Hidden homelessness: international evidence review

3. Background

3.1 Homelessness and 'hidden' homelessness

Homelessness can be defined in multiple ways depending on policies and laws of each government. The European Federation of National Organisations Working with the Homeless (FEANTSA) defines homelessness as having three domains:

• physical (having access to an adequate and secure physical living space)
• legal (having legal right to occupy this space)
• and, social (having access to privacy in this space) (Bretherton & Pleace, 2018).

Through the combinations of these domains it recognises four main types of homelessness: rooflessness (sleeping rough), houselessness (living in temporary accommodation, supported accommodation or institutionalised), insecure housing (sofa surfing, at risk of eviction or at risk of harm) and inadequate housing (living in inadequate mobile homes, condemned buildings or over-crowded circumstances) (Johnson, Ribar, & Zhu, 2017).

The homelessness definition used by each government may vary in the weight given to each of these dimensions. This has budgetary, legal, and political implications, and affects what support people are entitled to. Depending on the definitions of homelessness used and methods used to collect data, certain populations may be missing from counts. This is also known as 'hidden homelessness'.

The term 'hidden homelessness' is often used to mean both that someone is homeless but physically hidden from public view and also to describe people who may be experiencing homelessness or are at risk of homelessness, who meet the legal definition of homelessness and have a right to access support, but do not appear in official homelessness statistics. Hidden from view and hidden from the official counts are two separate but related situations, as people experiencing homelessness but hidden from public view can also be missed by certain types of homelessness counts. This evidence review takes both into account.

The method/s used to collect data on the prevalence of homelessness in an area or specific population is guided by the definition of homelessness employed. By prioritising some dimensions over others, counts will inevitably capture some experiences more than others. Because it is not feasible nor cost-effective to collect data on every single aspect of any phenomenon, this can unintentionally lead to gaps in data collection.

Internationally, people experiencing hidden homelessness were present in a variety of circumstances including:

• rough sleeping in less visible sites
• staying in unsafe, overcrowded or insecure accommodation
• sofa surfing (staying with a series of different friends or relatives where this is not reasonable)
• sharing accommodation with another household on a long-term basis because they cannot secure their own home
• staying in or moving between temporary and/or shared accommodation (e.g., hostels)
• staying in refuges
• sleeping in cars, tents or other unsuitable non-residential accommodation
• living in unsafe circumstances- like those in situations of domestic abuse
• or where people do not have a legal right to live in their dwelling, like squatting

Some situations listed above, such as temporarily staying with a friend or relative, are not necessarily problematic, and may be the person's or household's preference. Other situations are more clearly neither suitable nor preferred.

The literature on hidden homelessness points out that people missing from homelessness counts could be people who do not approach their local authority/support services for help; people who approach their local authority/support services but exit the process prematurely; those who find alternative temporary 'solutions' (sofa surfing, rough sleeping in less visible places, etc.) and those who do not see themselves as homeless.

3.2 Homelessness in Scotland

Section 24 of the Housing (Scotland) Act 1987, as amended, defines homelessness for the purposes of the Act as follows. A person is homeless if they have no accommodation in the UK or elsewhere. A person is also homeless if they have accommodation but cannot reasonably occupy it, for example because of a threat of violence. A person is potentially homeless (threatened with homelessness) if it is likely that they will become homeless within two months. A person is intentionally homeless if they deliberately did or failed to do anything which led to the loss of accommodation which it was reasonable for them to continue to occupy (The Scottish Government, 2022c).

Since 2001 there have been multiple laws passed in Scotland that have a direct impact on homelessness policy. The Housing (Scotland) Act 2001 amends the 1987 Act and requires councils to provide a minimum of temporary accommodation, advice and assistance to all applicants assessed as homeless, regardless of whether they have been assessed as being in priority need.

The Homelessness etc. (Scotland) Act 2003 outlined the provision that, by 2012, anyone finding themselves homeless through no fault of their own must be entitled to settled accommodation in a local authority or housing association tenancy or a private rental. As a result, through the Homeless (Abolition of Priority Need Test) (Scotland) Order 2012 from 31 December 2012, all unintentionally homeless households are entitled to settled accommodation.

In recent years, a focus on prevention has been developed to provide support to people to avoid them going through homelessness. In November 2022, the Homeless Persons (Suspension of Referrals between Local Authorities) (Scotland) Order 2022 came into force. It removed the need for a local connection and since, local authorities have had to provide assistance where the person presents and they cannot be referred to other local authorities. This allows people to access support where they choose.

Nevertheless, it is possible that some people might not approach their local authorities when experiencing homelessness. This could be for multiple reasons. Yet, in order to provide the best support and fulfil Scottish Government's commitment to end homelessness, it is necessary to be able to include them in the homelessness statistics to understand who they are and how support can be best provided.

See Annex one for more information on the stages of the application process.

3.2.1 Scotland's homelessness statistics

The Scottish Government's homelessness statistics are based on administrative data generated by local authorities in the course of processing homelessness applications. This data is reported twice a year through the Homelessness in Scotland statistics bulletin and later with a Homelessness in Scotland update six months into the financial year. These publications present data on homelessness trends; circumstances previous to a homelessness episode; households currently in temporary accommodation and the scale of rough sleeping; among other data points. Scotland's homelessness statistics also include measures on sofa surfing and overcrowding, generated by asking applicants the property type from which they became homeless.

Data for the statistical publications is collected using the homelessness application (HL1) and temporary accommodation (HL2 and HL3) data returns by local authorities. These allow local authorities to track households as they go through the homeless system. Each person is given a unique identifying number, so that previous homeless applications can be linked and avoid duplication.

