Defining Homelessness in the Rural United States

Rural homelessness in the United States is an understudied phenomenon. Among those studies which do address the issue, there exists no uniform or consistent definition for rural homelessness. In this review of the literature, we look at rural homelessness and consolidate the literature into four main groups based on the definitions currently in use. We recommend a comprehensive definition for rural homelessness that looks at this phenomenon on a spectrum of needs, populations, and periodicity. We further recommend that current homeless count methodology be improved by using a more detailed survey of homeless situations, not only in the rural United States, but in urban areas as well.


Introduction
The United States (U.S.) is an increasingly urban-centric society; however, rural areas represent more than 70% of the nation's land (Census Bureau, 2017;Lichter & Ziliak, 2017). Cities are centers of growth and development, cultural incubators, and, in recent decades, have dominated the conversation about culture, politics, and economic issues (Lichter & Brown, 2011;Lichter & Ziliak 2017). Due to this urban-centric focus, many social problems such as poverty and homelessness are viewed through an "urban lens." Not only are the experien ces of rural poverty and urban poverty very different, but rural poverty has been persistently high throughout the early twenty-first century while metropolitan poverty is on the decline (Tickamyer, Sherman & Warlick, 2017). One of the social problems closely connected with persistent poverty is homelessness. Though homelessness in the U.S. has been studied widely in the literature (Gkartzios & Ziebarth, 2016), many rural scholars express a concern over the lack of such studies specifically focused on the rural U.S. (Bruce, 2006;Edwards et al, 2009;Fitchen, 1992). Currently no comprehensive definition exists that could be used consistently while developing homelessness initiatives or building on previous homelessness reduction programs. The lack of a consistent, broad-based definition makes it difficult to assess the extent of this phenomenon in rural areas and makes it difficult for policymakers to prioritize resources for communities most in need.
There is a lack of 'visibility' associated with rural homelessness in that rural homelessness does not always appear the way the public imagines homelessness to be (Fitchen 1992). Homelessness is often only recognized when it is 'visible' to the eye, such when individuals sleep on city streets or reside in homeless shelters. This is in part because such situations are fairly simple for a researcher or the public to identify as clear-cut cases of homelessness (Edwards 2009;Fitchen 1992). However, situations such as sleeping in a car or couch-surfing might not be immediately recognized as homelessness (Edwards 2009;Waegemaker Shiff et al, 2016). Some rural, low-income residents have housing that is extremely inadequate or is unstable or temporary and so can be considered homeless. Individuals in such situations are not as obviously homeless in the eyes of the public when compared to individuals who are sleeping on the streets (Fitchen 1992). Neither scholarly sources nor media coverage typically mention rural homelessness in the United States (Fitchen 1992). Furthermore, some rural homeless individuals live in remote areas that are geographically isolated, thus making it difficult for the rest of the community to notice or consider them to be homeless (Claus 2012;Fitchen 1992;Waegemaker Shiff et al, 2016). In these ways, rural homelessness is less visible than homelessness in urban areas.
Given the lack of such visibility for homeless individuals in a rural area compared to the urban homeless, understanding the extent of this phenomenon is a challenge. One of the problems in conducting such rurally-focused studies of homelessness is the lack of a consistent definition. While scholars have used concise definitions pertaining to rooflessness and more comprehensive definitions focusing on housing instability, to date, no comprehensive definition exists to specifically address the issue of rural homelessness.
The contribution of this study is twofold. First, literature analysis and synthesis is important because it brings coherence to a research area, contributes to knowledge by integrating studies, and informs researchers and practitioners about the current state of the field (Bland;L N Meurer;G Maldonado, 1995;Turner, Gonzalez & Wood, 2008). Second, we propose a new comprehensive way to define rural homelessness. The purpose of this research is to advance discourse within the scholarly community toward the goal of establishing a useful, comprehensive definition of rural homelessness.

What is the Rural United States?
One of the challenges in defining the rural United States is to create distinct geographical boundaries of where it begins and where it ends (Johnson, 2017). There are multiple definitions of "rural," and they can be classified based on administrative, land-use, or economic concepts, leading to variation in socio-economic status, and the well-being of the population measured (Cromartie & Bucholtz, 2018).
The Census Bureau's definition of rural focuses on land-use. According to the Bureau, 19% of the U.S. population lived in rural areas during the 2010 decennial census. The Bureau's definition of "rural" is predicated on its definition of "urban." The Census Bureau identifies two types of urban areas: Urbanized Areas (UAs) of 50,000 or more people and Urban Clusters (UCs) of at least 2,500 but less than 50,000 people. Its definition of "rural" encompasses all population, housing, and territory not included within an urban area. If the population size threshold is increased from 2,500 to 50,000 under the land-use definition, the rural population increases from, 21 to 32 percent in the 2000 decennial census. However, if the threshold is lowered for an economic definition from 50,000 to 10,000, rural population decreases from 17 to 7 percent (Cromartie & Bucholtz, 2018).
Another widely used definition is based on whether a county is metropolitan or nonmetropolitan (Nolan, Waldfogel & Wimer, 2017;Johnson, 2017). There are 1,090 metropolitan counties in the US out of a total of 3,141 counties. Those counties not considered metropolitan, though different from each other, are lumped together as simply nonmetropolitan (Johnson, 2017). Since a variety of definitions exists to define "rural" in the U.S., estimates of population range from 17 to 49 percent, depending upon the definition used (Cromartie & Bucholtz, 2018).

