Research

Depression, poor educational attainment, poor health, and unemployment form a correlated web that reduces human flourishing, but the causal links between them are unclear. We estimate the causal effect of depressive symptoms on educational and labour market outcomes by triangulating four research designs across economic, clinical, and genomic data. In UK longitudinal data (305,686 person-wave observations), a one standard-deviation (s.d.) rise in depressive symptoms lowers income in the following wave by 0.036 s.d. and employment by 0.044 s.d.; instrumenting symptoms with bereavement gives larger but less precise short-run estimates. Two genetic designs capture lifetime liability: two-sample Mendelian randomization on summary statistics from up to 2.1 million individuals, and a polygenic index instrument in the genotyped UKHLS subsample. In these designs, genetic liability to depression lowers income by 0.24–0.49 s.d., educational attainment by 0.14–0.46 s.d., and occupational status by 0.36–0.65 s.d. The education and occupation penalties fall almost entirely on women, while most of the income penalty for both sexes runs through employment rather than pay. The economic cost of depression accumulates over the life course, so treatment before labour-market entry may yield the largest economic returns.

Life satisfaction is U-shaped over the life cycle, and roughly 40% of the variation in wellbeing is heritable. We ask how genetic predispositions for wellbeing interact with ageing. Matching a wellbeing polygenic index (PGI) to panel data from the British Household Panel Survey and the UK Household Longitudinal Study, we establish three results. First, the PGI predicts substantial and statistically significant differences in life satisfaction at every point in the life cycle. Second, the U-shaped age profile of wellbeing holds within both high- and low-PGI groups. Third, individuals with a lower PGI face a more pronounced U-shape, so the genetic wellbeing gap is widest in middle age and narrower at the tails. The first two findings are robust to a within-family design that controls for parental PGIs, while the third fails to replicate in the within-family sample. A mediation analysis shows that declining physical health accounts for roughly one-third of the age variation in the gap.

Does the child penalty reinforce genetic inequality? Using Norwegian administrative data linked to molecular genetic information, we estimate gene–environment (G×E) event studies around first birth. Women below the median polygenic index for educational attainment experience an employment penalty 1.5 percentage points larger initially, widening to 4 percentage points after ten years. The gap persists controlling for education, implying mechanisms beyond schooling. UK household panel data corroborate the pattern. Policies reducing the child penalty generate a double dividend: improving gender equality and attenuating the economic consequences of the genetic lottery.

We develop a structural model of health and human capital formation in which genetic endowments shape both the utility cost of health investments and the technology of health production, and hence the incentives to invest in healthy habits. Taking body mass index as the measure of health and physical activity and food intake as the investments, we study how these choices interact with variants in FTO and other loci associated with BMI in genome-wide association studies, using cohort data on British adolescents and US adults.