Who Lives in Vehicles and Why? Understanding Vehicular Homelessness in Los Angeles

Abstract

Homelessness continues to grow and to affect the lives of an increasingly diverse group of individuals. Many scholars have studied people living in homeless shelters and outdoors in tents. An overlooked population is the growing number of the unhoused living in vehicles. We draw on data from the Los Angeles Homeless Services Authority’s Homeless Demographic Survey to understand the characteristics of people living in vehicles and the extent to which they differ from the nonvehicular unhoused population. Compared to those living in tents, in makeshift shelters, and in public spaces, people living in vehicles are more likely to be women and to live in larger households with children, and are less likely to be chronically unhoused. These findings will help effectively target policies and services. Safe parking programs can provide temporary relief to those living in vehicles and, if done well, the interventions necessary to transition into permanent housing.

Keywords:

• Vehicular homelessness
• transportation
• safe parking
• housing
• Los Angeles

Vehicular homelessness—referring to individuals and households living in cars, vans, recreational vehicles (RVs), and other vehicles—is one of the fastest growing segments of the unhoused population (Ho, 2019; Lyons-Warren & Lowery, 2020). Estimates range in size and, due to the difficulty of counting those who live in vehicles, are likely undercounted. In San Francisco, 35% of the unhoused population lives in a vehicle compared to 50% in Seattle (All Home, 2020; Applied Survey Research, 2019c) and almost 60% in Los Angeles (Los Angeles Homeless Services Authority, 2020b). In some other California counties, the reported percentage of people living in vehicles ranges from 10% to 29% (Applied Survey Research, 2019a, 2019b, 2019d, 2019e, 2019f, 2019g). Although there is a growing body of research on the unhoused and some of its sub-populations, very few studies focus on the substantial and growing segment of this population that lives in vehicles.

To address this scholarship gap, we analyzed survey data collected by the Los Angeles Homeless Service Authority (LAHSA). LAHSA is the Continuum of Care (CoC) provider for Los Angeles County (hereafter LA CoC).¹ LAHSA’s survey is a unique government-produced data set that provides demographic, socioeconomic, systems involvement, and other lived experience data on a sample of individuals experiencing homelessness. In partnership with the University of Southern California (Henwood et al., 2020), LAHSA voluntarily administers the survey in conjunction with their annual point-in-time (PIT) count. The U.S. Department of Housing and Urban Development (HUD) requires PIT counts for CoCs to be eligible to receive funding for services and housing but does not require agencies to survey their unhoused populations.

Using these survey data, we compared the characteristics of individuals living in vehicles to individuals living in tents, makeshift shelters, and other public spaces (hereafter termed nonvehicular unhoused). We ask three questions:

1. What are the characteristics of those living in vehicles?

2. How do these characteristics differ from those of people living in tents, on sidewalks, or in makeshift shelters?

3. Do the factors that help predict homelessness differ between these two groups?

The answers to these questions are critically important to local agencies working to deliver services and provide permanent housing to the diverse needs of their unhoused populations.

We start by reviewing the small body of research on vehicular homelessness. We then describe our data and methodology, followed by our findings. Finally, we discuss short- and long-term interventions specifically targeted to people living in vehicles.

Vehicular Homelessness and People Living in Vehicles

Vehicular dwelling has a long and mixed history in the U.S. Mobile homes emerged in response to the Great Depression and the Second World War to meet the need for low-cost housing (Hall & Muller, 2018). Many people viewed them negatively as both cheap and transient relative to permanent housing firmly rooted in communities (Beamish et al., 2001; Hall & Muller, 2018; Kusenbach, 2009; MacTavish, 2007). Consequently, mobile residents were stigmatized, a prevalent attitude captured by the phrase “trailer park trash.” Mobile homes are a relatively large segment of the housing market in rural areas (Aman & Yarnal, 2010), where poverty rates are high and fewer local zoning restrictions limit their location (Meeks, 1988).²

Mobile homes can be found in urban areas too. A study of mobile homes in Los Angeles County identified 601 mobile home parks, scattered throughout various land uses (Pierce et al., 2018). Los Angeles County zoning codes allow manufactured housing in most land-use zones. Although many people live in mobile homes, RVs and vans have been the darling of outdoor enthusiasts who live in these vehicles while traveling (Counts & Counts, 2001; Hall & Muller, 2018; Twitchell, 2014).

