Changes in Consumption in the Early COVID-19 Era: Zip-Code Level Evidence from the U.S.
One sentence summary: Spending on goods and services that can (cannot) be consumed at home has increased (decreased) amid COVID-19.
The corresponding academic paper by Hakan Yilmazkuday has been accepted for publication at Journal of Risk and Financial Management.
The working paper version is available here.
Abstract
Using monthly zip-code level data on credit card transactions covering 16 U.S. cities, this paper investigates changes in consumption at local commercial places during the early coronavirus disease 2019 (COVID-19) era. Since using aggregate-level data can suppress valuable information on consumption patterns coming from zip codes, the main contribution is achieved by estimating common factors across zip codes that are controlled for factors that are zip-code and time specific as well as those that are zip-code and sector specific. The estimation results based on common factors across zip codes show that relative consumption of products and services that can be consumed at home (e.g., grocery, pharmacy, home maintenance) has increased up to 56% amid COVID-19 compared to the previous year, whereas relative consumption of products and services that cannot be consumed at home (e.g., fuel, transportation, personal care services, restaurant) has decreased up to 51%. Similarly, after controlling for the corresponding factors, online shopping has increased up to 21%, while its expenditure share has increased by up to 16% compared to the pre-COVID-19 period.
Non-technical Summary
Consumption within the U.S. is reduced significantly due to the coronavirus disease 2019 (COVID-19). This reduction has been through both the direct impact of COVID-19 due to lockdowns or social distancing and its indirect impact through financial market shocks and their effects on the real economy. The nationwide consumption fall in the U.S. is also evident widely across sectors (except for grocery) and especially for products purchased through offline (rather than online) shopping.
Abstract
Using monthly zip-code level data on credit card transactions covering 16 U.S. cities, this paper investigates changes in consumption at local commercial places during the early coronavirus disease 2019 (COVID-19) era. Since using aggregate-level data can suppress valuable information on consumption patterns coming from zip codes, the main contribution is achieved by estimating common factors across zip codes that are controlled for factors that are zip-code and time specific as well as those that are zip-code and sector specific. The estimation results based on common factors across zip codes show that relative consumption of products and services that can be consumed at home (e.g., grocery, pharmacy, home maintenance) has increased up to 56% amid COVID-19 compared to the previous year, whereas relative consumption of products and services that cannot be consumed at home (e.g., fuel, transportation, personal care services, restaurant) has decreased up to 51%. Similarly, after controlling for the corresponding factors, online shopping has increased up to 21%, while its expenditure share has increased by up to 16% compared to the pre-COVID-19 period.
Consumption within the U.S. is reduced significantly due to the coronavirus disease 2019 (COVID-19). This reduction has been through both the direct impact of COVID-19 due to lockdowns or social distancing and its indirect impact through financial market shocks and their effects on the real economy. The nationwide consumption fall in the U.S. is also evident widely across sectors (except for grocery) and especially for products purchased through offline (rather than online) shopping.
However, such a nationwide observation 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. For example, when zip codes are considered, spending on a particular sector may increase in one zip code, whereas it may decrease in another, resulting in no significant impact at the aggregate level. Therefore, using data from more disaggregated areas is important to understand the changes in consumption patterns amid COVID-19.
Based on this motivation, this paper investigates sector-level as well as online versus offline consumption patterns within the U.S. by using monthly zip-code level data (covering 16 U.S. cities) on credit card transactions for local commercial purchases. The main strategy is to identify common factors across zip codes representing sector-level or online versus offline consumption patterns at the U.S. national level that do not suffer from an aggregation problem. This is achieved by estimating sector-time fixed effects or shopping channel-time fixed effects in the monthly zip-code level data, where factors that are zip-code and time specific as well as those that are zip-code and sector specific are controlled for.
The results based on the sector-level data show that consumption of products and services that can be consumed at home (e.g., grocery, pharmacy, home maintenance) have increased by up to 56% during the lockdown period starting from March 2020, whereas consumption of products and services that cannot be consumed at home (e.g., fuel, transportation, personal care services, restaurant) have decreased by up to 51%.
This result is analogous to the one that has been used to explain the reduction in economic activity, unemployment or social distancing experience by workers' ability of working from home. The difference in this paper is that it is consuming at home that can be connected to the sectoral heterogeneity in consumption changes amid COVID-19.
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.
The corresponding academic paper by Hakan Yilmazkuday has been accepted for publication at Journal of Risk and Financial Management.