1%.

The HPS is a national survey also designed to represent states and several large metropolitan areas. I limit the analytic sample to respondents in California who reported less than $25,000 of annual income in 2019, the lowest income category in the survey. The survey is not able to identify or represent Sacramento residents, specifically.

I code key demographic variables in the HPS as similarly as possible to the L&F Survey. To measure access to resources, I examine whether the respondent worked for pay in the past week (yes = 1, no = 0) and whether the respondent or someone in their household received SNAP benefits (yes = 1, no = 0). I also estimate the fraction of that received free food from community organizations: food pantries/banks, shelters or soup kitchens, or “other community programs” (yes = 1, no = 0). This measure helps assess the extent to which low-income housed Californians use meal services similar to the ones provided by Loaves & Fishes.

To measure economic impacts of the pandemic, I analyze the HPS’s question asking if the respondent or anyone in their household lost “employment income” since March 13, 2020 (yes = 1, no = 0). Finally, I estimate stimulus payment receipt in spring of 2020 (yes = 1, no = 0) using an earlier wave of the HPS (collected July 16–21) because later waves no longer asked this question.

Results

L&F survey comparisons

Table 2 presents descriptive statistics for the L&F Survey, the 2019 PIT count of homelessness in Sacramento, and the HPS of low-income housed Californians. The L&F Survey’s demographic composition was generally similar to the 2019 PIT survey, but slightly older on average. Black people are around a third of the L&F Survey sample and the PIT survey in Sacramento, nearly three times their representation in Sacramento’s total population (Baiocchi et al., 2019). White people (both Hispanic and non-Hispanic) were nearly half of all people experiencing homelessness in Sacramento in the PIT survey. When limited to non-Latinas/os, White people were less than one-third of the L&F Survey. In both surveys, Latinas/os were underrepresented among people experiencing homelessness compared to Sacramento’s population.

Table 2. Comparison of L&F Survey with Sacramento’s point-in-time (PIT) survey of homelessness and the Household Pulse Survey (HPS) of low-income housed Californians.

	L&F Survey (Oct. 2020)	Sacramento PIT Survey of Homelessness (Jan 2019)	HPS of Low-Income Housed Californians (Oct. 2020)
Demographics
Mean Age	50	39	45
% Male	69%	62%	41%
% White ^a	30%	47%	30%
% Black ^a	31%	34%	7%
% Latina/o ^a	16%	18%	53%
% Unmarried Adults without Children	79%	73%	44%
Conditions Related to Homelessness
% Lived in Sacramento at least One Year	93%	90%
% Unsheltered	78%	70%
% Homeless at least One Year	74%	59%
% Physical or Mental Disability	53%	26%
% Severe Mental Health Condition	46%	21%
Access to Resources
% with Work for Pay	6%		32%
% Receive SNAP/CalFresh	42%		34%
% Receive Food from Community Org.^b	100%		14%
Sample Size	198	525	445
Estimated Population ^c	732	5,570	2,966,614

Notes: The main study sample is unweighted. The Sacramento PIT survey and HPS apply survey weights. “Low-income” in the HPS includes less than $25,000 for 2019 annual income, the lowest income category in the survey.

^a Unlike the L&F Survey and HPS, the PIT survey results do not differentiate Hispanic and non-Hispanic White or Black respondents.

^b The entire L&F Survey sample received free food by definition.

^c The population for the L&F Survey is the estimated number of unique individuals across all meal services during data collection (footnote 1) multiplied by the percentage of all survey respondents who identified as currently homeless. The population for the PIT Survey is the total PIT count for Sacramento County. The estimated population for low-income housed Californians is based on the author’s calculations with the 2019 American Community Survey.

Most notably, the L&F Survey included a higher fraction of people experiencing homelessness for at least a year compared to the PIT survey. The difference could reflect selection into to the L&F Survey or actual growth in durations of homelessness during the pandemic. Higher fractions in the L&F Survey reported a work-limiting disability or severe mental health condition (e.g. major depression, bipolar disorder, schizophrenia) compared to the PIT survey.²

Unsurprisingly, both the L&F Survey and 2019 PIT survey compositions substantially differed from the HPS of low-income housed Californians. For example, Latinas/os are the majority of low-income housed Californians but are underrepresented among people experiencing homelessness in Sacramento.

Two comparisons with the L&F Survey are particularly informative. First, any work for pay was much less common in the L&F Survey (6%; 95% CI: 3.3, 10.6) compared to low-income housed Californians (32%; 95% CI: 23.7, 39.6). However, rates of SNAP/CalFresh receipt were more similar between surveys. Second, 14% (95% CI: 7.5, 21.1) of low-income housed Californians reported receiving free food from some form of community organization, which would include organizations like Loaves & Fishes. Around half-a-million housed, low-income people in California may have used similar food services as the population in this study.

Self-reported pandemic impacts

Table 3 presents results for self-reported impacts of the COVID-19 pandemic in the L&F Survey, along with 95% confidence intervals and sample sizes for each estimate.

Table 3. Self-reported impacts of the coronavirus pandemic in the L&F Survey, with 95% confidence intervals (CI) and sample sizes (N).