Rural Homelessness
Similar challenges exist when assessing the extent of the homeless population in the rural U.S. The official count of homeless individuals in the U.S. is obtained by the annual national homelessness count coordinated by the Department of Housing and Urban Development (HUD) using the Point in Time (PIT) survey. The PIT count uses unduplicated 1-night estimates of both sheltered and unsheltered homeless populations. The 1-night counts are conducted by the Continuum of Care (CoC) 1 nationwide and occur during the last week in January of each year. HUD collects and reports this data in Annual Homelessness Assessment Report (AHAR) every year. In 2017, 17 people per 10,000 were homeless nationally. The total homeless population in 2017 was 553,742 homeless people with an increase of less than one percent from 2016 to 2017 (AHAR, 2017).
The data is further divided into estimates by states and CoCs along with the national estimate. Such detailed reporting provides an insight into annual trends in homeless in the US and provides estimates by urban, suburban and rural areas. HUD reports data on CoCs using three main definitions. Major city CoCs cover the 50 largest cities in the United States. In case of Phoenix and Mesa, AZ, and Arlington and Fort Worth, TX, two large cities were located in the same CoC. Smaller city, county, and regional CoCs are jurisdictions that are neither one of the 50 largest cities in the United States nor Balance of State or Statewide CoCs. The balance of State (BoS) or statewide CoCs are typically composed of multiple rural counties or represent an entire state. Thus, the primary definition used to identify a rural CoC is the Balance of the State definition. Using this definition, figure 1 shows the trend in homelessness in the U.S. Figure 1 Note: Authors' calculation of AHAR data.
1 CoC are the local planning bodies responsible for coordinating full range of services in a geographic area, which may cover a city, county, metropolitan area or entire state. (AHAR,2017) 4 The figure shows the rural homeless count from 2007-2017 and the proportion of rural homeless as compared to the total homeless count. Rural homelessness oscillates between 13-17 percent. It important to note that not all states have BoS CoC. Hence, some states are not included in the computation 2 . Some of these states such as NY, PA have large rural areas. However, given the CoC definition in the AHAR report, they were not included in BoS homeless counts as reported in figure 1. This also indicates the limitation of current HUD definition to count homeless in rural areas.
To compute the PIT estimate, HUD defines homelessness as those who are literally homeless, at imminent risk of homelessness, homeless under federal statutes, and fleeing or attempting to flee domestic violence situations. One of the challenges of using HUD's definition of homelessness is that the definition is influenced by the perceived urban nature of homelessness. Homelessness in rural areas is considerably less visible and it can be experienced as unstable housing conditions such as moving from one overcrowded and extremely substandard housing situation to another (Rural, 2014). Those living in mobile homes would likely not be included in the HUD definition (Salamon & MacTavish, 2006). Similarly, those living with violent housemates would not be considered homeless under the HUD definition until they attempt to flee (Forchuk Montgomery, Bernan, Ward-Griffen, Csiernik, Gorlick, Jensen, & Roisterer, 2010). Individuals experiencing domestic violence live in precarious situations and are often on the verge of homelessness.
A significant means of calculating the number homeless youths in the rural U.S. come from the reporting of schools as required in the McKinney-Vento Homeless Assistance Act. The number of homeless students in public schools has doubled from to 2013(US Department of Education, 2016. From 2013-2014, there were approximately 1.3 homeless youths enrolled in public schools (US Department of Education, 2016). However, the data easily available to the public is reported online at the state level. This makes it difficult to distinguish between rural and urban/suburban school data on homeless youth. Despite this limitation, the definition used in the McKinney-Vento act is more expansive than the PIT count and assists school districts in having a more accurate picture of homeless youth in their district.
Policies and scholarship regarding homelessness are primarily influenced by the perceived urban nature of this problem (Trella, 2014). However, over the years, scholars have expressed concerns over urban-focused homelessness research (Bruce, 2006;Edwards et al, 2009;Fitchen, 1992). The realities of poverty and homelessness in the rural U.S. are quite different from those in urban areas. For example, rural poverty has declined rapidly since the 1960s but remains persistent in recent years. Using a metropolitan vs. non-metropolitan definition, Weber and Miller (2017) find that rural poverty rates exceed metropolitan poverty rates consistently. More than 15-percent of rural counties are in persistent poverty compared to 4-percent of metropolitan counties. Not only are the experiences of rural poor are different from those of urban poor, rural poor are also less visible than the urban poor (Tickamayer, Sherman & Warlick, 2017). This lack of visibility remains a persistent problem for the rural U.S. and needs to be assessed specifically. Hence, it is important to focus on this urban-rural differentiation (Gkartzios & Ziebarth, 2016;Mullins, Wilkins, Mahan, & Bouldin, 2016;Forchuk et al, 2010) to explore the differences between the urban and rural U.S. and create policy interventions to specifically address challenges facing the rural U.S. (Rollinson, 2006).
