Homelessness | Rapid Evidence Assessment

MARCH 2019

Homelessness Causes of Homelessness and Rough Sleeping

Rapid Evidence Assessment

+++++++

Page 1 of 59 Homelessness | Rapid Evidence Assessment

Homelessness Causes of Homelessness and Rough Sleeping

Rapid Evidence Assessment

Page 2 of 59 Homelessness | Rapid Evidence Assessment

Table of Contents

1. Non-technical summary ...... 4 2. Review aims and objectives ...... 8 3. Review Methods ...... 11 Search Strategy ...... 11 Inclusion criteria ...... 12 Search results ...... 13 Quality assessment ...... 13 4. What are the key factors, drivers and causes of homelessness and rough sleeping in the UK? ...... 16 Subgroups ...... 20 5. What are the strongest drivers, structural and individual, of homelessness and rough sleeping? ...... 25 Subgroups ...... 28 Rough Sleeping ...... 30 An overview of the strength of evidence ...... 30 6. How do these drivers vary over time & geography (national contexts and regional within the UK)? ...... 34 Time ...... 34 Geography ...... 35 7. How do these drivers vary across types of homelessness and rough sleeping? ...... 39 Family homelessness ...... 40 Single homelessness ...... 40 Young people ...... 41 Rough Sleeping ...... 41 8. What are the pathways in and out of homelessness in the UK? ...... 43 Pathways into homelessness ...... 43 Pathways out of homelessness ...... 45 9.Conclusion ...... 48 References ...... 49 Appendix A ...... 52 Appendix B ...... 58

Page 3 of 59 Homelessness | Rapid Evidence Assessment

a suite of models that will inform 1.Non- homelessness policies. To build models which are useful and technical robust, it is important to have a grasp of the theory underlying summary homelessness, which is where this review fits in. This review aims to summarise the existing evidence on the causes of We conducted a search of papers, homelessness in the UK and assess reports and books relevant to the strength of the evidence. It also answering the review’s research seeks to identify any weaknesses and questions (i.e. main causes of gaps in the current evidence base. homelessness, how causes vary for different subgroups, how causes vary Alma Economics was commissioned among different types of to carry out a one-off study consisting homelessness or rough sleeping, how of three elements – to summarise do causes vary geographically and knowledge on the causes of over time) using a set of search terms homelessness and rough sleeping and and inclusion criteria. provide advice on possible next steps In total we screened the titles and towards developing a suite of abstracts of 144 studies. From this quantitative, predictive models of total number of studies, information homelessness and rough sleeping in from 58 was included in the review. England. While the focus was on studies Accompanying this report is a second concerning the UK, studies that report “A review of homelessness examined other countries – such as models”, which categorises and US, Australia, Germany – were also describes a range of models used to included in the review. inform policy decisions on When considering the quality of homelessness. The report also evidence reviewed there are two key assesses the relative strengths and dimensions of interest – firstly about weaknesses of different models. the standard of each research paper (i.e. its integrity and what it adds to The third and final report is a feasibility knowledge generally). Secondly, there study that considers existing evidence is the question of relevance to the about review questions. We used the • the causes of homelessness, Evidence for Policy and Practice Information and Co-ordinating • models applied to measure and predict homelessness and Centre’s (EPPI-Centre) Weight of Evidence framework to assess the • data availability to identify a range quality of evidence available. Broadly of options for the development of the evidence reviewed was of

Page 4 of 59 Homelessness | Rapid Evidence Assessment

medium-high quality in terms of was limited in its usefulness to inform research and relevance to answering prevention policies. For example, the review questions. However, in eviction is present in most lists of terms of using the evidence to inform reasons and rank highly across all model development, there were a types of homelessness and various number of gaps that are highlighted in subgroups. However, there is less the report. information available on factors that Most research divided the causes of lead to eviction (e.g. employment homelessness into structural and issues, rent increases, inability to pay individual factors, though the report a deposit on another rental property) discusses the criticism of this and how policy could effectively target established categorisation. Recent these issues. literature uses a hybrid approach The limited evidence we read on the which acknowledges that structural causes of homelessness over time factors create the conditions within showed the end of an assured which homelessness occurs and shorthold tenancy has become an people with personal problems are increasingly common reason for more vulnerable to these adverse homelessness in the past ten years. social and economic trends than Again, from a policy perspective, more others. In terms of important causes of detailed information would be helpful overall homelessness, papers often to understand the reasons behind this cited affordability of housing, change. relationship breakdown and poverty. In terms of the causes of While there was recognition that some homelessness across the three causes of homelessness do interact, different types of homelessness there was limited detail on how a set considered in the report – statutory of causes interact or any dynamic homelessness,1 single homelessness effects of different causes on and rough sleeping – the research homelessness. indicated that structural factors were We saw relatively little quantitative more important in explaining family work attempting to measure the homelessness. People sleeping rough relative strength of different causes of were more likely to have individual homelessness, with Bramley and factors contributing to their reasons for Fitzpatrick (2017) being the exception. being homeless (e.g. mental health Consequently, often analysis of the and relationship breakdown). reasons people became homeless

1 putting this report into context, we use the term 1 Under the HomelessnessHomelessness Reduction Reduction Act Act 2017, 2017, the definition of statutory homelessness has been recently extended to include all homeless people (including single homeless‘statutory andhomelessness’ those in hidden to refer homelessness) to the former who the definition of statutory homelessness has been official definition (i.e. homeless households in recentlyturn to Local extended Authorities to include for allhomelessness homeless people and rough sleeping services. For ease of reference and to avoid any confusion when putting this report into context, we usepriority the term needs ‘statutory that apply homelessness’ to LAs for temporary to refer to the (including single homeless and those in hidden accommodation), which is still universally used in homelessness)former official definitionwho turn to(i.e. Local homeless Authorities households for in priority needs that apply to LAs for temporary accommodation), which is still universally used in the literaturethe literature on homelessness on homelessness and rough and sleeping rough sleeping in homelessness and rough sleeping services. For in England. easeEngland. of reference and to avoid any confusion when

Page 5 of 59 Homelessness | Rapid Evidence Assessment

Family Homelesness Strong evidence Medium evidence Weak evidence

Domestic Individual financial abuse or related factors violence Overstaying welcome/could Financial Anti-social issues no longer be behaviour accomodated Relationship Asked to leave breakdown or evicted

Overcrowding Individual personal factors CAUSES OF FAMILY HOMELESSNESS

Lack of affortable housing Physical health Loss of job/ Housing unemployment in poor condition Loss Mental health of rented Poverty or tied housing Substance misuse Welfare reforms Individual health related factors Structural factors

Single Homelesness Strong evidence Medium evidence Weak evidence

Women Domestic abuse or Individual financial violence Leaving related factors prison Financial issues Bereavement Institutional discharge Leaving Asked to leave parental home Relationship or evicted after breakdown arguments Discharge from armed forces Individual personal factors CAUSES OF SINGLE HOMELESSNESS

Lack of affortable housing Physical health Loss of job unemployment

Mental health Poverty Structural factors Substance misuse Welfare reforms Individual health related factors

Page 6 of 59 Homelessness | Rapid Evidence Assessment

Rough Sleeping Strong evidence Medium evidence Weak evidence

Violence, harassment Individual financial or abuse related factors

Bereavement Institutional Financial discharge issues

Relationship Asked to leave breakdown or evicted

Leaving prison Individual personal factors CAUSES OF ROUGH SLEEPING Structural factors General population

Loss of job unemployment Mental health Housing in poor Substance Lack of condition misuse affordable housing Lack of emergency Individual health accommodation related factors Young people

Page 7 of 59 Homelessness | Rapid Evidence Assessment

2. Review The structure of the report is as follows: aims and • Section 3 outlines the review objectives methods • Section 4 answers “what are the The review aims to summarise the key factors, drivers and causes of existing evidence on the causes of homelessness and rough sleeping homelessness in the UK and assess in the UK?” the strength of the evidence. It also seeks to identify any weaknesses and • Section 5 answers “what are the gaps in the current evidence base. strongest drivers, structural and individual, of homelessness and The specific research questions the rough sleeping?” review answers are: • Section 6 answers “how do these • What are the key factors, drivers drivers vary over time and and causes of homelessness and geography (i.e. national contexts rough sleeping in the UK? and regional within the UK)?” • What are the strongest drivers, • Section 7 answers “how do these structural and individual, of drivers vary across types of homelessness and rough sleeping? homelessness and rough sleeping?” • What is the strength of evidence underpinning this assessment? Is • Section 8 answers “what are the evidence stronger for certain drivers pathways in and out of than others? homelessness in the UK?” • Can the strength of each driver be • Section 9 concludes the review, quantified, for the purposes of discussing whether the strength of model development? each driver can be quantified for the purposes of model development. • How do these drivers vary over time and geography (i.e. national Alma Economics was commissioned contexts and regional within the to carry out a one-off study consisting UK)? of three elements – to summarise knowledge on the causes of • How do these drivers vary across homelessness and rough sleeping and types of homelessness and rough provide advice on possible next steps sleeping? towards developing a suite of • What are the pathways in and out quantitative, predictive models of of homelessness in the UK? homelessness and rough sleeping in England.

Page 8 of 59 Homelessness | Rapid Evidence Assessment

Accompanying this report is a second report “A review of homelessness Terminology models”, which categorises and For the purposes of this report, we describes a range of models used to break down homelessness into three inform policy decisions on categories using the following homelessness. The report also definitions: assesses the relative strengths and weaknesses of different models. • Homeless families – this could be a single parent or couple with The final report is a feasibility study dependent children in their care that considers existing evidence about i) the causes of homelessness, ii) • Single homeless people – the models applied to measure and identifying feature of this group is predict homelessness and iii) data that they do not have dependent availability to identify a range of children. This group includes a options for the development of a suite couple without children in their care of models that will inform and the hidden homeless homelessness policies. population While the evidence review aims to • Rough sleeping – those who are investigate the causes of homeless and bedded down on the homelessness in the UK, the majority streets including people new to the of the UK-based research was only streets and people entrenched on relevant to England. We found no the streets with multiple, complex evidence for Northern Ireland and needs Wales and limited evidence covering • With regards to the terms causes, Scotland. Consequently, one gap in drivers, and triggers, we try to the evidence reviewed is around replicate the terminology used in understanding of homelessness each study. The terms risk factors causes across the devolved nations. and predictors were also We note the focus of the later encountered during our review, but feasibility study is for modelling less frequently. Table 1 contains homelessness in England. common definitions for these terms, though it is important to note that the terms are not explicitly defined in most papers we reviewed, and hence there may be occasions where the authors use them differently.

Page 9 of 59 Homelessness | Rapid Evidence Assessment

Terms and descriptions

Term Description

Act, condition, circumstance or characteristic that leads to Cause and/or sustains homelessness.

