“I know a guy”: Examining homeless income generation and spatial mobility

This study explores the variety of income generating activities, along with corresponding distances that homeless and marginally housed persons travel to engage in the activities. Altogether 24 individuals are interviewed, and ethnographic field observations occur over several months to increase rapport. The qualitatively gathered data were used in ArcGIS to calculate travel distances. These income generating activities that persons engaged in were commonly forms of shadow work; the varied types of income generating activities and corresponding mobility considerations are discussed.


Introduction
Although popular discourse often refers to "the homeless" as a homogeneous entity, the term has been used to describe people in a wide variety of situations. Peters & Robillard (2009), for example, considered individuals who had no place to live, but stayed with friends, family, or others in temporary housing. The National Alliance to End Homelessness (2012) also included people fleeing domestic violence households with limited housing alternatives, while Thomson's (2016, p.7) focus was on persons who "did not have a place of their own where they could expect to stay for more than 30 days and if they did not pay rent." The homeless have also been defined to include persons in temporary housing such as boarding houses, transition houses or shelters, persons staying at someone else's residence, persons in hospitals or transitioning into or out of jails with no fixed address (Thomson, 2016), as well as those living in "unstable housing" in shelters, on the street, in temporary accommodations or in single room occupancy (SRO) hotels (Corneil et al, 2006). Homelessness may also refer to people in "precarious housing," which includes persons in "marginal accommodation" such as SROs or persons who have experienced two or more periods of homelessness in the previous 12 months (Somers et al., 2016).
There are also many known common features to homelessness. For example, numerous studies commonly detail victimization concerns (Roy et al., 2014), encampments (Chamard, 2010), and crime (Fischer et al., 2008;Garland et al., 2010). We also know that the homeless are forced into public spaces more frequently than domiciled persons, and that this influences homeless mobility in several ways, even without considering transportation disadvantage experienced by the poor (Hine & Mitchell, 2001). Homeless mobility is affected by urban planning (Braverman, 2010), legislation (Doherty et al, 2008;May, 2003;Mitchel, 1997), other factors such as physical CrimRxiv • "I know a guy": Examining homeless income generation and spatial mobility 3 disability, need of restrooms and special services (Iwata & Karato, 2011;Rahimian, Wolch & Koegel, 1992), and by criminal justice response (Kaufman, 2021;Wardhaugh, 1996). Persons living in SROs and shelters are also noted as sometimes having affected mobility as a direct result of their housing situation (Nair, 2016;Rankin, 2015;Rollinson, 1990). All these situations acknowledge one central issue: that the homeless do not have their own space to conduct daily activities.
As a consequence, the homeless often must engage in private activities in public settings. When use-conflicts arise in public settings, urban planning and legislation is used to help regulate and control these places so as to impose "what is properly public" (Mitchell, 1997;Reyes, 2016). Properly public is a distinction made by policy makers and persons less negatively affected by the consequences of regulation in these spaces. This in turn means that everyday actions such as sleeping, changing clothing, going to the washroom, storing one's belongings, and so forth, can become criminalized when conducted in public places (Mitchell, 1997). In some places this may trigger efforts to engage in "spatial cleansing" (Amster, 2003) where sidewalks become privatized (Amster, 2003;Kohn, 2004) and law enforcement force the homeless to move along (Snow & Mulcahy, 2001).