In this review, we focus on identifying major themes in rural homelessness definitions. Many of the studies we use are focused on the rural U.S., but we also use international articles focusing on rural homelessness in England, Canada, and other western nations. We have organized the definitions of homelessness into three categories with several subcategories: (a) The simplified definition of homelessness, (b) descriptions of homelessness without specific metrics to evaluate episodes of homelessness (descriptions of possible rural homeless situations and critiques of existing definitions without providing an alternative definition), and (c) specifically, defined levels of homelessness (3-levels of homelessness and 4 or more levels of homelessness).

A View from Previous Studies
Defining homeless is "problematic, contested, and emotionally charged" (O'Sullivan, 2006). There is little to no agreement on how homelessness should be defined (Chamberlain & MacKenzie, 1992). Fitchen (1992) argues that a comprehensive and agreed upon definition is necessary to ensure a "systematic attempt to measure the problem." In contrast, Aron (2006) asserts that it is less important to have a comprehensive definition of homelessness and more important to ensure that restrictive definitions do not limit people's eligibility for homelessness assistance programs. Aron's focus is on providing assistance and not being caught only in debates about comprehensive vs restrictive definition. At worst, defining homelessness can reduce homelessness to simply not having a house . There is also a risk where researchers or policymakers may project their own personal representations of "acceptable" forms of homeless . Aron (2006) and Fitchen (1996) warn that having broad definitions, such as characterizing the "doubled up" as homeless, may cause resources to be spent on "doubled up" populations and not on those who are in more dire circumstances (e.g. sleeping on the streets). Both emphasize that the perhaps broad definition is less important in itself. Policy needs is a definition that focuses on subsections of homelessness so policymakers can focus resources on specific programs for specific groups (Fitchen, 1996). Some have concluded that attempting to define homelessness is meaningless (Chamberlain & MacKenzie, 1992).
On the other side, Hanninen (2006) finds that the debate over definitions brings life to the topic of rural homelessness. However, he cautions that homelessness is a relational consequence of different practices and there is no one-size-fits-all perspective. All studies should examine as many different types of homeless individuals as possible and make clear who is included and excluded from these studies. They should also post results for each segment of the population separately (Burt, 1996). Having clearly defined data would help with the goal of having a systematic attempt to measure rural homelessness because the inconsistent definitions of homelessness mean some studies are missing different sections in their counts (Burt, 1996;Fitchen, 1991). Further, although a perfect homelessness definition would not solve political or agenda based questions, a comprehensive definition puts forth a new reference point for future debates on homelessness policy (Chamberlain & MacKenzie, 1992).

The simplified definition of homelessness
Despite all the debate over which definition is the best, the number of studies that use a very limited definition of rural homelessness is staggering. The Point-In-Time (PIT) survey is the most widespread method used to count the number of homeless people in a specific geographic area. The PIT survey involves researchers going out into the field to physically count homeless people sleeping in the streets. The count is conducted nationwide on the coldest night in January and is used to count the number of homeless people in a given area. The Annual Homeless Assessment Report (AHAR) uses the following definition: "Homeless describes a person who lacks a fixed, regular, and adequate nighttime residence" (p. 2). Though this definition is useful in describing one type of homelessness, it does not cover other situations that could be considered as homelessness episodes. For example, rural homelessness is often a hidden situation where people live with friends or stay in structures that are not meant for human habitation such as barns (Gkartzios & Ziebarth, 2016). Living in a car, in a parking lot, or camping in the woods can be part of rural homelessness (Fitchen, 1992). Rural homelessness is also experienced through precarious housing conditions and constant moves from one substandard, overcrowded, expensive situation to another (Rural Research Note, 2014). Staying with violent housemates and being forced to use nonlocal resources can also contribute to rural homelessness (Forchuk et al, 2010). Rural homeless individuals who move to cities can be considered part of rural homelessness (Cloke, Milbourne, & Widdowfield 2003;Edwards et al 2009;O'Sullivan 2006). All these situations are very unstable, but probably would not be assessed as homelessness by the Annual Homeless Assessment Report.
Aman and Yarnal (2016) study mobile homes in the rural U.S. and provide evidence that residing in mobile homes may lead to a state of "quasi-homelessness." The residents are quasi-homeless because they have a place to live but are at constant risk of homelessness (Salamon & MacTavish 2006). This is not to be confused with Lawrence's definition of homelessness which considers "quasi-homelessness" to be those living in an improvised shelter like in tents, cars, or abandoned structures (Lawrence, 1994). Living in mobile homes creates a unique set of vulnerabilities for the mobile home residents because they reside in a structure that is in between a house and a vehicle. According to Housing Assistance Council, 14.4 percent of rural households live in a mobile home compared only 6.5 percent nationwide (HAC, 2012). Mobile homes appear to be secure, affordable housing but in reality, put the homeowner at a higher-than-average risk of losing the asset because most mobile homes are financed through personal property loans and subprime lenders, which are less favorable to consumers (Geisler & George 2006). Further, often the mobile home owner owns the mobile home but not the land underneath. This means they are still subject to the whims of a property owner who owns the land where the mobile home is located. This landlessness of owning a mobile home but not the land underneath puts many mobile home residents in a state of quasi-homelessness (Salamon & MacTavish, 2006).
