The Schumpeterian lag

Michael Mandel writes (WSJ),

But does e-commerce destroy more jobs than it creates? So far the answer seems to be no. From the third quarter of 2015 to the third quarter of 2017, brick-and-mortar retail full-time-equivalent jobs fell by roughly 123,000, or about 1%, according to my think tank’s analysis of the latest Labor Department data.

Over the same two-year stretch, the e-commerce industry has added some 178,000 jobs in fulfillment centers and electronic shopping firms. In addition, express delivery companies and other local couriers boosted their full-time-equivalent workers by another 58,000.

If you take this on its face, then the retail industry as a whole is getting much less efficient. There has been a net increase in labor required to deliver goods from manufacturers to consumers.

I think what is happening is what might be called Schumpeterian lag. Schumpeter suggested that in the process of creative destruction, optimism over new methods combined with a reluctance to let go of old methods to create a boom. But eventually some firms realize that they are wrong, as they suffer losses. Then you get the bust. I expect that is the way it will play out in the case of retail.

12 thoughts on “The Schumpeterian lag

  1. Alternatively, this reflects some capture of whatever is the opposite of consumer surplus: people are paying others so that they don’t have to drive to stores, find things on shelves, wait in line to check out, and so forth.

    My gut says that isn’t quite enough to explain all the new jobs, but it has never terribly reliable.

  2. Interesting way of looking at it. The missing data is how much people are buying. If the total amount of goods sold is the same, then, yes, retail is becoming less efficient. If people are buying more because of the added convenience and the world of new products that are now available to them, the story is a bit more complicated. My personal experience is more in line with the latter. I’m buying things that I normally wouldn’t have bought either because I wouldn’t know where to find them or didn’t want to spend what little free time I have shopping around for.

    • An example might be if Uber actually added more full-time driver equivalents than were “lost” by taxi companies but rider-miles increased by more than the increase in drivers.

      Also, labor should be measured in dollars rather than full-time equivalents as not all labor is equal. Maybe, the e-commerce jobs are going to lower-skilled cheaper labor freeing higher-skilled labor previously tied up in retail to higher value uses. Probably not a big factor in this case, but there are many positive supply shocks beyond cutting full-time labor equivalents.

      • Right! Tho I’d say labor should be measured BOTH in wage amounts as well as worker counts.

        I’ve read that there are huge numbers of Dollar Stores opening…

  3. One reason not to take it on its face is that Mandel is comparing only the final portion of the brick-and-mortar distribution chain to the entire e-commerce distribution chain.

  4. This reminds me a bit about “did NAFTA create or destroy more jobs in the US”. People add up what they can see regarding workers in the directly conflicted companies, and they ignore the massive value created for consumers. Part of that value is availability of some goods and services that were not previously available. But there is also the simple productivity gain* overall which translates into a large increase in “economic rent” for consumers, who then spend this rent on more goods and services, some of which will be in industries that are not directly related to the new Schumpeterian competition.

    This last category of consumers buying new goods and services also contains workers, and those workers are essentially always left out of the “was there a net gain or loss of jobs” equation.

    *The overall productivity gain in my opinion is what matters the most, not jobs in and of themselves. If we want widespread prosperity (e.g. rising overall wages), then we need widespread productivity gains. So the most pressing question isn’t about jobs per se, but about “does the economic issue at hand increase productivity or not”. Unfortunately this largely seems to be ignored in the public discussions of such issues.

  5. Closer to the truth, investment creates jobs before it destroys them, that is, creation of a new system entails considerable amount of work that will disappear once in place. In this, it is not retail jobs that will be lost in the aftermath, they are already being lost, but ecommerce jobs that have been successfully automated away, staff that will be shed once it is no longer expanding. The construction of a house entails a lot of work, but once built, requires minimal continued effort.

    • Or, to put it another way, the optimism is not among retail, that is all gloom and doom, the optimism is among ecommerce that expects to grow to the heavens to which reality will eventually catch up.

  6. I think these jobs figures aren’t strictly comparable. The jobs lost only counts retail employees (cashiers, stockroom employees, etc.), but not people who make the deliveries to the stores and those sorts of jobs, whereas the e-commerce job count effectively encompasses many of those roles. Again, I believe, if someone actually knows please chime in. This also doesn’t account for the job growth we would have expected to see; how has retail+e-commerce employment grown relative to overall employment?

Comments are closed.