, , , ,

The Death of Retail Is Greatly Exaggerated

The last few years have seen a bevy of high-profile retailers either go bankrupt or close stores across the country, raising concern that brick-and-mortar retailers are doomed. It’s undoubtedly true that consumers are buying more and more goods online, and that’s putting pressure on some retailers. However, it’s important to put this in context: So far, the overall decline of brick-and-mortar retail is a relatively minor change in the economy, and not a major structural disruption.

It’s not hard to see why the “death of retail” story has gotten so much traction in the anecdote-driven news world. It’s easy to find examples across the country of empty malls, vacant big boxes, and abandoned retail strips that create the impression that the demand for brick-and-mortar shopping has cratered. However, the data tell a different story.

Still near historical high

In the aggregate, it’s true that employment has flat-lined at brick-and-mortar retailers, which we define as NAICS codes 44 and 45 minus non-store retailers. However, despite the weak recent growth, brick-and-mortar retail employment is still close to a historical high at 15.3 million, falling only 22,000 jobs below the peak reached in 2017.

The historically high number of retail jobs overall does mask one negative trend: The growth of retail has not kept up with the rest of the economy. At the peak in the mid-1980s, retail made up 11.8% of overall employment. By 2017, it had fallen to 10.4%. So is this the much trumpeted death of retail? This is overblown for two reasons.

First, this brings retail employment’s share of the labor market just back to 1970s levels. Retail may not be keeping up with the rest of the economy, but historically speaking we aren’t entering new territory, either.

Second, it’s useful to place this into a broader context by comparing it to two other industries that illustrate what a major structural change in the economy actually looks like: the decline of manufacturing and the rise of healthcare. As a share of employment, retail looks relatively flat over the past 50 years compared with these industries, generally fluctuating between 10% and 12%. In contrast, automation and globalization have pushed the manufacturing share of employment from 25% in 1970 to less than 9% today. On the other side of the ledger, the growth of healthcare spending has pushed healthcare employment from 5% of jobs to 13%.

The changes retail is undergoing also look relatively modest if we look at the trends since 2000. Over this period, the brick-and-mortar retail share of employment has declined by 0.8 percentage point, while healthcare has increased by 3.6 percentage points and manufacturing has declined by 4.6 percentage points.

The decline in brick-and-mortar retail has been stronger among some segments, like department stores. However, this has not been mostly about a contraction in employment, but a shift to e-commerce. Employment gains in e-commerce are visible in warehousing and nonstore retailers, the latter of which includes e-commerce sellers like Amazon. Over the last decade, non-store retailers have added 157,000 jobs and warehousing has added almost 369,000, which combined more than offset the job losses of 392,000 in department stores.

Despite the positive aggregate statistics, it is undoubtedly true that there are some parts of the country that have seen a decline in retail. These are the places with vacant malls and empty retail strips that drive the “death of retail.” For example, in a quarter of the metro areas we examined, retail employment has declined by 5% or more over the last decade, and 6% of metro areas have seen double-digit declines. If e-commerce is not killing retail, then what is happening in the places that are losing retail jobs?

Look at fundamentals

The change in retail employment at the metro level over the past decade is explained by fundamentals. Using the change in population alone, we can explain more than half of the variation in retail job growth over the last decade at the metro level. Including other economic measures, like total non-retail job growth and the change in lagged unemployment rate, improves the fit slightly more.

In fact, the pattern of retail job loss and growth across metro areas were basically what would be expected based on past trends. To show this, first the change in retail employment from 1997 to 2007 is regressed on population growth, lagged unemployment rate change, and non-retail job growth over the same period. Using this model we can predict metro level retail growth from 2007 to 2017 using the coefficients estimated from 1997 to 2007 combined with the actual population growth and economic conditions from 2007 to 2017. This predicted change in retail employment has a 0.74 correlation with the actual change in retail employment, and can explain 54% of the variation in it. In other words, retail has fallen where you would expect given changes in the underlying fundamentals.

This analysis suggests that when there is an empty mall or abandoned big box store, there is a good chance that it can be explained by lack of population growth or a weak economy. While e-commerce is undoubtedly growing and has been a factor in the closing of some retailers, the brick-and-mortar retail sector overall is not undergoing the kind of structural job loss seen, for example, in manufacturing. Instead, retail is undergoing a process of gradual change that is necessary in a dynamic economy, but it is not yet a major disruption or cause for concern.