Can a basic income help address homelessness? A Titmussian perspective

Abstract

Homelessness is a worsening problem across the developed world and existing policy responses are failing to have an impact. This article considers whether a basic income (BI) can play a role in radically overhauling prevailing homelessness policy and interventions. Drawing on Richard Titmuss’ classical arguments about the value of universalist welfare, I argue that a BI can play a role, but only as part of a suite of universalist measures that includes large-scale social housing investment and rent controls. I highlight how a BI can help address the ‘income side’ of the housing affordability problem driving homelessness, but must be coupled with other measures that address housing cost and supply. I also consider how a BI can reduce stigma arising from targeted homelessness measures. I conclude by arguing that addressing homelessness requires us to transform the logic of welfare provision and that a BI can help do this.

Homelessness is an entrenched problem throughout the developed world. It is estimated that at least 1.9 million people are homeless across the OECD (Organisation for Economic Co-operation and Development), with recent increases reported in one third of member countries (OECD, 2020). This is a significant issue, as homelessness has profound impacts on individuals and communities. People experiencing homelessness die up to ‘up to 30 years earlier than the general population on average’ (OECD, 2020, p. 3). Homelessness also places a significant burden on health, justice, and social service systems, as evidenced by the disproportionately high rates of engagement with (and thus cost to) these systems among people experiencing homelessness (Culhane, 2008).

Importantly, homelessness and its negative effects remain entrenched despite significant intervention from government and non-government actors. In Australia, for instance, homelessness has increased despite growing investment in (predominantly crisis-oriented) specialist homelessness services (Pawson et al., 2018). This indicates that prevailing policy approaches and welfare systems are failing to adequately address the factors driving homelessness, signalling a pressing need to consider radical alternatives to the status quo. This article contributes to addressing this need by critically analysing the capacity of a basic income (BI) to form part of such an alternative approach.

At a time when social policy is dominated by targeted and conditional welfare approaches (Watts & Fitzpatrick, 2018), a BI proposes to provide ‘a periodic cash payment unconditionally delivered to all on an individual basis, without means-test or work requirement’ (BIEN, 2016, n.p.). BI proponents promote it as a remedy to the rising insecurity and inequality plaguing 21st-century capitalist societies, and to the paternalistic and punitive approach of contemporary welfare states to managing this insecurity (Klein et al., 2019). They argue that a BI has the potential to ensure a basic level of economic security and freedom for all, and a more equal distribution of societal wealth (Standing, 2020).

While there is an increasingly voluminous literature on BI per se, there is little consideration given to its possible impacts on homelessness specifically. There is only one study dedicated to the topic known to this author (Kerman, 2021); and while there is often passing reference to homelessness in the BI literature (e.g. Malleson & Calnitsky, 2021), limited attention is given to its structural and policy drivers, and how a BI might ameliorate these. Given that homelessness is one of the most acute manifestations of the insecurity and inequalities that BI proponents hope to address, developing a more in-depth understanding of the affordances of a BI vis-à-vis homelessness, and prevailing responses thereto, is clearly necessary.

While homelessness is a multidimensional phenomenon (Somerville, 2013) that is associated with a broad array of individual problems and challenges (such as mental illness and addiction), it is almost always underpinned by a household's inability to access stable and affordable housing. Homelessness is thus fundamentally a ‘housing problem’ (Colburn & Aldern, 2022), and any attempt to address it must grapple with the structural processes that render housing unaffordable and inaccessible to an increasing number of people (Parsell, 2018). This article will therefore consider how a BI can help address these structural processes.

I will focus specifically on how a BI can help transform responses to homelessness in ‘liberal welfare regimes’ (Esping-Andersen, 1990), like Australia, Canada and the United States. Drawing on the international research literature, I show how homelessness interventions in liberal regimes are highly targeted and preoccupied with the support needs of particular homeless subgroups, leaving the structural drivers of homelessness untouched (Evans & Baker, 2021; Evans et al., 2021). I also present a case study of homelessness in Australia to illustrate these structural drivers and to consider the capacity of a BI in addressing them. Australia is an illustrative case, as its housing system and homelessness policies exemplify the residualising logic of liberal welfarism and its consequences (Clarke et al., 2022); and because of its long history of, and pioneering role in, the use of residual/targeted welfare measures (Saunders, 1999).

Drawing on Titmuss’ (2006 [1968]) classical insights on the optimal relationship between universalist and targeted welfare, I argue that the utility of a BI lies in its capacity to contribute to a suite of universalist welfare measures that can both ameliorate the structural drivers of homelessness and provide a basis upon which more targeted interventions can build. In relation to ameliorating structures, I contend that a BI can help address the income side of the housing affordability crisis driving homelessness; however, its effectiveness will ultimately depend on whether it is coupled with other universalist measures, like increased social housing and rent control, that address the cost and supply side of the problem. I also suggest that it can contribute to ameliorating the stigma attached to targeted homelessness interventions, either by directly replacing them or by decoupling necessary targeted measures from the provision of basic resources.

