The 2020 US presidential election was won by Democrat challenger Joseph Biden over Republican incumbent Donald Trump and was contested during the global novel coronavirus (SARS-CoV-2) pandemic. The pandemic and policy responses to curb SARS-CoV-2 cases in the US sharply increased unemployment. We investigate whether there was economic voting in the 2020 election, that is, whether increased unemployment rate during the pandemic negatively affected Trump’s vote. We examine the eight closely contested ‘battleground states’ where Biden won: Arizona, Georgia, Michigan, Minnesota, Wisconsin, Pennsylvania, Nevada, and New Hampshire. Using triple-differences (DDD) models by state (that is, examining changes in a candidate's county-level vote-share depending on the change in unemployment and historical Republican-Democrat partisanship), we find that there was a significant effect of unemployment change in a number of battleground states. We then use these models to predict what would have happened if the increase in unemployment rate in spring 2020 had been half of the actual value. Although Biden would have still secured enough electoral votes to win the presidency according to the predictions, the result would have been far closer.