Mobile homes are often the most feasible housing option in Native American lands due to a lack of building material suppliers, increased building costs, and a lack of financial investment in Native lands (Geisler & George 2006). As previously stated, mobile homes can be high-risk investments because they are less desirable loans for consumers and have subprime lending practices. Further, some Native American communities face a severe shortage of affordable housing which makes overcrowding in substandard housing a significant problem (Geisler & George, 2006). The 2000 census showed a crowding rate in Native American and Alaskan Native households at 18 percent, three times the national average. Overcrowded living spaces can put extra strain on utilities and appliances, shortening their lifespan which is an added expense for the household Geisler & George, 2006). About 18 percent of homeowners in Native American areas are cost burdened (Geisler & George, 2006). Cloke and Widdowfield (2002) describe how urban definitions of literal homelessness do not accurately reflect rural homelessness situations but fail to elaborate on rural homelessness. They also critique that studies of homelessness focus on literal homelessness which is very visible (e.g., people sleeping on sidewalks and panhandling at street corners). This hypervisibility of literal homelessness is part of why it dominates academic attention and policymakers priorities when considering homelessness.
Claus (2012) focuses on a subpopulation of homeless individualshomeless veterans in rural areas. Rurality increases the barriers rural homeless veterans face because of the geographic isolation of many rural areas. Geographical distance between services available through Veterans Affairs (VA) offices and homeless veterans can delay services and treatments these veterans need, increasing their risk of becoming homeless. Veterans who suffer from PTSD are at an increased risk of homelessness, and delayed treatment can exacerbate the disorder. Lack of public transportation is a barrier to treatment by denying access to social services for veterans in need. Claus does not offer any specific definition of veteran homelessness, although it appears that his focus is on literal homelessness.
The policy brief written by the National Advisory Committee on Rural Health and Human Services (2014) provides a basic definition of homelessness. The committee uses the definitions of unspecified federal agencies and includes homeless individuals who live in shelters and those who are temporarily staying with friends or family members. The National Alliance to End Homelessness defines homelessness through an urban lens and focuses mainly on literal homelessness. The Alliance defines homeless as people sleeping outside and in shelters (National Alliance to End Homelessness, 2016). The Alliance also defines chronic homelessness as those who have a disability and are homeless for repeated episodes or long-time periods.
Mullin (2016) focuses on the implementation of the McKinney-Vento Act and uses the comprehensive definition used in the Act. The McKinney-Vento Homeless Assistance Act defines a homeless person as: (1) an individual who lacks a fixed, regular, and adequate nighttime residence; and (2) an individual who has a primary nighttime residence that is -A) a supervised publicly or privately operated shelter designed to provide temporary living accommodations (including welfare hotels, congregate shelters, and transitional housing for the mentally ill); (B) an institution that provides a temporary residence for individuals intended to be institutionalized; or (C) a public or private place not designed for, or ordinarily used as, a regular sleeping accommodation for human beings.
However, this definition disregards those who live in substandard housing, a common occurrence among the rural homeless population (National Coalition for the Homeless, 2007).
Forchuk et al (2010) use a brief description of a possible rural homelessness situation without defining rural homelessness or giving metrics for their study. They describe rural homelessness in Canada as hidden from public view and includes situations such as living in substandard housing, staying with violent housemates, temporarily living with friends or family, and using non-local resources as rural homelessness. Gkartzios & Ziebarth (2016) describe rural homeless as those who live in precarious housing conditions or stay in structures not suitable for humans while those at risk of homelessness double up with others and share substandard housing conditions.
All of these works study rural homelessness or housing insecurity and contribute to the field but do not provide objective metrics of rural homelessness or the specific homeless population they focus on. The lack of an agreed-upon definition makes it difficult to measure the rural homeless population and thereby may limit the ability of policymakers to provide the right support for the homeless in rural areas. While both Aman and Stahre focus on housing insecurity, they don't define housing insecurity or the specific population they studied, which makes it difficult to assess if both articles focus on the same demographic or different ones. Would the quasi-homeless living in mobile homes qualify for Stahre's study of the housing insecure and chronic health? Forchuk et al would consider someone living with an abusive domestic partner as homeless, but those individuals would not be counted in the PIT survey. This lack of consistency of definition makes it difficult to assess the extent of this phenomenon in the rural area.

Descriptions of Possible Rural Homeless Situations
Studies in this group provide a detailed description of rural homeless situations, which could be illustrated as unstable living situations. These unstable living arrangements can be overlooked if the traditional, urban-centered definition is used. Cloke et al (2003) describe the way rural homeless people in England are easily labeled as "not homeless." The homeless are easily ignored because they are hypermobile and tend to disperse across a vast area. The traveling homeless are considered vagrants, hippies, or beggars. When homeless youths move to cities they are seen as 'out migrants' and when homeless people move to rural areas they are not considered as the responsibility of that rural area. Rural areas often label homeless people as those in housing need rather than as homeless people. Labeling the homeless as those in housing need disregards the one factor that connects all of the various groups with housing needsthey are all homeless.  consider living in dangerous substandard housing, on a friend's floor, and sleeping in barns to be episodes of rural homelessness. Rural homeless are hypermobile and often move through rural areas and from urban areas for temporary relief from urban homelessness. There are few homeless shelters in rural towns yet one study found about one-third of applicants to a city shelter came from rural areas (Cloke et al, 2003). Often rural residents assume rural homeless will move to the cities to find jobs or cheap apartments, but there is little research to support that claim (Cloke et al, 2003). Though some homeless people have moved into urban areas from rural areas, this is not the norm. Many rural homeless people move around within rural areas and from the original rural area to a new rural area. The rural homeless highly value anonymity and the ability to blend with their surroundings (Cloke et al, 2003).  also emphasize the invisibility of rural homelessness due to the dispersion across small towns in large geographical spaces. There are not places where homeless people congregate and become visible such as store doors or homeless shelters. This invisibility makes it even more difficult for the rural homelessness to be a part of the dialogue of non-homeless rural residents.