My approach in this article is to focus on the sociological and theoretical dimensions of the BI–homelessness question. I am therefore more concerned with conceptualising the implications of a BI for the social structures and governmental logics that shape homelessness, and responses thereto, than I am with technical questions about BI design or implementation. This is appropriate given then venue for this special issue (Journal of Sociology) and because the dearth of existing work on homelessness in the BI literature means that laying out the theoretical stakes is a logical first step, one which can lay the groundwork for future technical studies. Technical questions are not wholly avoidable, however, for some (like the size of BI payments) have a qualitative impact on the nature and implications of a BI and thus significant theoretical import (Malleson & Calnitsky, 2021). Technical questions will thus be invoked to the extent that they have non-trivial implications for the conceptual issues discussed in the article.

Responses to homelessness in liberal welfare states

Responses to homelessness in the Anglophone world typically reflect the logic of what Esping-Andersen (1990) called ‘liberal welfare regimes’. These are welfare systems that encourage engagement with the market as the default source of individual welfare and which limit state provision to populations unable to do this (due to poverty, disability, etc.) through means testing, needs-based targeting and ‘less eligibility’ (ensuring state welfare is less attractive than market offerings). They are informed by a belief that engagement with the market represents independence and personal responsibility, whereas state provision promotes dependence and moral degradation.

While liberal welfare states vary in important respects (Deeming, 2017), such as their levels of social housing (Clarke et al, 2022), existing research outlines some important consistencies in their responses to homelessness. As Evans et al. (2021) have recently argued, this liberal logic manifests as ‘reticence’ on the part of the state to intervene to address homelessness beyond what is deemed absolutely necessary to ensure basic survival, lest these interventions encourage dependency and discourage market engagement. Homelessness responses are therefore highly targeted based on calculations about people's type, intensity, and duration of need. The provision of stable housing, typically through residualised social housing systems or ‘housing first’ voucher programs, is limited to a small subgroup of homeless people whose needs are deemed so complex that they are unable to access market housing, and whose ‘chronic’ homelessness places a disproportionate burden on homelessness services and mainstream support systems (e.g. hospitals, police, etc.) (Clarke et al, 2022; Baker & Evans, 2016). By contrast, most people experiencing homelessness are supported through short-term crisis interventions, such as temporary accommodation in shelters and refuges (Coleman & Fopp, 2014; Evans et al., 2021). Here, the provision of support is often coupled with paternalistic interventions that focus on individual behaviour and deficiencies (i.e. addiction, mental illness, employment-related skills, or financial literacy) (Bullen, 2015; Stuart, 2016). In terms of preventative measures, liberal regimes prefer targeted rental subsidies aimed at sustaining market engagement over robust social housing systems or regulated rents, despite the effectiveness of such subsidies being contingent on market fluctuations and their susceptibility to landlord exploitation (Colburn, 2021; Desmond, 2016).

Importantly, these liberal-welfarist approaches construct homelessness as a discrete problem that is unconnected to the broader structures of contemporary capitalist societies (Farrugia & Gerrard, 2016; Willse, 2015). They presuppose that homelessness is associated with the circumstances and characteristics of specific individuals or groups – people with addictions and/or severe mental illnesses, people exiting institutions (prison, psychiatric facilities), women fleeing domestic violence, etc. (Bullen, 2015). This is reflected in the tendency of policy makers and researchers to become preoccupied with developing ever more fine-grained understandings of the needs of different homeless and ‘at-risk’ subgroups, while obscuring the social structures within which these needs emerge (Clarke et al, 2022; Farrugia & Gerrard, 2016). While attention to people's individual vulnerabilities can lead to more caring practice interventions (Clarke et al., 2020; Parr, 2022), it also perpetuates stigmatising representations of the homeless as deficient or pathological individuals (Gowan, 2010; Parsell et al., 2021). This stigma discourages many homeless people from accessing needed services (Parsell et al, 2021), at the same time provides authorities with a justification for coercing service engagement for people's ‘own good’ (Stuart, 2016).

Structural processes are, of course, integral to the production and persistence of homelessness. Recent work from the United States shows that housing market conditions, such as rent levels and vacancy rates, are far better predictors of differences in rates of homelessness between regions than non-housing (typically individual-level) risk factors like mental illness, drug use, and individual-level poverty (Colburn & Aldern, 2022). Labour market conditions also play a role, as do welfare reforms that reduce payments and services (including public housing) or promote conditionality and sanctions (Bramley & Fitzpatrick, 2018; Mitchell, 2011; Parsell et al, 2021; Peterie et al., 2020).

Given that homelessness responses in liberal welfare regimes overlook the role played by structural processes, they have largely failed to make lasting inroads into reducing homelessness, despite producing a seemingly endless stream of policy and practice innovations (Evans & Baker, 2021) and receiving increasing public investment (Pawson et al., 2018). Indeed, even seemingly progressive, evidence-based interventions, like the much celebrated ‘housing first’ approach, have been constrained by the liberal preoccupation with targeting and treating homelessness as a discrete/individualised problem (Evans & Baker, 2021; Willse, 2015). Housing first holds that housing is both a right and a precondition for people experiencing homelessness to address associated issues like mental illness and addiction (Padgett et al., 2016). Housing first programs thus focus on ‘ending homelessness’, rather than simply managing its symptoms and effects, by rapidly rehousing people in long-term, independent housing (instead of crisis shelters or transitional programs) and then provide them with voluntary, client-led support to sustain their tenancy and work on personal issues.

