This paper explores the government service interactions for a cohort of formerly homeless youth in Aotearoa New Zealand, using linked administrative data. We report the rates of service interactions pre- and post-housing for a cohort of 69 youth, aged 18–25 years old, who received housing and support from a Housing First provider. 60.9% were women, and 69.6% were Māori. Very high rates of service interactions were seen both before and after they were housed. The evidence shows promising potential improvements in young people’s lives when housed. The most significant changes we saw in the one and two years post-housing were in incomes from both wages/salaries and social welfare benefits; there was a significant increase in total income for our cohort. There was also a promising reduction in hospitalisations and emergency deparment admissions. Overall, Housing First shows promising outcomes for youth.

Introduction

Youth in Aotearoa New Zealand (henceforth referred to as Aotearoa) experience disproportionately high rates of homelessness. Of the 41,644 people who were experiencing homelessness at the 2018 Census, half were 25 years old or younger (Amore et al. 2020). Furthermore, 60% of Māori experiencing homelessness were under 25 (Amore et al. 2020). The experiences of these more than 20,000 youth are under-researched and poorly understood. Existing literature on youth experiences of homelessness in Aotearoa is limited. We build on current international literature and utilise Aotearoa specific data to explore young people’s government service interactions before and after being housed by a Housing First (HF) provider.

Youth homelessness is a complex and multi-faceted phenomenon with many intersecting issues such as ethnicity, age, gender, and sexual orientation (Page 2017; Shelton and Bond 2017; Wirihana and Smith 2021). International research indicates poverty as one of the key contributing factors that lead to, and maintains, youth homelessness (O’Grady and Gaetz 2004; Sharam and Hulse 2014; Ferguson et al. 2015). Youth who experience homelessness are frequently reported to struggle with obtaining, and maintaining, an income – in large part due to their age (O’Grady and Gaetz 2004; Ferguson et al. 2012, 2015). Furthermore, the international literature also indicates that experiencing homelessness is associated with a loss of income support (i.e. welfare benefits) (Zivanovic et al. 2016). Liberal welfare states (such as those seen in Australia, Aotearoa, and the United States of America) see high levels of poverty due to neoliberalism and its subsequent decreases in state welfare support (Stephens and Fitzpatrick 2007). Liberal welfare states see increased poverty, income inequities, and homelessness than more protective social democratic welfare states (such as those seen in Nordic countries) (Stephens and Fitzpatrick 2007). In addition to the high rates of poverty they face, youth experiencing homelessness are at risk of poor mental and physical health, substance use, imprisonment, victimisation, and risky sexual behaviour (Thompson et al. 2010; Johnson and Chamberlain 2011; Dawson and Sinwell 2012; Chamberlain and Johnson 2013; Kozloff et al. 2013; Heerde et al. 2015; Britton and Pilnik 2018; Krahn et al. 2018; Robinson 2018). These risk factors are heightened for people of diverse gender and sexual orientation identities, and racial minorities (Page 2017; Shelton and Bond 2017).

In the Aotearoa context, a recent study found that 29% of youth aged between 12 and 18 had experienced housing deprivation at least once during the past year (Clark et al. 2021). Furthermore, research has also found that due to a limited range of options being available to them, youth in Aotearoa are often rendered homeless to escape abuse and low quality of life in their family homes (Groot et al. 2017). These experiences often lead youth into the state’s care, either through child welfare or youth justice. In Aotearoa, this is more commonly experienced by Māori, who make up 69% of those taken into the state care system (Atwool 2020; Mills et al. 2021). Along with perpetuating fractured relationships with familial networks, time spent in state care has been linked to an increased risk of cyclic homelessness in youth and into adulthood (Johnson and Chamberlain 2008; Mendes et al. 2008; Shah et al. 2017). Beyond this, there is limited literature on youth homelessness specific to Aotearoa. This literature shows the need for Aotearoa–specific research focusing on youth’s experiences of homelessness and the effectiveness of the policies intended to address the issue.

HF is a holistic model of housing support, predicated on the Right to Housing (Tsemberis 2010). It removes barriers to entry, such as sobriety, and instead focuses on housing people and linking them with appropriate support (Tsemberis 2010). The model has been widely adopted internationally, and funding has been increased for HF in Aotearoa by successive governments (Pierse et al. 2019). For this paper, we use data from our research partners, The People’s Project, who were the first large scale HF provider in Aotearoa. We explore the change of interactions with government services amongst youth in a HF cohort both before and after receiving housing and support from TPP. Our analysis utilises linked administrative data to explore health, income, mental health, and justice data of youth who have been housed by TPP.