Act, condition, circumstance or characteristic that Driver contributes to and/or sustains homelessness, but does not necessarily cause it

Act, condition, circumstance or characteristic Predictor systematically predates homelessness events, but is not necessarily a causal factor

Act, condition, circumstance or characteristic that increases the probability of homelessness, i.e has a causal Risk factor impact at population level, but may or may not have an impact in the case of any particular individual/household

Act, condition, circumstance or characteristic that Trigger immediately predates a homelessness event but does not constitute a fundamental cause

Page 10 of 59 Homelessness | Rapid Evidence Assessment

vary among different types of 3. Review homelessness or rough sleeping, how do causes vary geographically, over Methods time etc.). The databases and online libraries searched included: Google Scholar, JSTOR, Science Direct, SpringerLink, SSNR eLibrary, IDEAS, Search Strategy NBER, PsychInfo and Cochrane. The search was conducted using We conducted a search of papers, combinations of different search reports and books relevant to terms. answering the study’s research We provide a list of the search terms questions (i.e. main causes of we used for the REA in the following homelessness, how causes vary for table: different subgroups, how do causes

Rapid evidence assessment search terms

Primary terms General secondary terms Specific secondary terms

Homeless Causes Poverty Homelessness Drivers Substance misuse Rough sleeping Risk factors Housing benefit Transitional housing Structural causes Welfare Statutory homelessness Predictors Local housing allowance Sofa surfing Pathways Universal Credit Temporary accommodation England Housing benefit No fixed abode UK Discretionary Housing Hidden homeless Triggers Payments Single homeless Protective factors Benefit cap Supported housing Transitional housing Austerity Non-statutory homeless Transitional housing Supporting People Multiple exclusion Housing (affordable) Chronic homeless Mental health Family breakdown Relationship breakdown Unemployment Labour market Complex needs Disability House prices Assured short-hold tenancy Migrant Immigrant

Page 11 of 59 Homelessness | Rapid Evidence Assessment

In addition, the references listed in from 58 were included in the review. some studies retrieved were screened While the focus was on studies in order to identify additional articles covering the UK, studies that for possible inclusion in the REA. In examined other countries – such as total we screened the titles and US, Australia, Germany – were also abstracts of 144 studies. From this included in the review. Only studies total number of studies, information written in English were included. Inclusion criteria

Terms and descriptions Criteria Review scope

Exposure of interest • Homelessness and rough sleeping causes

Participants • Statutory homelessness/families who are homeless - families with dependent children and households containing a vulnerable person

• Single homeless people - all those who fall outside the statutory definition and could be staying in hostels, or sofa surfing. This category also includes the hidden homeless population • Rough sleeping - those who are homeless and bedded down on the streets, including people new to the streets, and people entrenched on the streets with multiple, complex needs

Peer review • Prioritised peer reviewed evidence but we did include technical reports

Geographic location • UK and international – a focus on UK but we also considered relevant research from comparable developed countries (e.g. Germany, US, Australia, Canada)

Dates of research • Prioritised research from 2005 but some important, often cited papers pre-dated 2005 so these were included • Date of publications were considered on a case by case basis

Research methods/ • All methods study design

Language • English only

Type of publication • Peer reviewed journal articles; non-peer reviewed academic outputs (reports, working papers, etc.); government commissioned research; publications by other research organisations (e.g. charity bodies in the Homelessness Sector); practitioner and provider evidence • Includes evidence reviews and original studies • Excludes editorials/newspaper articles

Page 12 of 59 Homelessness | Rapid Evidence Assessment

Search results Quality assessment

A total of 58 papers were included in When considering the quality of the evidence assessment. Of these evidence reviewed there are two key studies, 7 referenced international dimensions of interest – the standard comparisons. The following types of of each research paper (i.e. its integrity papers were included: and what it adds to knowledge generally) and the relevance to the • Reviews of the literature (16 papers) review research questions. We use the • Qualitative primary research (16 Evidence for Policy and Practice papers) Information and Co-ordinating Centre’s (EPPI-Centre) Weight of • Quantitative primary research (10 Evidence framework to assess the papers) quality of evidence along both these • A combination of the above - most dimensions. As outlined in Gough often a review followed by (2007), the Weight of Evidence qualitative primary research (16 framework gives a tool for making papers) separate judgements about generic There were many review questions and review-specific criteria. and a limited amount of time to carry We have gone through each paper out the study – this led to judgements and made an assessment about the being made about the relevance of following four dimensions – Evidence papers, based on screening the titles A, Evidence B, Evidence C and and abstracts. We recognise this Evidence D.2 The descriptions below review is not comprehensive and there are adapted from Gough (2007). is more relevant material that we did not have time to include.

22 For more details on the assessment of each paper, the interested reader may refer to the appendix A. For more details on the assessment of each paper, the interested reader may refer to the appendix A.

Page 13 of 59 Homelessness | Rapid Evidence Assessment

Weight of Evidence A – General understanding causes of judgement on coherence and homelessness. integrity There was some overlap between This is a generic and therefore non- Weight of Evidence A and B in the review specific judgement about the assessment. However certain coherence and integrity of the reoccurring factors created variations evidence on its own terms. It may such as i) when the research method align with the generally accepted was appropriate, but the sample came criteria for evaluating the quality of the from country outside the UK (where evidence by those who use and Weight of Evidence A>Weight of produce it. In making this judgement, Evidence B), and ii) where the small we used a modified NICE checklist sample was appropriate for generic designed to assess qualitative or research but for the sake of the review quantitative evidence and think about specific questions (i.e. understanding how research questions are answered. the causes of homelessness), a larger è 37 papers received a ranking of sample would have been desirable to high answer the research questions (again here Weight of Evidence A>Weight of è 17 papers received a ranking of Evidence B). medium è 39 papers received a ranking of è 4 papers received a ranking of low high The papers that ranked low were è 11 papers received a ranking of either very outdated, or from short medium summary reviews with some detail missing that could not be found è 18 papers received a ranking of elsewhere. low

Weight of Evidence B – Specific Weight of Evidence C – Specific judgement on forms of evidence judgement on relevance This is a review specific judgement This is a review specific judgement about the appropriateness of certain about the relevance of the scope of forms of evidence for answering the the evidence for this review. For review questions, in other words example, a research study may not whether the methodology used is have the type of sample, the type of suited to answer the review research evidence gathering or analysis questions. For example, the evaluation required to answer the review of a pilot programme which seeks to research questions or it may not have improve long term outcomes for been undertaken in an appropriate homelessness individuals is not likely context from which results can be to be the most useful method for extrapolated. For example, a case study with a small sample may not

Page 14 of 59 Homelessness | Rapid Evidence Assessment

provide robust information on causes Papers ranked as low quality for this of homelessness in the UK. specific review were based entirely out of the UK and their subject areas were There may also be issues of propriety not focused on causes of of how the research was undertaken, homelessness but other aspects of such as the ethics of the research, homelessness. The two papers that that could impact on its inclusion and included primary data collection and interpretation (Pawson et. al. 2003). were ranked low had very small Studies that were marked as low in samples, so it would be difficult to say this area were those that discussed how relevant the findings were to the homelessness generally without UK homeless population as a whole. having a focus on causes, whether set These assessments of the quality of in the UK or internationally, and those evidence, while based on structured focusing only on specific types of frameworks, require making many homelessness. Studies that were subjective judgements. For example, marked as medium were studies how useful is research asking a small about causes of homelessness but unrepresentative sample of homeless outside of the UK context. Studies that people in England to understanding were marked as high in this area were the causes of homelessness in the those about the causes of UK? In these cases, we have given a homelessness in the UK context, medium overall assessment, where including for different types of the research is of good quality. homelessness, different geographic areas (within the UK) and including è 31 papers received an overall international comparisons where ranking of high England was part of the set of è 24 papers received an overall countries examined. ranking of medium è 30 papers received a ranking of è 3 papers received an overall high ranking of low è 19 papers received a ranking of Overall the quality of research was medium medium-high to answer the research è 9 papers received a ranking of low questions outlined earlier.

Weight of Evidence D – Overall judgement This is a review specific judgement which combines the previous assessments to give an overall weighting.

Page 15 of 59 Homelessness | Rapid Evidence Assessment

addiction challenges. Relationship 4. What are problems can include domestic abuse and violence, addiction, the key mental health problems of other family members and a lack of factors, financial resilience. Even though this division of factors is drivers and well established in the literature, we did see some issues with the structure. Some incorrectly interpret causes of individual factors as issues of personal agency (i.e. the individual is culpable), homelessness though there are individual factors that are outside a person’s control and rough (Bramley and Fitzpatrick, 2017). In recent literature, most academic sleeping in the commentators attempt to weave together consideration of both UK? structural and individual factors when studying homelessness – a blended or Traditionally, the factors that cause hybrid approach. This leads them to a homelessness, in the UK and position described by Pleace (2000) as internationally, have been divided the ‘‘new orthodoxy’’. The key between structural and individual assertions of this approach are the factors (Bramley and Fitzpatrick, 2017; following: structural factors create the Busch-Geertsema et al., 2010; conditions within which homelessness Fitzpatrick, 2005; Neale, 1997): will occur and people with personal • Structural factors are wider societal problems are more vulnerable to these and economic issues that affect adverse social and economic trends opportunities and social than others. Hence, the high environments for individuals. This concentration of people with personal includes unfavourable housing and problems in the homeless population labour market conditions, reduced can be explained by their susceptibility to structural forces and not solely by welfare and benefits rising levels of poverty and the growing their personal circumstances fragmentation of the family. (Fitzpatrick, 2005). • Individual factors apply to the However, it should be added that the personal circumstances of a hybrid approach – while providing a homeless person. These factors more rounded explanation for may include personal crisis, homelessness than the individual and traumatic events, mental health or structural accounts that preceded it –

Page 16 of 59 Homelessness | Rapid Evidence Assessment

is, according to Fitzpatrick (2005), for thinking about homelessness unsatisfactory from a theoretical point causes, which identifies causal of view, since: mechanisms on four levels: • the individual/structural division of • economic structures – social class the explanations for homelessness interacts with other stratification is unhelpfully crude and reflects the processes and welfare policies to discredited concept of a strict generate poverty and to determine agency/structure dichotomy in poor individuals’ and households’ sociological theory (Stones, 2001). non-access to material resources such as housing, income, • it fails to convincingly host a whole employment and household goods. range of factors that could contribute to homelessness – • housing structures – inadequate especially when structural factors housing supply and a deterioration are limited to macroeconomic social in affordability can squeeze out and economic forces and individual those on lower incomes; tenure and factors are limited to personal allocation policies, coupled with the behaviours. For instance, the collective impacts of private experience of poor parenting is choices, can lead to residential neither a macro-structural nor a segregation and spatial behavioural issue. concentration of the least advantaged groups. • there are many factors that could be interpreted as operating either at • patriarchal and interpersonal a structural or individual level. For structures – can lead to the instance, is the breakdown of a emergence of domestic abuse and homeless person’s marriage the violence, child neglect, weak social result of personal behaviours or the support, relationship breakdown, result of a structural trend towards etc. growing family fragmentation. • individual attributes – personal • the “new orthodoxy” is difficult to resilience can be undermined by account for the cases of mental health problems, substance homelessness that are the result of misuse, lack of self-esteem and/or acute personal crises where confidence. structural factors may seem almost In what follows, we present a absent. summary of some key papers that use • it lacks a clear conceptualisation of the structural/individual factor split to homelessness causation. discuss the causes of homelessness. Furthermore, it might also be useful to Bramley and Fitzpatrick (2017) find note that Fitzpatrick (2005) proposed that – in the UK at least – the odds of an alternative “conceptual framework” experiencing homelessness are

Page 17 of 59 Homelessness | Rapid Evidence Assessment

systematically structured around a set Wilson and Barton (2018) also of identifiable individual and structural highlight the structural and individual factors, with most of the latter being factors that could lead to outside the control of those directly homelessness. On the one hand, affected. They state that factors such structural factors include the lack of as demographics, housing and labour available affordable housing in markets, poverty and teenage England, the recent decline of the experiences are important drivers of social housing sector as a proportion homelessness and demonstrate their of all housing, tighter mortgage relative importance (see the next regulation and requirements for higher section). Their findings mainly support deposits from first-time buyers. On the the structural analysis of other hand, individual factors include homelessness, without discounting relationship breakdown, mental illness the possibility of wholly individual and addiction, discharge from prison causes in specific cases. They also and leaving the care system. The recognise the potentially protective authors also mention that the impact of strong social support circumstances in which families networks. become homeless tend to differ from those of single homeless individuals, The National Audit Office (2017) with the latter being more likely to analysed levels of homelessness since experience more ‘chaotic’ lifestyles. 2004-5 to understand its causes. The However, no further details were research found that the causes differ provided. for individual households, who can become homeless for many different Fitzpatrick et al. (2018) conclude that reasons (e.g. affordability of housing, the causes of homelessness are relationship breakdown, parents no complex, with no ‘necessary’ or longer being willing or able to house ‘sufficient’ single ‘trigger’ event that children in their own homes). The one can point to. They also mention report shows that, while it is possible that individual, interpersonal and for anybody to become homeless, the structural factors all play a role and risk is highest for households who live interact with each other. To give an in centres of economic activity and example of interactions take the who are on the margins of being able example of poverty – Fitzpatrick and to pay market rents for their homes. Bramley (2017) find that childhood Throughout the analysis period, the poverty is a powerful predictor for risk of homelessness was highest in adult homelessness. While poverty London and other urban centres. can lead to homelessness though a Since 2010-11, the risk of person not being able to pay for homelessness was lowest outside of housing, Fitzpatrick and Bramley urban centres in the South East and (2017) note other ways poverty can East of England. cause homelessness. They note that poverty has a strong causal effect on