Housing first programs produced impressive outcomes, typically enabling upward of 80% of their clients on average to successfully exit homelessness (Padgett et al., 2016). The problem, however, is that they are typically highly targeted, meaning that these benefits reach only a small proportion of the homeless population (Baker & Evans, 2016). Those targeted tend to be the ‘chronically homeless’: long-term homeless individuals suffering severe mental health issues who make up a fraction of the overall homeless population but who consume a disproportionate amount of homelessness services and other resources (Culhane, 2008). The provision of housing to this group is thus justified based on the ‘cost-savings’ that derive from ending their homelessness and associated service use (Willse, 2015); and on the fact their complex needs limit their capacity for (market-based) independence (Clarke et al, 2022; Parr, 2022).

Housing first is thus typically delivered through discrete, targeted programs (Parkinson & Parsell, 2018) that do little to address the structural drivers of homelessness (Willse, 2015). This has meant that the promises of ending homelessness that accompanied the (sometimes large-scale) roll-out of housing first initiatives in liberal welfare regimes have not been fulfilled (Evans & Baker, 2021). It has also constrained the effectiveness of housing first programs, as they struggle to gain access to stable housing in a timely manner in cities where housing markets are increasingly unaffordable and social housing systems increasingly overburdened (Anderson-Baron & Collins, 2019; Clarke & Parsell, 2020).

The foregoing review of the international homelessness literature raises important questions for consideration of the value of a BI vis-à-vis homelessness. This includes whether a BI can help address the structural drivers of homelessness overlooked by liberal approaches and whether its universal quality can help mitigate the stigma and individualising blame such approaches produce.

Titmuss, universalism and basic income

This article suggests that the value of a BI as a means of addressing homelessness comes in large part from its universal quality and its subsequent potential to overcome the limitations of the liberal approach. While proposals/proponents vary on how universal a BI should/could be, a universalist tendency is central to its rationale (Calnitsky, 2016; Standing, 2017). As Standing (2017) explains, ‘It is often called a universal basic income (UBI) because it is intended to be paid to all.’ For Standing, this universal quality flows directly from the moral justification for a BI: it is a means of ensuring that everyone benefits from social and economic progress, rather than just the powerful few.

The universalism of a BI is also commonly justified on instrumental grounds. For instance, it is seen as a way of guaranteeing that everyone in a society has a basic level of economic security by making sure individuals have an income sufficient to meet their basic needs (Malleson & Calnitsky, 2021; Standing, 2020). It is also seen as a way of overcoming the stigma attached to receipt of welfare in liberal regimes, as payments would be made to everyone, unconditionally, and thus ‘discourage “us and them” divisions and confirm that all of us are of equal worth’ (Standing, 2020, p. 3).

To help make sense of how the universalist aspects of a BI can help overcome the limitations of liberal responses to homelessness, I draw on the classical arguments of Richard Titmuss (2006 [1968]) on the value of universalism in social welfare systems. For Titmuss (2006 [1968]), a predominance of targeted measures in a welfare system is highly problematic as it tends to promote a stigmatising logic of discrimination and deterrence. He argued that: ‘The fundamental objective of all such tests of eligibility is to keep people out’ (Titmuss, 2006 [1968], p. 45). Targeted welfare thus often positions people requiring support as ‘supplicants’, imbuing them with ‘both the sense of personal failure and the stigma of a public burden’ (Titmuss, 2006 [1968], p. 45).

However, Titmuss did not propose that all targeted services be dispensed with. Rather, he argued that some forms of ‘positive discrimination’ are necessary in welfare provision to ensure that the unique needs of certain groups are accounted for. Titmuss recognised that ‘one-size-fits-all’ universalism alone cannot address all the diverse challenges faced by people in increasingly complex modernising societies. But he argued that targeted responses to the unique needs of specific groups can only succeed if they are subtended by a robust suite of universal measures. As he put it:

In all the main spheres of need, some structure of universalism is an essential prerequisite to selective positive discrimination; it provides a general system of values and a sense of community; socially approved agencies for clients, patients and consumers, and also for the recruitment, training and deployment of staff at all levels; it sees welfare, not as a burden, but as complementary and as an instrument of change and, finally, it allows positive discriminatory services to be provided as rights for categories of people and for classes of need in terms of priority social areas and other impersonal classifications. (Titmuss, 2006 [1968], p. 46)

Put simply, universal welfare measures provide the practical, relational, and symbolic infrastructure necessary for more targeted interventions to work in a more effective and less stigmatising way. This point is further elaborated by Spicker (2005), who shows that residualisation and stigmatisation are not necessary outcomes of targeting, but rather depends on the rationale underpinning their use and what other measures they are combined with.