Vehicular homelessness is a relatively recent urban phenomenon, born out of necessity—individuals or families who lose their homes but manage to retain their vehicles for shelter. As the population experiencing homelessness has grown, so has the population of people living in vehicles (Pollard, 2018). For example, within the LA CoC, the number of persons living in vehicles increased by 55% from 12,200 in 2016 to 18,900 in 2020 (Los Angeles Homeless Services Authority, 2020c). Of this group, 49% lived in RVs and campers, 28% in vans, and 23% in cars (Los Angeles Homeless Services Authority, 2016, 2020b). Although the numbers are not yet available, vehicular homelessness will likely continue to grow in the wake of the ongoing COVID-19 pandemic (Guynn, 2021).

There are very few studies on vehicular homelessness, largely due to data limitations. HUD requires CoCs to report aggregate data on the number of unhoused individuals (Herring, 2021), prompting PIT counts, which provide information on their number and spatial location. HUD does not require the CoCs to report on the number of individuals living in vehicles, but some agencies—such as the LA CoC—collect these data. Although useful, these counts likely underreport the number of people living in vehicles, because people living in vehicles are mobile and often try to blend in with other parked vehicles to avoid the growing number of regulations either banning or restricting vehicle dwelling. Finally, other than counts and location, homeless agencies typically collect few other details about this population group.

However, the small—largely qualitative—body of research on this population group provides some insights into the characteristics of those living in vehicles and the factors associated with vehicular homelessness. Vehicles are expensive to own and operate (Frederick & Gilderbloom, 2018; Johnson et al., 2010; Lutz, 2014); therefore, vehicle dwelling is more likely to be an option for individuals and families with greater and more stable resources. In a study of people living in oversized vehicles (RVs) in Oakland, CA, most interviewees reported receiving regular income from either low-paying jobs or benefits (Pruss et al., 2022). A steady income may allow for the purchase of a vehicle or, more likely, may enable some unhoused individuals or families to hold on to their vehicles long after the loss of their homes.

Safety is an important motivation for living in vehicles, suggesting that the most vulnerable population groups might be more likely to live in vehicles compared to other groups. Many people living in vehicles claim that they are safer and provide more stable and secure lodging than living in public spaces such as sidewalks, underpasses, or parks (Lyons-Warren & Lowery, 2020; Pollard, 2018; Quinn, 2018). For women who are unhoused due to domestic violence, vehicles can provide safety for themselves and their children while allowing them the mobility necessary to take care of their families (Baker et al., 2010; Bassuk et al., 2001). Increasingly, cities are using municipal laws to criminalize homelessness; vehicle dwelling may better shield the unhoused from policing, fines, and criminalization (Darrah-Okike et al., 2018; Dozier, 2019; Herring, 2019; National Law Center on Homelessness & Poverty, 2019; Wakin, 2008).

Some people living in vehicles assert that vehicles are a less restrictive and safer alternative to homeless shelters. Researchers from the University of California San Francisco interviewed vehicle residents. They found that many participants “would not trade their vehicle residence for a place in a congregate shelter because they wanted to stay with their pets, partners, and property” (Pruss et al., 2022, p. 10). Many unhoused people are resistant to living in homeless shelters—particularly shelters located in downtown areas—based on concerns about crime (Donley & Wright, 2012). Finally, during the COVID-19 pandemic, vehicles provided opportunities for social distancing compared to overcrowded homeless shelters, reducing the likelihood of contracting the virus (Moody, 2020). In Los Angeles, the vaccination rate among the county’s unhoused population is low; as of July 2021, only 33% were fully vaccinated (Oreskes & Smith, 2021).