Methods

This paper extends earlier analyses on the service usage in the HF cohort that utilised the Integrated Data Infrastructure (IDI) (Pierse et al. 2019; Fraser et al. 2021). The IDI consists of microdata of services provided by Aotearoa government agencies, census and official surveys, and data collected by various non-governmental organisations (NGOs) and is maintained and regularly updated by Tatauranga Aotearoa Statistics New Zealand (StatsNZ). The IDI allows for linking de-identified administrative data across a wide range of government service interactions for the individuals in the HF cohort (Black 2016).

Results

We begin by comparing the demographics of the youth in the HF cohort with the youth of the same age bracket (aged between 18 and 25 on the reference date) in the GYP. Following that, the rate of lifetime service interactions amongst the HF youth and GYP is presented. We then present the rate of change in service interaction pre- and post-housing for the HF youth, with a table of comparison data from the GYP for the same periods as the HF cohort.

compares the demographics of our HF cohort with the GYP of the same age. Females (60.9%) and Māori (69.6%) are over-represented in the HF cohort compared to the GYP cohort. One-third of the HF cohort was aged between 21 and 22, and 17 and 18 years olds¹ were relatively less common in the HF cohort.

shows the percentage of each cohort who had interacted with government services before and after being housed; the data prior to being housed spans from birth to when they were housed by TPP. Of particular note, 61% of the HF cohort had been reported to be the victim in a child abuse case, compared to 9.1% of the GYP. 71.4% of women in the HF cohort had a maternity-related hospitalisation record, compared to only 18.6% of the GYP comparison. In addition, shows the rate of services interactions amongst the HF cohort in their first two years after being housed by the HF programme, as well as the two years post-reference date for the GYP. For every service except education, the HF youth had more interactions with government services prior to being housed. This higher rate of interaction was reduced after they were housed. However, for most domains, their interactions post-housing still remained higher than the GYP group’s lifetime interactions, as well as their post-reference period interactions. We note that, interestingly, victimisations increased in the two years post-housing (or post-reference date for the GYP) for both the HF and GYP youth. More research is needed to investigate why this is. The drop in HF youth who were taxed as a wage or salary earner between the lifetime pre-housing period and the two years post-housing period suggests that these youth drop in and out of employment over time. The similar number of GYP youth taxed in the lifetime period and two years post-reference date period suggests they experience less churn in employment compared to the HF cohort.

shows the rates of interactions with government services for the HF youth, and shows the same analysis for our GYP comparison group. Income and benefits data show that before being housed, income from wages/salaries dropped from $1,600 in the second to last year before being housed ($3,000 cumulative across the two years) to $1,400 in the year before being housed. There is an increase to an annual income from employment: an increase to $2,300 in the first year after being housed and a more significant jump to $3,800 in the second year ($6,100 cumulative across the two years). Welfare benefit data shows a similar trend, but with a slightly smaller increase in the two years post-housing compared to income. Overall, in the one year prior to being housed, HF youth had a mean income of $13,300 from both wages/salaries and benefits, which increased to a total mean income of $15,900 in the second-year post-housing. This shows a significant increase in total income once these youth received Housing First support. These incomes still remained low, which could be in part due to the young age of this cohort and the possibility of them still being supported by caregivers; however, as the data for the GYP shows, the HF cohort had lower incomes than the wider population so age alone does not explain these low incomes.

The GYP had considerably higher income rates than the HF cohort; in the two years after the reference date, the GYP had a mean total income from wages and salaries of $47,000 ($23,500 per year), in comparison to the HF cohort’s income from wages and salaries of $6,100 ($3,050 per year). The GYP youth had much lower incomes from social welfare benefits, while the HF youth still relied on it as their primary income source. The HF cohort’s income from both wages/salaries (p ≤ 0.001) and welfare benefits (p ≤ 0.001) had a statistically significant increase post-housing when compared to the GYP, as did the number of months receiving income from wages/salaries (p ≤ 0.001) and the number of months receiving an income from welfare benefits (p ≤ 0.001). When comparing the two years pre-housing with the two years post-housing, the HF cohort had a statistically significant increase in their incomes from wages/salaries (p ≤ 0.05), the number of months receiving income from a welfare benefit (p ≤ 0.05), and income from welfare benefits (p ≤ 0.005). Overall, the HF programme had a significant impact on incomes in the two years post-housing.