Page 18 of 59 Homelessness | Rapid Evidence Assessment

mental and physical health outcomes Thus, deteriorating economic and that entrenched poverty is also conditions in England could also be linked to serious forms of drug misuse expected to generate more individual and chronic offending. For research on vulnerabilities to homelessness over interactions and overlaps (also known time. as multiple disadvantage), see the summaries of Fitzpatrick et al. (2011) and the Bramley et al. (2015) report in the next section. Section highlights With respect to the main structural factors, Fitzpatrick et al. (2018) further In conclusion, the strongest evidence mention that international comparative we found was in the areas where there research and the experience of is a broad consensus in the literature previous UK recessions suggests that about i) the way to organise the housing market trends and policies various causes of homelessness (i.e. have the most direct impact on levels structural and individual), ii) the fact of homelessness. The influence of that a combination of structural and labour market change is more likely to individual factors can lead to be lagged and diffuse, and strongly homelessness, iii) which causes are mediated by welfare arrangements important (e.g. affordability of housing, and other contextual factors. relationship breakdown and poverty) and iv) that causes vary over a number Moreover, Loopstra et al. (2016) of dimensions. demonstrate that reductions in spending on social welfare by local There are two main gaps in the authorities and central government in evidence reviewed about overall the UK are strongly associated with causes of homelessness (without increased homelessness. considering specific subgroups or types of homelessness). Firstly, there Finally, Fitzpatrick et al. (2018) also is a lack of clarity around the discuss how individual vulnerabilities, importance of different factors (see support needs, and ‘risk taking’ next section) and secondly while we behaviours implicated in some saw acknowledgement that some people’s homelessness are causes of homelessness do interact, themselves often, though not always, these overviews did not attempt to rooted in the pressures associated gain an understanding about how a with poverty and other forms of set of causes interacts or any dynamic structural disadvantage. At the same effects of different causes on time, ‘anchor’ social relationships – homelessness. such as family bonds – that could act as a primary ‘buffer’ to homelessness, can be put under considerable strain by stressful financial circumstances.

Page 19 of 59 Homelessness | Rapid Evidence Assessment

that structural factors relating to Subgroups financial hardship, housing, and labour market pressures greatly contribute to The previous discussion demonstrates family tensions and conflict. This that there is much variation in indicates a difference in the order of explaining why some people become how structural and individual factors homeless. For example, individual interact compared to the “new factors play a more important role for orthodoxy”. For young people some people in falling into relationship breakdown tends to be homelessness (Clarke, 2016). It seems the initial driver of homelessness reasonable to expect that different (Watts et al., 2015) – though this types of individuals may experience particular driver can be influenced by homelessness for different reasons. To both structural factors (e.g. attempt to unpick the causes of employment) or other individual factors homelessness, we investigate the (e.g. mental and physical health available research on different issues). population subgroups, such as The research also examined the views different age groups (young and older of LAs and homelessness service people) and women. For a discussion providers on the impact of welfare on family homelessness, see section 7 benefit changes as introduced within which covers types of homelessness. Universal Credit and wider welfare reforms in the past year. Young People To be more specific, the findings of Even though the rate of youth Homeless Link (2018) strongly suggest homelessness is difficult to quantify, that the administrative changes and roughly half of the individuals in delays under Universal Credit, supported homeless accommodation including delayed payments, housing in England are between the ages of costs paid direct to claimant, monthly 18-24 (Homeless Link, 2017). payments in arrears, removal of Relationship breakdown between automatic entitlement to housing costs young people and their family, or their for 18-21 year olds, digital by default primary caregivers has consistently and youth obligation, adversely emerged as a leading cause of youth affected young people’s access to homelessness (Homeless Link, 2015 housing. and 2018). As for wider reforms – that is, Recent research by Homeless Link sanctions, capping of local housing (2018), reflecting on findings from allowance to shared accommodation surveys with 188 homelessness rate, benefit cap, abolition of the spare services and local authorities (LAs) room subsidy (i.e. bedroom tax) and across England as well as in-depth non-dependent deductions – the interviews with 25 young people who findings showed that both benefit have experienced homelessness, finds

Page 20 of 59 Homelessness | Rapid Evidence Assessment

sanctions and the capping of the local that 36% of young people at housing allowance to the shared responding providers and LAs were accommodation rate particularly homeless because their influenced young people’s housing parents/caregivers were no longer options. More specifically, benefit able or willing to accommodate them, sanctions and the capping of the local with a further 24% no longer able to housing allowance to the shared stay with other relatives or friends. accommodation rate were considered by about 50% and 40% respectively of Nevertheless, the authors noted that respondents to have influenced young the importance of family conflict in people’s ability to access and sustain forcing young people to leave the their accommodation. family home does not necessarily reveal the full story of the causes of Moreover, individual factors – such as youth homelessness and how they are mental and physical ill health or changing. Family conflict can be the substance misuse – played a key role result of individual problems and in them either choosing or being support needs, interpersonal problems asked to leave home. Some young or manifestation of childhood trauma. people also reported that experiences In addition, problems within families of domestic abuse and violence or may also be due to structural factors, neglect contributed to leaving the such as changes in housing/labour family home. market conditions. Watts et al. (2015) reviewed the Interestingly, some of the study’s key empirical evidence on the immediate participants suggested that increasing triggers of homelessness among financial difficulties – due mainly to young people. The authors reported welfare reform – started to play a key that relationship breakdown has been, role in causing family conflict and and continues to be, a main cause of youth homelessness, while another homelessness in the UK. However, participant discussed on how general they state that – according to the data socio-structural changes – in – this factor is even more important for education and labour market – created triggering youth homelessness. For a context in which certain groups of example, according to statutory young people were more exposed to statistics for Scotland for 2014/15, homelessness. Watts et al. (2015) 39% of the 16-24 year olds accepted noted that these comments, which as homeless cited being asked to suggest a shift towards structural leave as the immediate reason for causes of youth homelessness, align homelessness – compared to 25% with findings from other studies, such among all age groups – while a further as YHNE (2014) and Homeless Link 28% cited a dispute within the (2014). household as the immediate reason (Scottish Government, 2015). In Watts et al. (2015) also presented the addition, Homeless Link (2014) found available evidence on the main

Page 21 of 59 Homelessness | Rapid Evidence Assessment

characteristics of young people who Young LGBT experienced homelessness in the UK. Everything mentioned above regarding Part of the report looked at the young people also applies to the following topics: socio-economic young members of the LGBT background, age, gender, ethnicity, community. However, young people household type and who identify as LGBT experience more nationality/migration status. Regarding acute challenges and are more likely the socio-economic background, to find themselves homeless than their Watts et al. (2015) referred to the non-LGBT peers. results of the UK Poverty and Social Exclusion Survey 2012 (Fitzpatrick, et According to AKT (2015), being a al., 2013), which showed that lesbian, gay, bisexual or transgender increased risk of homelessness is young person increases the risk of linked with socio-economic experiencing parental rejection, disadvantage of various kinds, such as alongside familial physical, sexual and being a lone parent, living in material emotional abuse and/or violence, deprivation and living in a deprived discrimination, substance abuse, neighbourhood. The authors also mental health problems and sexual referred to evidence from Quilgars et exploitation.3 Cull et al. (2006) and al. (2008), that some black and TUC (2016) mention that young LGBT minority ethnic groups are at higher individuals experience high levels of risk of homelessness – a fact that homelessness as a result of reflects the increased risks of socio- homophobia experienced. economic disadvantage amongst these groups. Quilgars et al. (2008) Older people also conclude that the vast majority of young homeless people come from Drivers leading older people to disadvantaged backgrounds. Lastly, homelessness appear to be different citing Bramley and Fitzpatrick compared to the general population. (unpublished), the authors mention For example, Warnes and Crane that analysis of survey data in England (2006) focus on the causes of new and Scotland suggests that the strong episodes of homelessness in the UK association between young age and by examining information from 131 homelessness is explained by the people aged 50 years and over and 4 disproportionate experience of poverty their key-workers or case managers. among young people, rather than their Two thirds of the respondents had youth per se. become homeless for the first time in later life. The findings of the paper suggest that the reasons for

3 3 This This has has also also been been highlighted highlighted by Cochran by Cochran et al. et al. (2002) andhomeless Durso people and Gates assign (2012) key workers in the (or US. case managers) to assess a client’s problems and to advise (2002)4 and Durso and Gates (2012) in the US. The authors note that most hostels and projects for homelessand support people them, assign as was key the workers case for (or all case except managers) seven 4 The authors note that most hostels and projects for to assess a client’s problems and to advise and support them,respondents. as was the case for all except seven respondents.

Page 22 of 59 Homelessness | Rapid Evidence Assessment

homelessness among older people are impact of housing costs on the multiple, complex and for many deep- financial security of older people. seated. Furthermore, both sources indicated that fewer than half of the We note the UK evidence on the identified reasons were personal to causes of homelessness among older them, while a third pointed to service people is relatively dated. deficiencies. Women Crane et al. (2005) used – along with the UK data from Warnes and Crane Though not only affecting women, (2006) – similar information from 122 experiences of violence and sexual people aged 50 years and over in exploitation – both in their own Boston (US) and 124 in Melbourne personal relationships and in the (Australia). The findings indicate that relationships they witnessed growing the reasons why older people become up – play a particularly important role homeless were similar in the three in women’s pathways into cities, including mortgage or rent homelessness (Jones, 1999; May et problems, housing was sold, al., 2007). Pleace et al. (2008) drew on converted, or needed repair and data from 5 linked surveys on families relationship breakdown. It is further and 16-17 year olds accepted as 5 shown that previous experiences of homeless by English LAs, conducted homelessness were more common in 2005 by BMRB Social Research amongst men than women, and that and analysed by the Centre for men were significantly more likely than Housing Policy and found that: women to have been homeless for • family homelessness tends to be periods of more than three years. experienced primarily by younger Women are considered to be more women, who are socially and likely to have first become homeless economically marginalised lone after the age of 50 years (Crane and parents with young dependent Warnes, 2012). children. A recent survey in Australia revealed • there exists a strong association that the causes of homelessness between male domestic abuse and among older people include lack of violence and experience of family affordable housing, declining rate of homelessness, reflected in the home ownership, death of a spouse presence of a disproportionate resulting in reduced income, leaving a number of female lone parents who violent partner or spouse and inability had experienced violence or abuse. to live on the Government pension (Homelessness Australia, 2016). The • the largely female-headed lone survey also highlights the significant parent households who sought

5 5 The The total total number number of usable of usable interviews interviews from each from each survey was:survey 2053 was: (Survey2053 (Survey 1), 450 1), (Survey450 (Survey 2), 3502), 350 (Survey 3), 571 (Survey 4) and 180 (Survey 5). (Survey 3), 571 (Survey 4) and 180 (Survey 5).

