Risks of In-Person Voting During COVID-19

When the COVID-19 pandemic was declared a national emergency on March 13, the country was in the middle of primary election season. As states franticly adjusted plans for voting, Wisconsin’s state Supreme Court overruled Governor Tony Evers’ executive order to reschedule the election, just one day before the election was set to occur on April 7. In the weeks leading up to the election, the Wisconsin Elections Committee allowed County and Municipal Clerks to consolidate the number of polling locations, in a careful tradeoff between decreasing in-person events while increasing the density of people. As the Nov. 3 general election nears, state Elections Committees must now consider how best to prepare for in-person voting at a potentially larger scale.

To help understand the relationship between in-person voting and virus spread, economists at Ball State University and the University of Wisconsin – Oshkosh have examined the impact of in-person voting in the Wisconsin state primary election on the COVID-19 positive test rate and case counts in Wisconsin. While the Wisconsin Department of Health Services successfully traced dozens of cases to in-person voting, a subsequent statewide analysis found no COVID-19 surge attributable to in-person voting in Wisconsin’s primary. Cotti et al. argued this statewide analysis masked county-level variation in COVID-19 transmission, and ultimately failed to account for how counties in Wisconsin consolidated voting polls differently and have different population densities. Their working paper, released via the National Bureau of Economics Research, used county-level COVID-19 testing data and voting data to determine if there was an increase in the COVID-19 positive testing rate in the weeks after the election for counties that had a higher rate of in-person votes cast per voting location.

For all their models, the authors found that for counties with more in-person voters per voting location, there was a statistically significant increase in COVID-19 cases and the COVID-19 positive test rate. For every 10% increase in in-person voters per voting location, the authors found an associated 18.4% increase in the COVID-19 positive test rate 2-3 weeks after the election occurred. The authors used cellphone data to control for population density and level of social distancing. Even accounting for the Brown County outbreak linked to a meat-packing district, and the outlying population density in Milwaukee, the authors still found an increased in positive case rate in the weeks following the election.

How can the authors be sure that increased COVID-19 transmission was due to in-person activity truly “above and beyond” baseline in-person activity on April 7? They used data from GPS pings in smartphones to posit three reasons:

1) There were significant increases in in-person activity on April 7 in Wisconsin,

2) The activity was focused around polling locations, and

3) The activity was correlated with higher in-person voting density.

These findings have significant policy implications leading up to the general election on Nov. 3. While many states consider consolidating voting locations, research suggests this could lead to a surge in cases. It also reduces turnout, particularly in urban areas. Of note, this study was done before a statewide mask mandate, so additional precautions and mask enforcement may help curb a surge linked to in-person voting. In any case, robust investment in infrastructure related to number of polling locations, social distancing at polling locations, as well as absentee and mail-in voting will be essential to mitigating virus spread on Nov. 3.

---

Cotti, Chad D, Bryan Engelhardt, Joshua Foster, Erik Nesson, and Paul Niekamp. 2020. “The Relationship between In-Person Voting, Consolidated Polling Locations, and Absentee Voting on COVID-19: Evidence from the Wisconsin Primary” SSRN Working Paper 27187. https://doi.org/10.2139/ssrn.3597233.