Background Paid paternity leave could affect the mental health of both parents by giving fathers time with their families and facilitating shared childcare. There is limited evidence about relationships between paternity leave and parental mental health. The few existing studies report mixed findings. Better evidence may support parental leave policymaking.

Methods We examined the association between fathers’ uptake of paternity leave and mental health for fathers and mothers. Our sample was new parents from households with a father eligible for paternity leave, drawn from a national survey, the UK Household Longitudinal Study. Outcomes and exposure data from waves 2-10 (2010-2020) were linked with dataon parents’ prior characteristics in waves 1-9 (2009-2019). The primary outcome was the GHQ-12 score, measuring current symptoms of common mental illness. We used logistic regression to identify factors associated with taking paternity leave. To identify the association between paternity leave and mental health, we used inverse probability weighted regression adjustment (IPWRA); this creates doubly robust estimators. The fully-adjusted IPWRA model incorporated demographic, socioeconomic and health characteristics of both parents as well as the child’s age and season of birth.

Results Complete data were available for 1,405 fathers and 1,393 mothers, with babies aged on average 6.4 months. Among fathers, 78.6% took paternity leave; 78.5% of mothers had a partner who did so. Odds of taking paternity leave were higher with older maternal age; first-time fathers; more highly qualified fathers; and fathers born in the UK.

Fully-adjusted IPWRA model results showed no association between paternity leave and mental health for either parent. For fathers, the average treatment effect was a decrease of 0.16 points (95% CI -0.45-0.14; p=0.29) on the 12-point GHQ-12. For mothers the average treatment effect was a decrease of 0.14 points (95% CI -0.43-0.15; p=0.357).

Discussion We found no association between paternity leave and mental health in a sample of (similar after weighting) leave-eligible UK families. This may be because paternity leave makes no difference to mental health; its effects are offset by unobserved confounders; or effects are too small to detect. Null effects might also be explained by the UK’s short statutory paternity leave period or by its limitation to fathers in secure employment. A strength of this study is the adjustment for a rich set of potential confounders, including prior mental health of both parents. An important limitation is the restriction to families with a father eligible for statutory paternity leave.