Heterogenous effects of the Finnish partial basic income experiment on mental health

Abstract Background In Finland, a randomized trial of a targeted, partial basic income scheme of 560€ was conducted in 2017-2018 for the unemployed. No studies have investigated whether the effects of this trial on mental strain were moderated by sociodemographic characteristics of participants. This is an important limitation because some population groups may benefit more than others or some groups may be harmed, which would suggest a need for targeted programs. For example, people with precarious employment situations may benefit more as basic income would imply a reduced administrative burden of reporting income to social insurance institutions. Methods A phone survey conducted at the end of the trial yielded a response rate of 20 percent (treated n = 569, control n = 1028). Mental strain was measured by the MHI-5 instrument and dichotomized. Potential moderators included age, gender, education, prior employment status, household size and family type. We investigated heterogeneity in the effect on mental strain using three complementary statistical approaches: subgroup analysis (first level interactions), multilevel modelling (four level interactions) and Causal Forest machine learning approach (up to 8 level interactions). The models were adjusted for survey weights and background variables. Results The treated group showed a reduction in the odds of mental strain compared with control (adjusted OR 0.61 [95%CI: 0.45; 0.83], risk difference -0.08 %-points [95%CI:-0.13; -0.03]. Subgroup analysis indicated that younger participants benefited more than older groups. Using multilevel modelling and a causal forest approach, we found no evidence for subgroup-related differences in the effect of the trial on mental strain. Conclusions While these findings are limited by the low response rate of the survey, they indicate that basic income schemes improve average mental health of unemployed people and are unlikely to redistribute mental health across sociodemographic characteristics. Key messages • Partial basic income trial was associated with a reduction in mental strain among the unemployed in Finland. • We found no evidence for subgroup-related differences in the effect of the trial on mental strain.