Survival Strategies and Shadow Work
A challenge for the homeless is to provide for basic needs despite having fewer resources to achieve and acquire them; for example, not having adequate money for public transportation, food, shelter, and clothing. Those who are unemployed and living on the streets may employ other income gaining methods -some legal and some illegal -sometimes discussed as a survival skill (Lee et al., 2011) or as shadow work (Snow & Anderson, 1993). Types of shadow work include the unlicensed selling of goods on the street, panhandling, prostitution, selling drugs, selling one's blood, and stealing, among other activities (Conroy, 2001;Lei, 2013;Poremski et al., 2015).
Few studies have assessed the geography of shadow work and how these activities interact with other challenges the homeless face, although there are indications many kinds of shadow work are associated with a physical territory (Snow & Andersen, 1993). For example, panhandlers make claims to territory that produce more reliable income (Bajari & Kuswarno, 2020;Meir-Dviri & Raz, 1995), prostitution often occurs in particular "strolls" (Lowman, 2000) and drug dealing occurs in particular places (Perkins et al., 1992). And although a substantial literature also exists regarding geographical parameters of crime (e.g., Ackerman & Rossmo, 2015;Andresen et al., 2014), including activities some homeless may engage in -e.g., the purchase or sale of CrimRxiv • "I know a guy": Examining homeless income generation and spatial mobility 4 drugs (Johnson, Taylor & Ratcliffe, 2013) and shoplifting (Levine & Lee, 2013) --these studies commonly exclude homeless persons for methodological reasons (e.g., Ackerman & Rossmo, 2015).
As a consequence, little is known about such fundamental topics as how the urban homeless spend their time (Lee & Schreck, 2005;Wolch et al., 1993), move about the city (Kauppi et al., 2015), and how variation in those movement patterns is related to other aspects of the homeless and their life space (Peters & Robillard, 2009 (Kaufman, 2021) and interactions between the homeless and businesses, agencies, and the urban landscape. Knowing where homeless people go may also assist homeless services and advocates for the homeless to provide resources for this group of persons, at the very least by potentially allowing for more comprehensive homeless counts --a vital activity that contributes to understanding the scope of the problem, resource allocation for homeless service agencies, and outreach efforts (Thomson, 2016).

Study Setting and Design
An ongoing methodological challenge for research regarding homelessness is how to locate a criterion sample of individuals who are in that situation. Our research was conducted primarily in Vancouver's Downtown Eastside (DES), a historically disadvantaged portion of the downtown area that for decades has been associated with homelessness, high rates of alcoholism and addictions, and primarily petty criminal activity (Campbell, Boyd & Culbert, 2009), and consequently, where many of the city's homeless shelters are located.
A typical limitation of research on the homeless is the inability to contact the "hidden homeless" (Kauppi et al, 2015), persons who become skilled at using invisibility strategies and self-sufficiency to remain undetected and "successful" or "highfunctioning" as street homeless (Lee et al, 2011). Accordingly, following the suggestion CrimRxiv • "I know a guy": Examining homeless income generation and spatial mobility 5 of DeVerteuil et al. (2009, p.661), we sought to engage directly with the homeless themselves rather than through service providers or agency employees. This allowed us to identify a sample of individuals that included but also went beyond homeless persons who seek accommodation in shelters, given the previous finding that many homeless persons avoid shelters (Wasserman & Clair, 2009). Unsheltered homeless persons might also exhibit different mobility patterns and engage the most complex survival strategies, shadow work, and skills at living undetected in inclement weather for prolonged periods of time as compared to those who return to a shelter each night (Wolch, Rahimian & Koegel, 1993).
Homelessness for this study was defined as persons who reported they currently sleep on the street, in unstable housing or in city-provided shelters. Communication with participants commonly occurred around Vancouver's Downtown Eastside Markets (DES Markets) though some participants were interviewed in other parts of the downtown area. The DES Markets are a regulated city solution to the homeless selling second-hand products on sidewalks downtown. The first three persons interviewed were involved in some capacity with the DES Markets; snowball referrals to later participants included persons not involved with the markets. Ultimately, 24 persons were interviewed for this study. 1 We conducted semi-structured interviews with participants and engaged in ethnographic field observations of the settings in which the participants spent time.
Data collection occurred in two separate time periods, one in spring and the second over the fall and winter. In total, approximately 110 hours were spent in the DES in addition to approximately 36 hours of interviewing with participants. Interviews lasted 75-90 minutes, though subsequent contact with participants was common after the interviews occurred.
Interviews were aided by using a physical map of downtown Vancouver that facilitated discussions about mobility and locations in Vancouver. During interviews the emphasis was on locations that participants considered to be "routinely" travelled to, not necessarily daily. While semi-structured, each of the interviews included the following themes: daily mobility; restrictions on mobility; mobility for shadow work; income sources; factors affecting income.

