In a previous article we mooted the idea that the capital markets have been a relatively poor indicator of economic recovery around COVID-19. In this article we analyse real-time spending data from the US market to gauge if and how sectors have been showing signs of recovery. In doing so we hope to shed some light on the progress made. Computing the cross-correlation between spending data and mobility level changes allows us to understand (i) where there is still hesitation in spending in spite of reopening, (ii) which sectors are gaining the most benefit out of reopening, and (iii) what this could mean for the future of the economy, as mobility is increased but restraint from COVID-19 new normals will need to persist.
A. Introduction: US state reopening status', and snapshot of spending patterns post lock-down
Figure: State reopening map for US states as of June 5th 2020. Source: NY Times
Figure: Year on year change in restaurant spending patterns for US states as of 23rd May 2020. Source: Exabel
We further explore the evolution of spending trends across major sectors in the United States, and see how they have evolved dramatically leading up to- and after- their lockdown around mid March from COVID-19.
B. Results from the US market consumer spending data since lockdown to present day
Spending data has been sourced from Exabel, and represents live transaction data from credit and debit card usage. Mobility data is sourced from Google and Apple, blended in a ratio of 80/20 based on all major mobility types.
Figure: overall changes in spending by sector, and changes in mobility (MOB-) over the period from 20 February through to 25 May 2020. For spending data, baseline 100% levels are based on year on year differences, calculated daily; whereas mobility data baselines are based on period before the lockdown started ~February 2020. Source: Exabel (for spending data); Google and Apple (for mobility stats); team analysis. Note 7-day averaging applied to mobility stats to smooth out daily variations; combined spending data is also smoothed.
Overall spending is fortunately increasing with more mobility, and shows a relatively linear 1:1 correlative relationship. It is still however some 75% of previous index levels. It is hoped that as mobility increases back to 100% overall spending would return to previous baseline levels.
As expected, the airline and hotel sector were amongst the hardest hit by lockdowns - bottoming out at 0% and <10% of index respectively. To make matters worse, the airline sector especially is showing little to no signs of recovery. The hotel sector (as we will see later) is showing promising signs of positive traction, but starts from a very low base and therefore is expected to take a far longer time to recover.
Dining out is show slow recovery, this includes fine dining and casual dining which bottomed out at some 7% and 25% of index levels respectively. Casual dining specifically is showing some signs of recovery since reopening, now at 40% of index, however the growth is relatively slow if comparing to other spending categories such as apparel and fast food spending.
Meanwhile, the apparel sector is showing very positive signs of recovery with strong correlation to mobility, and so too are fast food and food delivery. It is clear that these are high-focus areas for consumers since the start of their new found freedoms with lockdowns being eased.
Not shown in this figure is online grocery spending, which is dipping now since reopening (declining trend vs increasing mobility) as consumers are able to reach stores directly for their purchasing. This topped out at some +400% of index levels during the peak of lockdown.
Figure: post lockdown cross correlations across spending and mobility categories. Correlation computation utilises day on day changes in spending and mobility. Period: 28 March to 25 May. Source(s): as above.
Most spending in the US since the lowest mobility period from lockdown has centered around apparel/clothes, dining out, food purchases, and hotels. All these spending categories show a distinct increase from their lowest point since 28 March
On the other hand, spending on groceries, food delivery, and big box retail has dropped in contrast to a rise in overall spending
Overall spending has increased in line with mobility changes. By end May it has risen from 50% at the lowest point, to 80% of index levels.
Figure: post lockdown correlations between spending by sector with 'composite' mobility level. Period: 28 March to 25 May. Composite mobility is combined across Google and Apple, based on our own internal assumptions. Source(s): as above.
Combined/overall spending is well correlated to increase in mobility (0.67x), with R2 fit of 0.955
Other positive correlations with good correlative fits are food delivery (1.34x), apparel (1.15x), casual dining (0.57x), fast food (0.96x), and hotels (0.25x)
Negative correlations are big box and groceries, consistent with a reduction to bring back spending levels to normal baseline pre-lockdown, as these forms of spending had shot up during the lockdown due to panic buying
Moving forward, as mobilities continue increasing, sectors that are expected to benefit are food and dining, as well as apparel spending.
C. What this could mean for the US economy and other economies around the world
There is still discernible apprehension amongst consumers to engage in spending within luxury categories and where there is possibly high inherent risk of infections. The airline and fine dining sectors still at <5% and <10% of index levels, and while the latter is showing shallow signs of improvement, spending in the airline sector is barely moving.
Rather surprisingly big box retail has been negatively correlated to increases in mobility since the bottom most point of lockdown, perhaps due to the sizeable excess stockpiles from panic buying during the lockdown period itself. It is also possible that consumers are opting to focus their new found freedoms on small ticket extravagances/rewards like apparel, casual dining and fast food spending instead.
Overall, the recovery in spending is still slower than what the index levels of capital markets are suggesting. While market indices such as the S&P 500 and Nasdaq index had increased to some 86% and 100% of pre-outbreak index levels respectively as at end May, overall spending was only at 80% versus index after having fallen to a low of 60% during lockdown.
And evidently, spending recovery is varied across the board, with some sectors such as big box retail, apparel, food and dining, and groceries showing high levels of spending; however, the airlines, hotel and fine dining sectors show weak spending levels and slow recovery.
One should also bear in mind that the US government has offered generous economic stimulus benefits to aid the recovery of their economy, which is reflected in the spending patterns we see here. Other countries may not be so fortunate.
What may also be evident is that countries that struggle with increasing mobility back to baseline levels will have a tough time restoring their economies back to previous levels. It is quite clear from this data that mobility is a critical factor in the majority of spending patterns.
For countries where transmission has yet to reach below epidemic levels while mobility is already low; they may inevitably be about to face both a massive healthcare and economic crisis that could last for a long time to come, e.g. Brazil, Mexico, Bangladesh, Indonesia, India and Philippines.
Update 8th June: As a further benchmark comparison, we have sourced data on the breakdown of household expenditure for major categories across a range of countries in the figure below. Sectors such as leisure and transport total some 20-25% of household spending; food, restaurants and hotels some 15-30%, while clothing/apparel and entertainment ranges between 4-10%.
It should be clear how inhibiting movement of people for too long can potentially have dire consequences on a local economy. Transportation alone constitutes some 10-15% of spending, while leisure and recreation another ~10%. Meanwhile, clothing is limited to only 3-8%, and basic food related spending (ex. restaurants and entertainment) is generally only 6-15%.
In short, the current trends we are seeing in the US as it reopens its economy may not be adequate to buoy the economy to a recovery back to previous levels within a short span of time.
Figure: household expenditure by major category for select countries