Mental health effects of Universal Credit during COVID-19: a difference-in-differences analysis of UKHLS data

Background Pre-COVID-19 pandemic evidence suggests that the UK’s introduction of Universal Credit (UC) was associated with a deterioration in mental health. Given pandemic-related changes to the benefits system, we sought to evaluate the mental health effects of UC during the pandemic. Methods We used nationally representative individual-level UK Longitudinal Household Panel Study data (2018–2021). We employed a difference-in-differences analysis with individual and time fixed effects comparing UC to legacy claimants (excluding Working Tax Credit claimants). The self-reported General Health Questionnaire (GHQ)-36 served as our main continuous outcome, while the GHQ-12 served as a supplementary caseness measure (≥3). Results Claiming UC during the COVID-19 pandemic reduced psychological distress compared with the comparison group, even after adjusting for confounders (−0.96 (95% CI −1.63 to −0.30)). After adding additional controls the estimates remain similar (−0.86 (95% CI −1.56 to −0.18)). When measured by GHQ-12 caseness, it was also associated with a reduction in the prevalence of psychological distress, but this was less certain due to wide CIs (−0.05 (95% CI −0.10 to 0.00)) and was more uncertain still after adjustment for further confounders (−0.04 (95% CI −0.10 to 0.01)). Event-study models showed that the greatest reduction in psychological distress was experienced during the first treatment year (−0.97 (95% CI −1.77 to −0.17)), without any significant effect 1 year post-treatment (0.12 (95% CI −1.23 to 1.46)). Conclusion Our findings suggest that during the COVID-19 pandemic, the psychological distress of UC claimants was reduced in the short-term compared with a group of legacy claimants—although with some uncertainty across alternative measures. This suggests that the pre-pandemic negative mental health effects of UC may be ameliorated by welfare reforms that aim to increase the generosity and reduce the conditionality of UC.