One of the most significant and unexpected features of the Arkansas economy during the COVID-19 pandemic has been the robust behavior of consumer spending. Despite sharp increases in unemployment and losses in earnings, Arkansas consumers maintained and even exceeded previous trends in overall spending—albeit with notable effects on the composition of purchases. This sustained spending profile contrasts with the behavior of national retail sales, which experienced historic declines in the spring of 2020.
In order to investigate the issue of consumer spending during the pandemic in more detail, researchers at the Arkansas Economic Development Institute have been refining a methodology for extracting retail sales data from reports on county sales tax receipts. The raw data, available monthly for all of Arkansas counties broken down by four-digit industry sectors (NAICS codes), is provided by the Arkansas Department of Finance and Administration: Local Tax Distribution by NAICS.
From the raw data, we extract measures of total retail sales corresponding to the components of the U.S. Retail Trade and Food Services accounts published by the Census Bureau. More details about the methodology are available here: Arkansas Retail Sales—A New Data Set from AEDI. The result is a data set of time series for statewide retail sectors, as well as for county-level totals. For comparison with seasonally-adjusted national data, we implement a rudimentary seasonal-adjustment technique to adjust the statewide data.
Figure 1 displays an overall comparison of consumer spending trends during the pandemic. Data for the U.S. and Arkansas are presented as indexes, normalized so that the second half of 2019 provides a baseline value of 100.
The U.S. Retail Sales data show a sharp decline in early 2020, with sales dropping 21.2% from pre-pandemic levels by April. The Arkansas data show declines in March and April, but the downturn was in the range of only 3 to 4 percent. By May and June, spending was recovering briskly both nationwide and here in Arkansas, but spending in Arkansas rose further above pre-pandemic levels than indicated by the U.S. data: Over the second half of the year, Arkansas retail sales were 8.7% higher than in 2019 while nationwide spending was up only 4.0%.
Spending by Industry Group
Table 1 compares growth rates for components of retail sales for the U.S. and Arkansas. Measuring from December 2019, U.S. Retail and Food Services sales declined 21.5% through April, while the downturn in Arkansas was only 3.1%. In every industry group except for Gasoline Stations and Health & Personal Care Stores, Arkansas retail sales outperformed the nation. By the end of the year, U.S. spending totals were higher than the previous year in 9 of 13 subsectors with ongoing weakness in Electronics and Appliance Stores, Gasoline Stations, Clothing Stores and Food Services. In Arkansas end-of-year sales were down from 2019 in only three sectors, Health and Personal Care Stores, Gasoline Stations and Food Services.
Figure 2 shows the relative sizes of the industry groups for the U.S. and Arkansas in 2019, before the onset of the COVID-19 pandemic. Among the larger subsectors, Arkansas spending tends to be more concentrated in industry groups 447 and 452, Gasoline Stations and General Merchandise Stores. On the other hand, Arkansans spend relatively less on Food and Beverage Stores, Clothing and Clothing Accessories Stores, and Nonstore Retailers (which includes internet shopping).
Figure 3a through 3m, below, compare seasonally-adjusted time series for the U.S. and Arkansas, sector by sector.
Figure 3a shows an important component of consumer spending during the pandemic: Motor Vehicles and Parts Dealers. As described in Arkansas Retail Sales—A New Data Set from AEDI, construction of this component is complicated by the per-vehicle nature of sales taxes at the county level in Arkansas. Consequently, for NAICS sector 441, the Arkansas measure combines two components: Taxable sales at motor vehicle and parts dealers excluding automobiles, plus an estimate of the total value of vehicle sales based on an index of the number of vehicles sold.
Figure 3a shows one of the sharpest divergences between U.S. and Arkansas data. While motor vehicle related sales plunged nationwide in March and April, the downturn in Arkansas lasted only through March. By April, sales in Arkansas exceeded pre-pandemic levels.
Figure 3b displays sales at Furniture and Home Furnishings Stores. This was one of the temporarily hardest-hit sectors nationwide, with sales falling more than 50% in April, then recovering to above per-pandemic levels by July. In Arkansas, the sales decline in April was less than 20%, and sales over the second half of the year have averaged 13.3% above levels of the previous year.
Similarly, sales at Electronics and Appliance stores suffered sharper springtime losses nationwide than in Arkansas, and the nationwide statistics continue to show more weakness in the second half of the year.
Building Materials, Garden Equipment and Supplies, is one of the subsectors that thrived during the pandemic (Figure 3d). Evidently, with additional time on their hands at home, people devoted more time and resources to home-improvement projects. Nationwide, spending increased by more than 15% relative to the previous year, while in Arkansas the sustained increase averaged over 25%, and a peaked at a year-over year increase of 45% in June.
Food and Beverage Stores also saw sales increase during the pandemic. As shown in Figure 3e, consumer stockpiling in March resulted in a surge of over 25% nationwide and 40% in Arkansas. Over the second half of the year, U.S. grocery store sales have remained 10% above pre-pandemic levels, with Arkansas sales trending even higher.
