Using Google Data to Estimate the Effects of Regional Mobility on Daily COVID-19 Cases: Evidence from Ontario Public Health Units

Abstract

From September 2020 onwards, COVID-19 cases have rapidly increased across Canada. This study estimates the effects of Google population mobility indicators on daily COVID-19 cases to evaluate the impacts of public movements across different regions in Ontario. We focus on Ontario as Google mobility data are available for Public Health Units (PHUs) for that province. Results based on pooled data from May 1st – November 15th imply that higher mobility at retail stores is significantly correlated with an increase in daily COVID-19 cases. However, empirical estimates from individual PHU level time-series models reveal regional differences, as these findings are based primarily on the relationship between retail mobility and daily cases for the Public Health Units (PHUs) of Toronto and Peel. These results support the implementation of region-specific lockdowns. Further, different specifications generate daily COVID-19 forecasts for Peel and Toronto that are on average, approximately 6%-9% different from actual values. The models of this research should be of value to local health authorities who are in search of simple models that are not computationally intensive and are capable of generating reliable forecasts for specific regions.