"Why did retail jobs start to decline in the early 2010s specifically, after holding steady for more than two decades after the dot com boom?"
Two decades? - internet access pre 1995 was primitive and throttled by the attempted AOL/Compuserve hegemony. Sure online sales started to grow with the arrival of Amazon after Win 95 made internet access a doddle, but exponential growth from zero looks linear for quite a while.
Retail only needed to pay attention after 2008 when pocketbook problems concentrated our attention on the benefits of online sales. So the decline in retail jobs started at that point - 5 years rather than 20.
What about the widespread adoption of mobile phones with unlimited data and ubiquitous shopping apps in the 2010's so that people can now shop wherever they are, instead of only In front of their computer?
I updated my own calculations for 2023 based on the figures released by the BLS employment projections team in August and it seems like occupational churn is still about 50% above 2010 levels, which definitely supports the theory that the 2020s could be at least as dynamic of a decade as the 1950s in terms of occupational change and possibly more.
Interestingly enough, the official BLS forecasts have been very conservative compared to the actual pace of change, which I wrote more about here:
The period 1980-2010 “saw growth in basic manual labor and personal services occupations, which don’t pay very well but are difficult for machines to automate.”
Do the data for ‘low-skill’jobs in the subsequent period 2010-to date suggest negative growth for occupations becoming easier to automate? In other words, to which extent are we confident that the technology is the driving factor behind changes at the low end of the occupation scale?
Concerning the high-end of the occupation scale, should not we expect the distribution of cognitive abilities (which have ceased improving in the absolute in the population: flatlining or reversal of the Flynn effect) to put limits to its growth?
"Why did retail jobs start to decline in the early 2010s specifically, after holding steady for more than two decades after the dot com boom?"
Two decades? - internet access pre 1995 was primitive and throttled by the attempted AOL/Compuserve hegemony. Sure online sales started to grow with the arrival of Amazon after Win 95 made internet access a doddle, but exponential growth from zero looks linear for quite a while.
Retail only needed to pay attention after 2008 when pocketbook problems concentrated our attention on the benefits of online sales. So the decline in retail jobs started at that point - 5 years rather than 20.
What about the widespread adoption of mobile phones with unlimited data and ubiquitous shopping apps in the 2010's so that people can now shop wherever they are, instead of only In front of their computer?
This is amazing! I did some similar research this spring and was likewise surprised by slow occupational churn was in the 2010s.
https://www.2120insights.com/p/how-quickly-have-machines-always
I updated my own calculations for 2023 based on the figures released by the BLS employment projections team in August and it seems like occupational churn is still about 50% above 2010 levels, which definitely supports the theory that the 2020s could be at least as dynamic of a decade as the 1950s in terms of occupational change and possibly more.
Interestingly enough, the official BLS forecasts have been very conservative compared to the actual pace of change, which I wrote more about here:
https://www.2120insights.com/p/can-you-actually-predict-what-jobs
I also followed up with a similar analysis about how much the 2010s decline in retail might have been foreseeable:
https://www.2120insights.com/p/how-unpredictable-is-technologys
Subscribing, I wish I had discovered your work sooner!
It would be interesting to see any correlation existing between the rise of social media during the period when retail sales jobs started dropping.
Fascinating findings. Allow me two questions:
The period 1980-2010 “saw growth in basic manual labor and personal services occupations, which don’t pay very well but are difficult for machines to automate.”
Do the data for ‘low-skill’jobs in the subsequent period 2010-to date suggest negative growth for occupations becoming easier to automate? In other words, to which extent are we confident that the technology is the driving factor behind changes at the low end of the occupation scale?
Concerning the high-end of the occupation scale, should not we expect the distribution of cognitive abilities (which have ceased improving in the absolute in the population: flatlining or reversal of the Flynn effect) to put limits to its growth?