Finally, individuals with specific needs may consider vehicle dwelling preferable to other options. A number of studies find that the unhoused face transportation barriers to accessing resources (Brallier et al., 2019; Crawley et al., 2013; Nickasch & Marnocha, 2009). Vehicles, therefore, may enable unhoused individuals to overcome these barriers and reach necessary destinations such as employment, social services, and healthcare (Pruss et al., 2022). Moreover, compared to living on the streets, vehicles can function as a form of stable shelter and stabilize or improve mental health outcomes by enabling the unhoused to avoid posttraumatic stress disorder (PTSD) triggers from traumatic experiences that they had experienced while living unsheltered in public spaces (Henwood et al., 2013).

To summarize, the number of people living in vehicles has grown substantially over the last few decades. Vehicular dwelling has some benefits relative to other shelter options; it can provide better safety, privacy, and access to benefits compared to sleeping in public spaces or shelters (Ivey et al., 2018; Pruss & Cheng, 2020; Wakin, 2014), suggesting that certain types of individuals and families may be more or less willing to live in their vehicles. However, there has been almost no research on the characteristics of individuals or households living in vehicles. A better understanding of this group is a necessary first step to improving their living conditions. Further, limited space, amenities, and protections make life very difficult for those who live in vehicles (Pruss et al., 2022). Therefore, a better understanding of this population is needed to develop services and programs to help them transition into long-term, stable housing.

Means and Methods

This study examines people living in vehicles in the LA CoC. We use data from the LAHSA 2020 demographic survey to test whether this population group differs from those living in tents, on sidewalks, or in makeshift shelters.

LAHSA conducts a demographic survey in conjunction with their annual PIT counts. The survey includes detailed characteristics of a sample of people experiencing homelessness, information that is not otherwise available from the PIT counts required by HUD. Unlike the PIT counts that rely on volunteers, these data were collected by a team of researchers and service provider surveyors over 3 months, from December 1, 2019, through March 1, 2020.

Surveyors asked respondents about their dwelling type and location. This information included the city where the survey took place, the community or city where they slept the previous night, and the type of place where they slept the previous night (i.e., shelter, home, street, tent, car, other vehicle, etc.). Surveyors then asked respondents where they had spent most nights in the past 30 days, repeating the same set of locations. Respondents who slept in a shelter in both the last 30 days and the previous night were ineligible for the survey. Respondents then answered 60 questions about their demographic and socioeconomic characteristics. They also reported on their mental health, physical health, and trauma-related experiences, including those associated with criminal justice, armed services, foster care, and domestic violence, and the number of years they had been unhoused. Finally, respondents identified up to 22 reasons why they were experiencing homelessness.

The survey included 3,873 total responses. We removed records with incomplete information and those of respondents who had spent some time in a sheltered location, resulting in a final count of 3,754 responses. Seventeen percent of survey respondents reported living in a vehicle (641) compared to 83% of respondents living in tents, on sidewalks, in makeshift shelters, and in public spaces (3,113; Los Angeles Homeless Services Authority, 2020a). In comparison, for the same year, the PIT count counted 18,544 people living in vehicles (60%) and 12,079 people living in encampments (40%). Among survey respondents living in a vehicle, 341 (53%) lived in cars, 69 (11%) lived in vans/SUVs, and 231 (36%) lived in RVs. The discrepancy in these figures suggests that the survey underrepresents people living in vehicles relative to the PIT count.

The lack of data on this population group in Los Angeles makes it difficult to determine the extent to which the sample represents the larger population group. However, the LAHSA used a variety of approaches to generate a representative sample for their demographic survey. According to a July 2020 methodology report that details sample selection for the demographic survey (Henwood et al., 2020), LAHSA and their partners identified 505 census tracts from which to survey a random sample of unhoused individuals. They selected these tracts based on local geographic knowledge from community partners at planning sessions, stratified hotspot mapping of PIT count data, and the location of youth hotspots and safe parking lots. By including the latter two groups—unhoused youth and people living in vehicles—the report states that the survey data was designed to be more representative of the total unhoused population of the LA CoC compared to their previous surveys (2016, 2017, 2018, and 2019).

A recent analysis of the spatial trends of vehicular homelessness shows that vehicular homelessness is randomly distributed and less concentrated than other forms of unhoused homelessness (Giamarino et al., 2022). Across LA County, there were only 17 safe parking program (SPP) sites, which served approximately 508 people—a small portion of the vehicular homeless population (McElwain et al., 2021). According to the demographic survey data, only 83 out of the 641 respondents living in their vehicles (∼12%) were surveyed at SPP sites, loosely reflecting the distribution of the population living in vehicles within and outside these sites.