Local authorities also collect case level data through the PREVENT1 form to monitor homelessness prevention through the housing options (provision) guidance, although this is not mandatory. This relates to the support local authorities provide to individuals who are at risk of homelessness to try to prevent the homeless episode from happening when possible.

A limitation of this approach is that data is not collected for any households who are homeless but do not engage with their local authority. The HL1 data return captures presentations but not the overall incidence of homelessness. This means, that despite having a robust method and a comprehensive framework of rights for people experiencing homelessness, official statistics could be unintentionally missing an unknown portion of the homeless population in Scotland.

Another critique of homelessness data collection has been the insufficient inclusion of equalities data, specifically for disability, sexual orientation or trans status. This makes it difficult to capture data on groups like LGBTI people and disabled people approaching their local authorities. Additionally, the HL1 captures high level information around eligibility for assistance based on nationality and immigration status but does not capture information on asylum seeker status specifically. The Scottish Government is currently undertaking a review of their homelessness data collection. This review aims to improve the data collection process conducted by local authorities as part of their delivery of their statutory duties around homelessness and homelessness prevention.

The Homelessness in Scotland: 2021/22 statistical bulletin notes that there were a total of 28,882 households assessed as homeless or threatened with homelessness in Scotland. 67% of these were single person households and 28% contained children; and 85% were of White ethnicity (including White Polish). These households contained a total of 32,592 adults and 14,372 children. 2021/22 figures are higher compared to the 2020/21 but are lower than pre-pandemic levels.

Of the 28,882 households assessed as homeless or threatened with homelessness 99% (28,513) were assessed as unintentionally homeless. The most common reason given for making a homeless application was 'household disputes' (35%) followed by being 'asked to leave' (26%) and 'dispute within household: violent or abusive' (14%). Most households became homeless from a 'family home' (28%), 'from friends and partners' (20%) and from a 'private rented tenancy' (15%) (The Scottish Government, 2022b).

3.2.2 Estimations of hidden homelessness in Scotland

There are several estimations of hidden homelessness levels conducted in the last two decades, some specific to Scotland and some for the wider UK. Due to the differences in housing policies and ways of collecting data on homelessness between Scotland and the other three nations in the UK, care needs to be taken when extrapolating or comparing data. Nevertheless, these estimates can still provide some indications of the potential scale of hidden homelessness for Scotland and who is more likely to experience it.

The Office of National Statistics (ONS) recently published an evidence review of the hidden homelessness data landscape across the four UK nations. This report identifies that women, people from a minority ethnic group and young people are at a higher risk of experiencing hidden homelessness in the UK (Office of National Statistics, 2023). This report also highlights Denmark's biennial homelessness counts (for more on it see section 4.4) and Australia's use of the census to capture information on hidden homelessness as valuable approaches to capturing hidden homelessness in national statistics (for more on it see section 4.2.2).

A topic report by Shelter Scotland indicated that ethnic minorities, migrants, women and people living in rural areas might be at a heightened risk of experiencing hidden forms of homelessness, but note that due to the very nature of the issue it is extremely difficult to measure the scale of it (Husbands, 2018).

An estimation published by CRISIS in 2021 as part of their 'Homeless monitor' series focused on five main circumstances that are grouped under the concept of 'core homelessness' which refers to people considered to be experiencing the most extreme and immediate forms of homelessness and who are effectively homeless at a point in time. It includes people rough sleeping, living in unconventional accommodation (sleeping in cars or tents), hostels, unsuitable temporary accommodation and sofa surfing (Watts, Bramley, Fitzpatrick, Pawson, & Young, 2021). The concept of core homelessness used by CRISIS overlaps with the understanding of both statutory homelessness used in the UK and with the hidden homelessness present on the literature included here, such as sofa surfing and living in unconventional accommodation – understood as unsuitable.

This study pulled data from multiple databases from Scotland, including the HL1 and PREVENT1 returns, to estimate the levels of 'core homelessness' there. In Scotland over the period 2012-2019 core homelessness was estimated at 14,250 homeless households. Based on data from the circumstances prior to a homeless application: 7,970 households were sofa surfing; 3,320 living in hostels; 1,180 in unsuitable temporary accommodation; 900 rough sleeping and 880 in unconventional accommodation (Watts et al., 2021, p. 78). This gives a rough indication of kinds and proportions of the different forms hidden homeless can encompass. However, it is understood by the use of PREVENT1 and HL1 data returns as part of the data sources that some or most of these households later approached their council to apply for housing support. This leaves the question of what were the particular living circumstances of the households that did not approach their local authority- as they would be the ones experiencing hidden homelessness- and if these living circumstances present in the same proportions as for this cohort experiencing core homelessness.

Palmer (2004) produced an estimate of the hidden homeless population in London in the early 2000s. It included multiple circumstances of homelessness or housing insecurity. Although the profile of London's population differs from that of Scotland, this exercise provided a valuable example of the use of multiple secondary data sources to estimate hidden homelessness. Sources used included both administrative and official data sources. This exercise concluded that in London in 2003 there were an estimated 140,000 severely overcrowded households; 15,000 people living in hostels, night shelters and refuges on a non-permanent basis; and 8,000 people squatting involuntarily (Palmer, 2004).

The London Assembly published an estimation in 2017 that there were 13 times more people in concealed homelessness than those visibly sleeping rough in London (London Assembly, 2017). LGBTI youth and those that are not eligible for support were signalled as the most at risk.

All these estimations provide insights and lessons to build from when trying to determine who could be missing from official counts. Each study approached this challenge in different ways, some by using administrative data sources or routinely collected data and others by introducing bespoke data collection tools. The following chapter focuses on evidence regarding methods to enumerate homelessness and hidden homelessness.