|
Related Topics
|
Category: Covid-19, consumer spending and economic analysis
Date: 17 May 2020 In this article we take a look at a research paper from the Bureau of Economic Analysis (BEA) in which they take a look at the impact of Covid-19 on consumer spending using card transactions.
Overall, we find the largest effects of this pandemic for sectors such as accommodations and restaurants, which by the second week of March, show declines of around 80 percent and 70 percent, respectively. " |
Introduction into the research paper
Since it started in March 2020, the COVID-19 pandemic has dramatically impacted the U.S. economy. In the month after its declaration, as social distancing policies have taken effect, businesses have closed, and over 22 million workers have sought unemployment benefits. To help policymakers develop a course for economic recovery, it is important to use the most timely and accurate economic information available. This is especially pertinent given that pandemic-driven economic changes are measurable in days and weeks, rather than months and years. We present estimates of consumer spending using card transaction data that are updated daily with a lag of about 3 days. The underlying data used to construct these series are collected by Fiserv, one of the largest card intermediaries in the country. Each observation in the data corresponds to a single card swipe (for example, debit card, credit card, or gift card), although all data are aggregated to the state and national levels and thus anonymized. While alternative data cannot offer a replacement for carefully conducted surveys, these data offer an important complementary source of information to traditional survey source data.
We use daily card data from the Fiserv series to measure the reduction in revenue around the time of the pandemic. This includes both the immediate effects of the pandemic and measures of the total effects for March. Overall, we find the largest effects of this pandemic for sectors such as accommodations and restaurants, which by the second week of March, show declines of around 80 percent and 70 percent, respectively. However, these declines were partly offset by the 100 percent immediate increase in food and beverage store sales (likely a result of households stocking up on goods in early March). In aggregate, for the month of March, we find about a 5 percent decline for the combined retail and food services sector, roughly matching the estimates of the advance MRTS for March for similar categories. The decline for March appears to be even larger for other nonretail categories, including categories like accommodations, recreation, and ambulatory health care. However, the effects for spending in March understate the full effect of the pandemic, as the main effects didn’t occur until the middle of the month. We estimate an aggregate “pandemic effect”—the effect of the pandemic on consumer spending after mitigation measures have had time to take hold—of around –27.8 percent.
We use daily card data from the Fiserv series to measure the reduction in revenue around the time of the pandemic. This includes both the immediate effects of the pandemic and measures of the total effects for March. Overall, we find the largest effects of this pandemic for sectors such as accommodations and restaurants, which by the second week of March, show declines of around 80 percent and 70 percent, respectively. However, these declines were partly offset by the 100 percent immediate increase in food and beverage store sales (likely a result of households stocking up on goods in early March). In aggregate, for the month of March, we find about a 5 percent decline for the combined retail and food services sector, roughly matching the estimates of the advance MRTS for March for similar categories. The decline for March appears to be even larger for other nonretail categories, including categories like accommodations, recreation, and ambulatory health care. However, the effects for spending in March understate the full effect of the pandemic, as the main effects didn’t occur until the middle of the month. We estimate an aggregate “pandemic effect”—the effect of the pandemic on consumer spending after mitigation measures have had time to take hold—of around –27.8 percent.
Hardest Hit Industries
The two images below show the effects on restaurants and accommodations, which are two industries hit particularly hard by the COVID-19 pandemic. Near the end of March, the full decline in sales is around 70 percent for restaurants and around 80 percent for accommodations. A feature that stands out for the restaurant category is a post-pandemic weekend effect. Prior to the pandemic, there was a spike in sales over the weekend, which is removed by our day-of-week controls. However, post-pandemic, these weekend sales fall flat, relative to pre-pandemic levels, leading the counterfactual change in sales to fall over the weekends.
Other relevant sectors
Not every industry was subject to declines such as those observed in restaurants and accommodations. The food and beverage sector (image below), which includes grocery stores, experienced a particularly large increase in sales of about 100 percent around the time of the event, followed by a rapid fall, with sales falling below expectations the last few days of March.
The effects on gas stations are more nuanced and appear to be marginal immediately following the pandemic, followed by a decline around a week after March 11, with lows of around 40 percent. About half of this decline is likely due to declining gas prices of around 20 percent in March, but the remaining decline is likely due to a reduction in travel, as more individuals remained home.
Food and Beverages |
Gas stations |
State level estimates
The Fiserv series also includes state-level data by industry. Using the state-level data, we can look at how consumer spending in the different states has been affected by the pandemic. To do this, we repeat the same procedure that we used for the national data but at the state level for each state independently. We have chosen to study NAICS industry 722 (restaurants), which showed large declines after the beginning of the COVID-19 pandemic at the national level. We cannot show point estimates and standard errors for each state in a convenient format, so instead, we show the average of the point estimates across all states for each period from February 23 to the most recent date for which we have data available. We also show the 10th and 90th percentiles across states for each time period (noting that the approximately 40 states in which estimates fall between these two lines may change from day to day). Finally, Washington and New York, two states known to have been affected by the COVID-19 pandemic more than other states, but at different times, are shown for reference below.
The image above demonstrates two things. First, it demonstrates how a majority of states in the country follow a similar pattern to the average state, as the 10th and 90th percentiles are relatively close to the mean. This pattern coincides with a dramatic increase in news coverage on the topic across several states in the country starting around March 11 (Baker and others 2020). Second, it shows how states such as New York and Washington differ from national trends. These states are of particular interest for different reasons. For Washington state, we may expect the economic impact of the pandemic to precede the impact on the rest of the country, as Washington was the first state to report a COVID-19 case. New York, on the other hand, is a high-density area that has been particularly hard hit by the pandemic, but it was not affected at a notably early date. Our preliminary findings are consistent with these patterns. Focusing on food service and drinking places, we find the dollars spent for these industries declined in Washington state prior to the rest of the country.