Page 23 of 59 Homelessness | Rapid Evidence Assessment

assistance under the homelessness law often tried to avoid using the statutory system. Women and female lone parents experiencing homelessness tended to exhaust informal options (i.e. staying with family, friends and acquaintances) prior to seeking assistance under the homelessness laws. Moreover, women and their children had often been homeless for some time before seeking formal help. Furthermore, a German study based on interviews with homeless women suggested that poverty, poor school education, violence and addiction in families were the main causes of homelessness (Enders-Dragasser, 2010).6 Large and Kliger (2013) note that women live longer than men but continue to earn less, due to lower wages and more part-time work which leads to lower savings, increasing the probability they end up homeless. McFerran (2010) highlights that the persons most likely to be tenants after the age of 45 are never-married people, sole parent households or those who have experienced the dissolution of a relationship.

6 6 ThisThis research research was was based based on aon small a small sample sample of 37 of 37 interviewsinterviews with women with women in Germany. in Germany.

Page 24 of 59 Homelessness | Rapid Evidence Assessment

• Scottish Household Survey from 5. What are 2001-2007 and 2010 • Poverty and Social Exclusion survey the strongest for 2012 • British Cohort Study – provides drivers, systematic data from birth to young adulthood on every individual born structural and in England, Scotland and Wales in one specific week in 1970 individual, of By examining surveys covering different samples of the UK homelessness population, the authors allow for an assessment of how consistent the and rough causes of homelessness are in different datasets. sleeping? According to the researchers, the The previous section presented some results of the Scottish Household findings on the causes of Survey analysis suggest that – homelessness. We next seek to demographics aside – housing market understand what the published conditions are the most important evidence tells us about the relative factor ‘explaining’ homelessness, importance of these causes. For each followed by labour market conditions relevant paper that adds to the and poverty. The analysis of the other literature, we briefly set out i) how the two datasets further reinforces the researchers identified the relative messages on the significant effects of contribution of each cause and ii) the poverty, employment, tenure and results, which are displayed as a family relationships. ranking of causes of homelessness. In sum, Bramley and Fitzpatrick (2017) Bramley and Fitzpatrick (2017) find that homelessness causes are determine the strength of various consistent across surveys. However, factors that cause homelessness by their ranking is different, probably analysing three large-scale UK because of differences in data surveys: availability within each survey and model specifications.7 The fact that the ranking of factors is not the same across surveys implies that the relative

77 ForFor example, example, variance variance in time in time – the – Britishthe British Cohort Cohort Study datageography were (Scottishcollected Household pre-2000, Survey while vs the British Scottish Cohort StudyHousehold data were Survey collected data pre were-2000, collected while the in Scottish late 2010 and geographyStudy) could (Scottish be very Householdimportant drivers Survey of this vs result.British Household Survey data were collected in late 2010 and Cohort Study) could be very important drivers of this result.

Page 25 of 59 Homelessness | Rapid Evidence Assessment

importance of distinct causes may variables block and finally the poverty differ across individuals, geographic variables block. Another sequence locations and over time. To be more might involve introducing the poverty specific, Bramley and Fitzpatrick block following demographics and (2017) use a logistic regression model then the labour market block, etc. for all three datasets mentioned above. In order to determine the There is consistency in the most strength of each factor, they calculate important cause of homelessness the net addition to the overall identified in the Scottish Household explanatory power of the model – in Survey, which is demographics or other words, how much of the individual characteristics. Regarding variation in homelessness is the remaining (non-demographic) ‘explained’ by each factor – by blocks of variables, if they were introducing the variables in blocks and completely unrelated, their in different sequences. One such percentages in the table below would sequence could for instance involve have been the same regardless of the adding to the model the variables order they have been introduced in the block covering demographics first, model. But, the fact that the followed by the labour market percentage of explanation changes so much is a clear indication that they

Scottish Household Survey

Sequence 1 Sequence 2 Sequence 3

% of Block of % of Block of % of Block of variables explanation variables explanation variables explanation

Demographics 38 Demographics 38 Demographics 38

Housing Labour Poverty 24 30 16 market market

Housing Labour market 27 Poverty 6 18 market

Labour Housing market 10 2 Poverty 5 market

Source: Bramley and Fitzpatrick (2017)

Page 26 of 59 Homelessness | Rapid Evidence Assessment

often coexist (e.g. unfavourable labour quantifying overlapping factors. market conditions coexist with According to the researchers, many poverty, making it tricky to speak people who have experienced MEH about the effect of one independently have also experienced one or more of of the other). the following additional domains of deep social exclusion: institutional Then, according to the table above – care, substance misuse or excluding demographics – housing participation in street culture activities market conditions appear to be the set (e.g. begging). The study is based on of factors that adds the most to the data gathered between February and explanatory power of the model (i.e. May 2010 and over 450 interviewers they are most closely correlated with with people who have experienced homelessness), followed by labour MEH. market conditions and poverty. For the analysis of the other two datasets, see Bramley et al. (2015) attempted to the appendix B. provide a statistical profile of a major manifestation of Severe and Multiple Fitzpatrick et al. (2011) examine Disadvantage (SMD) in England. In the ‘multiple exclusion homelessness’ study, the term SMD is used as (MEH) in seven urban locations across shorthand for representing “the the UK – Belfast, Birmingham, Bristol, problems faced by adults involved in Cardiff, Glasgow, Leeds and homelessness, substance misuse and Westminster. This paper is included as criminal justice systems in England” – it was one of the few examples with complicating factors, including

Overlaps between domains of deep social exclusion

Homelessness (98%)

15%

3% 5% 6%

Street Culture Institutional Activities (67%) Care (62%) 47%

< 1% 12% 6%

< 1% 4% < 1%

Source: Fitzpatrick et al.(2011) Substance Misuse (70%)

Page 27 of 59 Homelessness | Rapid Evidence Assessment

poverty and mental ill-health. Subgroups Among other things, the study found that: Family homelessness • SMD, as defined in the report, is As far as family homelessness in distinguished from other forms of England is concerned, Pleace et al. social disadvantage due to “the (2008) (see section 4) note that the degree of stigma and dislocation reasons for applying as homeless from societal norms that these include relationship breakdown, intersecting experiences represent.” eviction, overcrowded housing, outstaying welcome or no longer able • the people affected by SMD were to be accommodated, inability to pay mostly men aged 25-44 with long the mortgage or rent, anti-social histories of economic and social behaviour and mental or physical marginalisation and in most cases health problems. The study is fairly of childhood trauma. dated and it is possible that reasons • apart from general background for applying as homeless have poverty, SMD’s most prominent changed. Section 6 has a more detail early roots seem to come from very discussion on how causes of difficult family relationships and very homelessness have varied over time. poor educational experiences. Family homelessness - • the distribution of SMD cases varies reasons for applying as homeless widely throughout the country, and Reason8 % is widely concentrated in northern cities, some coastal towns and Relationship breakdown 38 central London Boroughs. Eviction or being threatened with 26 eviction • The quality of life reported by persons facing SMD is much worse Overcrowded housing 24 than that reported by many other Outstaying welcome/could no low-income and vulnerable people, 20 longer be accommodated particularly regarding their mental health and the sense of social Inability to pay the mortgage or rent 7 isolation. Harassment, anti-social behaviour or 4 crime

Mental or physical health problems 2

Source: Pleace et al. (2008)

8 Multiple responses were possible.

Page 28 of 59 Homelessness | Rapid Evidence Assessment

A more recent report by Wilson and asked and what information is Barton (2018) presented the reasons available from administrative data or for statutory homelessness recorded surveys. If data were collected by LAs in 2017/18 across England following a uniform procedure, we and involved 56,580 households. The could better identify the changes in findings related to loss of last settled causes of homelessness over time. home are outlined in table 6. Young people

Reasons for loss of last settled The results of the Homeless Link home (proportion of all households (2018) survey (see section 4) indicate accepted as homeless in 2017/2018) that young people accessing homelessness accommodation Reason % services became homeless due to a variety of reasons, as outlined in table End of assured shorthold tenancy 27 below. Parents no longer able to 15 accommodate Young people accessing Other relatives/friends no longer 12 homelessness accommodation services in able to accommodate August 2017 needed accommodation for the following reasons Relationship breakdown - violent 12 Reason % Loss of other rented or tied housing 6 Parents/caregivers were no longer 49 able or willing to accommodate Relationship breakdown - other 6 Drug or alcohol problems 31 Mortgage or rent arrears 1 Mental or physical health problems 26

Other 18 Leaving care 17 Source: Ministry of Housing, Communities and Local Government, Statutory homelessness live Anti-social behaviour or crime 17 table 774, 29 June 2018 Overcrowded housing 12 Even though homelessness causes Other debt-related issues 12 are also consistent across these studies, we should point out that their Unemployment 11 ranking is not the same. For example, inability to pay the mortgage/rent is Domestic abuse 11 found to rank 5th in Pleace et al. (2008) th Financial problems caused by and 7 in Wilson and Barton (2018). 9 benefits reduction Clearly, the ranking of causes is highly Source: Homeless Link (2018) dependent on which questions are

Page 29 of 59 Homelessness | Rapid Evidence Assessment

Rough Sleeping Being asked to leave or evicted was the most commonly reported overall According to the CHAIN (2018) annual category of reason for leaving last report, the people rough sleeping for settled base, cited by 36% (33% in the first time in the area of Greater 2016/17) of the people who were seen London from April 2017 to March sleeping rough for the first time. The 2018 reported various reasons for reasons that fall under the leaving their last settled base, as listed employment and education category in Table 8. account for 18% (down from 22% in 2016/17). Relationship breakdown Reason for leaving last settled was cited as reason for leaving last base settled base by 12% of the people who were seen sleeping rough for the Reason % first time (the same proportion as in 2016/17). Asked to leave or evicted 36

Employment and education 18 An overview of the Relationship breakdown/ death of relative or friend/ move nearer 15 strength of evidence family/friends/community

End of stay in short or medium-term 6 There is a large body of evidence on accommodation the causes of homelessness – the majority of the research covered in this End of stay in institution 5 report was from reviews of the Financial issues 4 literature and qualitative evidence. Qualitative research tended to be Victim of violence, harassment or 4 drawn from small scale surveys of abuse homeless individuals or staff working Housing conditions 1.1 with them. We recognise that, while the samples are relatively small, they Transient/travelling around 1.1 are dealing with a typically hard to reach population. Another source of Domestic violence – perpetrator 0.3 information was surveys of subject experts. For example, Baptista et al. Other 10.2 (2017) questioned leading academics to understand the causes Source: CHAIN (2018) homelessness in a number of countries. However, the results of such analyses should be interpreted

Page 30 of 59 Homelessness | Rapid Evidence Assessment

with caution as the definition of The quality of survey evidence is homelessness tends to vary across mixed. The main issue with the countries (Fitzpatrick and Stephens, questionnaire evidence is around i) 2007) and the available data are often asking the right questions, ii) sample incomparable (Baptista et al., 2017). representativeness and iii) sample sizes. The quality of the review papers was generally good. They also effectively In terms of asking the right question, answered the questions they were we frequently see that the reason for asking (e.g. what does the current homelessness reported by many is literature say about the causes of eviction. For example, one survey homelessness across countries?). asked homeless people their reasons However, this research was often not for being homeless and the most sufficient for model development common response was that they were purposes as it did not provide a solid evicted (over 35% of respondents) evidence base which quantified the while only 4% of respondents said different causes of homelessness. they had financial issues (CHAIN, 2018). While this finding is interesting, From the perspective of having a if the goal is to inform policy useful evidence base for model development this is only partially development, we considered the helpful – it would be more useful to quality of research which uses either understand the factors that led to the qualitative or quantitative data to eviction for the homeless population.10 estimate relative effects of the causes This is why the individual pathways / of homelessness. journeys research is important – Primary data collection about especially for designing policies to understanding the causes of prevent homelessness, though this homelessness is mostly from two work often has very small samples. sources: i) surveys designed On the sample, it is understandable specifically to understand the causes that access to homeless people who of homelessness - where people are are willing to share personal asked about how they became 9 information makes it difficult to collect homeless (e.g. CLG statistics ) and ii) large, representative samples. While broader surveys, which formed the learning about a single person’s basis for quantitative analysis, which homelessness experience is valuable, ask about homelessness experiences when building national statistical as well as many other areas (e.g. the models, it is ideal to understand Scottish Household Survey). information about the entire homeless

99 (Homeless Link, 2018). WeWe alsoalso sawsaw anan exampleexample ofof aa surveysurvey ofof thosethose who worked at accommodation projects recording 10 whoinformati workedon aboutat accommodation individual characteristics projects of those making useThe ofintroduction the shelter of (Homeless the new data Link, collection 2018). recording10 information about individual system HCLIC will be useful for this purpose. characteristics The introduction of those of the making new data use collectionof the shelter system HCLIC will be useful for this purpose.