Although the survey underestimates the number of people living in vehicles, the proportion of people living in cars compared to RVs and campers is more similar between the two data sources (Figure 1). The differences between the survey and PIT count percentages may arise because most SPP locations in Los Angeles do not permit RVs.

Figure 1. Distribution of survey respondents and overall population living in cars and vans compared to recreational vehicles (RVs) and campers.

Therefore, we cannot know whether the people surveyed are representative of the vehicular unhoused population. However, the extensive efforts of LAHSA to collect a representative sample, the geographic dispersion of respondents outside of SPP sites, and the similarity in the distribution by vehicle type between the survey and the PIT count are all promising signs. Given that this survey is one of the only surveys that includes people living in vehicles, we find these data to be a useful, albeit limited, source of information.

We use these data to examine the association between vehicular homelessness and characteristics in four different, but related, domains that research suggests might be associated with vehicular dwelling: sociodemographic, housing, trauma-related, and mobility. First, with respect to sociodemographic characteristics, we expect that people who live in vehicles may be more likely to be female because of the safety benefits of living in a vehicle (Wakin, 2014). Because of this gender-related hypothesis, we theorize people living in vehicles also will be more likely to live with children. We also hypothesize that such women may have become unhoused due to domestic abuse, one of the pathways identified in a review on gender and homelessness (Phipps et al., 2019).

The second set of hypotheses are predicated on the value of the vehicle as an asset. Because people living in vehicles possess this asset, we suspect they are also likely to have higher incomes and employment rates. Because they have held onto this asset, they may have experienced homelessness for shorter periods and, therefore, would be less likely to be chronically unhoused. HUD (U.S. Department of Housing and Urban Development, 2022) defines chronic homelessness as an individual living unhoused for 12 consecutive months and falling into one of three categories: living with a disability, recently stayed in an institutional care facility, or serving as a head of household.

Finally, given the ease of automobile travel relative to other modes, we expect vehicle ownership to be positively associated with greater travel distances. We measured travel distance by estimating the distance between where the respondent had slept within the past 30 days and where the respondent slept the previous night before they were surveyed. We calculated this as the straight-line distance between the geographic center of the city or neighborhood for these two locations using geographic information systems.

We tested these hypotheses in two ways. First, we produced descriptive statistics, comparing survey respondents who reported living in a vehicle versus survey respondents who reported living in a tent, on the sidewalk, or in a makeshift shelter. We tested for significant differences using t-tests (continuous variables) and Chi-squared tests (binary variables). Due to multicollinearity, we excluded certain variables from our logistic regression model. For example, identifying as White or Black or reported years unhoused and falling into the definition of chronic homelessness were highly collinear. Second, we developed a logistic regression model (logit model) to better understand the relationship between the likelihood of living in a vehicle relative to being unsheltered, controlling for a set of independent characteristics. The model takes the following form: $$\mathit{logit}\mathrm{}\left(p\right)=\ln \left[\frac{p}{\left(1-p\right)}\right]={\beta }_0+{\beta }_1{X}_1+\mathrm{}\dots +{\beta }_m{X}_m+\epsilon$$ where p is the probability that the respondent is vehicular homeless, p/(1 − p) is the odds ratio, β is a vector of coefficients, m is the number of explanatory variables, and ε is the random error term. Due to missing data, the model included 1,533 observations (down from 3,753). Most of the missing data relates to self-reported monthly income (1,789 missing responses), which we chose to keep in our model given our second asset-based hypothesis that people living in vehicles are likely to have a higher monthly income and to be employed.³

Who Is Living in Vehicles in Los Angeles and Why?

The survey data showed that people who lived in vehicles in Los Angeles were different from the nonvehicular unhoused across two of our four survey response categories. We found statistically significant differences in sociodemographic characteristics and housing variables. Table 1 presents the descriptive statistics for three groups: the full sample, the nonvehicular unhoused, and those living in vehicles. The data are organized by our four domains of analysis, and the table includes statistical tests for differences in the characteristics of the vehicular and nonvehicular respondents.