Importantly for my argument here, one way that universal measures help underpin targeted interventions is by ensuring that structural problems are not overlooked. For Titmuss (2006 [1968]), universal measures recognise, and contribute to ameliorating, the structural conditions that create or exacerbate social ‘diswelfare’ (hardship, need, etc.). He saw measures like universal healthcare and social security as recognition of the ‘social costs and social insecurities which are the product of a rapidly changing industrial-urban society’:

They are part of the price we pay to some people for bearing part of the costs of other people's progress; the obsolescence of skills, redundancies, premature retirements, accidents, many categories of disease and handicap, urban blight and slum clearance, smoke pollution, and a hundred-and-one other socially generated disservices. (Titmuss, 2006 [1968], p. 44)

Similar reasoning underpins BI proposals, particularly those that foreground the impact of structural processes like automation, globalisation and deindustrialisation on people's capacity to guarantee income for themselves through work (Klein et al., 2019). The structural processes that render housing unaffordable to an increasing number of people are another example of such ‘socially generated disservices’. Targeted measures alone are typically unable to address structural sources of diswelfare, not least because their focus on testing individual need obscures and ignores the impact of structural processes.

Another way that universal measures help underpin targeted welfare interventions is by contributing to a sense of social solidarity and enhancing the legitimacy of public support per se. They do this by giving a broader range of socio-political constituencies a stake in public welfare, as more people have the possibility of benefiting from social protections, which in turn helps generate a sense that everyone is involved in supporting everyone else. According to Esping-Andersen (1990), it is this process that explains relative popularity and durability of Northern Europe's ‘social democratic’ welfare states, where universal measures predominate. Moreover, according to Bauman (2005), it is precisely the lack of a shared stake in welfare, driven by limited universalism and the dominance of targeting, that helps explain the erosion of support for the welfare state in liberal democracies. Here, measures like means testing have enacted a separation between those who pay for welfare measures and those who benefit from them, leading the former to view the latter as failures and a public burden in precisely the way that Titmuss warned.

As Standing (2020) argues, the universalism of BI offers a means of attenuating these kinds of ‘us and them’ divisions. Empirical studies of BI-like experiments show that broad-based payments are perceived as more ‘normal’ and legitimate than targeted welfare measures, even where these are not completely universal (Calnitsky, 2016). This is because such payments eschew distinctions between deserving and undeserving populations and thus do not involve the invasive and stigmatising procedures used to determine deservingness. If we follow Titmuss’ reasoning, this sense of legitimacy should also extend beyond universal measures to positively discriminating ones, allowing them to be ‘provided as rights for categories of people and for classes of need in terms of priority social areas’ (Titmuss, 2006 [1968], p. 46). This is because the universalism of the BI can contribute to ‘a general system of values and a sense of community’ (Titmuss, 2006 [1968], p. 46) and to a view that all forms of support, including targeted support, are an expression of these values. As Standing (2020, p. 3) puts it, a BI will ‘strengthen social solidarity, including human relations: it would be an expression that we are all part of a national community, sharing the benefits of the national public wealth created over our collective history’. I show below that a BI may also help decouple people's personal problems from their structurally induced disadvantage (e.g. experiences of homelessness), thus enabling them to seek targeted support for the former with less stigma and blame.

However, we should not expect that a BI will fulfil the promises of Titmussian universalism on its own. Many BI proponents recognise that it will be most beneficial if coupled a robust suite of welfare services and regulations (Malleson & Calnitsky, 2021; Standing, 2017, 2020). As Malleson and Calnitsky (2021) have recently shown, the impact and value of a BI on economic security is mediated by the level and type of public services and regulations that are implemented alongside it. This is because a BI, in the absence of such measures, leaves BI recipients to rely on the market to meet their needs; and markets consistently fail to meet many important social needs, including stable housing for low-income households and vulnerable groups like the homeless (Madden & Marcuse, 2016). Indeed, Kerman (2021, p. 8) states in her analysis of the BI/homelessness question that ‘UBI must not be framed as a poverty reduction panacea, nor it is one for preventing and ending homelessness’, noting the need to also address the ‘the lack of investments to build new affordable housing’ among other things. Thus, following Malleson and Calnitsky (2021), I will specify the value of a BI in responding to homelessness, not in isolation, but as part of a ‘hybrid’ that also comprises key services and regulations.

Moreover, while a BI has the potential to enhance social solidarity, it is unlikely to mitigate divisions and stigmas surrounding public support completely. BI policy is necessarily redistributive in nature, meaning that there will always be net recipients and net contributors, the latter of whom may look down on the former. Moreover, it is possible that some recipients may see themselves as more deserving of BI payments (e.g. retirees) than other groups (e.g. the homeless). Therefore, the BI-services-regulation hybrid should not be seen as a technocratic panacea for the problem of stigma. It will need to be coupled with ongoing political advocacy to address the stigmatising cultural myths surrounding poverty and social inequality. Yet, by reducing the administrative reinforcement for these divisions and stigmas, a BI will certainly lighten the load for this advocacy work.

The structural drivers of homelessness

This section will lay out the structural drivers of homelessness, drawing on published evidence from the Australian case to illustrate my claims. As noted earlier, Australia's housing system is emblematic of liberal welfarism in many respects, including the privileging of market provision and the stringent targeting of state support (Clarke et al, 2022; Pawson et al., 2017). This case study will lay the foundation for consideration in the following section of the capacity of a BI to help address these structural issues.