Results
CrimRxiv • "I know a guy": Examining homeless income generation and spatial mobility 6

The Sample
The sample was a mostly white, mixed-gender group of individuals between 27 and 70 years of age. Table 1 provides descriptive information about participants including the mean age, ethnicity and gender, current housing type, drug use, childhood difficulties, institutional supervision and communication with family. The current housing type of participants varied despite all persons having recently been street homeless or in homeless shelters in Vancouver prior to their current housing situation.
[Insert Table 1] Drug use was reported by nearly all participants. Only two individuals out of 24 did not report drinking alcohol to intoxication or consuming other substances on a weekly basis during the past 30 days. Many individuals participated in polydrug use, and consumed drugs daily. Childhood difficulties -coded as adverse childhood experiences (ACEs; e.g., physical, psychological or sexual abuse; see Felitti et al, 1998) -were common, with 89.5% of the sample reporting one or more. Institutional supervision was also something many of the participants had experienced at some point in their lives. Communication with Family varied considerably but in most cases was either nonexistent or minimal. Lastly, more than half of the 24 persons interviewed reported having disabilities. Among the 24 respondents, only one (a male) was a military veteran.

Mobility and Shadow Work
Homeless shadow work, operationalized as an active means of acquiring income, was very diverse. It is useful to conceptualize these strategies as proactive due to 22 of 24 research participants receiving some form of government financial aid, whether via pension, disability, welfare, housing or transportation vouchers, and so on. Only one man and one woman did not receive such aid. Table 2 shows the 23 different survival strategies encountered, grouped thematically. All strategies are mutually exclusive; for example, an individual "selling heroin outside of the market" was not counted also as "selling goods not at market" unless that same individual additionally sold goods besides heroin outside of the market. Additionally, in the table there are parentheses distinguishing how many persons were male or female for each strategy.
[Insert Table 2] The survival strategies outlined in Table 2 involved different levels of mobility in order to implement them. For example, volunteering at the market or at one of the DES shelters involved little to no mobility, while persons who were involved in occasional employment or scavenging would go much further afield. Other activities such as selling drugs or alcohol could occur almost anywhere, with participants travelling to wherever their market was. The third column in Table 2 shows distances travelled for the specific activities reported by our respondents and thus are not necessarily indicative of the distance that would be involved for others who might engage in the same activities. Distances were computed by asking respondents where they slept, or slept most frequently, and computing distances from that location to the location of their reported survival strategies.
Results for distance to strategies are difficult to easily quantify and present. The decision was made to classify "supported by partner/friends" as an active strategy because this strategy is only available to individuals who cultivated relationships in such a manner to benefit financially from them. But this was not the case for most interview participants, as indicated by the low number of instances of that strategy.
However, mobility is not applicable in the same sense to this strategy as others and thus was not quantified. Distance for "drug carrier/[drug] storage/[drug] lookout" is similarly inappropriate given the lack of consistency for travel distance. Similarly, the carpentry strategy varied so much so that a set distance was not established.
Moreover, "boosting" (theft) involves multiple strategies. It requires the individual to travel somewhere to steal something and then to travel somewhere else to sell it. So the distance for that strategy is 2.7km to the theft location and then subsequent 2.6km to where the participant indicated preferring to sell. However, the final destination is only 3.3km from the individual's starting home location, not the 5.3km to the theft area then to the sell area. More important than difficulties in expressing total distances to strategies, however, is that all persons travelled distances less than four kilometres except for five experienced street homeless persons who favoured several long-range strategies.

Getting Around
While 13 persons normally walked everywhere, five most commonly took the bus and four rode bicycles. Unsurprisingly, persons who travelled further tended to use the bus. Only one person indicated they enjoyed using the SkyTrain (a public highspeed transit in Vancouver). Only one individual owned a car when the interviews were conducted, though two others indicated having owned vehicles during periods of street homelessness in the previous three years. The individual who owned a car nonetheless indicated he walked to most places. CrimRxiv • "I know a guy": Examining homeless income generation and spatial mobility 8 There were a few alternative transportation methods due to irregular mobility concerns. Paige is a middle-aged woman who has to rely on disability vans and cabs to travel everywhere. She explained, I had open-heart surgery four months ago, seriously. I can't take buses anymore, I tried and just get so sick, it's the worst…I don't use a bike or walk because I was once hit by a hit-and-run driver. It was before I became homeless but it broke my spine, bleeding on the side of my brain. When I was hit I flew 20 feet on the sidewalk, spent weeks in the hospital. I'm scared of it happening again.
Meanwhile, Jack, an SRO occupant in his 50s, enjoyed bicycling most places but struggled to find areas to secure his bicycle because the place he stayed did not allow tenants to bring bicycles into the building. "I'm pretty sure it's not legal what they do there -not just the no bike policy but also charging tenants money to bring guests there. It's not right," Jack indicated.