Table 1. Characteristics of total, nonvehicular unhoused, and vehicular unhoused individuals.

Individual characteristics	Total	Tents, sidewalks, makeshift, and public spaces	Cars, RVs, vans, and campers	Significance
Sociodemographic
Gender
Male	68%	71.4%	53%	***
Female	30%	26.2%	46%	***
Transgender	1.6%	2%	.8%	**
Gender nonconforming	.2%	.2%	.1%
Missing data	.2%	.2%	.1%
Race/ethnicity
White	33%	31%	36%	**
Black	32%	32%	32%
Asian	2%	2%	1%
American Indian/Native Hawaiian	2%	2%	2%
Multiple races	3%	3%	3%
Hispanic	25%	26%	23%
Missing data	3%	4%	3%
Senior (over 62 years old)	11%	10%	15%	***
Living with children	5%	2%	18%	***
Monthly income (mean)	$743	$567	$1,458	***
Employment status
Employed	14%	13%	19%	***
Unemployed (looking, not looking, or student)	56%	29%	30%	***
Retired	2.1%	2.2%	3.6%	**
Missing data	12%	13%	8%
On disability	17%	16%	21%	***
Housing
Chronically unhoused	48%	50%	38%	***
Years unhoused (median)	4	4	3	***
Years unhoused (mean)	6.5	7	5	***
Physical disability	43%	42%	53%	***
Mental health	51%	52%	48%
Justice system	78%	80%	68%	***
Household with children	2%	1%	6%	***
Main reasons that led to loss of housing (avg. # reported)	1.65	1.51	1.48	**
Domestic violence	9%	9.2%	10.7%
Family/personal	20%	20.4%	19.1%
Housing	12%	12.5%	12.5%
Incarceration	6%	5.3%	3.6%	**
Mental and physical health	16%	17.4%	12.1%	***
Unemployment	31%	30.2%	32.4%	*
Trauma-related experiences
Domestic violence (%)	36%	27%	31%
Veteran/armed services (%)	7%	7%	7%
Mobility
Slept in different neighborhood than last night in the past 30 days	11%	11%	9%
Distance between last 30 days and last night (median)	2 miles	2 miles	2 miles
Total survey responses (n)	3754	3113	641

Note. RVs = recreational vehicles.

*p < .1; **p < .05; ***p < .01 for t-test (continuous) or Chi-squared test (binary) significance between vehicular and nonvehicular respondents.

Compared to nonvehicular respondents, people living in vehicles identified as female, White, and older at higher rates. The vehicular unhoused were also more likely to be in households with children. People living in vehicles reported higher employment rates and were more likely to be actively looking for work. Seattle is one of the few cities that has reported on the demographic differences between people living in cars relative to other people experiencing homelessness. Their data show some similar results. People living in vehicles were more likely to be female (42% vs. 70%), White (55% vs. 41%), and older (97% over age 25 vs. 45%) than people experiencing homelessness not living in vehicles (All Home, 2020).

In terms of housing and unhoused status, we find that people living in vehicles experienced less chronic homelessness than people living in tents and on sidewalks. They also reported lower numbers of reasons that led to their loss of housing. Among people living in vehicles, unemployment was the main reason for becoming unhoused. People living in tents, on sidewalks, and in public spaces were more likely to have reported that mental and physical health or incarceration were the main reasons that led to their loss of housing.

Consistent with previous ethnographic research on people living in vehicles (Wakin, 2014), but in contrast to our hypothesis, we found that survey respondents who slept in their vehicles and people who slept in tents and on sidewalks were relatively similar in terms of their residential mobility. A small percentage of respondents in both groups had slept in a different neighborhood the past 30 days compared to where they slept the night before they were surveyed. Among those who had slept in a different neighborhood the night before they were surveyed, both groups had the same median distance between the two locations. This finding is in line with recent surveys of people living in RVs in Oakland; researchers found that study participants parked in particular places to preserve connections with employment opportunities and social networks (Pruss et al., 2022).