These descriptions are useful but limited. An objective definition of rural homelessness and the target group is very important because when each scholar uses a different definition to study a different section of the homeless, we are unable accurately or objectively combine and compare data from each study. Cloke et al ( , 2003 focus their studies on the hypermobility of the rural homeless. Other scholars focus on rural homeless who face different barriers, such as Aman and Yarnal's (2016) study of quasi-homelessness in mobile homes. Claus (2012) studies rural veterans facing literal homelessness. Each aspect of rural homelessness is in need of further research and academic attention, however the lack of metrics and definitions in these works makes it difficult to identify the target population.
The International Handbook of Rural Studies states that rural homelessness is often a hidden situation where people live with friends or stay in structures not meant for human habitation such as barns (Gkartzios & Ziebarth, 2016). Rural individuals who are at risk of homelessness often share housing with others in substandard and overcrowded living conditions. Patton (1988) studies housing insecure families and emphasizes instability as a factor in rural homelessness. He considers individuals homeless if their housing situation is temporary, unstable and the only option for the individual. However, he does not think all who live in substandard housing should be labeled homeless. He views instability as the essential criteria of rural homelessness and as the factor that separates the rural homeless from the rural poor who live in substandard housing. However, Waegemaker, Schiff, Schiff, and Turner (2016) study rural Canada and consider the rural poor living in substandard housing as a part of the rural homeless demographic. Waegemaker et al do not provide a definition of homelessness but describe rural homeless situations such as living in inadequate housing unfit for human habitation and doubling up in overcrowded living arrangements. The vast geographical isolation of rural communities, the harsh weather in Canada, and the lack of social service providers in rural Canada create unique challenges such as extreme danger for those sleeping outside and fewer services to help them get back on their feet (Waegemaker, et al, 2016). Edwards et al (2009) describe how rural geography in the US often has poorer residents living in the most inaccessible and remote areas which are the farthest from the social services. Those most in need are farthest away from the service providers, which makes aid difficult to obtain. Claus (2012) observes a similar problem for homeless veterans. Within these vast geographic spaces are tight-knit towns with little social anonymity. There are more services available in the town but less privacy. Thus, rural homeless are likely to stay out in the remote rural areas to avoid the entire town knowing their personal situation.
The Rural Research Note (2014) contrasts typical urban homelessness with rural homeless situations and describes the challenges in counting the rural homeless. The Note defines literal homelessness as living on streets and in shelters. But also states that rural homeless can also be living in precarious housing arrangements, frequent moving from one substandard housing situation to another, and doubling or tripling up with friends and family. These rural homeless situations are less visible which is part of why there is a lack of service providers and resources for such rural homeless individuals.
Several notable trends in rural homelessness appear in the literature. One theme is the overrepresentation of vulnerable communities, particularly indigenous peoples, among rural homeless populations. Argent and Rolley (2006) note overrepresentation of Aboriginal and Torres Strait Islander peoples in Australian rural homeless populations. Maori people are over-represented amongst rural homeless in New Zealand (Kearns 2006). Geisler and George (2006) focus their studies on how rural homelessness affects Native American communities. O'Sullivan (2006) finds that Irish immigrants are overrepresented among the homeless in the United States and England. Beer (2006) notes that there is an overrepresentation of youth under age 25 in rural Australian homeless populations. Edwards et al (2009) studies rural homeless youth in the United States and finds three specific paradoxes that affect rural youth homeless people. First, large geographical distances create small compact towns that lack anonymity found in cities. Second, cultural conservatism and emphasis on independence/self-sufficiency can cause rural residents to ignore or disapprove of rural homeless individuals yet also motivate others to help the homeless. Third, the lack of social services means organizations must collaborate and compete for people to provide services. Historical policies regarding land ownership have an enormous effect on modern-day rural homeless such as the historical land theft from Native Americans affecting contemporary rural homelessness in Native American communities (Geisler & George, 2006). Irish land policies allowed only one child to inherit land, which yielded mass outmigration of Irish rural residents to the US and the UK (O'Sullivan, 2006).

Critique of Existing Definitions
Many scholars review homeless definitions as being inadequate but do not provide the definition they use in their study. Aron (2006) is not impressed with the never-ending debate over how to define rural homelessness and finds that it is more important to focus on segmenting the homeless population into meaningful subgroups so policymakers can focus resources on specific groups. Greisler and George (2006) emphasize that rural homeless may have a place to stay but they reside in such a precarious state that they should not be considered simply 'housed." Edwards et al (2009) were deliberate in including both homeless sleeping in tents and homeless staying in motels and couch surfing.
