I have consistently argued that ecommerce is boosting employment by creating jobs at fulfillment centers. For example, over the past year, ecommerce jobs have risen by 61,000, while brick-and-mortar retail has fallen by only 7,000. That sounds like a counter-intuitive result, given that ecommerce is supposedly more productive than brick-and-mortar retail.

But the increase in paid jobs is much easier to understand if you realize that shopping for goods is actually the result of two inputs: paid market work by employees and unpaid time by households, in the form of driving to the store, parking, wandering through the aisles, checking out, driving home.

According to the American Time Use Survey from the Bureau of Labor Statistics, in 2015-2016 Americans spent .645 hours per day on average shopping for consumer goods or traveling to shopping, or 4.5 hours per week. Since there are 260 million Americans aged 15 and over, that means Americans spend approximately 1.2 billion hours a week shopping for consumer goods or traveling to shopping (that’s the equivalent of 30 million full-time jobs).

By comparison, in 2006-2007 Americans spent 4.75 hours per week shopping for consumer goods or traveling to shopping, or 0.25 more. That extra quarter hour corresponds to 64 million extra hours per week (260 million x .25). So because of the increase in ecommerce over the past 9 years, American households save 64 million hours per week, or the equivalent of 1.6 million full-time jobs.

Some of these jobs are being moved into the market sector: The fulfillment center workers do the aisle-cruising that shoppers used to do themselves, the truck drivers take the place of the consumers driving back and forth to the mall.

This also implies that retail productivity is being unmeasured, since we’re not counting the reduction of household hours. I don’t have an estimate yet, but the undermeasurement could be substantial.

 

When I was young, oh so many years ago, my parents would take me shopping at Korvette’s, a chain of discount department stores originally based in New York City. But alas, Korvette’s went bankrupt in 1980, just one of hundreds of names on Wikipedia’s long list of defunct department stores. Then, of course, Wikipedia has an equally long list of defunct retailers of the United States, including such stalwarts as Robert Hall (whose jingle is still stuck in my mind*).

Now, of course, we have apparently entered the era of the “retail apocalypse,” a newly minted calamity which has its own Wikipedia entry. Suddenly all of our shopping malls and big box stores are supposed to disappear, washed away by the ever-rising tide of ecommerce.

But at least so far, the data shows very little signs of an apocalypse. According to the latest BLS employment report, employment at brick-and-mortar retail is down by only 7,000 jobs over the past year, a mere rounding error for an industry that employs over 15 million workers. At the same time, employment in the ecommerce sector has rise by 61,000 over the past year, more than making up for any erosion in brick-and-mortar retail.

Indeed, if we look at three month averages, brick-and-mortar retail looks more or less flat. Job losses in some retail sectors, such as general merchandise and clothing stores, is being partly offset by employment gains in motor vehicles, auto parts, building materials, and health stores. Companies announce store closings, hoping to get a better deal out of landlords–but for the same reason, stores that are doing well don’t announce that they are expanding.

So far this looks like an evolution rather than a revolution: Some retailers are closing, while others (including online sellers) are expanding. The big news is that fulfillment centers pay 30-40% more than brick-and-mortar jobs in the same area. But more about that in another post.

 

*”Where the values go up, up ,up

And the prices go down, down, down

Robert Hall will show you

The reason they give you

High quality, economy!”

I’ve been writing a lot about ecommerce lately. I would be remiss if I did not address the Amazon offer to buy Whole Foods. I have two thoughts.

First, I’m not an antitrust economist. But it strikes me that under ordinary antitrust logic, Amazon’s purchase of Whole Foods would ring no alarm bells. Despite the name recognition of Whole Foods, the company’s $16 billion in sales last year pales next to Kroger’s $115 billion in sales or Albertson’s roughly $60 billion in sales. If anything, the purchase increases competition in the grocery business.

Second, what is Amazon likely to do with Whole Foods? I’m not privy to Amazon’s strategic thinking, of course. But the main point of ecommerce, when done right, is to reduce the time consumers spend shopping—the drive to the store, the search for parking, the endless trudging through the aisles. In turn, ecommerce companies have been shifting those unpaid hours of household labor into the market sector, creating decent-paid jobs in fulfillment centers and electronic shopping operations.

By my calculations, the shift to ecommerce over the past nine years has saved American households roughly 64 million hours per week in reduced shopping time, or the equivalent of 1.6 million full-time jobs. (yes, that million is right…calculations below).

Some of those hours of unpaid household labor were shifted to the market sector. Over the same period, the number of ecommerce jobs rose by almost 400,000, as fulfillment center workers and drivers took over the tasks that consumers used to do. Brick-and-mortar jobs dropped by 76,000, so that was a net plus for job creation.

If these trends continue, Amazon and other ecommerce companies will try to make shopping for meals easier and less time-consuming for American households, and generate more paid jobs in the process.