We present the model results for the likelihood of living in a vehicle in Table 2 and coefficient plots in Figure 2. Controlling for other characteristics, compared to nonvehicular unhoused individuals, the largest predictor of someone living in a vehicle was being in a household with children. This variable was 5 times more predictive than the next highest contributor. Other significant and strong predictors included being retired, female, employed, older (over age 62), and receiving disability payments. In contrast, our model predicted that respondents who were unhoused and lived in tents or encampments were more likely to be chronically homeless, have physical or mental health reasons that caused them to be unhoused, or have lower incomes. Within the chronically homeless category, people living in tents or encampments were more likely to have been involved in the criminal justice system and less likely to have been in a household with children.

Figure 2. Graphic representation of logit model results.

Table 2. Logit model results—likelihood of living in a vehicle.

	Odds ratio	SE	p > |z|
Gender (excluded category: Male)
Female	2.10	0.32	***
Transgender	0.52	0.35
Race/ethnicity (excluded category: Black)
White	1.65	0.28	***
Asian	1.65	1.11
American Indian/Native Hawaiian	0.92	0.43
Multiple races	0.74	0.30
Hispanic	1.14	0.22
Senior (> 62 years old)	1.73	0.39	**
Living with children	11.47	3.07	***
Employment/occupational situation (excluded: Unemployed)
Employed	1.98	0.37	***
Retired	2.30	0.91	**
Monthly income	1.00	0.00002	**
Receives disability benefits	1.69	0.31	***
Chronically unhoused status	0.64	0.10	***
Main reasons for becoming homeless
Violence	0.96	0.20
Family/personal	0.89	0.14
Housing	1.14	0.20
Incarceration	0.53	0.19	*
Mental/physical health	0.72	0.13	*
Unemployment	0.94	0.13
Domestic violence victim	0.93	0.14
Veteran status	1.04	0.28
Slept in different place within last 30 days	0.70	0.17
Distance between last two places slept	1.01	0.01
Constant	0.14	0.03

Note. SE = standard error.

*p < .1. **p < .05. ***p < .01.

People who were living in vehicles in Los Angeles had different needs than those living unsheltered in tents and public spaces. They were newly unhoused and less likely to be experiencing mental and physical health issues or to have been involved in the criminal justice system. Comparatively, those dwelling in vehicles were more likely to identify as female, have children, be an older adult, be on disability and/or retired, or work and/or have a monthly income. The growing population of people living in vehicles presents unique challenges and opportunities for developing targeted policy solutions to provide safe places to park, access to life-sustaining services, and transition into permanent housing.

Discussion, Conclusions, and Policy Implications

The data confirm that the population of people who live in vehicles is significantly different from those who shelter in tents, in makeshift shelters, on sidewalks, and in other public spaces. Taking these differences into account when crafting policy responses enables policymakers to better tailor programs and services to people’s unique needs. As we discuss below, site-specific interventions can help address both the short- and long-term needs of people living in vehicles.

Although people living in vehicles are comparatively “better off” than their unsheltered peers, this temporary form of shelter is still precarious. Vehicular residents have reported a life of constant mobility in finding safe places to park, a lack of ability to cook for or clean themselves, and a fear of ticketing, towing, and potentially the loss of their vehicle (Pruss et al., 2022; Wakin, 2014). SPPs are short-term strategies to reduce the harm of living in a vehicle. Often organized in partnership with local community organizations such as churches, SPPs provide people living in vehicles a legal place to sleep overnight, an interim solution until they can transition into permanent housing. SPPs primarily rely on partnerships with organizations that are willing to transform their parking lots into quasi-trailer park communities. Los Angeles County has an SPP (SafeParkingLA, 2021b); however, the program is not scaled to meet the magnitude of the population. From the 2020 survey, only a small sample (12%) of people surveyed were living within SPP sites. SPPs traditionally operate from 7 p.m. to 7 a.m., so it may have been difficult to survey relevant respondents during the day. In Los Angeles County, there are 17 safe parking sites with 439 spaces available that can serve approximately 508 people, a fraction of the 11,124 vehicles counted in the 2020 PIT homelessness count (County of Los Angeles, 2021). The LA County SPP is a step in the right direction but could be improved to better serve the population of people living in vehicles, given the current barriers to applying for and participating in Los Angeles’ SPP (SafeParkingLA, 2021b). People eligible to participate in SPPs need to have operational vehicles. They must also possess a valid driver’s license and prove that they are the vehicle’s owner via insurance and registration. After completing an application, their approval is subject to vehicle type and lot availability. Most lots do not permit RVs/campers, yet 36% of people living in vehicles reside in these vehicle types in Los Angeles. With one exception (a parking lot at the Veteran Administration Building which is available only to veterans), most safe parking lots do not accept RVs, which currently comprise half of the vehicular homeless population in Los Angeles (Los Angeles Homeless Services Authority, 2020a). This requirement helps to explain the small number of vans, SUVs, RVs, and campers among the respondents sampled at SPP sites. Only 11 of the 83 respondents (13%) lived in vans or SUVs and only three (4%) lived in RVs or campers.

