Is the United States in a recession? We asked our data scientists to look at foot traffic trends as of the end of June 2022 in light of their correlation with traditional economic indicators of recession. Here’s what we found.
With fresh foot traffic data as a proxy for spending and the financial crisis of 2008 and 2009 as a point of reference, we begin our 2022 ground truth analysis with a 2014 report from the U.S. Bureau of Labor and Statistics.
The starting point: Correlating spending and foot traffic
In a 2014 issue of its Beyond the Numbers publication, the United States Bureau of Labor Statistics (BLS) analyzes consumer spending by sector before, during and after a recession, using the period of 2007 to 2013 as a timeframe.
The objective of the report is to assess the relative state of recession at a given point in time by using core economic indicators – we’ll use this same established baseline as a jumping-off point for our analysis here.
In our analysis, the actual value of the measurement is not particularly important. It is a unitless ratio measure, and each is formatted so that a declining value indicates a worsening of conditions. In addition, the ratios have been differenced year-over-year to remove potential seasonal effects and summarized at a national level.
Of the potential recession indicators noted by BLS, there are three we can easily analyze with Unacast data, where foot traffic acts as a proxy for spending. Those indicators are:
- Eating in vs. Eating Out
- Full Service vs. Limited Service Restaurants
- New vs. Used Cars
Let’s begin looking at the relative importance of each indicator and examine how things went the last time the United States was in a recession.
BLS Relative Importance (RI) of Category in Consumer Price Index
As measured by the BLS study, the relative importance (RI) of the food spending component of the CPI was up about half a point during the period of recession. This returned to previous levels within about two years.
Food at home (grocers) grew a full point during the period of recession, while only declining a half point in the wake of the recession. Food away from home (restaurants), on the other hand, lost strength during the last recession and had not yet recovered two years later, though limited service establishments did recover better than full service restaurants.
BLS foot traffic correlations and an exclusion
Here are the correlated results displayed over a 30 month period vis a vis the 2014 BLS data.
New vs. Used Cars
Though our correlation with BLS was strong on this indicator, we’ve decided to exclude it from results, at least for now. The global supply chain crisis and resultant shortage of chips available has crippled the current automotive industry, placing a premium on used vehicles. The same condition was not true in 2011, so we weren’t comfortable comparing apples to apples.
There is also the matter of a sharp rise in traffic as indicated at the end of July 2022, which does not seem reflective of the performance of the other indicators we studied. We suspect this inclination will flatten when we look at it again in a 30-day lag data set.
For the record, the BLS range as we measured for the New vs Used Cars was 2.63 to 1.67, whereas the normalized Unacast’s range is 2.40 to 1.29. While the correlation is strong, we are parking this metric until such time as we can vet it more fully.
Full Service Restaurants vs. Limited Service Restaurants
The normalized 2014 BLS range is 1.25 to 1.17. The top end of the Unacast range (1.57) is well-correlated, but the bottom end (0.51), seems a bit low. This may have something to do with a Covid-era variance in restaurant dwell times where a reduction in dwell time did not necessarily mean a reduction in spending, or the inverse.
Eat In (Grocery Stores) vs. Eat Out (Restaurants)
This is an important indicator. Average annual food-at-home prices were 3.5 percent higher in 2021 than in 2020. For context, the 20-year historical level of retail food price inflation is 2.0 percent per year—meaning the 2021 increase was 75 percent above average.
Restaurant prices are also rising as the supply chain is squeezed. So, it’s more expensive to eat at home, but it’s also more expensive to eat out. Which are Americans doing more of as we are (or are not) entering a recession?
After Inversion, the BLS range for this indicator was 0.81 to 0.65, whereas the Unacast range is 0.83 to 0.42. While the lower end of the range differs, BLS and Unacast agree that eating-in (grocery stores) are on the rise vs. restaurants in mid-2022.
Is that changing foot traffic patterns for grocers? Based on further analysis, it seems so.
Groceries and Income Level
How much we earn has a lot to do with where we shop for groceries. The plot below looks at the average income level of shoppers (non-workers) at selected grocery stores. Stores were chosen based on being in at least 40 of the 50 states.
The metric should not be taken to be a measure of actual income in dollars. It is instead a unitless measure on a 0-5 scale of the income level. The levels are as follows:
- Income < $24,999
- $25,000 to $49,999
- $50,000 to $74,999
- $75,000 to $99,999
- $100,000 to $124,999
- More than $125,000
The equation for the average calculation based on our data schema is:
The general idea behind this measure is based on the combination of two factors. Firstly, income is bracketed by buckets and secondly, grocery stores are also bucketed into levels of cost. Given these two factors, we hypothesize that in times of economic crisis, this measure of income should increase across all levels of cost.
Consider this simplified example – Store A is an expensive store and Store B is a discount store. Rich customers are given a value of 2, middle class customers a value of 1 and poor customers a value of 0. The score is the simple average of the row. Before the recession, the scores are as follows:
After an economic downturn, customers will likely begin to shop at the discount store in order to conserve money. However poor and middle class customers are more likely to feel the impact of the recession stronger and therefore move to the discount store. After the recession the breakdown looks like this.
One of the rich, most of the middle class, and all of the poor customers have moved to the discount store. The income score has increased for BOTH of the stores. Therefore seeing the income level increase across ALL STORES can be seen as a potentially negative signal.
Retail Dwell Times
The time spent inside of stores is a pretty close approximation for the likelihood of a customer purchasing something. Increasing time spent in stores is a positive sign, decreasing time spent is negative.
The locations chosen for inclusion in the plot below are limited to only those NAICs codes that are strongly associated with retail.
While the tier of visits with the shortest visit duration remains flat as of the end of July 2022, dwell times at retail stores among the median (p50) and longer visit (p75) tiers are both inclining.
Used a simple proxy then, foot traffic patterns indicate spending at retail stores is up – a negative indicator of recession. That said, increasing shopper income level may be a negative indicator especially at discount stores
So is the United States in a recession? The signals are mixed.
Our correlation analysis based on the BLS data from the 2014 report shows strength for grocers in Eat in vs. Eat out, a Full Service / Limited Service restaurant foot traffic ratio that is trending negative, and New vs. Used Car results we don’t trust yet to make a call.
Average income levels of Grocery Store visitors are up across the board – remember that we hypothesize that this is a negative indicator resulting from more people being concerned about food costs.
Finally, retail dwell times are trending up, if slowly, indicating that either a) a recession has yet to really hit the U.S., or b) consumers have yet to make some of the spending shifts associated with a state of recession.
- National Bureau of Economic Research (NBER)
- United States Bureau of Labor Statistics
- Ground Truth | Foot Traffic Data
- Can Foot Traffic Predict Quarterly Revenue?
- Location Intelligence and Mobility Data
- Retail Site Selection and Location Data
- Consumer Trends and Economic Growth
- Unemployment Rates and Gross Domestic Product (GDP)