Page 31 of 59 Homelessness | Rapid Evidence Assessment

population. We acknowledge the specific interactions. Much research composition of this population is acknowledges that the set of causes changing, which alongside data do interact, though we only saw one collection issues make quality paper that measured the overlaps empirical evidence difficult to obtain. between homelessness, institutional care and substance misuse (amongst Overall the research that discussed other factors). directions of causality or aimed to statistically establish causality of Our interpretation of the evidence factors was very limited. reviewed is that there was more agreement about the relative We reviewed a smaller evidence base importance of different factors the around homelessness pathways. more specific a subgroup considered Pathways will explicitly attempt to map is (e.g. young single mothers, older individuals’ journeys, telling the story homeless men). of all the experiences that led to homelessness. The pathway to homelessness for each individual is Section highlights unique. However, research does There are many gaps in our indicate some similar features in the knowledge, most notably a consensus pathways of some subgroups. For on the relative size of different drivers example, for homeless single parents, and an agreed ranking of specific experiencing domestic abuse and factors. violence is often a common experience. We have seen small We saw relatively little quantitative sample examples of mapping work attempting to measure the homelessness journeys. Expanding relative strength of different factors this research for larger samples will be with Bramley and Fitzpatrick (2017) important to drill down to further being the exception. This paper was of understand both interactions between high quality and more research in this causes and the order of events that space is needed – for example, can lead to homelessness for different extensions for different types of subgroups of the population. homelessness or different subgroups. One area where further information The analysis of the reasons people would be useful is around became homeless was limited in its understanding the interactions usefulness. The way the questions are between different homelessness asked of homeless people was often drivers – this is currently a gap in the unhelpful from a policy perspective. literature. Beyond the consensus that For example, eviction is present in housing pressures cause some most lists of reasons and rank highly vulnerable individuals to be at risk of across all types of homelessness and homelessness, we saw limited the subgroups mentioned above, it is evidence explaining and quantifying problematic to view this as a

Page 32 of 59 Homelessness | Rapid Evidence Assessment

fundamental reason behind in the homeless population should homelessness. Clearly, the more lead to an increased understanding of pertinent question for the policy maker the routes into homelessness. trying to reduce homelessness would be about what led to the eviction (e.g. For example, what proportion of the employment issues, landlord homeless population have increasing rent). experienced mental health issues and unsustained employment compared to As with the previous section, the the proportion who have only quantitative work looking at the experienced mental health issues, different reasons for homelessness did recognising that not all who not consider how factors relate to experience these drivers will become each other or overlap. Understanding homeless. how multiple disadvantage is present

Page 33 of 59 Homelessness | Rapid Evidence Assessment

6. How do Time Disentangling whether the causes of these drivers homelessness have actually changed, or whether our understanding of vary over time causes has developed can be difficult. & geography With regards to the first point, while the DCLG (2012) report mentions that the immediate causes of (national homelessness have remained fairly constant over the years, Fitzpatrick et contexts and al. (2018) note that their balance differs over time. This is largely verified regional within in a recent report by the National Audit Office (2017), where they examined trends in the causes of statutory the UK)? homelessness in England (see figure 1 below).11

Figure 1. Reason for loss of last settled home for statutory homelessness in England

Source: National Audit Office analysis of the DCLG 's P1E data

1111 uploads/attachment_data/file/721292/Acceptanc The latest statistics are available at the following link: es_and_Decisions.xlsx followinghttps://www.gov.uk/government/uploads/system/uploads/attachme link: nt_data/file/721292/Acceptances_and https:/_Decisions.xlsx/www.gov.uk/government/uploads/system/

Page 34 of 59 Homelessness | Rapid Evidence Assessment

The proportion of households that lost during the 1980s, as research their last settled home due to the repeatedly demonstrated the non- ending of private sector tenancies has housing problems experienced by increased dramatically, becoming the many single homeless people, biggest single reason given for particularly with regards to mental statutory homelessness in the last few health, drugs and alcohol (Pleace, years. Before this, homelessness was 1998). This led to the hybrid approach primarily driven by more personal outlined previously. factors – such as relationship breakdown and parents no longer being willing or able to house children Geography in their own homes. Both Wilson and Barton (2018) and the National Audit Our reading of the literature around Office (2017) identify the end of the the causes of homelessness in assured shorthold tenancy as the different geographic areas is that, defining characteristic of the increase broadly, the general causes of in homelessness that occurred since homelessness are the same as those 2010. outlined for overall homelessness in In terms of the second point, section 4, though the characteristics according to Fitzpatrick (2005), up of distinct geographic areas means until the 1960s, homelessness in the that different causes have varying UK was explained with emphasis on degrees of importance within each individual pathology, often focusing on area. the ill-health and/or substance dependencies of homeless people. Rural areas However, the latter part of the decade Some research finds that the drivers of saw a shift from individual factors rural homelessness are often similar to dominating explanations for becoming those in urban areas, including homeless towards a focus on more relationship breakdown, being a victim structural factors. of domestic abuse, becoming This was reinforced by a series of unemployed or losing a source of academic studies which forcefully put income (Snelling, 2017). the case that homelessness was the However, while drivers of result of social and economic forces. homelessness may be similar across The influential report by Drake et al. areas, the characteristics specific to (1981) attributed homelessness some local areas can contribute to primarily to an insufficient supply of explaining homelessness. For affordable accommodation for those in example, the lack of affordable weak economic positions. housing is a cause of homelessness The housing market-based account of generally. However, house price homelessness quickly ran into trouble pressure in some rural areas comes

Page 35 of 59 Homelessness | Rapid Evidence Assessment

from characteristics specific to rural alongside lower wages, as well as a areas. These include: general decline in employment in the agricultural sector (Fitzpatrick et al., • high demand for second homes 2000). A lack of employment and holiday lets from an affluent opportunities was a key explanation population moving into the area behind young people’s homelessness • restrictions on further property in geographically isolated areas, so development (e.g. to protect the this issue may extend beyond rural greenbelt and preserve the areas to include coastal areas (see characteristics of rural villages) Homelessness Link, 2018). This research demonstrates the interaction These issues leave a gap between between geography, labour market housing supply and demand conditions and personal relationships (Fitzpatrick et al., 2000; Snelling, leading to youth homelessness. 2017). According to DEFRA (2018), house prices are less affordable in Finally, the lack of transportation has predominantly rural areas than in also been associated with predominantly urban areas (excluding homelessness in rural areas, as it London).12 impedes the rural homeless’ access to jobs, services, healthcare, education Our review found very limited evidence and affordable housing (Fitchen, that individuals in rural areas are more 1992), though we acknowledge this likely to rely on support from their research is dated. The geography of social network. Robinson and Coward rural areas compounds these (2003) found that in urban centres, problems as large distances must be such as London, 69% of individuals travelled to reach these services, and had stayed with family and friends often there is limited or no public since becoming homeless, rising to transportation available. 72% in Sheffield and up to 77% in the mainly rural Craven. Understanding We note that while rural areas have the factors explaining these findings their own characteristics that affect would be useful. different causes of homelessness, homelessness remains more prevalent We have seen that labour market in urban centres. The National Audit conditions are an important structural Office (2017) found the risk of determinant of homelessness homelessness was higher in London generally. There are aspects of labour than other cities consistently over market conditions specific to rural time. areas. Notably, a high level of part- time or seasonal employment

1212 7.4 times in predominantly urban areas (excluding In In 2017,2017, thethe averageaverage lowerlower quartilequartile househouse priceprice was 8.6 times the average lower quartile earnings in was 8.6 times the average lower quartile earnings London), 15.1 times in London and 9.1 times in predominantly rural areas. This compares with 7.4 times in Englandpredominantly as a whole. urban areas (excluding London), in15.1 predominantly times in London rural andareas. 9.1 This times compares in England with as a whole.

Page 36 of 59 Homelessness | Rapid Evidence Assessment

problems in their childhood). Repeat Coastal Areas homelessness was common, with Coastal areas have some similarities many respondents moving between with rural areas with respect to two different forms of accommodation and homelessness drivers i) housing different types of homelessness over affordability and ii) labour market the course of their housing pathways. conditions. There is pressure on house The researchers identified four prices due to migration from retirees subgroups with similar homeless and holidaymakers and there is a lack journey features – single men, young of available affordable, mainstream people (all single women), homeless accommodation in coastal areas. Also, families and highly vulnerable these areas have seasonal labour migrants. In addition, many causes of markets and have seen employment homelessness in the specific area and household incomes fall due to the were common to the general causes decline of the tourism industry of overall homelessness described in (Fitzpatrick et al., 2000). section 4. However, there are likely to The above evidence suggests that the be some important local factors (e.g. availability of affordable housing is the some areas may have higher migrant main driver of homelessness in rural populations). areas, while the labour market conditions in those areas (e.g. Section highlights seasonal nature of employment, low- wages) further contribute to the The end of an assured shorthold problem. tenancy as a reason for homelessness has increased dramatically in recent Local research years. Again, from a policy perspective more information would be helpful – is We saw unpublished research from a the issue growing because of a LA who commissioned a qualitative general increase in private rentals in research of the causes of the UK or is it a change in terms of homelessness in that area. The study rental accommodation (e.g. increase in was based on interviews with 24 rent, reduction in support)? individuals. The key findings are that the balance of factors at play in The strength of evidence about rural causing homelessness varied greatly and coastal areas is fairly weak – we between different homeless groups. did not include many papers in the The structural factor of the housing review and those we found were fairly market was important in all cases, dated. However, the research did find while personal and interpersonal that some characteristics of rural areas factors played a much greater role in can contribute to causes of some cases (i.e. young women all homelessness through structural attributed their homelessness to factors such as limited housing stock and seasonal low-paid employment.

Page 37 of 59 Homelessness | Rapid Evidence Assessment

International evidence in causes of homelessness for older As well as examining the causes of people in UK and US, notably marital homelessness for different areas within breakdown, job termination, death of a the UK, we briefly considered the spouse and eviction for rent arrears. causes of homelessness in comparable developing countries, such as Canada and the US. Regarding international comparisons of causes of homelessness, Fitzpatrick and Stephens (2007) note that caution is required. Comparing homelessness across countries is complicated due to variation in research traditions and ideological assumptions, as well as actual differences in homelessness. They go on to discuss the differences in the causes of homelessness in England, Canada and the US. The authors find that in all three countries there seems to be a consensus that structural factors – especially the shortage of affordable housing and cutbacks in social programmes – are the fundamental drivers of the overall scale of homelessness, while personal problems and ‘trigger’ events – such as relationship breakdown, mental health problems, and substance misuse – increase an individual’s vulnerability to homelessness. They also mention that in England and Canada, housing affordability is particularly emphasised in the more prosperous regions (i.e. London and the South of England). Housing affordability in the US seems to be a more national-level concern.13 Warnes and Crane (2006) also find similarities

13 13 See See also also Hanratty Hanratty (2017) (2017).