****

Calculations

According to the American Time Use Survey from the Bureau of Labor Statistics, in 2015-2016 Americans spent .645 hours per day on average shopping for consumer goods or traveling to shopping, or 4.5 hours per week. Since there are 260 million Americans aged 15 and over, that means Americans spend approximately 1.2 billion hours a week shopping for consumer goods or traveling to shopping (that’s the equivalent of 30 million full-time jobs).

By comparison, in 2006-2007 Americans spent 4.75 hours per week shopping for consumer goods or traveling to shopping, or 0.25 more. That extra quarter hour corresponds to 64 million extra hours per week (260 million x .25). So because of the increase in ecommerce over the past 9 years, American households save 64 million hours per week, or the equivalent of 1.6 million full-time jobs.

Note: I get an almost identical result based on the magnitude of the decline in  brick-and-mortar’s share of retail sales.

 

Michael Mandel, chief economic strategist at the Progressive Policy Institute, joined Alphachat to talk about his report, “The Coming Productivity Boom”, co-authored by Bret Swanson of Entropy Economics.

Mandel argues that the decades-long productivity stagnation will end once companies in the “physical” industries — transportation, construction, manufacturing, healthcare, wholesale and retail trade — start investing in information technology the way that companies in the digital industries have.

From the summary of their paper:

The digital industries, which account for around 25% of U.S. private-sector employment and 30% of private-sector GDP, make 70% of all private-sector investments in information technology. The physical industries, which are 75% of private-sector employment and 70% of private-sector GDP, make just 30% of the investments in information technology.…

Information technologies make existing processes more efficient. More importantly, however, creative deployment of IT empowers entirely new business models and processes, new products, services, and platforms. It promotes more competitive differentiation. The digital industries have embraced and benefited from scalable platforms, such as the Web and the smartphone, which sparked additional entrepreneurial explosions of variety and experimentation. The physical industries, by and large, have not. They have deployed comparatively little IT, and where they have done so, it has been focused on efficiency, not innovation and new scalable platforms. That’s about to change.

Mandel tells me why he believes that productivity pessimists such as Robert Gordon are wrong to extrapolate from current trends, why he isn’t worried by the increasing market concentration in the leading companies in many economic sectors, and how public policy can play a role in facilitating this forthcoming productivity boom. And in a speed round, we also discuss the specific ways in which physical industries can be transformed by more IT investment, plus which technologies hold the most promise.

For the full interview, listen here at Financial Times. 

On June 7,  Axios  journalist Chris Matthews wrote a piece “Big Companies, Fewer Workers”  that said:

The five most valuable companies in the U.S. are all technology firms that employ far fewer workers than their industrial predecessors.

He echoes a common complaint. But is it really true?

I decided to  compare employment at today’s most valuable  tech/telecom companies with employment at the most valuable industrial giants of the past.  My point of reference is 1979, the all-time peak year for manufacturing employment in the United States. At the end of 1979, these were the top 10 industrial companies (ordered by market cap), and their total employment.

---

Table 1: Top 10 Most Valuable US Industrial Companies, 1979*
	Jobs(thousands)
IBM	337
General Motors	853
General Electric	405
Eastman Kodak	126
DuPont	134
3M	88
Dow Chemical	56
Merck	31
Xerox	116
Johnson & Johnson	72
Total	2218
Data: Siblis Research , corporate annual reports, Progressive Policy Institute

---

So as of 1979, the ten most valuable U.S. industrial companies had a total employment of 2.2 million.*

By comparison, I took today’s most valuable tech/telecom companies, ordered by market cap as of June 10. Here’s the list, plus their total employment.

---

Table 2: Top 10 Most Valuable US Tech/Telecom Companies, 6/10/2017
	jobs(thousands)
Apple	116
Alphabet	72
Microsoft	114
Amazon.com	341
Facebook	17
AT&T	268
Verizon	161
Comcast	159
Oracle	136
Intel	106
Total	1490
Data: Annual reports, PPI

---

So the ten most valuable US tech/telecom companies today employ 1.5 million workers, roughly two-thirds as many as the 2.2 million employed by the ten most valuable US industrial companies in 1979.

However, a look at the two lists shows something interesting:  Take General Motors out of the 1979 list, and the size distribution in 1979 doesn’t look that much different than the size distribution in 2017.   Industrial leaders such as Kodak, Dupont and Xerox employed between 100K and 150K workers in 1979, roughly the same workforce as Apple, Microsoft, Verizon, Comcast, Oracle, and Intel today. GE and IBM in 1979 employed roughly the same number of workers as Amazon today. And Merck in 1979 wasn’t that much larger than Facebook today.

Conclusion: In terms of employment, the major difference between the industrial giants of 1979 and the tech/telecom giants of 2017 is one superstar company, General Motors, whose 1979 workforce dwarfed any other company on either the 1979 or the 2017 list.

 

To be continued…

 

*My definition of ‘industrial’ includes non-energy manufacturing companies. The 1979 list should include Procter and Gamble, but I could not locate their employment data in their annual report.  There’s no reason to think that substituting P&G for J&J would make a significant difference in the list. Ford Motor’s stock price underwent a steep plunge in 1979 that took it out of the top 10 industrial companies by market cap.