The Early Impact of COVID-19 on Local Commerce: Changes in Spend Across Neighborhoods and Online

Abstract

We document a number of striking features about the initial impact of the pandemic on local commerce across 16 US cities. There are two novel contributions from this analysis: exploration of neighborhood-level effects and shifts between offline and online purchasing channels. In our analysis we use approximately 450 million credit card transactions per month from a rolling sample of 11 million anonymized customers between October 2019 and March 2020. Across the 16 cities we profile, consumers decreased spend on the set of goods and services we define as "local commerce" by 12.8% between March 2019 and March 2020. Growth in all 16 cities was negative. Consumers shifted a substantial share of local commerce spend online, such that year-over-year growth in online spend was small, but positive, at 1.5%. With respect to grocery and pharmacy purchases, online spend grew at least three times as fast as offline spend. Overall spend declines were uniform across neighborhoods of differing median household income, though lower-income neighborhoods experienced the highest proportion of extreme negative declines. We also find evidence that many low-income neighborhoods are increasing spend on online grocery slower than others, but increasing their use of online restaurants the fastest. Consumers in low-income neighborhoods also tend to live farther from the grocery stores at which they shop. Compared to their counterparts in higher-income neighborhoods, consumers in low-income neighborhoods have not been more likely to shop at grocery stores closer to where they live since the onset of the pandemic. * We thank James Duguid and Bryan Kim for their substantial contributions to the production and analysis of the research presented here. We also thank colleagues at the JPMorgan Chase Institute for their comments and suggestions. This research was made possible by a data-use agreement between Lindsay E. Relihan and the JPMorgan Chase Institute (JPMCI), which has created de-identified data assets that are selectively available to be used for academic research. All statistics from JPMCI data reflect observations based on at least 100 customer accounts with medians reported with small errors to protect privacy where appropriate. The opinions expressed are those of the authors alone and do not represent the views of JPMorgan Chase & Co. Contact info for Lindsay E.

Changes in Consumption in the Early COVID-19 Era: Zip-Code Level Evidence from the U.S.

“…The results based on online versus offline shopping show that online shopping has increased by up to 21%, while its expenditure share has increased by up to 16% compared to the pre-COVID-19 period. These results are consistent with those in Relihan et al (2020) who have used an earlier version of the dataset used in this paper and shown that the increase in online shopping has been only about 1.5% in March 2020 (with respect to March 2019). Nevertheless, different from Relihan et al (2020) who have focused on aggregate-level data in the U.S. to obtain this measure, this paper has shown that, after controlling for factors that are zip-code and time specific as well as those that are zip-code and sector specific, the common factor across zip codes representing online spending within the U.S. has increased by about 7.4% in March 2020 (with respect to March 2019).…”

“…These results are consistent with those in Relihan et al (2020) who have used an earlier version of the dataset used in this paper and shown that the increase in online shopping has been only about 1.5% in March 2020 (with respect to March 2019). Nevertheless, different from Relihan et al (2020) who have focused on aggregate-level data in the U.S. to obtain this measure, this paper has shown that, after controlling for factors that are zip-code and time specific as well as those that are zip-code and sector specific, the common factor across zip codes representing online spending within the U.S. has increased by about 7.4% in March 2020 (with respect to March 2019). This difference can again be attributed to estimating common factor across zip codes in this paper as opposed to using aggregate-level data.…”

“…Regarding online versus offline consumption, studies such as by Carvalho et al (2020) have shown that offline shopping has declined much more than online shipping, whereas studies such as by Chen et al (2021) have shown that offline shopping has reduced by about 70% following COVID-19. Similarly, Relihan et al (2020) have shown that the increase in online shopping has been only about 1.5% in March 2020 (with respect to March 2019). Different from these studies focusing on the (geographically) aggregate level data covering online versus offline shopping or the corresponding heterogeneity, this paper focuses on common factors across zip codes to avoid any aggregation bias.…”

“…Consumption within the U.S. is reduced significantly due to the coronavirus disease 2019 . As shown by studies such as by Baker et al (2020b) or Relihan et al (2020), this reduction is evident widely across sectors (except for grocery) and especially for products purchased through offline (rather than online) shopping. However, these observations based on nationwide data can easily suppress valuable information on consumption patterns coming from more disaggregated areas as their effects may cancel each other out during the aggregation process.…”

The dynamics of food shopping behavior: Exploring travel patterns in low-income Detroit neighborhoods experiencing extreme disinvestment using agent-based modeling

“…With lowincomes, limited private automobile access, reliance on public transit for commuting over long distances, and the rapid spread of COVID-19 infections across Detroit, there was a severe overall impact on accessing stores, and this was reflected in local commercial spending. Following the COVID-19 outbreak, the City of Detroit confronted the fourth sharpest decline among major U.S. cities in local shopping, behind San Francisco, Chicago and New York, reflective, in part, on the overall mobility, sociodemographic composition, and health vulnerabilities of the population [57].…”