As noted, homelessness is fundamentally a housing problem (Colburn & Aldern, 2022; Parsell, 2018). This does not mean that housing is the only important factor; individual-level poverty, health and social support needs, social capital, and other structural processes like labour market dynamics have all been shown play a role (Batterham, 2017; Bramley & Fitzpatrick, 2018). However, the impact of these other factors is mediated by the availability and accessibility of safe, stable, and affordable housing. Indeed, if the housing first literature teaches us anything, it is that people's need for stable housing must be met before they can be expected to address other (particularly individual-level) issues (Padgett et al., 2016). It also shows that when barriers to housing access are removed, even people with complex support needs can permanently exit homelessness. Put simply, access to affordable housing is a necessary, if not sufficient, condition to enable someone to avoid or exit from homelessness.

Key to addressing homelessness, then, is addressing the housing affordability crisis that is leaving increasing numbers of low-income households unable to access and/or sustain suitable housing (Madden & Marcuse, 2016; Wetzstein, 2017). Australia provides a striking example of this crisis. Those falling in the bottom two quintiles of the income distribution (the standard definition of ‘low income’ in Australia) have seen their housing costs increase from 23% of their income on average in 2011 to 29% in 2018 (Pawson et al., 2020). The proportion of people on low incomes experiencing rental stress (defined as paying more than 30% of income in rent) has increased from 40% to 45% over roughly the same period, and reached 48% in capital cities (Pawson et al., 2020).

There are two sides to the housing affordability problem (Wetzstein, 2017). The first concerns the cost of housing itself and the supply of housing affordable to people on low incomes. Liberal welfare states like Australia have traditionally promoted home ownership (through tax breaks, subsidies, etc.) as the preferred method of helping people meet their housing needs. However, the cost of purchasing a home has increased drastically in recent decades, driven by policies encouraging financialisation and asset-based welfare, meaning home ownership is now out of reach for many people (Adkins et al., 2021). This translates into more higher-income households renting, placing downward pressure on the rental market. Rental vacancy rates are thus at record lows (1% nationally), with the impact felt most by people with low incomes (Anglicare Australia, 2022).

Also contributing is the abandonment of large-scale social housing provision, which had helped ensure access to affordable housing for many lower-income households in the post-war years (Jacobs, 2019). A mixture of under-investment, the sale of stock, demolition, and the intensified targeting of what remains has rendered social housing a highly residualised and stigmatised tenure that is out of reach for all but the neediest households; and even these groups face extensive wait times and no guarantee of accessing a tenancy. In Australia, social housing declined from 6.2% of all dwellings in 1991 to 4.2% in 2018 (Pawson et al., 2020; Productivity Commission, 2019). In this context, housing providers are developing ever more fine-grained methods of targeting social housing at the neediest applicants, engendering a situation where homelessness or risk thereof no longer guarantees timely access to a property (Clarke et al, 2022).

The other side of the affordability crisis is the decline in the incomes of those at the lower end of class structure (Adkins et al., 2021; Wetzstein, 2017). Real wages have been deliberately suppressed because of concerns about runaway inflation through the disempowering of trade unions, the weakening of labour laws and the subsequent casualisation and precaritisation of work (Adkins et al., 2021). A recent analysis of rental affordability in Australia shows that, as of April 2022, only 1.6% of advertised rental vacancies were affordable for a single person earning minimum wage, ¹ and only 0.7% if that person is a single parent (Anglicare Australia, 2022). The situation is somewhat better for a couple earning minimum wage, who can afford 15.3%, although this has declined from 30% in 2012, reflecting ‘the stark drop in affordability over time as wages have stagnated but rents continue to rise’ (Anglicare Australia, 2022, p. 11). The situation is further complicated by the fact that an increasing number of people are in insecure, casualised work, meaning that their incomes may fluctuate and/or that their employment is precarious (Anglicare Australia, 2022).

At the same time, welfare state retrenchment and restructuring has rendered income support payments both inadequate and insecure. In Australia, unemployment benefits are today paid at a rate that is well below the poverty line, following a period of over twenty years where they have not been increased in real terms (Parsell et al, 2021). People reliant on income support payments thus find it increasingly difficult to find affordable rentals within the private market. It is estimated that there are close to 0% (eight in 45,000) of advertised rental vacancies are affordable for a single adult receiving unemployment benefits in 2022, and only slightly more for people receiving the aged pension or disability support payments (Anglicare Australia, 2022). It is unsurprising, then, that the vast majority (81%) of people at risk of homelessness in Australia are reliant on income support payments (Batterham, 2021).

Income support recipients also face intensified welfare conditionality, which renders the incomes of many insecure and intermittent. ‘Conditionality’ refers to a form of welfare targeting where access to benefits is rendered conditional upon behaviour (e.g. attending case management appointments, applying for a certain number of jobs) and where failure to comply is met with sanctions (e.g. payment suspensions) (Watts & Fitzpatrick, 2018). Pawson et al. (2018) report that the imposition of such sanctions have risen considerably in Australia in recent years, following policy changes introducing greater conditionality into the welfare system. Homelessness sector stakeholders interviewed by Pawson et al. (2018) reported that these measures are pushing people into homelessness. They also noted that conditionality measures have the capacity to further entrench the material disadvantage experienced by the homeless, and that homelessness creates barriers to meeting compliance requirements (e.g. difficulty keeping appointments).