Health and Personal Care Stores (NAICS 446) is one of the few subsectors that experienced a sharper initial downturn in Arkansas than nationwide (although the difference was not significant). As shown in Figure 3f, Arkansas sales surged in the summer months and U.S. sales increased to a level about 5% higher than pre-pandemic levels.
Calculation of sales at gasoline stations is complicated by the fact that gasoline sales are not subject to sales tax. The data from the Local Tax Distribution by NAICS accounts record all non-gasoline sales at gasoline stations. To calculate a statewide measure that includes gasoline, we add non-gasoline sales to the total statewide gasoline sales measure that is constructed as part of the ATSIG data; namely, data on total gallons of motor fuel sales from DF&A is combined with average price per gallon from the Oil Price Information Service to yield an aggregate measure of gasoline sales. The results of these calculations, displayed in Figure 3g, show that gasoline station sales in Arkansas have been running even with the nationwide trends. Although the downturn in gasoline expenditures in 2020 is significant, it reflects lower gasoline prices as much as it does fewer gallons purchased. To some extent, the reductions in gasoline sales frees-up purchasing power for consumers to spend in other sectors.
Nationwide, the subsector that experienced the sharpest declines in the spring of 2020 was Clothing and Clothing Accessories. As shown in Figure 3h, sales in that subsector dropped by more than 80% during the springtime. In Arkansas, the drop was also substantial – over 50%. And while nationwide sales still lag behind the previous year, sales in Arkansas have returned to previous levels.
A collection of specialty stores, NAICS 451 provides another example of a subsector that has benefited from the pandemic, at least over the second half of 2020. Consisting of Sporting Goods, Hobbies, Musical Instrument and Book Stores, sales in subsector 451 declined by 45% nationwide in April but were down less than 10% in Arkansas. As shown in Figure 3i, U.S. sales in this subsector were up significantly from the previous year (between 10% and 20%) during the second half of 2020, but not by as much as in Arkansas.
General Merchandise stores, which include department stores along with “Other” general merchandise stores, appears to have benefited from the initial phase of consumer stockpiling in March, but quickly dropped off. Shown in Figure 3j, general merchandise stores nationwide have experienced somewhat stronger sales over the second half of the year, and have generally been even stronger in Arkansas (except for one outlier month).
Industry Group 453, Miscellaneous Store Retailers, is another category that combines several smaller specialty stores, including florists, office supply stores and used merchandise stores (among others). As shown in Figure 3k, sales in subsector 453 for the entire U.S. followed a pattern of plunging sharply in the spring but recovering back to pre-pandemic levels by the end of the year. In Arkansas, on the other hand, sales were up by approximately 25% by the end of the year, after showing no notable weakness at all during the early months of the year.
Industry group 454, Nonstore Retailers, displays an interesting pattern. As shown in Figure 3l, U.S. sales at nonstore retailers showed no downturn at the onset of the pandemic, and surged 20% to 30% over the course of the year. In Arkansas the 2020 increase was in the range of 40 to 50%. The explanation for this remarkable performance is in the composition of subsector 454: It includes “Electronic Shopping and Mail Order Houses.”
Consumer sales via internet merchants soared during 2020. The data suggest that the share of online shopping before the pandemic was smaller than the U.S. average [see Table 1], but it experienced much sharper increases as Arkansas shoppers caught up.
In Figure 3l, the sharp increase in sales over the second half of 2019 requires some explanation. In July 2019, Act 822 took effect, mandating that out-of-state retailers collect and remit sales and use taxes on internet sales. The associated surge in tax collections doesn’t represent an increase in retail sales; rather, it reflects a broadening of the tax base. The effect was not limited to NAICS sector 454, but it is the most evident there.
The impact of Act 822 creates a discontinuity in the Arkansas Retail Sales series that complicates interpretation of the data, but is largely unavoidable. As part of the seasonal adjustment process, we estimated the magnitude of a July 2019 trend shift for all subsectors. In many cases there was no significant effect, but several sectors did show statistically significant increases. In the case of sector 454, the magnitude of this “internet sales tax effect” was a whopping 64%. Calculated for Total Arkansas Retail and Food Service Sales, the effect is estimated to be 4.8%. Although it is reasonable to expect that actual retail sales did not change after July 2019, counties and municipalities experienced an increase in sales tax revenues that amounted to nearly 5% of retail-related tax collections. Using the same methodology, we estimate that total county and municipal tax collections were boosted by 4.2%. These estimates are preliminary, and will be explored in more detail in further research.
The final subsector to consider is Food Services and Drinking Places. As shown in Figure 3m, sales at restaurants and bars declined sharply in the spring of 2020. U.S. sales were down over 50%, while the decline in Arkansas sales was less than half as large at 24%. Even by the end of the year, nationwide restaurant and bar sales remained significantly lower than a year earlier. In Arkansas it was one of the few sectors that had not recovered to pre-pandemic levels by the end of the year, but the year-over-year comparison was only -2.6%.