Page 38 of 59 Homelessness | Rapid Evidence Assessment

be homeless because of structural 7. How do factors, most significantly housing market pressures. We have also seen these drivers limited evidence that many families are more likely to be hidden homeless and not seek out support until they have vary across been homeless for some time. That is not to say that interpersonal factors types of are not important. The most prevalent final trigger for homelessness was homelessness reported to be relationship breakdown, which was more important than eviction/threat of and rough eviction. sleeping? As noted previously, the circumstances in which families We have seen that there are become homeless tend to differ from differences in the causes of those of single homeless individuals homelessness for different types of (Wilson and Barton, 2018). Family individuals. We have also seen that homelessness tends to be an issue local characteristics can lead to primarily related to housing market variations in the acuteness of different pressure and affordability (Pleace et causes. Therefore, if the individual al., 2008). On the other hand, characteristics or geographic locations relationship breakdown plays a differ across groups, we may expect stronger role in the case of single that the drivers of different types of homelessness. The lack of affordable homelessness to vary. On this issue, housing as well as relationship Watts et al. (2015) note that breakdown were identified as key “the structural economic and housing drivers of rough sleeping in Greater market context may be important for London (CHAIN, 2018). Young single some particular groups experiencing homeless people are also more likely homelessness, with personal factors to be rough sleeping (Homeless Link, playing a more minor role, whereas for 2018). other groups in other contexts, Research in England by Reeve (2011) interpersonal and individual factors found that hidden homelessness – a may be more important.” phenomenon that is very difficult to Our overview of the research supports measure given its nature – was highly this statement. To summarise the prevalent: 62% of the 437 single research which is outlined in more homeless people surveyed were detail below, families are more likely to hidden homeless (see also Fitzpatrick

Page 39 of 59 Homelessness | Rapid Evidence Assessment

et al., 2000). Women in particular have 2008).14 Further, Busch-Geertsema reported purposely remaining hidden and Fitzpatrick (2008) point out that while sleeping rough, and as a result reasons relating to individual ‘personal’ are less likely to appear in official problems, including drug, alcohol or rough sleeping counts and estimates mental health problems, were reported (Hutchinson et al., 2014). by a very small share of homeless families.15 The results above tend to be support Family homelessness by research in other European countries. Baptista et al. (2017) Research into homeless families examine family homelessness in 13 EU suggests that structural factors, such countries, including the UK. They as housing market pressures, are conclude that, in most countries most important in causing surveyed, family homelessness was homelessness (see below). There can reported as being more likely to be also be difficulties in labour market caused by structural factors, such as participation driven by childcare costs lack of affordable housing, poverty and (Baptista et al., 2017). the increasing gap between rent levels and welfare benefits. They describe Evidence from England indicates that the group of homeless families as structural factors – especially a most likely to be single female parents, shortage of affordable housing – are often victims of domestic abuse and the underlying drivers of the overall violence, who are frequently excluded scale of homelessness (Busch- economically Geertsema and Fitzpatrick, 2008). However, regarding the immediate ‘triggers’ for homelessness, the data available from a nationally Single homelessness representative survey on statutory homeless families in England revealed There is a range of established triggers that the most important factor was that can lead to homelessness for relationship breakdown (accounting for single people including (Fitzpatrick et 38% of homeless families), followed by al., 2000): eviction or threatened eviction (affecting 26% of all homeless • Leaving the parental home after families), overcrowding and arguments overstaying welcome (Pleace et al., • Marital/relationship breakdown

1414 15 However,However, PleacePleace etet al.al. (2008)(2008) mentionedmentioned thatthat the last two reasonsNotice that often these seemed two studies to reflect are a fairlybreakdown dated in informalthe last twoarrangements reasons often entered seemed into toafter reflect losing a settled accommodation.and hence it is possible that reasons for which . 15breakdown in informal arrangements entered into families apply as homeless could have changed Notice that these two studies are fairly dated and hence it is possible that reasons for which families apply after losing settled accommodation. as homeless could have changed.

Page 40 of 59 Homelessness | Rapid Evidence Assessment

• Widowhood a split from a partner, no longer being • Discharge from armed forces able to stay with friends or extended • Leaving care family, overcrowding and tenancy • Leaving prison ending. • Sharp deterioration in mental heath • Increase in substance misuse • Financial crisis/mounting debt Rough Sleeping • Eviction People who are sleeping rough have With respect to single homeless often spent time as hidden homeless people in England, Fitzpatrick et al. (McDonagh, 2011) and exhausted (2000) state that the available (mainly their options. qualitative) research, suggests that relationship breakdown is the Among the main causes of why predominant factor. This is supported people are sleeping rough in Greater by Busch-Geertsema and Fitzpatrick London include eviction, (2008) who report that a major survey unemployment, relationship of single homeless people in England breakdown, end of stay (in – found in Anderson et al. (1993) – accommodation or institution) and showed that family/relationship violence, harassment or abuse (see difficulties was the most common section 5 for more details). reason given by respondents for leaving their last settled home. Furthermore, Fountain et al. (2003) found that drug and alcohol users who had experienced rough sleeping reported other reasons for falling into Young people homelessness, such as problems with parents (58%) or partners (34%), In order to examine young peoples’ financial challenges (49%), mental experience of rough sleeping and sofa health issues (21%) and problems with surfing, Clarke (2016) used an online the police (44%). Homeless Link survey of 2,011 people aged16-25 in (2018) also note that a lack of the UK. The survey showed that the affordable housing and emergency main reasons for rough sleeping and accommodation are key drivers of sofa surfing were the negative home rough sleeping and youth environment and young people being homelessness (illustrating the overlap asked to leave by their parents. Pleace between the two groups). et al. (2008) concluded that for 16-17 year olds relationship breakdown with parents or step-parents was by far the most important reason for applying as homeless. Clarke (2016) also found that other important reasons included

Page 41 of 59 Homelessness | Rapid Evidence Assessment

Fitzpatrick et al. (2011) – summarised in McDonagh (2011) – identified four phases associated with multiple exclusion homelessness, a subgroup that forms part of the rough sleeping population: Stage 1 – substance misuse - the earliest experiences were around alcohol, drugs or solvents abuse. Leaving home or care was also part of this initial stage. Stage 2 – transition to street lifestyles – experiences that indicated worsening problems were often the next phase such as survival shoplifting, survival sex work, spending time in prison, anxiety or depression. Stage 3 – confirmed street lifestyles – these experiences occurred in the middle-late phase of individual journeys and included sleeping rough and begging, intravenous drug use. Being admitted to hospital with mental health issues, bankruptcy and getting divorced occurred at this stage. Stage 4 – “official” homelessness – this set of experiences occurred late in an individual’s multiple exclusion homelessness sequence – applying to councils as homeless, staying in temporary accommodation, death of a partner or being evicted/repossessed.

Page 42 of 59 Homelessness | Rapid Evidence Assessment

higher risk of homelessness such as 8. What are care leavers, young offenders, individuals with childhood experiences the pathways of abuse and/or neglect, poor education experiences and mental health problems. We have seen in in and out of earlier sections that many of these factors interact, which points to homelessness possible pathways into homelessness including some combination of these in the UK? factors. We note that each factor or combinations of factors is not necessary or sufficient to cause homelessness – and many Pathways into experiencing them will never be homelessness homeless. Research outlined previously shows While we have seen in previous that the experiences of single mothers sections there is a trigger event that have some similar features, they often causes a person to become experience domestic abuse and homeless, there are many, often inter- violence, are economically excluded related, reasons that have led to that (i.e. cannot work due to high childcare point. For example, many homeless costs) and tend to engage with people when questioned about why services later on in their homeless they are homeless will say they were journey (see section 4 for more on evicted. While the cause of women’s homelessness). homelessness was eviction, the question of interest is what led to that The difference between thinking about eviction? There could be a number of causes and pathways is the time interacting stages in someone’s life dimension. Pathways will explicitly try experience that led to that eviction to map individuals’ experiences to tell (e.g. low financial resilience caused by the story of all the experiences that led poor education outcomes linked to to homelessness. Mackie and Thomas abuse in childhood). (2014) explore the experiences of single homeless people across Great The pathway to homelessness for Britain by adopting a three-stage, each individual is unique. However, multimethod design. The initial phase research does indicate some similar explored LA implementation of features in the pathways of some subgroups. To mention an example, Watts et al. (2015) note that certain groups are at

Page 43 of 59 Homelessness | Rapid Evidence Assessment

statutory homelessness duties.16 due to a violent or non-violent dispute Stage two of the study sought to decreases. On the other hand, the examine experiences and perceptions percentage of those who become of homelessness services, both from homeless after leaving an institution the perspectives of service users and increases. providers. In total, 480 single homeless people completed a The study further shows that 10% of questionnaire across 16 LAs and 14 respondents never lived in permanent individuals completed a telephone housing during their adult life and interview. Finally, stage three of the almost 80% had slept rough. The study used in-depth interviews with 30 vulnerability of young people is single homeless people in order to prominent: a quarter of them have explore their homelessness pathways. never lived in permanent housing. Mackie and Thomas (2014) found that Nearly 75% of people experienced homelessness began at an early age, more than one period of often due to a relationship breakdown homelessness and more than 50% at home. Many people then faced a had faced three or more experiences. vicious cycle of recurrent In addition, about one-third of people homelessness, with most having first became homeless in a different experienced rough sleeping. The UK local authority to the one where earlier a person becomes homeless, they most recently faced the greater the chance of repeatedly homelessness. facing homelessness. Petersen and Parsell (2014) group the Nearly 50% of respondents became pathways into homelessness for older homeless when they were 20 or people in Australia. A framework younger. In addition, 44% of them first proposed includes i) those who have became homeless when they left their been homeless for many years (called parental/family home, 21% exiting the long term) and ii) those who become social rented sector and 11% leaving homeless later in life (first time). The the private rented sector. The main authors find that homelessness in reasons why people left their older women is more likely to come accommodation during their first from a family crisis such as domestic episode of homelessness were: a non- abuse and violence, separation or violent dispute (41%), a violent dispute widowhood, compared to older males (19%), being given notice by a landlord for whom work-related challenges are (15%), and discharge from an usually the main reason (e.g. loss of institution (12%). After the first employment). experience of homelessness, the Jones and Petersen (2014) identify five percentage of people leaving housing topics highlighted by the research

1616 They had data from 207 local authorities (51%).

Page 44 of 59 Homelessness | Rapid Evidence Assessment

characterising pathways into older-age could be required to obtain a deeper homelessness in Australia: i) structural Busch-Geertsema and Fitzpatrick disadvantage experienced during the (2008) study the downward trend of course of life, ii) vulnerabilities homelessness which was at that time associated with older age (e.g. poor or observed in England and Germany. declining physical and/or mental health The authors claimed that targeted and cognition, reduced income etc.) preventative interventions were to are directly related to homelessness some extent responsible for the for many elderly people iii) the frequent decline. occurrence of triggers or ‘critical incidents’ that resulted in actual or Concerning England, Busch- imminent homelessness (e.g. falling Geertsema and Fitzpatrick (2008) behind with the rent or unaffordable report that the central government has rent increase, breakdown of a family greatly increased the emphasis on relationship, and death of a partner). 17 preventative approaches since 2002. The authors also report that some The Homeless Act (2002) placed a studies emphasised the role of public statutory duty on LAs to produce a policy and aspects of the housing strategy for preventing and alleviating system in shaping pathways into homelessness. Also, substantial homelessness for older Australians – funding – dependent on LAs see for example Batterham et al. committing to goals related to (2013) and Morris et al. (2005). homelessness – was provided to support local preventative activities. The authors also noted the national target to cut the number of statutorily Pathways out of homeless households in temporary homelessness housing in half from 2004 to 2010. This new focus on preventing homelessness seems to have been Targeted preventative and alleviative widely welcomed by LAs. interventions, in England at least, are found to contribute to getting people As noted in Pawson et al. (2007) and out of homelessness. This finding Busch-Geertsema and Fitzpatrick however, does not seem to be backed (2008), common preventive measures up by international evidence, possibly taken by LAs, include ‘enhanced’ due to the small number of studies housing advice, rent deposit and and the heterogeneity of interventions, related schemes, family mediation, methods and outcome measures. domestic abuse and violence victim Further and more systematic research