There is a general decrease in health service utilisation rates after the HF cohort received Housing First support, with some increases in the two years post-housing. Hospitalisations and pharmaceutical dispensing showed a drop in the first year they received HF support, then remained steady in the second year post-housing. Notably, there is a sustained decrease in emergency department presentations. However, at the same time, a significant increase is seen in the rate of secondary care (outpatient) service utilisation in the two-year post-housing period. There was a significant difference in hospitalisations (p ≤ 0.001) and emergency department interactions (p ≤ 0.05) for the HF cohort in the two years pre- and post-housing. Additionally, there was a significant difference in hospitalisations (p ≤ 0.05) between the HF cohort and the GYP in the two years post-housing/post-reference. The GYP cohort shows a slightly increasing trend in the health service utilisation within the reference period except for the injury claims.

In the justice domain, there is an initial drop in interactions across all four areas (police offences, criminal charges, major corrections events, and victimisation) in the first year post-housing for the HF cohort, and then a slight increase in the second year post-housing. For all except victimisation, the differences between the two-year pre-housing and two-years post-housing period are a decrease; victimisation is increased by 9%. However, all changes are only slight, and our sample size is small. Therefore, further research is needed to see how these results might differ across a larger sample of youth who have received housing support. For the GYP, all justice interactions were significantly lower than the HF cohort, although they did see a very slight rise in major corrections events.

Discussion

In comparison to the GYP, the youth in the HF cohort had higher levels of interactions with government services prior to becoming housed, with improvements seen across multiple domains once they received HF support. Consistent with existing data on demographics of homeless populations in Aotearoa, our cohort was more likely to be Māori and female (Amore et al. 2020). The data we have demonstrated that these youth presented to a range of different government services leading up to engaging with TPP. Despite this high interaction rate, they still became homeless and required TPP’s services. This indicates that state support systems, particularly the social welfare system, but also the child and adolescent health sector, child protection services, the education sector and the youth justice system in Aotearoa are not fully meeting the needs of youth at risk of experiencing homelessness.

A positive finding of our results was outcome related to income and benefits data. Income from both wages/salaries and benefits increased significantly for the HF cohort. It shows the effect of stable housing and comprehensive support systems. These youth had the housing stability necessary to allow them to focus on other needs, and the support from TPP to ensure that they received the most appropriate form of financial aid. This is consistent with international findings from both Aubry et al. (2016) and O’Campo et al. (2016), who reported improvements in finances of those stably housed by HF programmes.

However, even with the positive effect of being stably housed, it is worth noting that overall income rates for the HF cohort remained much lower than the GYP comparison group. The HF cohort still lives in poverty despite their incomes rising after becoming housed. It is likely due to a myriad of reasons: primarily low social welfare benefit rates, low incomes, and high cost of living, pay inequities across genders, and pay inequities between Māori and Pākehā. A Government-mandated Welfare Expert Advisory Group (WEAG) review of the social welfare system found that welfare rates are far too low, benefits are increasingly difficult to access, and young Māori women with children are particularly affected by the social welfare system in Aotearoa (Whakamana Tāngata 2019). As an example of institutional racism, Māori women, particularly young Māori women with children, often face benefit sanctions or are penalised rather than properly supported by welfare providers (Cram et al. 2021). This can subject them to further poverty, jeopardising their housing stability (Cram et al. 2021). Policy change is needed to ensure that the welfare state adequately serves those at risk of homelessness.

Normative Pākehā-oriented moral attitudes towards young parenthood have contributed to stigma and inadequate support for young parents, particularly young mothers (Ware et al. 2017). Conservative and discriminatory values have shaped social support for young parents; with the inadequacy of the Sole Parent benefit being a key example (Ware et al. 2017; Gray and Crichton-Hill 2019). Housing instability is a significant issue for young parents, and being separated from children places additional pressure on parents who are experiencing homelessness (Meadows-Oliver 2006). Additionally, young wahine Māori have higher rates of fertility than young Pākehā women. As has been observed by Leonie Pihama (2011), Pākehā values have dominated discourse and policymaking about youth pregnancy, usually framing it negatively. She argues that Māori views on youth pregnancy need to be contextualised within Māori frameworks that place high value on fertility, and within existing social contexts (Pihama 2011). Adequate social support, including housing and adequate welfare levels, as well as enabling opportunities to access education and practical parenting support, contribute to positive outcomes for young parents and their children (Craig et al. 2004; Pihama 2011; Scanlen and Hooper 2019).