Factors Influencing Mobility
Gender. The overall results presented in Table 2 mask considerable variation among participants that was related to several other demographic and strategic factors. There were noticeable discrepancies between genders with respect to the quantity of strategies used, with males in the sample pursuing a greater variety of strategies to gain income than females. The range of strategies for women was 1-3 with a mean of 1.81 strategies; among males, the range was 2-4 with a mean of 3.07 strategies.
During the months that the interviews occurred, eight of the nine people who said they were currently street homeless were males, which may explain why males were more likely to engage in more income gaining strategies.
Disability. Different forms of disability also influenced where participants would go in the course of their day. Brad was in his 50s with prostate problems, and shared that, "Bathrooms are one of the rarest things in this city," often having to consider where he went to have access to one. Even when they do exist, facilities available to the homeless may be less than adequate. In two separate occurrences interviewees had attempted to use automated public bathrooms downtown and became trapped inside them for over an hour, a phenomenon that has previously been documented in Vancouver (CBC News, 2011). Participants subsequently travelled to avoid having to use the public bathrooms. CrimRxiv • "I know a guy": Examining homeless income generation and spatial mobility 9 Kenny, meanwhile, could not walk very far without being in pain after having been stabbed twice in his spine, once with a knife and once with a screwdriver. Kenny's pain was a daily consideration, he indicated as much when he said, If I have to travel more than about 10 blocks I just can't. I have to try and get a cab or if I have gas-money I'll drive. Everyday walking sucks when your back aches, it affects my legs sometimes and I lose feeling. 2 Those with mental health issues had their mobility affected in other ways. Dennis moved to Vancouver from Ontario, was in his 30s and was currently street homeless.
Unlike most of the other participants whose mobility was constrained by social stigma of being homeless or from physical disability, Dennis required buildings where he purchased food to have: wood flooring. It symbolizes that the owner supports local business and recognizes the importance of flooring in helping keep food fresh. Otherwise the [store's] produce is picked by underpaid people from fields somewhere else.
Shelter Policies. Several participants indicated that having to wait in line for several hours at food banks restricted their ability to travel and engage in various kinds of shadow work. Similarly, shelter check-in requirements were often such that persons needed to remain close to particular shelters if they were to reserve housing for the night.