Similarly, Peterie et al. (2020) show how Australia's compulsory income management (CIM) schemes put welfare recipients at greater risk of homelessness. These schemes entail the quarantining of welfare payments on government-issued debit cards that can only be used for approved purposes (generally excluding the purchase of alcohol, cigarettes, and gambling). Peterie et al. (2020) show how CIM schemes have created technical and logistical difficulties for paying rent, contributing to late payments and failed transfers that placed tenancies at risk. CIM also exacerbates other factors that contribute to homelessness, including financial deprivation (e.g. by preventing access to second-hand markets and cheap goods from stores not registered for the debit card), poor health (e.g. through increasing stress and making access to some mental health practitioners difficult) and stigma (use of the debit cards identifies people as welfare recipients in public settings).

Conditionality also exists in homelessness service provision, where it inhibits people's ability to access housing and other basic resources (Clarke, Watts et al., 2020; Clarke, Parsell et al., 2020). Most homelessness services (except for housing first) require compliance with behavioural rules, such as engagement with case management/therapeutic programs, actively searching for employment or private rental accommodation, and abstinence from alcohol and drugs. Homelessness service systems also often operate on what's termed the ‘staircase model’ (Padgett et al., 2016), where access to permanent housing is conditional upon successfully transitioning through crisis and then transitional housing programs, paying rent and complying with their rules, and thus demonstrating ‘readiness’ for permanent housing. It is often difficult for people experiencing homelessness to comply with these conditions, especially if they have complex needs, and many thus fail to access resources they need. Part of this difficulty arises from the structure of crisis programs themselves. They typical involve congregate accommodation with shared amenities, meaning they concentrate people with complex health and behavioural problems together, creating environments that people experience as degrading, dangerous, and/or counterproductive to their recovery goals (e.g. abstinence is notoriously difficult in homeless shelters) (Parsell & Clarke, 2019).

Addressing the structural drivers of homelessness

What role can a BI play in addressing these structural drivers? As a guaranteed regular cash payment, a BI has the capacity to address the income side of the housing affordability crisis. BI payments would supplement the incomes of people in low-wage employment, helping mitigate the insidious conjunction of wage stagnation and rising rents. It can also assist people on income support payments to meet their housing costs, although its effectiveness here will be contingent on the level at which it is paid – more so than for low-wage earners. In Australia, a BI would need to be paid at a much higher rate than current unemployment benefits, which are below poverty levels, and to be indexed appropriately to keep pace with rising housing and living costs. As Malleson and Calnitsky (2021) point out, the size of BI payments is a non-trivial matter that conditions both its practical and normative import, and this is clearly the case regarding the question of housing affordability/homelessness.

A BI would also help address the income side of housing affordability by ensuring income security for people at risk of homelessness. As noted above, people on casual employment contracts can often face fluctuating incomes and employment insecurity, affecting their ability to meet their housing costs (Anglicare Australia, 2022). This is significant given that loss of employment is common precursor to homelessness (Bramley & Fitzpatrick, 2018). People on income support payments are also at risk due to the income insecurity created by welfare conditionality regimes and sanctions (Pawson et al., 2018; Peterie et al., 2020). As Kerman (2021) highlights, a BI is an unconditional payment that people receive regardless of their willingness/capacity to meet behavioural conditions, meaning that it would effectively eliminate the risk of homelessness that arises from conditionality and sanctions. It is also intended to be paid to each individual in cash, meaning that it would also circumvent the homelessness risks associated by CIM (Peterie et al., 2020).

Notwithstanding these affordances, a BI alone does not address the housing cost and supply side of the housing affordability problem. Indeed, implementation of a BI in isolation would render its impact on homelessness contingent upon the state of housing markets (cf. Malleson & Calnitsky, 2021). This means that gains achieved by a BI on the income side could easily be undone by rising market-housing costs; and such cost rises are likely, as the growing concentration of home ownership pushes more higher-income earners into the private rental market, driving up rents and making it increasingly difficult for low-income households to compete (Adkins et al., 2021; Anglicare Australia, 2022). A BI may even drive-up housing costs at the lower end of the market by increasing the capacity of low-income households to bid up rents on the small number of affordable properties.

Hence, for a BI to be effective in addressing homelessness, it must be implemented as part of a ‘BI-services-regulation hybrid’ (Malleson & Calnitsky, 2021). Central to the ‘services’ component of this hybrid would be an expanded social housing sector to boost the supply of affordable housing and reduce competition for low-cost private rentals. Social housing provides far greater levels of housing stability and affordability than the private rental market and has been shown to prevent homelessness for disadvantaged households (O’Donnell, 2021). However, the increasingly residualised nature of the social housing sector in liberal/liberalising welfare states means that these benefits are only available to an increasingly narrow group of high-needs households (Clarke et al, 2022). In Australia, this situation means that many people experiencing or at risk of homelessness are either unable to access social housing or must wait for extended periods – sometimes years – to do so (Clarke et al, 2022; Clarke & Parsell, 2020). Expanding the supply and loosening the targeting of social housing is thus essential to addressing homelessness – and the closer it can become to a universal (i.e. untargeted) service, the greater the impact it will have.