Our health findings show that the young women in the HF cohort were much more likely to have a maternal hospitalisation than the young women in the GYP comparison group. Our previous research has explored the service usage of all women (i.e. of all ages, not just youth) housed by TPP, finding that 81.4% of women in the wider HF cohort were mothers, that they were more likely to have multiple children, and more likely to be receiving a Sole Parent benefit, the amount of which has been repeatedly identified as inadequate (Fraser et al. 2021). Considering the high number of young women with maternal hospitalisation records whose service interactions we explore in this paper (and whose data was included as part of these earlier analyses), we predict that they too are likely to have multiple children and be receiving a Sole Parent benefit.

Furthermore, our health results also suggest that the HF cohort utilises more outpatient health services than the GYP but experience fewer hospitalisations and emergency department presentations once housed. This is a notable, positive physical health outcome, which we previously observed in the wider HF cohort (Pierse et al. 2021). Furthermore, this is consistent with international evaluations of HF programmes which show improvements in healthcare-related outcomes (Patterson et al. 2013; Aquin et al. 2017; Baxter et al. 2019). Outpatient health service consists of specialised secondary care often delivered in a hospital setting. It is critical to the management of chronic health conditions and the prevention of ambulatory sensitive hospitalisations and acute presentations in the emergency department. Appointment-based healthcare services have remained difficult to access for people experiencing homelessness (Chelvakumar et al. 2017). Additionally, people experiencing homelessness often present to emergency departments with non-urgent issues as they do not have the resources or capacity needed to engage in appointment-driven healthcare (Chelvakumar et al. 2017; Ramsay et al. 2019). With wraparound support based on stable housing, young people in the HF cohort are engaging more with the management of ongoing health concerns, which in return reduced the acute health service utilisations.

Our data shows that the HF cohort had markedly higher rates of interactions with the justice system than the comparison GYP. These findings suggest that youth experiencing homelessness are more likely to be involved in the justice system than their non-homeless counterparts. Any time spent in the justice system is disruptive to housing stability and adds complexity to future housing and employment options (Pierse et al. 2019, 2021; Fraser et al. 2021). When options are limited, these youth may struggle to find solutions and government support for their needs. The stress of survival in difficult situations may result in heightened levels of risk-taking (Auerswald and Eyre 2002; Ferguson et al. 2015; Gluckman 2018; Piche et al. 2018). As a result, survival instincts such as theft become a common means for youth to gain resources during periods of homelessness (Evans and Forsyth 2004; Hein 2011). To prevent this, the government needs more accessible options, such as HF, that youth can access when their whānau and support systems are limited. Furthermore, we expect that these high rates of encounters with the justice system are related to the well-established over-representation of Māori (particularly Māori women) in the justice system (Hess 2011; Webb 2017; Fernando 2018). As our cohort is predominantly Māori and female, we expect that they, too, have experienced the effects of a colonial justice system that criminalises Māori at higher rates than their non-Māori counterparts (Hess 2011; Webb 2017; Fernando 2018).

The findings we have presented are broadly consistent with international evaluations of the impacts of HF, particularly for youth. A systematic review of HF evaluations reported a number of studies that showed improvements in; residential stability, reduced participation in the justice system, reduced emergency department use and inpatient hospitalisations, and improvement in quality of life – including financial wellbeing (Woodhall-Melnik and Dunn 2016). Our data also shows that the HF cohort had markedly higher rates of reported child abuse as a victim than the GYP. These findings are consistent with international research that shows a strong association between adverse childhood events and an increased risk of later life homelessness (Fitzpatrick et al. 2013). Our results support the need for adequate mechanisms and policies to prevent family harm that has contributed to youth homelessness in our communities and the need to detect and address the lifelong impact of child abuse in Aotearoa (Felitti et al. 1998).

The literature on HF for youth particularly highlights the importance of HF in providing housing stability, which has flow-on effects into other domains of life (Kozloff et al. 2016; Wang et al. 2019). Gaetz et al. (2019) note that the key performance indicator most commonly used in evaluations of housing programmes is whether a person is housed or not. Instead, they posit that evaluations must also look to outcomes such as health and wellbeing, school and career goals, community connectedness, and beyond (Gaetz et al. 2019). While the administrative data we have utilised does not allow us to look at community connectedness or personal goals, we have been able to report positive outcomes relating to health, justice interactions, and most significantly, income. The results we have presented show that HF has led to a reduction in service interactions in some domains for this cohort. However, due to the complex lives many have already lived by age 25, most gains in wellbeing are likely to take longer than the two years of focused support to effect significant change in service usage.

