104 Chapter 4 the agency’s assessors who determine eligibility for subsidised supported housing, and, importantly, who vary in terms of their leniency [14-16] .An application who was assigned to a strict assessor may be determined ineligible for supported housing, while the same application would have been eligible if assigned to a lenient assessor. We estimate local average treatment effects (LATE) of eligibility for supported housing among individuals whose eligibility was affected by the random variation in assessor leniency using an instrumental variables design. Our estimates suggest that, among individuals whose eligibility depends on assessor assignment, supported housing eligibility increases the probability of moving into supported housing by 31.6 percentage points (pp) (standard error (se): 5.7) in the next calendar year, and decreases the likelihod of using home care by 13.4 pp (se: 6.4). Average total care expenditure increases by 20,017 euros (se: 5,006), mostly at the expense of supported housing expenditure (11,883 euros; se: 2,634), and potentially also mental health treatment costs (7,698 euros; se: 6,440). We find little evidence of substantial offsetting cost savings on other types of care. We also find that eligibility for supported housing decreases personal income by 1,470 euros (se: 748) in the short-run and 2,017 euros (se: 747) in the long-run, as well as income from work by 2,149 euros (se: 717) and 3,410 euros (se: 1,488), and likely extensive-margin labour participation by 7.0 pp (se: 3.6) and 7.5 pp (se: 5.5.), respectively. On the other hand, both mothers and fathers of applicants increase their labour participation in the fourth year after their child is granted eligibility to supported housing, by 15.1 pp (se: 7.2) and 17.8 pp (se: 7.6), respectively. Our results are consistent with previous findings in the literature which suggest that a stable living environment supports mental health treatment adherence [17, 18]. We do not find evidence that eligibility for supported housing environment decreases mental health treatment costs overall, as suggested by previous studies which argued that supported housing may reduce costly service use, such as acute hospitalizations, by preventing severe mental health deterioration [19-21]. The positive effects of supported housing eligibility on parental extensive-margin labour supply may reflect that supported housing is a substitute for informal care by family members. Disability has been found to negatively impact parents’ labour market outcomes, both in terms of participation and earnings [22, 23], with smaller effects on fathers compared with mothers [24-26]. Our study adds to the existing evidence by reporting causal effects for the broader population with mental health problems, and not focusing on specific groups such as the homeless living with mental disorders or patients undergoing deinstitutionalization after long-periods living in institutions [11]. Another key strength of our design is that our estimates apply to individuals for whom supported housing eligibility is determined by the leniency of the assessors. This provides evidence relevant for policy-makers because determination of eligibility for supported housing, rather than the individual decision to apply, is the primary policy lever for affecting uptake of supported housing in the
RkJQdWJsaXNoZXIy MjY0ODMw