1717 leading to overcrowding, a sudden deterioration These These critical critical triggers triggers may may also also include include receiving a notice to vacate the premises, a worsening disability in the quality of the dwelling, and a violent receivingmaking an a noticeaccommodation to vacate theinaccessible, premises, a change in family circumstances leading to overcrowding, a episode in shared housing. worseningsudden deterioration disability making in the qualityan accommodation of the dwelling, and a violent episode in shared housing. inaccessible, a change in family circumstances

Page 45 of 59 Homelessness | Rapid Evidence Assessment

support and tenancy sustainment. programmes. However, a lesson from England in particular, revealed that Busch-Geertsema and Fitzpatrick attention should be paid to any (2008) also attempted to assess the perverse incentives generated by contribution of targeted preventive prevention programmes. In addition, interventions to reducing experience from both countries further homelessness in England and suggested that legal duties to provide Germany and reflected on the lessons 18 temporary housing for homeless of their analysis. households can be a key policy driver A noteworthy result of this paper is for improved preventive interventions. that homelessness can be reduced by The expense and political targeted prevention policies and that embarrassment of having many positive results can be achieved even households in temporary housing, has in the face of unfavourable structural acted as an important stimulus to find trends. Successful prevention policies more proactive ways of preventing at secondary and tertiary levels19 homelessness. should be carefully targeted at the key Altena et al. (2010) presented an homelessness causes, which may international review of quantitative differ to some extent between studies for effective interventions for countries. Relationship breakdown homeless young people. Only 11 and eviction often seem to be the studies published between 1985 and most prominent causes targeted – see 2008 were identified for evaluation for instance Fitzpatrick and Stephens according to predefined criteria. Four (2007) and Pleace et al. (2008). The of these studies were of fair quality, authors also highlight that such while the rest were poorly rated. No policies must be supported by study received a ranking of good. appropriate resources (e.g. the case of Almost all intervention studies were England) and have an effective developed and conducted in the US, governance framework for and two were carried out in Canada implementation (e.g. the case of and South Korea. Germany). A strong steer by central government/umbrella organisations The authors reported that there is no can also be helpful. compelling evidence that specific interventions are effective for the Evidence from both countries showed homeless, due to the poor study that local administrations may have a quality and the small number of positive attitude towards prevention

19 1818 TheThe paperpaper alsoalso uunderlinesnderlines thethe profoundprofound impactimpact of transnational Secondary institutional prevention and concerns conceptual interventions differences focused on people at high potential risk of homelessness ofon transnational understanding institutional homelessness and conceptual and preventing it, warningbecause of the dangersof their characteristics of international or in comparisonscrisis situations that differencesdo not pay onenough understanding attention tohomelessness national contexts. and which are likely to lead to homelessness in the near preventing19 it, warning of the dangers of Secondary prevention concerns interventions focused onfuture. people Tertiary at high prevention potential refers risk to of measure homelessnesss targeted international comparisons that do not pay enough at people who have already been affected by because of their characteristics or in crisis situations which are likely to lead to homelessness in the near attention to national contexts. homelessness. future. Tertiary prevention refers to measures targeted at people who have already been affected by homelessness.

Page 46 of 59

19 Secondary prevention concerns interventions focused on people at high potential risk of homelessness Homelessness | Rapid Evidence Assessment

intervention studies. The conclusions that can be drawn from the studies are limited by the heterogeneity of the interventions, participants, methods and outcome measures. Many interventions focused on reducing substance abuse, while other important outcomes, such as quality of life, received little attention. The most convincing, but still marginal, were results of interventions based on approaches of cognitive behaviour, which revealed some positive effects on psychological measures.

Page 47 of 59 Homelessness | Rapid Evidence Assessment

9.Conclusion We have seen that there is a large most important. evidence base on the causes of homelessness. Generally speaking, This review has summarised the the quality of evidence is medium-high evidence on causes and discussed in terms of understanding causes in gaps with respect to an evidence base broad terms. The quality of evidence for designing prevention policies. on the causes of homelessness was assessed on the basis of understanding causes generally and not made in relation to using the research for model development specifically. One issue encountered investigating the causes of homelessness is that identifying a cause will depend on the perspective of the researcher. For example, some will want to understand the final trigger that led to someone becoming homeless, while others will want to think about issues from an individual’s past that may have increased the risk of them becoming homeless (e.g. low education outcomes). It is important when asking about causes to be clear about the intended purpose for understanding them. If one is interested in predicting homelessness, then it is important to identify leading indicators that can be used as early warning signals of changes in homelessness – but this will not require understanding all structural factors that can cause homelessness. However, if one’s focus is around designing prevention policies, then information that establishes causal factors that can be changed by policy levers will be the

Page 48 of 59 Homelessness | Rapid Evidence Assessment

Cauce, A.M., 2002. Challenges faced by homeless sexual minorities: comparison of gay, References lesbian, bisexual, and transgender homeless adolescents with their heterosexual counterparts. Altena, A.M., Brilleslijper-Kater, S.N., Wolf, American Journal of Public Health 92, 773–777. J.R.L.M., 2010. Effective interventions for https://doi.org/10.2105/AJPH.92.5.773 homeless youth. American Journal of Preventive Medicine 38, 637–645. Crane, M., Byrne, K., Fu, R., Lipmann, B., https://doi.org/10.1016/j.amepre.2010.02.017 Mirabelli, F., Rota-Bartelink, A., Ryan, M., Shea, R., Watt, H., Warnes, A.M., 2005. The causes of Anderson, I., Kemp, P., Quilgars, D., 1993. Single homelessness in later life: findings from a 3- homeless people. HMSO, London. nation study. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences Baptista, I., Benjaminsen, L., Pleace, N., Busch- 60, 152–159. Geertsema, V., 2017. Family homelessness in https://doi.org/10.1093/geronb/60.3.S152 Europe: 7 EOH comparative studies in homeless. Crane, M., Warnes, A.M., 2012. Homeless Batterham, D., Yates, E., Mallett, S., Kolar, V., people: older people, in: International Westmore, T., 2013. Ageing out of place? The Encyclopedia of Housing and Home. Elsevier, impact of gender and location on older Victorians San Diego. in homelessness: a pilot study. Hanover Welfare Services, Melbourne. Cull, M., Platzer, H., Balloch, S., 2006. Out on my own: understanding the experiences and needs Bramley, G., Fitzpatrick, S., 2017. Homelessness of homeless lesbian, gay, bisexual and in the UK: who is most at risk? Housing Studies transgender youth. 33, 96–116. https://doi.org/10.1080/02673037.2017.1344957 Department for Communities and Local Government [DCLG], 2012. Making every contact Bramley, G., Fitzpatrick, S., unpublished. The count - A joint approach to preventing social distribution of homelessness: impacts of homelessness. labour markets, housing markets and poverty in the UK. Department for Environment, Food and Rural Affairs [DEFRA], 2018. Statistical digest of rural Bramley, G., Fitzpatrick, S., Edwards, J., Ford, D., England - September 2018. Johnsen, S., Sosenko, F., Watkins, D., 2015. Hard Edges: Mapping severe and multiple Drake, M., O’Brien, M., Biebuyck, T., 1982. Single disadvantage in England. Lankelly Chase and homeless. HMSO, London. Foundation, England. Durso, L., Gates, G., 2012. Serving our youth: Busch-Geertsema, V., Edgar, W., O’Sullivan, E., findings from a national survey of service Pleace, N., 2010. Homelessness and homeless providers working with LGBT who are homeless policies in Europe: lessons from research. or at risk of becoming homeless. Williams European Consensus Conference on Institute, Los Angeles. Homelessness. Enders-Dragasser, U., 2010. Women and Busch-Geertsema, V., Fitzpatrick, S., 2008. homelessness in Germany. Homelessness in Effective homelessness prevention? Explaining Europe, Spring, 12–14. reductions in homelessness in Germany and England. European Journal of Homelessness 2, Fitchen, J.M., 1992. On the edge of 69–95. homelessness: rural poverty and housing insecurity. Rural Sociology 57, 173–193. CHAIN, 2018. CHAIN annual report - Greater https://doi.org/10.1111/j.1549- London: April 2017 - March 2018. 0831.1992.tb00462.x

Clarke, A., 2016. The prevalence of rough Fitzpatrick, S., 2005. Explaining homelessness: a sleeping and sofa surfing amongst young people critical realist perspective. Housing, Theory and in the UK. Social Inclusion 4, 60–72. Society 22, 1–17. https://doi.org/10.17645/si.v4i4.597 https://doi.org/10.1080/14036090510034563

Cochran, B.N., Stewart, A.J., Ginzler, J.A., Fitzpatrick, S., Johnsen, S., White, M., 2011.

Page 49 of 59 Homelessness | Rapid Evidence Assessment

Multiple exclusion homelessness in the UK: key Sydney. patterns and intersections. Social Policy and Society 10, 501–512. Large, J., Kliger, B., 2013. Ageing and women’s https://doi.org/10.1017/S147474641100025X homelessness: overcoming the bag lady syndrome. Women’s Property Initiatives, Fitzpatrick, S., Kemp, P., Klinker, S., 2000. Single Melbourne. homelessness: an overview of research in Britain. Policy Press, Bristol. Loopstra, R., Reeves, A., Barr, B., Taylor- Robinson, D., McKee, M., Stuckler, D., 2016. The Fitzpatrick, S., Pawson, H., Bramley, G., Wilcox, impact of economic downturns and budget cuts S., Watts, B., 2013. The homelessness monitor: on homelessness claim rates across 323 local England 2013. Crisis, London. authorities in England, 2004–12. Journal of Public Health 38, 417–425. Fitzpatrick, S., Pawson, H., Bramley, G., Wilcox, https://doi.org/10.1093/pubmed/fdv126 S., Watts, B., Wood, J., 2018. The homelessness monitor: England 2018. Crisis, London. Mackie, P., Thomas, I., 2014. Nations apart? Experiences of single homeless people across Fitzpatrick, S., Stephens, M., 2007. An Great Britain. Crisis, London. international review of homelessness and social housing policy. Department for Communities and May, J., Cloke, P., Johnsen, S., 2007. Alternative Local Government, London. cartographies of homelessness: rendering visible British women’s experiences of ‘visible’ Fountain, J., Howes, S., Marsden, J., Taylor, C., homelessness. Gender, Place & Culture 14, 121– Strang, J., 2003. Drug and alcohol Use and the 140. link with homelessness: results from a survey of https://doi.org/10.1080/09663690701213677 homeless people in London. Addiction Research & Theory 11, 245–256. McDonagh, T., 2011. Tackling homelessness and https://doi.org/10.1080/1606635031000135631 exclusion: understanding complex lives. Joseph Rowntree Foundation., York. Gough, D., 2007. Weight of evidence: a framework for the appraisal of the quality and McFerran, L., 2010. It could be you: female, relevance of evidence. Research Papers in single, older and homeless. Homelessness NSW, Education 22, 213–228. Woolloomooloo. https://doi.org/10.1080/02671520701296189 Morris, A., Judd, B., Kavanagh, K., 2005. Hanratty, M., 2017. Do local economic conditions Marginality amidst plenty: pathways into affect homelessness? Impact of area housing homelessness for older Australians. Australian market factors, unemployment, and poverty on Journal of Social Issues 40, 241–251. community homeless rates. Housing Policy https://doi.org/10.1002/j.1839- Debate 27, 640–655. 4655.2005.tb00969.x https://doi.org/10.1080/10511482.2017.1282885 National Audit Office, 2017. Homelessness. Homeless Link, 2018. Young and homeless. Department for Communities and Local Government, London. Homeless Link, 2017. Single homelessness support in England: annual review. Neale, J., 1997. Theorising homelessness: contemporary sociological and feminist Homeless Link, 2015. Young and homeless. perspectives, in: Homelessness and Social Policy. Routledge, London. Homeless Link, 2014. Young and homeless. Pawson, H., Netto, G., Jones, C., Wager, F., Homelessness Australia, 2016. Homelessness Fancy, C., Lomax, D., 2007. Evaluating and older people. homelessness prevention. Department for Communities and Local Government, London. Jones, A., 1999. Out of sight, out of mind?: women’s experience of single homelessness. Petersen, M., Parsell, C., 2014. Older women’s Crisis, London. pathways out of homelessness in Australia. Mercy Foundation, Sydney. Jones, A., Petersen, M., 2014. Older people, in: Homelessness in Australia. NewSouth Publishing, Pleace, N., 2000. The new consensus, the old

Page 50 of 59 Homelessness | Rapid Evidence Assessment

consensus and the provision of services for people sleeping rough. Housing Studies 15, 581– 594. https://doi.org/10.1080/02673030050081113

Pleace, N., 1998. Single homelessness as social exclusion: the unique and the extreme. Social Policy & Administration 32, 46–59. https://doi.org/10.1111/1467-9515.00085

Pleace, N., Great Britain, Department for Communities and Local Government, 2008. Statutory homelessness in England: the experience of families and 16-17 year olds. Department for Communities and Local Government, London.