Our use of administrative data is not intended to be a full proxy for need within this cohort, but rather a proxy for what government services know about need. For instance, our previous research has dispelled the notion that many, including previous New Zealand Governments, have held about people experiencing homelessness being ‘hard to reach’ (Pierse et al. 2019). Our usage of administrative data enabled us to show that the wider HF cohort had over 200,000 individual interactions with government services prior to being housed (Pierse et al. 2019). This allows us to emphasise that government agencies do, in fact, know a considerable amount about these populations and their needs. This has important implications for the delivery of government services. The data we have presented in this paper enables us to see where service interactions are declining or increasing. We have been able to show that, for instance, the youth in this cohort have incredibly low incomes (despite improvements made post-housing), and are in need of increased financial support in order to receive liveable incomes. While the data we have presented will not be a full picture of need amongst this population, it does allow us to see that these youth had a disproportionately high level of interactions with government services both pre and post-housing, which suggests that a high level of continuing support is needed to ensure their long-term wellbeing.

The main limitation of this paper is the small size of our cohort. While our results reflect government agencies’ poor ability to act as a safety net to prevent homelessness – where 69 youth needed the housing support from TPP – our cohort size is too small for us to apply our findings to all youth who experience homelessness in Aotearoa. However, we expect that all homeless youth in Aotearoa will have elevated levels of government service interactions. Furthermore, although the IDI is useful for its population data, it can lack important detail and context. For example, we do not know what the experiences of these service interactions were like: what was useful for them and what was not. Additionally, it is not possible to fully capture the needs and aspirations of this cohort will not be able to be captured by government administrative data alone, and would be best captured using complementary qualitative research; there are always going to be needs that are not met by government services. This cohort is, by definition, a cohort that was failed by our social safety net; however, it is possible that some services worked well and prevented youth from becoming homeless in many other cases.

Future research that explores the service interactions and long-term outcomes of youth across different forms of housing support is necessary to ensure that funding is directed towards the most equitable and effective programmes. There are numerous forms of housing support available in Aotearoa, including HF, as well as transitional housing, community housing providers, and Kāinga Ora.² Yet, little is known about the effectiveness of each (Ministry of Housing and Urban Development 2014, 2020). Both qualitative and quantitative research are needed in this area so that youth experiencing homelessness have their voices heard and their needs met.

In terms of policy, the cross-agency Homelessness Action Plan (HAP) sets out a series of actions and priority areas for addressing homelessness. Youth are identified as a priority group within the HAP, with the document noting the need for targeted programmes and support for youth. Research such as that presented in this paper serves to broaden and deepen the existing knowledge base that shapes how the HAP should be implemented. By exploring the services used by youth who had required stable housing and support from TPP, we can better understand how best to place effective interventions, reduce the number of youths experiencing homelessness, and ensure that support programmes improve youth wellbeing.

Conclusion

Youth experiences of homelessness are under-researched in Aotearoa. This paper provides an important contribution to the existing literature by providing the first quantitative study in Aotearoa of government service interactions for a cohort of youth housed by HF. Our results show that these youth had rates of interaction much higher than the GYP comparison group throughout their lives. There were significant positive outcomes suggesting improvements in health and wellbeing among these results, although these young people maintained constrained incomes compared to other non-homeless youth. At two years housed, this group are likely still stabilising from periods of homelessness. Future research and policy are needed to further explore youth homelessness in Aotearoa so these youth can be supported to the best of our abilities.

Data sharing policy

Data was accessed with the permission of Statistics New Zealand under their ‘five safes’ framework. All data are de-identified and only accessible via a secure connection from approved datalabs. Data and results must be aggregated and anonymised in accordance with Statistics New Zealand protocols. All results are checked for confidentiality by Statistics New Zealand prior to their release from the secure environment.

Statistics New Zealand disclaimer

These results are not official statistics. They have been created for research purposes from the Integrated Data Infrastructure, which is carefully managed by Statistics New Zealand. For more information about the IDI, please visit https://www.stats.govt.nz/integrated-data/. The results are based in part on tax data supplied by Inland Revenue to Statistics New Zealand under the Tax Administration Act 1994 for statistical purposes. Any discussion of data limitations or weaknesses is in the context of using the IDI for statistical purposes and is not related to the data’s ability to support Inland Revenue’s core operational requirements.

Acknowledgements

We acknowledge the young people who are discussed in this paper, each of whom has their own unique story to tell. We also acknowledge the staff of The People’s Project for their invaluable role. Without their permission and support, this work would not be possible. Ngā mihi nui.

Disclosure statement

No potential conflict of interest was reported by the author(s).

This paper was supported by funding from the New Zealand Ministry of Business Innovation and Employment, Endeavour Fund.