The 2020 demographic survey added SPP sites to the survey protocol to collect a more representative sample of people living in vehicles across Los Angeles. We assessed whether there are differences within the vehicular homeless population based on where people were surveyed. We found half as many differences between survey sites in the vehicular homeless population, compared to the number of differences between the vehicular homeless and nonvehicular homeless population. People surveyed in SPP sites were significantly different in terms of more identification as multiple races, less identification as Hispanic, fewer people living with children, more employed, and less involvement with the criminal justice system among the chronically unhoused population. A significantly greater percentage of people living in SPP sites were veterans and fewer had slept in a different location the previous night (Appendix A).

SPPs in Eugene, Oregon (Safe Sleep Site) and Vancouver, British Columbia, Canada (safe parking zone) provide spaces for RVs that allow two-person households and families to park and access hygiene infrastructure, internet, services, and opportunities for housing (City of Eugene, 2022; City of Vancouver, 2022). Following a two-year evaluation that included over 100 interviews with people living in safe parking sites in San Diego (Lewis et al., 2022), the City of San Diego approved three SPP sites to operate 24 hours a day and specifically target newly unhoused individuals, senior citizens, and households with children (The City of San Diego, 2022). SPP policies should be designed to ensure that families living in vehicles have a place to be during the day and that RV dwellers can find safe and legal spaces to park.

SPPs should strive not only to be safe places to park, but also to serve as hubs to connect the unhoused community to resources. Many people who shelter in vehicles, especially those living in cars and vans, purposely hide in plain sight, making it difficult for outreach workers to target their efforts. This is especially true for people living in cars and vans which are harder to spot than large, oversized vehicles such as RVs. SPPs could serve as one point of contact. Outreach or case workers could help participating families enroll in low-income government assistance programs and link them to other community resources. Lower-income households often own older vehicles that are more likely than newer vehicles to have mechanical problems (Johnson et al., 2010; Lutz, 2014). For example, safe parking LA allocated over 70% ($79,526.31) of its 2021 budget to assisting clients with car repairs at SPP sites (SafeParkingLA, 2021a). SPPs should help with vehicle repair.

Ideally, SSPs can provide opportunities to rapidly rehouse people living in vehicles so that they avoid becoming chronically unhoused. People living in vehicles were more likely to work and to have experienced homelessness for shorter durations, and were less chronically unhoused. This group may respond well to rapid-rehousing assistance as they face fewer barriers to being rehoused relative to other unsheltered populations. Santa Barbara’s New Beginnings serves as a national example of a successful SPP (Wakin, 2014). Over the last 15 years, the organization placed approximately 1,000 people into permanent housing through rapid rehousing (New Beginnings, 2021). We caution, however, that rapid rehousing alone may not be sufficient given prior evidence that rapid rehousing without ongoing housing assistance may force families back into homelessness (Gubits et al., 2018). The approach to more permanently house people living in vehicles must, most likely, include a combination of rapid rehousing and housing subsidies.

Finally, we recommend that federal government-mandated data collection of people experiencing homelessness expand to include counts of people living in vehicles and consider opportunities to also systematically collect survey data to accompany PIT counts. Although it is costly to do so on an annual basis, LAHSA’s survey provides important insights into the unique characteristics and needs of people experiencing homelessness. Additional outreach is necessary to ensure that the survey includes a random, representative sample of people living in vehicles. Many other CoCs throughout the U.S. could benefit from including a survey as part of their homeless data collection efforts.