Quilgars, D., Johnsen, S., Pleace, N., 2008. Youth homelessness in the UK: a decade of progress? Joseph Rowntree Foundation, York.

Robinson, D., Coward, S., 2003. Hidden homelessness: your place, not mine. Crisis/The Countryside Agency, London.

Scottish Government, 2015. Youth homelessness ad hoc analysis.

Snelling, C., 2017. Rethinking homelessness in rural communities. Institute for Public Policy Research, London.

Stones, R., 2001. Refusing the Realism— Structuration Divide. European Journal of Social Theory 4, 177–197. https://doi.org/10.1177/13684310122225064

The Albert Kennedy Trust [AKT], 2015. LGBT youth homelessness: a UK national scoping of cause, prevalence, response and outcome.

TUC, 2016. Housing, homelessness and young LGBT people.

Warnes, A., Crane, M., 2006. The causes of homelessness among older people in England. Housing Studies 21, 401–421. https://doi.org/10.1080/02673030600586027

Watts, B., Johnsen, S., Sosenko, F., 2015. Youth homelessness in the UK: a review for the OVO Foundation. Heriot-Watt University, Edinburgh.

Wilson, W., Barton, C., 2018. Statutory homelessness in England (briefing paper No. 01164). House of the Commons Library.

Youth Homelessness North East [YHNE], 2014. Youth Homelessness in the North East: survey findings. Newcastle.

Page 51 of 59 Homelessness | Rapid Evidence Assessment

Appendix A

Author(s) Year Title Weight of Evidence

A B C D

General Specific Specific Overall judgement on judgement judgement judgement coherence on forms of on and integrity evidence relevance

Altena, A.M., 2010 Effective Medium Low Low Low Brilleslijper- interventions for Kater, S.N., homeless youth Wolf, J.R.L.M.

Baptista, I., 2017 Family High High Medium High Benjaminsen, homelessness in L., Pleace, N., Europe: 7 EOH Busch- comparative studies Geertsema, V. in homeless

Batterham, D., 2013 Ageing out of place? High Medium Medium Medium Yates, E., The impact of gender Mallett, S., and location on older Kolar, V., Victorians in Westmore, T. homelessness: a pilot study

Bramley, G., 2017 Homelessness in the High High High High Fitzpatrick, S. UK: who is most at risk?

Bramley, G., 2015 Hard Edges: High High High High Fitzpatrick, S., Mapping severe and Edwards, J., multiple Ford, D., disadvantage in Johnsen, S., England Sosenko, F., Watkins, D.

Busch- 2010 Homelessness and High High Low High Geertsema, V., homeless policies in Edgar, W., Europe: lessons from O’Sullivan, E., research Pleace, N.

Busch- 2008 Effective homeless- Medium High Low Medium Geertsema, V., ness prevention? Fitzpatrick, S. Explaining reductions in homelessness in Germany and England

CHAIN 2018 CHAIN annual report High High High High - Greater London:

Page 52 of 59 Homelessness | Rapid Evidence Assessment

April 2017 - March 2018

Clarke, A. 2016 The prevalence of High High High High rough sleeping and sofa surfing amongst young people in the UK

Cochran, B.N., 2002 Challenges faced by Medium Medium Medium Medium Stewart, A.J., homeless sexual Ginzler, J.A., minorities: Cauce, A.M. comparison of gay, lesbian, bisexual, and transgender homeless adolescents with their heterosexual counterparts

Crane, M., 2005 The causes of High High High High Byrne, K., Fu, homelessness in R., Lipmann, later life: findings B., Mirabelli, from a 3-nation study F., Rota- Bartelink, A., Ryan, M., Shea, R., Watt, H., Warnes, A.M.

Crane, M., 2012 Homeless people: High High High High Warnes, A.M. older people

Cull, M., 2006 Out on my own: Medium Medium Medium Medium Platzer, H., understanding the Balloch, S. experiences and needs of homeless lesbian, gay, bisexual and transgender youth

DCLG 2012 Making every contact High High High High count - A joint approach to preventing homelessness

DEFRA 2018 Statistical digest of High High Medium High rural England - September 2018

Durso, L., 2012 Serving our youth: Medium Low Medium Medium Gates, G. findings from a national survey of service providers working with LGBT who are homeless or

Page 53 of 59 Homelessness | Rapid Evidence Assessment

at risk of becoming homeless

Enders- 2010 Women and Low Low Medium Medium Dragasser, U. homelessness in Germany

Fitchen, J.M. 1992 On the edge of Low Low Low Low homelessness: rural poverty and housing insecurity

Fitzpatrick, S. 2005 Explaining High High Medium High homelessness: a critical realist perspective

Fitzpatrick, S., 2011 Multiple exclusion High High High High Johnsen, S., homelessness in the White, M. UK: key patterns and intersections

Fitzpatrick, S., 2000 Single High High High High Kemp, P., homelessness: an Klinker, S. overview of research in Britain

Fitzpatrick, S., 2013 The homelessness High High High High Pawson, H., monitor: England Bramley, G., 2013 Wilcox, S., Watts, B.

Fitzpatrick, S., 2018 The homelessness High High High High Pawson, H., monitor: England Bramley, G., 2018 Wilcox, S., Watts, B., Wood, J.

Fitzpatrick, S., 2007 An international High High Low Medium Stephens, M. review of homelessness and social housing policy

Fountain, J., 2003 Drug and alcohol High High High High Howes, S., Use and the link with Marsden, J., homelessness: Taylor, C., results from a survey Strang, J. of homeless people in London

Hanratty, M. 2017 Do local economic High Medium Low Medium conditions affect homelessness? Impact of area housing market

Page 54 of 59 Homelessness | Rapid Evidence Assessment

factors, unemployment, and poverty on community homeless rates

Homeless 2018 Young and homeless High High High High Link

Homeless 2017 Single homelessness High High High High Link support in England: annual review

Homeless 2015 Young and homeless High High High High Link

Homeless 2014 Young and homeless High High High High Link

Homelessness 2016 Homelessness and Medium High Medium Medium Australia older people

Hutchinson, 2014 Rebuilding shattered High High Medium Medium S., Page, A., lives Sample, E.

Jones, A., 2014 Older people Medium High Medium Medium Petersen, M.

Large, J., 2013 Ageing and women’s Low Low Medium Medium Kliger, B. homelessness: overcoming the bag lady syndrome

Loopstra, R, 2015 The impact of High High High High Reeves, A., economic downturns Barr, B., and budget cuts on Taylor- homelessness claim Robinson, D., rates across 323 McKee, M., local authorities in Stuckler, D. England, 2004–12

Mackie, P., 2014 Nations apart? High High Medium Medium Thomas, I. Experiences of single homeless people across Great Britain

May, J., Cloke, 2007 Alternative Medium Low High Medium P., Johnsen, cartographies of S. homelessness: rendering visible British women’s experiences of ‘visible’ homelessness

Page 55 of 59 Homelessness | Rapid Evidence Assessment

McDonagh, T. 2011 Tackling High High Medium Medium homelessness and exclusion: understanding complex lives

McFerran, L. 2010 It could be you: Medium Low Low Low female, single, older and homeless

Morris, A., 2005 Marginality amidst Medium Low Medium Medium Judd, B., plenty: pathways into Kavanagh, K. homelessness for older Australians

National Audit 2017 Homelessness Medium High High High Office

Neale, J. 1997 Theorising Medium Medium Low Medium homelessness: contemporary sociological and feminist perspectives

Pawson, H., 2007 Evaluating High Medium Low Medium Netto, G., homelessness Jones, C., prevention Wager, F., Fancy, C., Lomax, D.

Petersen, M., 2014 Older women’s Medium High Medium Medium Parsell, C. pathways out of homelessness in Australia

Pleace, N. 2000 The new consensus, Medium High High High the old consensus and the provision of services for people sleeping rough

Pleace, N. 1998 Single homelessness Medium Medium High Medium as social exclusion: the unique and the extreme

Pleace, N. 2008 Statutory High High High High homelessness in England: the experience of families and 16-17 year olds

Quilgars, D., 2008 Youth homelessness High High High High Johnsen, S., in the UK: a decade Pleace, N. of progress?

Page 56 of 59 Homelessness | Rapid Evidence Assessment

Reeve, K. 2011 The hidden truth High High High High about homelessness: experiences of single homelessness in England

Robinson, D., 2003 Hidden High High High High Coward, S. homelessness: your place, not mine

Scottish 2015 Youth homelessness High High High High Government ad hoc analysis

Snelling, C. 2017 Rethinking High High High High homelessness in rural communities

AKT 2015 LGBT youth Low Medium High Medium homelessness: a UK national scoping of cause, prevalence, response and outcome

TUC 2016 Housing, Medium Medium Medium Medium homelessness and young LGBT people

Warnes, A., 2006 The causes of High Medium Medium Medium Crane, M. homelessness among older people in England

Watts, B., 2015 Youth homelessness High High High High Johnsen, S., in the UK: a review Sosenko, F. for the OVO Foundation

Wilson, W., 2018 Statutory High High High High Barton, C. homelessness in England (briefing paper No. 01164)

YHNE 2014 Youth Homelessness Medium Medium Medium Medium in the North East: survey findings

Page 57 of 59 Homelessness | Rapid Evidence Assessment

The British Cohort Study on the other Appendix B hand measures different information from the previous two, as it has richer According to the analysis conducted data regarding childhood poverty on the data of the Poverty and Social indicators and individual factors, Exclusion Survey, demographics including teenage and adult life aside, the factor most adding to the experiences and experiences. overall explanatory power of the model appears to be poverty, followed by In this analysis, poverty is most closely health and life events, housing and correlated with homelessness, labour market conditions. followed adult economic situation, teenage experiences, adult family and life events and geography.

British Cohort Study Sequence 1

Block of variables % of explanation

Demographics 1.6

Poverty 52

Geography 6

Teenage experiences 15

Adult economic situation 16

Adult family and life events 10

Source: Bramley and Fitzpatrick (2017)

Poverty and Social Exclusion Survey

Sequence 1 Sequence 2

Block of variables % of explanation Block of variables % of explanation Demographics 22 Demographics 22

Poverty 54 Health and life events 35

Labour market 5 Poverty 30

Housing market 8 Labour market 5

Health and life events 11 Housing market 7

Source: Bramley and Fitzpatrick (2017)

Page 58 of 59 Homelessness | Rapid Evidence Assessment

f

Homelessness Causes of Homelessness and Rough Sleeping

Rapid Evidence Assessment

Page 59 of 59