Comparing the Mobile and Non-Mobile
As the previous sections showed, mobility among the homeless mirrors broader social relations, with women and those with physical and/or mental disabilities less likely to be mobile and thus subject to the influences and constraints of living in shelters in the Downtown Eastside. With greater mobility, in contrast, came greater opportunity, and typically also, greater income. Overall, it seemed that lesser mobility came with few advantages and many disadvantages, while greater mobility was associated with more advantages and fewer disadvantages.
The Less Mobile. Being less mobile involved having most of one's life space in the core of the Downtown Eastside with its higher densities and greater visibility. Some of the participants reported being told to move along by law enforcement or being harassed by bus drivers. Among persons who viewed law enforcement favourably there was a consensus that things had improved compared to how they were in Vancouver many years ago, though 2010 appeared to be a bad year for nearly all participants. CrimRxiv • "I know a guy": Examining homeless income generation and spatial mobility 10 During the 2010 Winter Olympic Games participants indicated being fined and highly controlled, given one-way bus passes, picked up and dropped off outside of town and forced into temporary shelters with inadequate accommodations -consistent with the literature on forced mobility and spatial expulsion (e.g., Kaufman, 2021).
It was during that time that one of Kenny's friends was fined $3,000 for selling cigarettes in Pigeon Park, a central meeting place in the DES and home of various ongoing markets. Having had his possessions confiscated to displace Kenny off of the sidewalk, he remarked that "we already don't got shit, then the cops come in and take what little you do have…it's just not the right thing to do." Kenny's opinion of law enforcement mirrored Troy's, that "there is a two-tier system for how laws are applied and people are treated." The density of the DES also meant that many of the survival activities that respondents participated in had informal territorial boundaries that one failed to recognize at one's peril. Brad revealing that "trying to collect money in someone else's turf'll get you hurt. I've seen a knife pulled on another person over nothing. Over three dollars." He continued by noting that the most important thing about being homeless downtown was being able to identify who would try to take advantage of you, saying that "everyone who doesn't know how it works [in the Downtown Eastside] gets takenripped off. If anyone down here says 'trust me,' start running." Two activities that were well-served by the density and limited space of the DES were drug dealing and prostitution, both of which benefit from potential customers knowing where to find you. Angela's comment regarding prostitution territory was that, "territory is hard to figure out, you just know it when you see it. [Sex workers] dressed differently or someone being out of place." Despite this, Angela attracted most of her clientele within a block of her residence, lacking concerns of anonymity in her activities and remaining a non-mover (Wolch et al., 1993). Indeed, among the less mobile participants, Angela's income was among the highest reported: "middle class women, like me, charge 80-100 dollars for an hour but it depends what kind of service they want", with some activities costing more than others. While five of the 10 women in this study had engaged in prostitution at some point in their lives, only Angela had recently done so in Vancouver.
The More Mobile. The more mobile participants in our research could be described as successful street homeless. These persons remained street homeless for longer than one year and had diverse strategies that caused them to travel much farther than the rest of the sample. They took pride in their ability to construct their own shelters and CrimRxiv • "I know a guy": Examining homeless income generation and spatial mobility 11 had survival strategies that varied based on the season. For example, Cindy expressed pride in her ability to "make one hell of a cardboard condo," and indicated the value she placed on the skill of being resourceful to make shelter. The ingenuity and resourcefulness demonstrated by persons in this sample are comparable to similar reports of successful homeless elsewhere (e.g., Kelly, 2017).
Most notable was the way in which greater mobility allowed some of our participants to partake in the fringes of the wage economy where they could earn a few dollars when a contact they had made needed some extra hands. The lack of documentation meant that pay was in cash and thereby untaxed. For employers, it gave them ready access to a handy workforce who would take minimal wages and would not be looking for benefits because they never reached the status of "employees." For example, Marvel had contacts with a warehouse company who contacted him when they needed someone to do repairs inexpensively.
I know a guy who works down there [at the warehouses] and he knows I need some money and use to be a contractor so when they have stuff that needs done and isn't too hard, he phones me and pays a bit of cash.
Another homeless individual who unloads trucks, Jack, explained his situation: I've been helping unload trucks for a company for over a year now. It's only once a week but it helps. I find a few able-bodied guys, normally I look for good guys who can use a few bucks, and they help me unload the trucks up by the port.
Scavenging refers to foraging for other sorts of goods to sell such as clothing or appliances. In some cases, mutually beneficial relationships are created. Two of the three people who participated in scavenging made a temporary deal with a thrift store.

As Troy and Vanessa explained,
The Salvation Army we go to is always getting donations after it's closed and after we asked they said we could just take anything left by the front doors. We clean it up and try to sell it at the markets. When we sell it, it's cheaper than if the Salvation Army does.
This symbiotic relationship provided the homeless pair with merchandise to sell on the sidewalk or in the DES Markets and it aids the thrift store by keeping the premises uncluttered. Though rare among the participants, this sort of relationship also was found to have occurred in some cases outside of survival strategies, such as when a business owner thanked a homeless individual for sleeping at night in front of the CrimRxiv • "I know a guy": Examining homeless income generation and spatial mobility 12 establishment's doorway. This happened to Alison, the French-Canadian participant, who recalled that, "He used to like when I slept in the doorway of his shop because it meant, he thought, his shop wouldn't have a burglary happen there." Other activities can be very time consuming but less lucrative. An example would be binning, which refers to foraging for recyclables such as cans to turn in for money.
Reflecting on his income, Marvel noted that, "[I] may walk around looking for things to recycle for six or seven hours but only get 30…35 bucks tops at the [recycling] depot." While participants placed varying degrees of importance on different skills, it was clear that these skills were not all intuitive or easy to master. Marvel's decision to live next to the University of British Columbia (UBC), for example, was done to allow him to benefit from the "culture of waste" he observed among affluent and foreign university students. Living there also meant he could travel through more affluent neighbourhoods while travelling to downtown. Further, residing next to the relative isolation of the Endowment Lands (a forest preserve) beside the university allowed him to avoid hassles with police, find places to sleep, identify safe places and avoid fights with others, all survival skills identified by Lee (2011).

