Government Approval and the Effectiveness of Non-Pharmaceutical Interventions Against COVID-19: Time-Series Cross-Sectional Regressions Upon Regions in England, U.K.

Background: Studies have demonstrated the effectiveness of non-pharmaceutical interventions for curbing the spread of COVID-19. However, an important factor has been overlooked: politics. As recent studies have revealed, political beliefs could shape citizens’ perceptions about the pandemic. Methods: We conducted time-series cross-sectional regressions with lagged dependent variables, regional fixed effects, regional time-trends, and cluster-robust standard errors using England’s data on confirmed COVID-19 cases from 1 March to 30 June. We examined the effects of government disapproval on the time-dependent reproduction number, Rt. Disapproval was interacted with a lockdown dummy to assess whether citizens’ support for the government still affected the Rt under harsh interventions like lockdowns. As sensitivity tests, we used different lags (four, five, six, and seven days) on the main variables, ran the analysis without the interaction term, and used an alternative dependent variable. Findings: A percent point increase in disapproval resulted in a 0·012 (95% CI 0·005 to 0·019; P=0·014) increase in the time-dependent reproduction number (Rt). Under lockdown, a percent point rise in disapproval increased the Rt by 0·008. Interpretation: Non-pharmaceutical interventions are more effective when governments are popular. Because governments both design and execute these interventions, people are less likely to abide by the restrictions designed by the government when they do not believe the government is doing a good job. Funding: No funding to report. Declaration of Interests: There are no conflicts of interest to report. Ethics Approval Statement: No approval was necessary as the study did not involve human subjects.