With no agreed upon definition of rural homelessness, it is difficult to accurately count homelessness in the rural U.S. (Fitchen 1991). Rural homeless people may live in a shed or an old shack with no running water. The rural homeless are also often widely dispersed and not congregated in homeless shelters which makes them difficult to locate to count (Fitchen 1991). There needs to be a systematic way to measure rural homelessness (Fitchen 1992). Fitchen (1992) endorsed Patton's definition of homeless as those with unstable and temporary housing who are unable to secure adequate housing.

Three Levels of Homelessness
Studies in this group define rural homelessness into three subsections or homelessness as a spectrum of housing insecurity ranging from sleeping in the woods to less severe but still precarious situations such as couch-surfing or living in substandard housing. Coleman-Jensen and Steffen (2017) divide homelessness into three levels: homeless who completely lack housing; housing insecure who lack adequate stable and permanent housing; and people facing housing cost burden. Housing costs are considered a burden if individuals/families spend 30 percent or more of income on housing. Geisen and George (2006) state that 32 percent of Native Americans living on Native Lands are cost burdened.
According to Trella and Hilton (2014), rural homelessness is understudied because homeless people in rural areas are difficult to count. Physically counting rural homeless individuals is difficult because small communities isolated in large geographic areas lead to invisibility for the homeless living outdoors. Many homelessness researchers coordinate with homeless shelters to conduct counts but there are very few homeless shelters in the rural U.S. (Aron, 2006;Aron & Fitchen, 1996;Cloke & Widdowfield, 2000;First, 1994;Fitchen 1992;Trella & Hilton, 2014).
Rural homeless individuals tend to rely on their social networks such as friends and family members to find places to stay and reside in structures that are not suitable for human habitation like cars and abandoned buildings. Trella and Hilton (2014) strengthened this finding. They interviewed rural homeless individuals. However, it is not quite clear if they had a specific definition of homelessness besides self-reporting. They put out fliers requesting interviews but the vast majority of their homelessness sample came from collaboration with social services agencies and they also used snowball sampling. They categorized the homeless participants as staying in a homeless shelter, staying with family or friends, and staying outdoors or in cars. Thus providing three levels of homeless counts. Burt (1996) divides homelessness into literal homelessness and at imminent risk of homelessness. He suggests that the institutionalized population that does not have any other homes should also be included in homeless counts as a subsection of the literal homeless. Aron and Fitchen (1996) define homelessness as a continuum and argue that individuals can lie anywhere on the continuum of housing distress. Alongside the continuum of housing distress, the McKinney-Vento Act definition of homelessness is adequate so long as it is applied with flexibility (Aron & Fitchen, 1996). Post (2002) uses an expanded definition, which is used by the Bureau of Primary Health Care (PBHC) of the Health Resources and Services Administration. PBHC defines "homeless" as a person without permanent housing who may live on the street, in a homeless shelter, a single room facility, an abandoned building, a vehicle, or any other substandard or temporary situation. This also includes those who are doubled up and those released from institutions with no stable housing. Post emphasizes instability as a key for defining rural homelessness. Rollinson (2006) divides homelessness into three subsections or patterns, temporarily homeless, episodically homeless, and chronically homeless. He uses these divisions because there is no one form of homelessness and these patterns distinguish different subsections of homelessness. Milbourne and Cloke (2006) provide another perspective. They focus on the geographies of homeless individuals. They divide rural homelessness into a spectrum of homelessness. Firstly, a spectrum between local and in-migrating forms of rural homelessness, secondly, a spectrum of settled and transient forms of rural homeless and thirdly, a spectrum of visible and invisible forms of rural homelessness.
When considering a spectrum of homelessness, different variables need to be studied. There is a spectrum of housing insecurity, a spectrum of homelessness from temporary to chronic, a spectrum from invisible to visible, a spectrum from localimmigration, and from settled to transient. Even when considering only three subsections of homelessness, many different variables such as mobility, stability/instability, cost burden, eviction status, seasonality (of seasonal workers) and more must be studied and more research is needed.

Four or More Levels of Homelessness
Different government agencies define homelessness differently. The Point in Time (PIT) survey defines homelessness as anyone lacking a fixed, regular, and adequate nighttime residence. HUD defines homelessness as those who are literally homeless, at imminent risk of homelessness, homeless under federal statutes, and fleeing or attempting to flee domestic violence situations. The 2000 Census Special Report on Emergency and Transitional Shelter Population defined homelessness as housing units where people are doubled up with family or friends, housing units where those completing a "Be Counted questionnaire" used a friend's or family member's address as their usual residence place, foster care serving children, emergency and transitional shelters, halfway houses, jails, group homes, worker dorms, and nonsheltered outdoor locations (Smith & Smith, 2001). The Census Bureau defines a housing unit as: "... a house, an apartment, a group of rooms, or a single room occupied or intended for occupancy as separate living quarters. Separate living quarters are those in which the occupants do not live and eat with other persons in the structure and which have direct access from the outside of the building or through a common hall...Vacant seasonal/migratory mobile homes are included in the count of vacant seasonal/migratory housing units." (Definitions, nd, p. 3).