Discussion
This study's results were largely consistent with prior research. Motivations to travel were found to be similar to domiciled citizens and related to work, medical services, eating and social activities (May, 2003;Wolch et al, 1993). Travel for some was influenced by respondents' social networks (Iwata & Karato, 2011), though only three persons' mobility was greatly influenced by family.
While this sample's mobility was concentrated in the Downtown Eastside as anticipated (Somers et al., 2016), there was a clear dispersion when the sample was separated between street homeless and everyone else. This was the case due to street homeless persons seizing opportunities farther away than their more spatially limited peers. The only survival strategies that have average travel distances greater than 3.2km are strategies used by the five street homeless travellers; travelling distances of 3.2km moves participants well outside of the Downtown Eastside. The out-of-town migrants in this sample have many of the same attributes as previous studies have found (e.g., Kauppi et al, 2015;Lindquist, Lagory & Ritchey, 1999). By and large, the more mobile participants in our study were more resourceful, had some degree of organizational skill, and had social skills that allowed them to negotiate mutually beneficial arrangements with businesses and other contacts. The five Movers (Wolch et al., 1993) in this study --street homeless who travelled much farther than others and outside of the downtown area -were distinct from the non-movers. The five were different in terms of all being street homeless and having been street homeless for more than twelve consecutive months. For those five there was a relationship between travel distance and income, which is dissimilar to previous findings (Jocoy & Del Casino Jr, 2010, p. 1957) as their farthest distances were opportunistic and more lucrative than many of the shorter distanced survival strategies others used. However, many of the participants in Jocoy & Del Casino Jr.'s (2010) study were traveling to search for jobs, which was not the in case for the 24 participants interviewed for this research.
The income of respondents for this study also was quite different than found by Jocoy andDel Casino Jr. (2010, p.1952); their study had four respondents, of 124 total, who received 1,666 to 3,000 dollars monthly, although the median income for the sample as a whole was a mere $396 dollars monthly. Their median income was similar to other studies (e.g., Peters & Robillard, 2009). In our study, all participants received between $650 and $1400 dollars monthly. Reasons for this may be a greater access to financial government assistance in Canada compared to the United States, or more profitable income strategies. Location can play a large role in availability and cultural acceptance of some strategies (Okamoto, 2007) and this, too, may be why there is this discrepancy in income.
While individuals reported travelling to an average of seven places regularly, persons in this sample typically moved 11 kilometers per day compared to prior research finding an average of 22.5 (Jocoy & Del Casino Jr, 2010). The five participants with the unusually long trip distances in this study, however, such as Marvel, Troy, or Russel, regularly journeyed up to 35 kilometres. Additionally, participants in this study most commonly travelled to public places like parks or libraries for leisure and for social services (Jocoy & Del Casino Jr, 2010, p. 1955. While not travelling as far, both persons who panhandled had to travel, as did binners, due to a territory associated with those strategies. Rollinson (1998) noted homeless travel at times to isolate themselves from their threatening environment, and this study found that as well.
Specifically, this study found that many of the street homeless would opt to sleep in distant (in relation to city center) parks where they believed they were less likely to be robbed.
Participants who engaged in illegal activities tended to be the ones who were homeless longer and/or unable to find lawful employment that compared financially (Garland et al, 2010;Snow & Anderson, 1993). Angela engaged in prostitution because she was terminated from her job; Russel boosted out of desperation to sustain his drug habit while attempting to pay off those to whom he owed money, and Troy sold alcohol on the beach because of how lucrative it was and because it maximized the benefit of traveling to the beach to engage in binning, a legal survival strategy. Brad, who sold crack pipes, grappled existentially against his situation as someone who believed he was born to work a legitimate job but was unable to due to his extensive physical disabilities. The persons who sold goods outside of the market -an activity that has increasingly become criminalized (Kane, 2015) -mostly did so out of opportunistic convenience. For example, one participant reported taking a bus from one location to another and selling something to a stranger who became interested in what the participant was transporting.
The homeless involved in this study recognized the significance of certain places for facilitating crime or other activities. Accounts of bathrooms outside of a large community center were comparable to those discussed by Felson et al (1996); an area that should be avoided due to it being high in crime where large groups of persons actively engage in the sale and use of drugs. One participant, Chris, considered automated bathrooms as an ideal place to smoke meth due to the privacy afforded by their single service design (Hodgetts et al, 2010).
Shelters were identified universally as crime generators (Brantingham & Brantingham, 1995) where the risk of victimization via theft or assault was a constant concern because of the concentration of strangers, often with mental illness or substance abuse issues. Our findings suggest this may be because persons with physical and mental disabilities and the fewest social skills are more likely to be reliant on shelters and subject to its many disadvantages and constraints in return for the reliability of housing each night.
This study experiences a number of limitations worth mentioning. First, this study does not assess the economic viability of various income generating activities (e.g., Conroy, 2001;Reinhard, 2021). Activities are unlikely to be equally lucrative, and the economic yield of certain activities may incentivize engaging in some activities more frequently than others. Secondly, while qualitative interviewing provided a nuanced understanding of the everyday experiences of the homeless, a quantitative approach may have provided a larger sample of respondents with more diverse activities and mobility concerns. However, such a sample may have required using shelters as a CrimRxiv • "I know a guy": Examining homeless income generation and spatial mobility 15 sampling frame (e.g., May, 2003;Rollinson, 1998) or relied on participants substantially different from the persons involved in this study.