	Percent of L&F Survey	95% CI	N
COVID-19 Exposure
Believes has had COVID-19	3%	[1.4, 6.6]	198
Believes has been exposed to someone with COVID-19	11%	[7.1, 15.9]	196
Has been tested for the coronavirus	61%	[54.0, 67.7]	195
Has avoided shelters due to fear of coronavirus	27%	[21.2, 33.8]	192
Economic Impacts
Income in October much/somewhat lower than in February, 2020	33%	[26.5, 39.8]	192
Been laid off or lost a job because of the pandemic	16%	[11.5, 22.0]	193
Received a stimulus check from the federal government	45%	[38.1, 52.0]	198

Consistent with Sacramento’s relatively low COVID-19 case rates, very few respondents believed they had COVID-19 or were exposed to someone that had it. Though the three percent of the sample who believed they had COVID-19 translates to 300 cases per 10,0000 people, it includes only six survey respondents.³ The 95% confidence interval also includes the percentage that would result from Sacramento’s cumulative case rate (Table 1). Relatively few respondents believed they had been in contact with someone who had COVID-19. While administering the survey, many respondents emphatically replied that they had not directly encountered the coronavirus itself in any way. A few people explicitly stated the coronavirus was “not a problem around here.”

Most people in the L&F Survey were tested for the coronavirus at some point. Relative to all people experiencing homelessness in Sacramento, the L&F Survey sample likely had especially high access to testing because Loaves & Fishes also administers coronavirus tests.

More than one-in-four respondents avoided shelters because of the coronavirus. Many respondents elaborated on their answer beyond the survey question. Some answered “no” because they already avoid shelters for other reasons, like being too crowded or having too little privacy (consistent with other surveys about shelter usage, e.g. Applied Survey Research, 2019). Others answered “no” because the risk of COVID-19 was preferable to remaining unsheltered. For example, one middle-aged female respondent told me she continued to use shelters because, “I just can’t be out there in all that.”

Self-reported economic impacts of the pandemic were more common than self-reported disease impacts. About one-third of the sample reported that their income was lower in October than in February 2020. One-quarter (95% CI: 19.4, 31.7) said their income was “much lower” and 8% (95% CI: 4.7, 12.6) said “somewhat lower.” Only 5% (95% CI: 2.8, 9.4) of the sample said their income had increased between February and October. Most of the sample, 62% (95% CI: 54.9, 68.6), said their income remained “about the same.”

About 16% reported losing a job due to the pandemic (Table 3). Of those reporting job loss, 65% (95% CI: 45.8, 79.6) said their income declined between February and October. In comparison, 27% (95% CI: 20.5, 34.5) of those who did not report job loss said that their income declined. Overall, 39% (95% CI: 32.3, 46.3) of people in the L&F Survey reported a job and/or income loss since February 2020.

Employment/income losses were much more pronounced in the HPS of low-income housed Californians, in which 58% (CI: 47.6, 67.8) said that they or someone in their household “experienced a loss of employment income since March 13, 2020.” However, the lower fraction reporting income/job losses among the L&F Survey largely reflected very low employment and incomes in the first place. Very few people in the L&F Survey reported any earnings from full- or part-time work (6% in Table 2). The most commonly reported income types in the L&F Survey, “SNAP/CalFresh” and “SSI or SSA (Social Security),” were not contingent on the pandemic.

Many people in the L&F Survey did not report declines in monthly income because they had little room to fall. Two-thirds (95% CI: 59.7, 73.0) in the L&F Survey reported a monthly income of $500 or less in October and 32% (95% CI: 25.5, 38.7) reported “none.” Some respondents explicitly told me their income did not change during the pandemic because they “never really had any anyway.”

Finally, most people in the L&F Survey did not report receiving a stimulus check from the federal government (Table 3). Those with October incomes over $500 were more likely to say they received the stimulus, 70% (95% CI: 59.1, 81.5), compared to those with incomes of $500 or lower, 31% (95% CI: 23.2, 39.3). In contrast, 86% (95% CI: 78.9, 92.5) of low-income housed Californians reported receiving a stimulus check in an earlier wave of the HPS (collected July 16–21). These patterns suggest that the pandemic relief funds failed to meet those with the lowest incomes both between the housed and unhoused and among the unhoused.

Discussion

This study examined self-reported impacts of the COVID-19 pandemic for people experiencing homelessness in Sacramento, California. In collaboration with Loaves & Fishes, a homelessness service provider, I administered a face-to-face survey of almost 200 people experiencing homelessness on October 26–28, 2020. The survey captured self-reported impacts of the COVID-19 pandemic with respect to disease exposure and economic resources.

Results from the survey suggested limited exposure to the coronavirus for people experiencing homelessness in Sacramento, and most had access to testing. Concerns about shelter safety due to the coronavirus were fairly common, but other orientations toward shelters were as or more important for many people.