Importantly, pairing a BI with existing forms of targeted housing assistance, such as rental subsidies and housing vouchers, is unlikely to achieve the same gains as pairing it with social housing. Subsidies and vouchers tend to be deployed in lieu of efforts to address the cost and supply of affordable housing, and thus suffer similar problems to a stand-alone BI. That is, they presume the efficacy of the market in meeting the needs of at-risk groups, meaning their capacity to address homelessness is contingent on market conditions. Anderson-Baron and Collins (2019) showed that the use of housing vouchers by Canadian housing first programs was rendered ineffective when housing markets tightened and costs rose above what vouchers could effectively mitigate. Similarly, Australian research shows that rental subsidies for low-income households, paid under the Commonwealth Rental Assistance scheme, did little to improve housing affordability in a context of rising rents, as illustrated by the fact that 46% of households that receive the subsidy are still in housing stress (Anglicare Australia, 2022).

Pairing a BI with an expanded social housing sector would have several advantages vis-à-vis the task of addressing homelessness. On the one hand, access to social housing would limit the extent to which BI recipients are exposed to the vicissitudes of the private market. On the other hand, a BI would help ensure that social housing providers (non-profit organisations, state or local governments) have a stable rental revenue base – a constant problem where rents are set as a proportion of incomes (Jacobs, 2019) – as tenants would have a higher and more secure minimum income. It would also help alleviate the ongoing poverty experienced by many formerly homeless households that enter social housing, many of which rely on charity to supplement their meagre incomes (Parsell et al, 2021). Universal BI payments would mean this group had sufficient income to cover basic non-housing necessities (e.g. food, energy bills) and limit their reliance on charity, which many experience as shameful and degrading (Parsell et al, 2021).

The hybrid could also usefully include regulatory measures to control that rate of rental increases. Rent control is a much maligned, but also much misunderstood, policy proposal, particularly in liberal welfare states where intervention in the market is erroneously seen as reducing housing quality and supply (Slater, 2021). However, they remain a commonly used housing policy tool across Europe and elsewhere (Kettunen & Ruonavaara, 2021); and most contemporary policies and proposals take account of the impact on landlords’ capacity to provide and maintain rental properties (Slater, 2021). While this is not the place to expound the varieties and nuances of rent control policy, it is worth noting that that well-designed rent controls could help guarantee the effectiveness of a BI in preventing homelessness. They would do this by ensuring that improvements in income levels brought about by a BI are not undone by runaway rental cost inflation. This will be especially important in contexts where access to social housing remains limited, or where its expansion is slow, and the bulk of low-income households remain reliant on the private rental sector.

Addressing the stigma of targeted interventions

Beyond addressing the structural drivers of homelessness, I argue that a BI can also help increase support for, and reduce the stigma and blame arising from, responses to homelessness generally. One way a BI will do this is by replacing targeted measures with a universal one that benefits, and thus will enjoy the support of, a broader political constituency. As colleagues and I have argued elsewhere, governments – and by extension publics – tend to support actions to address homelessness when these actions produce benefits for the broader community (Parsell et al., 2020). This can be seen in the willingness of even conservative governments to invest in expensive housing first programs because these programs are shown to produce overall savings to taxpayers (Culhane, 2008; Parsell et al., 2017). It can also be seen in the large-scale investment in accommodating the homeless during the Covid-19 pandemic because this group was identified as a potential source of virus transmission and thus a threat to public health (Parsell et al, 2020). Given a BI would be paid to everyone on an individual basis, more people would have a stake in its continuation, unlike the targeted measures that support people experiencing or at risk of homelessness alone.

A BI is also less likely to attract the stigma that is attached to existing targeted measures for addressing homelessness. As noted above, research on BI-like measures that involve broad-based payments show that these are perceived as more ‘normal’ by publics than targeted measures, because they do not mark out their recipients as uniquely needy or deficient (Calnitsky, 2016). A BI would thus constitute an important break with prevailing liberal responses to homelessness, which (a) are highly targeted and restricted to minimise the possibility of dependence and market disengagement (Evans et al., 2021); and (b) treat homelessness as the product of individual pathologies or proclivities, and thus attract substantial stigma (Farrugia & Gerrard, 2016; Mitchell, 2011; Willse, 2015).

In this way, a BI is superior to existing measures used to address the income side of the housing affordability problem in liberal welfare regimes, such as rental subsidies and vouchers (see also Kerman, 2021, pp. 6–7 on this point). These measures are typically highly targeted (through means tests and assessment of individual needs/pathologies) and thus mark out their recipients as problem individuals, leading to stigmatisation and discriminatory treatment. Indeed, research shows that users of housing vouchers, such as the Section 8 housing voucher in the United States, experience significant stigma and discrimination from private landlords (Desmond, 2016). Voucher holders are perceived as ‘riskier’ tenants and thus struggle to compete with other renters in tight rental markets, despite having the same capacity to pay. A BI is also superior to simply increasing existing income support payments (although this would certainly be a positive development). For, such a move does little to combat the divisions and resultant stigma produced by these measures and which Bauman (2005) argues are a core reason why they have deteriorated to such an extent in the first place.

