The protective effect of smoking against COVID-19: A population-based study using instrumental variables

A low prevalence of smokers among the confirmed patients with COVID-19 has been reported by multiple hospital-based studies, and this observation gave rise to a hypothesis that smoking has a protective effect against the novel coronavirus. We test this prediction in a population-based study across the US states and use an instrumental variable approach to address the endogeneity of smoking rates. We find that a higher prevalence of smoking has a significant negative effect on the spread and the severity of the COVID-19 pandemic across the US state: it decreases the per capita number of registered cases, the case fatality rate, and the excess mortality. The protective effect is more pronounced in subgroups of the population that are more likely to be smokers: men of all ages and females of the older cohort. Our findings are robust to the inclusion of a broad range of control variables, exclusion of outliers, and placebo tests. Despite the protective effect against the COVID-19, smoking remains detrimental for health in the long-term, and we show that states with a higher rate of smoking also have higher mortality in the year before the outbreak.