Self-reported economic impacts (i.e. income and job losses) were more pronounced than disease exposure for people experiencing homelessness. Income/employment losses were even greater for low-income housed Californians. However, lower economic losses for people experiencing homelessness reflected persistent economic deprivation since before the pandemic. Compared to housed Californians, fewer L&F Survey respondents lost jobs or income because they frequently had little to lose in the first place. Moreover, far fewer people in the L&F Survey received an economic stimulus check compared to low-income housed Californians, perpetuating their economic deprivation.

The L&F Survey may have important limitations that affected these results. First, the sample likely did not include anyone participating in COVID-19 temporary housing programs, potentially biasing estimates of perceived exposure downward. This bias may have been small though, because Sacramento County reported extremely few positive cases among people experiencing homelessness (Sacramento County, 2020).

Second, sampling people through a homelessness service provider could have systematically missed people experiencing homelessness but with more resources. For example, the L&F Survey included more older people and people with disabilities or mental health challenges than the 2019 PIT survey for Sacramento. The L&F Survey also included a higher fraction of people experiencing homelessness for at least one year compared to the PIT survey. This difference could be due to different rates of disabilities and mental health challenges, greater service use with chronic homelessness, or real growth in typical lengths of homelessness during the pandemic. Employment rates might also be higher for people that were systematically not observed in this sample. I cannot know how self-reported impacts of the pandemic may differ for this unobserved portion of the Sacramento population experiencing homelessness. But if this unobserved portion of the population was slightly more economically advantaged than the study sample, their pandemic impacts may have been somewhere between those in the L&F Survey and the HPS of low-income housed Californians.

Finally, this study’s results have implications for future research and practice. Consistent with experiences in other cities, like Dallas (Benavides & Nukpezah, 2020) or San Diego (Marquez et al., 2020), proactive and widespread coronavirus testing and isolation may have been crucial for protecting people experiencing homelessness. These results also suggest that economic and housing supports for people experiencing homelessness should not treat the pandemic as a new set back, as it might be for many otherwise middle-income people. Instead, the pandemic mainly compounds the enduring marginalization that many of the study participants experienced, particularly given the complex recovery needs for people high rates of disabilities and mental health challenges (Padgett et al., 2016). This marginalization calls for long-term, not short-term, support.

Consistent with that goal, California’s short-term pandemic housing program, Project Roomkey, is transitioning to a longer-term transitional housing program, Project Homekey. The California Department of Housing and Community Development will spend up to $600 million for local governments to obtain vacant hotels, motels, and other housing units for people experiencing homelessness (HCD, 2020). However, the program would be far more effective for this study’s participants if it expanded the earlier Project Roomkey’s eligibility restrictions, which mainly targeted to those at high risk of coronavirus infection or severe illness (Berry-Jester & Hart, 2020). Project Homekey could also consider a broader definition for how people experiencing homelessness have been impacted by the pandemic, beyond just disease exposure or vulnerability.

Finally, economic relief efforts should proactively address access barriers due to housing status. The second round of stimulus payments in January 2021 likely failed to reach many people experiencing homelessness in the same manner as the first round, as will any further payments without some adjustment. First, greater outreach efforts could inform people experiencing homelessness that they were eligible for stimulus payments even without filing 2019 taxes. Second, the IRS could allow recipients to opt into payment debit cards in place of direct deposit or physical checks. A very small fraction of people to receive these debit cards in the first round of stimulus payments, but only at the discretion of the Treasury Department. These and additional efforts to overcome the banking and identification requirements for stimulus payments could partially remedy unequal access by housing status.

Acknowledgements

I thank Joe Smith, advocacy director at Loaves & Fishes, for his leadership of the survey administration and open collaboration. I also thank the volunteers who helped collect the survey data and the generous survey participants who shared their time and personal experiences. This article reflects the views of the author and does not necessarily reflect the views of Loaves & Fishes. The editor and anonymous reviewers provided supportive and constructive feedback.

Disclosure statement

No potential conflict of interest was reported by the author.

Additional information

Funding

This research was conducted in collaboration with Loaves & Fishes in Sacramento, CA.

Notes on contributors

Ryan Finnigan

Ryan Finnigan, Ph.D., is an associate professor of sociology at the University of California, Davis.

Notes

1 I roughly estimated the number of unique individuals at the five meal services using several pieces of data. First, Loaves & Fishes tracked the number of meals served at each service, totaling 1,363 meals. Second, the L&F Survey showed that 88% of respondents receive both breakfasts and lunches from Loaves & Fishes. Using these data, I estimated that 88% of the breakfast guests on the first day of data collection returned for lunch and were not new unique individuals for that meal service. Third, the L&F Survey showed that 75% of respondents use Loaves & Fishes services “more than once a week.” Many people told me they went every day. Using these data, I estimated that 75% of guests from meal services in the previous day returned on the second and third days of data collection, and they were not new unique individuals. The total number of estimated unique persons across all meals was 732.

2 Though not formally measured, I encountered very few respondents who were unable to understand or complete the survey. Only three respondents did not finish the survey because they were not able to understand or answer the questions. These responses were excluded from the final sample.

3 Of those six, one described symptoms to me that are inconsistent with COVID-19. Another described having symptoms similar COVID-19 but in November/December of 2019, before the first COVID-19 diagnosis in the United States.