118 Chapter 4 procedure (extended or abridged) and half-year period; type of application (regular application or other routes such as priority cases) and the responsible for its submission (long-term care provider, social worker or other); and whether the applicant was eligible for other types of long-term care in the last month. Individual-level controls include age, gender, migration background (Dutch, western or non-western first or second generation migrant), living situation (alone, with partner, with parents, in an institution or other), prior use of specialist mental health care, prior health care expenditure, prior working and prior personal income (all for the calendar year before the application). Standard errors are clustered at the assessor level. We explore subgroup effects of eligibility to supported housing on individual outcomes by age at the time of the application, prior personal income quintile, and mental disorder diagnosis. For parental outcomes, we consider heterogeneity for parents living with the child ahead of her application to supported housing. Robustness analyses include testing for the leniency instrumental variable approach, choices in the instrument construction, alternative 2SLS specifications and reporting findings on the individual outcomes for the subpopulation studied for parental spillovers. RESULTS We examine the effects of being eligible to supported housing using the variation in assessor´s leniency as an instrument in a 2SLS regression approach. Results of the first stage show that assessors´ leniency is highly predictive of whether an individual is deemed eligible to supported housing. An increase of 10 pp in the residualised instrument increases the probability of being eligible to supported housing by 9.8 pp (se: 0.5). Individual Outcomes Being granted eligibility increased the probability of moving into supported housing by 31.6 pp (se: 5.7) in the calendar year after application (i.e., in the 12-month period starting between 1 month - application in December, and 11 months - application in January, after the application) (Table 3). This effect shows that not every marginal individual granted eligibility will convert it into to admission during the following calendar year, and it compares with 17% of non-eligible individuals that are admitted to supported housing in the same period, after re-applying12. Supported housing eligibility further decreased the likelihood of using home care by 13.4 pp (se: 6.4), relative to an average utilization of 30% among the non-eligible. This is expected given that home care is an alternative form of support granted by assessors. We find no impact of supported housing eligibility on the likelihood of using mental health care services and on all-cause mortality. 12 The estimate in Table 3 gives the effect on admission in the next calendar year. Admission into supported housing in the 90, 180 and 365 days following application increased by respectively 33.8 pp, 35.2 pp, and 38.6 pp. The increasing trend indicates that the share converting eligibility into admission rises more strongly over time for the eligible compared to the non-eligible that get admitted after re-applying.
RkJQdWJsaXNoZXIy MjY0ODMw