Many people included in the census definition of homelessness would not be counted in the PIT survey of homelessness in urban areas much less in rural areas. The McKinney-Vento Homeless Assistance Act of 2009 defines homeless as: A person or family without a fixed, regular, and adequate nighttime residence, those with a residence not ordinarily used for regular sleeping accommodation like an abandoned building or camping ground, persons living in homeless shelters, persons who live in shelters not meant for human habitation and is leaving an institution where they temporarily lived, persons who would imminently lose their housing if they missed paying rent, are living in hotels/motels and lack resources to stay more than 14 days, persons within 14 days of eviction, a person has no identified residence and lacks the resources/support network to obtain permanent housing, unaccompanied youth and homeless families with children, those who experience long periods without permanent housings, those who experience instability as shown by frequent moves and expects to continue in this state for extended periods of time because of chronic disabilities or health conditions, domestic violence or childhood abuse, or multiple barriers to employment. Any person or family fleeing violence such as domestic violence or child abuse and lacks resources/support networks to obtain permanent housing is also considered homeless. This definition is more comprehensive than other government agencies such as the HUD or the Point in Time survey.
There is no continuity of definitions across governmental organizations in the U.S. In Finland, there has been a survey of homelessness carried out every year since 1986 (Hanninen, 2006). This survey includes those living outdoors, in temporary shelters or homeless shelters, living in institutions due to lack of housing, people soon to be released from institutions with no other housing, people living with friends and relatives due to lack of housing, and families who are separated or living in temporary housing due to lack of housing. The PIT survey would be more inclusive of different experiences of homelessness if it used a more comprehensive definition such as in case of Finland's annual homeless survey.
The Chamberlain and MacKenzie (1992) definition of homelessness is used by multiple scholars. Chamberlain and MacKenzie differentiate between the inadequately housed and the homeless. The inadequately housed are those living close to the minimum standard. The minimum standard is a small apartment with a kitchen, living room, bedroom, and a bathroom. In homelessness, there are three subcategoriespeople living in single rooms without a kitchen or bathroom, people living in a temporary shelter or couch-surfing, and people living without conventional shelter. Chamberlain and MacKenzie note there is a considerable dispute over the distinctions of inadequately housed and whether to consider people living in a single room to be homeless (1992). They also specifically do not consider those living in institutions such as dormitories or seminaries as homeless, although they technically live below the minimum standard. Argent and Rolley (2006) use Chamberlain and MacKenzie's (1992) continuum of homelessness. This continuum includes marginally housed, those living in houses near the minimum standard, tertiary homeless, those living in a single room without their own bathroom or kitchen and have no security of tenure, secondary homelessness, people moving from temporary shelter to temporary shelter like friends' houses or homeless shelters, primary homelessness, and people without conventional accommodation like living on the streets or in improvised dwellings. Beer, Delfabbro, Natalier, Oakley, Packer, and Verity (2006) also use the Chamberlain and MacKenzie definition of homelessness but use only primary, secondary, and tertiary homelessness and omit the marginally housed (2006). Beer et. al also emphasize that often young people in rural Australia shift from one form of homelessness to another. First, Rife, and Toomey (1994) included people in their survey of homelessness if they slept in a shelter or in no shelter for any length of time, slept in a shelter or mission provided by organizations which charged no or minimal fees, slept in or planned to stay in cheap motels for 45 days or fewer, and slept in other situations where they planned to stay 45 days or fewer, such as with friends or family. First et al. emphasize that any temporary stay is important because many rural counties do not have official homeless shelters. Lawrence (1995) defines homelessness as on the street, quasi-homeless (living in improvised shelters like cars or tents), living in shelters, doubling up, and near homelessthose who would be homeless without entitlements like fuel and rent assistance. Lawrence emphasizes the "culture of chronicity" which plagues the near homeless because their circumstances are so precarious that they can rapidly fall into another form of homelessness and are unable to become self-sufficient (1995). Meert and Bourgeois (2005) define homelessness as rooflessness, houselessness, living in insecure accommodation, and living in inadequate housing. "Rooflessness" counts those sleeping outdoors or on the streets; "houselessness" counts those who are in shelters but do not have resources to re-integrate into everyday society; and "living in insecure accommodation" refers to those living with an uncertain tenure of a dwelling or short-term living arrangements such as those staying with family, in extreme circumstances, or people at risk of domestic violence. People "living in inadequate housing" include those who live in substandard housing, those living in very overcrowded conditions, and those who live in a caravan or a boat. Vissing (1996) finds the McKinney Vento Homelessness Act definition of rural homelessness to be urban-focused and does not find it useful for her study of rural homeless children. Rural homelessness goes beyond a lack of shelter to sleep in; it includes cultural and social dimensions as well. She expands her definition to include different types of housing distress. Often children and families have a roof over their head but do not have a home, such as couch surfing. Vissing defines five subtypes of housing distress: Young families unable to close gap between housing costs and total income, individuals working but with too little income to afford housing, women unable to work due to a lack of childcare or limited skills, men who are older and with few social supports, and disabled people without social networks and supports.
The HEARTH Act expanded the McKinney Vento Act's definition of "at risk of homelessness." It provides the following guidelines.
An individual/family who earns less than 30% of median family income for the area and has insufficient funds and no social supports to stop them from living in an emergency shelter or other literal homelessness and has moved for economic reasons twice or more in the past two months before the application, or is living in another's home because of economic hardship, or has been given an eviction notice within 21 days after application, or lives in a hotel/motel not paid for by charitable organizations, or lives in an efficiency apartment with more than 2 people or a larger unit with over 1.5 people per room, or is leaving a funded institution, or otherwise lives in housing that is unstable with increased homelessness risks as identified in applicant's Con Plan.