Overall, this analysis highlights the heterogeneity of people experiencing homelessness. A better understanding of this diverse population group—including the places that they shelter—is necessary to effectively target interventions. If done well, homeless policies and services—such as SPPs—can meet the immediate needs of those experiencing homelessness while also facilitating their transition to permanent and stable housing.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Notes

Additional information

Funding

Research presented in this paper was made possible through funding received by the University of California Institute of Transportation Studies (UC ITS) from the State of California through the Public Transportation Account and the Road Repair and Accountability Act of 2017 (Senate Bill 1). The UC ITS is a network of faculty, research and administrative staff, and students dedicated to advancing the state of the art in transportation engineering, planning, and policy for the people of California. Established by the Legislature in 1947, the UC ITS has branches at UC Berkeley, UC Davis, UC Irvine, and UCLA.

Notes on contributors

Christopher Giamarino

Christopher Giamarino is a doctoral student in urban planning at UCLA.

Evelyn Blumenberg

Evelyn Blumenberg is a Professor of urban planning at UCLA.

Madeline Brozen

Madeline Brozen is Deputy Director of the Lewis Center for Regional Policy Studies at UCLA.

Notes

1 The LA CoC includes all cities in Los Angeles County except Glendale, Pasadena, and Long Beach.

2 Data from the 2017 National Household Travel Survey show that 40% of all recreational vehicles were located in rural areas and another 27% in small towns (Federal Highway Administration, 2018).

3 We tested a model using the larger sample by excluding the monthly income variable. The results are largely similar; however, the R² declines to 0.11.

Appendix A

Table A1. Characteristics of total vehicular respondents, vehicular respondents living on public streets, and vehicular respondents living in SPPs.

Individual characteristics	People living in vehicles (total)	Not surveyed at safe parking	Surveyed at safe parking	Significant differences
Sociodemographic
Gender
Male	53%	53%	55%
Female	46%	46%	41%
Transgender	0.8%	0.5%	2%
Gender nonconforming	0.1%	0.1%	0%
Missing data	0.1%	0%	1%
Race/ethnicity
White	36%	36%	31%
Black	32%	31%	35%
Asian	1%	1%	1%
American Indian/Native Hawaiian	2%	2%	5%
Multiple races	3%	2%	7%	**
Hispanic	23%	24%	12%	**
Missing data	3%	2%	8%
Senior (over 62 years old)	15%	14%	20%
Living with childre	18%	19%	6%	***
Monthly income (mean)	$1,458	$1,293	$2,461
Employment status
Employed	19%	16%	39%	***
Unemployed (looking, not looking, or student)	30%	42%	28%	***
Retired	3.6%	4%	4%
Missing data	8%	9%	4%
On disability	21%	20%	27%
Housing
Chronically unhoused	38%	38%	39%
Years unhoused (median)	3	3	2
Years unhoused (mean)	5	5	2.5
Physical disability	53%	51%	63%
Mental health	48%	47%	50%
Justice system	68%	72%	41%	***
Household with children	6%	7%	3%	***
Main reasons that led to loss of housing (avg. # reported)	1.48	1.57	1.54
Personal violence	10.7%	16%	8%
Family/personal	19.1%	28%	17%	**
Housing	12.5%	18%	13%
Incarceration	3.6%	5%	2%
Mental and physical health	12.1%	16%	20%
Unemployment	32.4%	44%	49%
Trauma-related experiences
Domestic violence	31%	38%	44%
Veteran/armed services	7%	6%	16%	***
Mobility
Slept in different neighborhood than last night in the past 30 days (%)	9%	10%	4%	*
Distance between last 30 days and last night (median)	2 miles	2.23 miles	2.44 miles
Total survey responses (n)	641	558	83

*p < .1; **p < .05; ***p < .01 for t-test (continuous) or Chi-squared test (binary) significance between vehicular respondents on public streets and vehicular respondents in SPPs.