Background Universal Credit (UC) is the largest reform of the UK welfare system since the Second World War. It streamlines six legacy benefits and tax credits, seeking to simplify claims and encourage employment. Previous research suggests UC harmed mental health pre-pandemic, potentially due to increased conditionality and lower payments. During the COVID-19 pandemic, the UK Government increased UC generosity and reduced conditionality, including by suspending work search requirements and providing a £20 weekly uplift. We investigate if pandemic-related changes to UC modified its effect on mental health.

Methods We analysed data from the UK Household Longitudinal Study (Understanding Society), a household-based panel study that follows a representative sample of the UK population annually. Data were from 2018, the end of UC rollout, to 2021, when pandemic-related changes were withdrawn. The study population consisted of working age adults (16-65) who reported claiming legacy benefits or UC pre-pandemic, yielding 13,514 observations. We defined people exposed to UC pre-pandemic and compared trends in this group to people exposed to legacy benefits pre-pandemic. Observations were classed as exposed to pandemic-related reforms if sampled after March 2020 and unexposed otherwise. Mental health was measured using the General Health Questionnaire, a validated measure of psychological distress, producing scores from 0 to 36. A difference-in-difference analysis was conducted using a two-way fixed effects linear model, adjusted for confounders, including age, sex, long-term health and cohabiting status. Analysis was conducted in R.

Results Pandemic-related UC changes were associated with reduced psychological distress, -0.97 points (95%CI: -1.73, -0.22), relative to trends among legacy benefit recipients. Subgroup analysis showed no difference in effects for under- versus over -25s or between UK countries. Heterogenous effects were found across household structures. Single adults and coupled parents claiming UC experienced improved mental health relative to trends in legacy benefit claimants: -1.47 (-2.71, -0.24) and -1.75 (-3.25, -0.25) respectively. Lone parents and childless couples on UC experienced harms to mental health, although this is less certain due to smaller sample sizes: 0.32 (-1.36, 1.99) and 1.42 (-0.56, 3.39).

Conclusion Reduced conditionality and increased payments of UC during the COVID-19 pandemic are associated with improved mental health. While our analytical approach accounts for time-invariant confounders, residual confounding due to the pandemic affecting legacy benefit recipients differently from UC claimants remains possible. Our findings strengthen evidence to support reducing UC conditionality and increasing generosity as a means of improving mental health among socio-economically disadvantaged people.