The HEARTH act improved the homelessness definition by including those who live in precarious housing conditions, extremely substandard or overcrowded housing, or cost burdened housing situations, situations quite common in rural areas but previously not included for program eligibility under some federal agencies (Housing Assistance Council, 2012).

Discussion
Defining rurality in the U.S. is a challenge and various agencies use different definitions. Defining rural homelessness is even more difficult and scholars and federal agencies have used multiple definitions for study and policy implementation purposes. However, the lack of a consistent definition is a challenge when it comes to counting the number of homeless people in the rural U.S. and providing adequate services for them. The lack of visibility of homeless individuals in rural areas makes it difficult to accurately measure rural homelessness. However, based on the current literature, we, the authors, think some definitional changes can be helpful. Though there is no one definition of what rural means, we adopt the rural and small-town definition of the Housing Assistance Council (HAC). We prefer this definition because it takes into account both the population density and distance from the nearest metropolitan area. The literature suggests that rural areas near urban areas (nonmetropolitan counties) have done better economically and demographically compared to counties that are remote from urban areas (Johnson, 2017;Patridge & Rickman, 2008). It would be helpful for future studies to assess distance to economic opportunities and social services. Hence, we think that a definition that used both population density and distance to metropolitan areas would provide a better picture of rurality. HAC defines Census Tracts as rural or small towns based on the following classification: Rural tract -Less than 16 housing units per square mile (.025 housing units per acre). Small-town tract -Sixteen to 64 housing units per square mile (.025 to 0.1 housing units per acre), and a low degree of commuting to a metropolitan core area identified by a USDA ERS designated "Rural-Urban Commuting Area Code" (RUCA) score of 4 or higher.
We suggest rural homelessness assessment can be expanded by focusing on the following: First, include homelessness as a continuum from literal homeless or rooflessness to housing instability (which could be due to various reasons including, doubling up, housing cost burden, living in precarious housing situation). Such an assessment would point towards the main challenges facing these populations and hence specific policy and service solutions could focus on the need for these populations.
Second, define homelessness to include a chronic and episodic definition. We define chronic homelessness using HUD guidelines. (An unaccompanied homeless individual with a disabling condition who has been continuously homeless for a year or more OR (2) an unaccompanied individual with a disabling condition who has had at least four episodes of homelessness in the past three years). While episodic homeless can be defined as those who go in and out of the homeless, there would be an overlap of individuals and families who would fall on the continuum of homeless and go through either chronic or episodic homeless situations. However, such differentiation would help service providers to target resources more appropriately.
Third, provide a situation-based definition for different homeless populations such as families, children, and victims of domestic violence, veterans, and mental health patients. This would help researchers and policymakers to prescribe policy solutions to address the needs of this population in the rural area. An example is the continuum definition used by Chamberlain and MacKenzie (1992). Policymakers can focus on a specific situation such as those living on the streets or those sheltering at a friend's house, and they then can specifically identify which population they are focusing on. Thus, all policies would not provide one-size-fits-all solution towards different homelessness situations.
Fourth, annual homeless count needs to be supported by an expansive count like Finland, which would make possible to provide a better measurement of homeless in rural and urban U.S. differently. While we understand that such an endeavor in current policy climate may not be easy, a pilot measurement focusing on specific states or counties could be helpful to provide a better measurement of this phenomenon in rural areas. While there is no perfect solution to understand homelessness in the rural U.S, such expansion of definition would help policymakers to assess this problem specifically for the rural U.S.

Limitations and Future Research
There are some key limitations to the current study. Like many social programs, current definitions of homelessness and policies to deal with the issue do suffer from some paternalistic view. However, as the current federal policy stands in terms of providing funding for homelessness-related programs, these paternalistic definitions persist. In the future, involving homeless populations themselves in the policymaking process would be ideal. Future research can focus on techniques such as in-depth interviews with this population to understand their needs better. The aim of this study is to advance the scholarly discussion about a better definition for rural homelessness by providing a detailed view of the extant literature. Hence, this research does not consider alternative housing such as communal living as a viable housing option. Future research can provide further focus on alternative living arrangements as viable living situations.
In some ways, focusing on a definition instead of the real experiences of homeless population limits the conversation to definition rather than focusing on better policy prescriptions. Arguing over definitions can take the attention away from the goal of understanding and assisting struggling communities. Thus, there is a significant need for more research into the experiences of the rural homeless.
Defining rural homelessness would help policymakers not only to "count better," but it would also provide much better data on homelessness in the rural U.S. generally. Currently, a lack of such data has limited the ability of policymakers to provide suitable policy solutions for rural homelessness. Consequently, a better definition and improved data would help policymakers and practitioners to more efficiently allocate resources by specially focusing on this population.
Rhucha Samudra is an Assistant Professor of Public Administration at SUNY Brockport. She received her PhD in Public Administration from American University. Her work focuses on welfare reform, food and income security and other safety net programs designed for low income families and individuals in the US.
Amelia Yousey received her Masters in Public Administration with a concentration in Poverty Studies from SUNY Brockport. Her work focuses on homelessness and poverty in rural areas.