The data are necessarily less precise at the county level. Relatively small idiosyncratic factors (as well as the suppression of some data for reasons of confidentiality) introduce sources of random fluctuations in the individual components of county-level retail sales. Consequently, we construct only broad, not-seasonally adjusted measures of retail sales for Arkansas counties.
Two subsectors present specific issues: First, gasoline sales are constructed using statewide data on fuel taxes, so we have no individual county measures for that component. Second, although there is a full set of county data for automobiles, the tax information provides us with only an index of the number of vehicles sold, not their value. Moreover, the taxes on motor vehicles are paid to the county of residence, rather than the county of purchase, introducing an additional complication in interpreting the data.
Accordingly, we present two measures of Retail Sales for Arkansas counties: The first excludes Gasoline Stations and the second excludes Gasoline Stations and Motor Vehicles and Parts Dealers. As shown in Figures 4a and 4b, the statewide measures for these aggregates show the same basic patterns as Total Retail and Food Services Sales (See Figure 1): Namely, the downturn in Spring 2020 is far smaller for Arkansas than for the U.S., and the surge in sales during the second half of 2020 is larger for Arkansas than for the U.S.
Gasoline sales dropped sharply in April and have remained persistently low, so their exclusion results in aggregates with smaller downturn in early 2020 and stronger growth in the second half of the year. Auto sales in Arkansas declined in March 2020 but began accelerating in April. Hence, the exclusion of autos essentially eliminates the March downturn that is evident in Figure 1 for Total Retail Sales.
For both measures, there is little or no overall downturn in April, corresponding to the national retail sales trough, and very little indication of any downturn in either March or April.
Given the statewide performance of these two measures, it is not surprising that very few counties in Arkansas experienced any downturn in early 2020 at all. In particular, if we examine sales in April 2020 (the national trough) relative to prior values, nearly every county experienced significant growth
Figure 5 shows a map that reveals differences across counties in the growth of Retail Sales less Gasoline. The map depicts growth rates calculated as the change from the average for the second half of 2019 through April 2020 (2019:H2 is used as the base for calculating growth rates so as to avoid contamination by the change in internet sales tax regime in July 2019).
In Figure 5, only 8 counties showed negative growth over the period, with 67 counties showing positive growth—with five counties exceeding 20% growth over the period. The collection of counties that experienced negative sales growth is clearly not entirely random, with the subset including Pulaski, Sebastian, Washington, Benton and Craighead – all central counties of Metropolitan Areas.
Figure 6 shows growth rates of Retail Sales less Gas and Autos. A similar pattern is clear, but after exclusion of the surge in auto sales in April, several more counties show negative growth in sales (with 15 counties showing declines). The counties showing negative growth are not exclusively metropolitan counties, but there is a clear tendency for the declining counties to contain urban areas. Without the strong auto sales included, growth rates are generally lower, with only 3 counties showing growth rates of over 20%.
Over the second half of 2020, strength in consumer spending prevailed across the state. As shown in Figure 7, county growth rates of Retail and Food Service sales less Gas and Auto ranged from 4.5% in Franklin County to 38.9% in Woodruff County.
The COVID-19 pandemic, and efforts to contain it, have disrupted economic activity in many sectors of the economy. Most directly affected has been retail spending, where business closures, capacity reductions and social distancing have had unequivocal effects on consumer behavior and spending patterns. Nevertheless, consumer spending in Arkansas remained relatively robust.
During the spring of 2020, Arkansas experienced fewer restrictions on commerce than many other states. We were one of the few states that did not go through a period of mandatory stay-at-home or shelter-in-place orders. Consequently, while consumers and businesses had to adapt to changes on protocols, commerce continued. The contribution of federal support and stimulus payments has also helped support the consumer sector. Arkansas is a relatively low wage and low cost-of-living state, so direct payments and unemployment insurance extensions have given Arkansans’ an extra boost in purchasing power.
There is some evidence that consumer spending was slowing toward the end of 2020, but two federal Covid-relief bills were signed into law on December 27, 2020 and March 11, 2021. Additional stimulus payments and unemployment insurance expansions are expected to provide additional purchasing power to fuel continued strength in consumer spending. With inoculations proceeding and social distancing rules being relaxed, consumers should also find more opportunities for spending. Consequently, the outlook is for retail sales to continue to drive economic activity in coming months. We will continue to track Arkansas Retail Sales to monitor and document these trends.
# # #
Documentation of Methodology is available here: Arkansas Retail Sales—A New Data Set from AEDI.
Data for Arkansas Retail and Food Service Sales for July 2017 through December 2020 are available in an Excel Spreadsheet: Arkansas-Retail-Sales-Dec-2020.
The data set includes statewide aggregates and components, both seasonally adjusted and not-seasonally adjusted. County-level data for Total Retail and Food Service Sales less Gasoline are available on a not-seasonally adjusted basis.
A PDF File of this Report is available HERE.