Conclusion and Implications
The homeless persons in this sample engaged in many kinds of income generating activities that most often required little travel. Exceptions exist, particularly among the few persons who were street homeless for longer periods of time and generated income by traveling greater distances to seize perceived opportunities across the city.
Regarding the first research question, and perhaps unsurprisingly, the sample often volunteered at the market, sold goods at the market, or volunteered at downtown shelters or service agencies nearby that provided meagre financial compensation.
There were 23 kinds of proactive activities identified and persons engaged in between 1 and 4 strategies to generate money. Regarding the second research question, persons typically travelled less than four km for most income generating activities, though some activities prompted further travel, such as scavenging for items to sell, or the one person who sold alcohol in an area populated with university students.
This study finds a number of conditions that appear to influence mobility which may be relevant for homeless service providers. First, many of the unsheltered homeless had unfavourable perceptions regarding shelters. While homeless point-in-time counts commonly cannot count all shelter residents (Thomson, 2016, p.13), homeless persons may be more easily identified for outreach services in a shelter compared to isolated city parks. Additionally, participants who did use shelters wished systems existed that allowed them to utilize shelters without daily check-in times that required residents to wait nearby at particular times. Similarly, participants desired greater efficiency during food distribution efforts to allow them to spend more time looking for employment instead of waiting long times to receive food. Second, proactive income generating activities were a driving force behind participants' desire to travel, and it was commonly short distances. This study was supportive of agencies considering sleeping location and nearby employment location when empowering homeless persons with employment and other income generating opportunities. Previous studies have found that homeless persons would rather engage in legitimate employment than some kinds of illegitimate, given the opportunity (e.g., Bose & Hwang, 2002;Wamstad, 2007); assessing desire to work and matching persons with suitable employment could similarly be beneficial for homeless persons and employers. CrimRxiv • "I know a guy": Examining homeless income generation and spatial mobility 16 Notes 1. Pseudonyms were used to protect participant privacy for this study.
2. Kenny is the individual who owned a car during interviews. Despite saying that he would use multiple methods to get somewhere, he indicated his primary method was walking. He most often used his vehicle for sleeping.

Independent Employment
Selling self-produced goods d 2 (2/0) 0.9 Panhandling 2 (2/0) 2.95 Prostitution 1 (0/1) 0.2 a "Count" is the number of persons engaging in the activity. The total exceeds 24 persons because most people engaged in multiple activities for income generation. b The one transgender individual is counted as female because the individual was transitioning to being female. c "Volunteers" at the market were compensated financially for their time, though it was typically $3/hour, less than legally required for minimally paid employees at traditional businesses in the study location. d One individual drew comics and another was Indigenous and sold carvings he made.