However, a BI should not completely replace targeted measures. Following Titmuss’ (2006 [1968]) reasoning, deploying a BI as part of a suite of universalist measures may help augment the effectiveness of targeted interventions and reduce the stigma that they attract. It will do this by helping to decouple access to housing and other essential resources (food, hygiene facilities, etc) from engagement with programs aimed at addressing personal issues, like mental illness, addiction, financial literacy, etc. As described earlier, most services targeting the homeless, like crisis or transitional housing programs, render access to the resources and support they offer (including support to access permanent housing) conditional upon engagement with these kinds of programs. As such, they position a person's individual problems/behaviours as central to their homelessness and, by extension, to any failures or setbacks they might experience. This is deeply stigmatising, as it cultivates a ‘sense of personal failure’ while ignoring structural conditions and the ‘diswelfare’ that they produce (Titmuss, 2006 [1968]).

The BI-services-regulation hybrid described aims to ensure that people's basic needs – for affordable housing and a stable income – are guaranteed unconditionally, as a matter of right and on the basis that homelessness is a housing problem (Colburn & Aldern, 2022). This does not mean that people will not be able to access targeted programs to help address personal issues. Indeed, rather than replace targeted services, the hybrid will enable people to engage with such services on their own terms, without coercion, and without their personal problems being blamed for, or treated as the key to addressing, their homelessness. Given this, people will be able to access these services on the same basis as members of the housed population. Of course, this will not necessarily remove the stigma that society attaches to issues addressed by these services (e.g. mental illness or addiction). However, this stigma will no longer be bound up with homelessness, specifically. It will be the same stigma experienced by housed populations seeking support for the same problems and thus something to be addressed through broader societal struggles to combat that stigma.

In this way, the BI-services-regulation hybrid helps advance the logic of the housing first movement, while overcoming the limitations of targeted housing first programs. Housing first holds that access to housing is a right, and not something that should be conditional upon individual behaviour (Padgett et al., 2016). It also holds that access to housing is a necessary precondition for addressing personal problems. However, it only applies this logic to a very narrow group of chronically homeless people who have complex needs (Baker & Evans, 2016). It therefore fails to break the link between a person's individual characteristics and their access to the resources required to end homelessness (Willse, 2015). The BI-services-regulation hybrid helps overcome this limitation by extending the unconditional logic of housing first to all homeless or potentially homeless groups. Thus, while I agree with Kerman (2021) that a BI should be seen as complementary to the housing first approach (Kerman, 2021), I suggest it can also help highlight the necessity of universalising housing first principles.

Conclusion

This article has considered the potential of a BI to help address homelessness. Inspired by Titmuss (2006 [1968]), I have argued that a BI can indeed play a role, but only as one part of a suite of universalist measures that ameliorate the structural drivers of homelessness and the stigma attached to existing targeted responses. A BI should form part of a BI-services-regulation hybrid (Malleson & Calnitsky, 2021) aimed at addressing the housing affordability crisis that underpins worsening homelessness. A BI can address the income side of the housing affordability problem by guaranteeing that people have sufficient income to meet their housing costs; but its success in doing this is contingent upon it being coupled with an expanded and more universalistic social housing sector and well-designed rent controls to address the cost and supply side of housing unaffordability. I also argued that the BI-services-regulation hybrid will help limit the stigma arising from targeted homelessness interventions by either replacing them or mitigating the conditionality they entail.

While a BI is not a panacea for homelessness (Kerman, 2021) neither is it inconsequential. In many ways, it represents one aspect of the broader shift in the logics of welfare provision required to address homelessness in liberal welfare regimes. As Evans et al. (2021) argue, ‘ending homelessness will necessitate a biopolitical reprogramming of the state itself’. It is not enough to invest in a never-ending stream of new practice-level innovations, as the failure of the innovations of housing first to stem rising rates of homelessness shows (Evans & Baker, 2021). Nor is it sufficient simply to boost targeted measures, such as private rental support initiatives and other targeted benefits, while leaving their stigmatising and residualising logic intact (Bauman, 2005). Instead, we must continue to challenge the residualising and individualising logic of liberal welfarism itself through the promotion of renewed universalism and social rights. A BI can play an important role in this endeavour.

This is, of course, an ambitious goal and one that is difficult to achieve. However, it is not an impossible one. Welfare regimes, as they exist today, are the historically contingent artefacts of past social struggles (Esping-Andersen, 1990) – they were mutable then, and they remain mutable now. Indeed, as Kasza (2002) has shown, welfare states are not as static entities: their logics are modified over time in response to historical circumstances and changing ideological currents. Scotland's (partial and incomplete) shift away from liberal to a more social democratic approach to welfare post-devolution offers a promising recent example of such changes (Vampa, 2016). As growing social insecurity becomes more and more the defining feature of our contemporary historical circumstances, there is a strong case to be made for universal welfare measures generally (Greve, 2020) and a BI in particular (Standing, 2020).