If the hypothesis is revised to "increasing automation does not automatically reduce employment in the short term" that is feasible from the 1970-1990 dataset a

If the hypothesis is revised to "increasing automation does not automatically reduce employment in the short term" that is feasible from the 1970-1990 dataset as you've noted. I suggest that the ATM comparison is inadequate for this proof. Looking at a nominally favorable article from AEI (https://www.aei.org/economics/what-atms-bank-tellers-rise-robots-and-jobs/ https://www.aei.org/economics/what-atms-bank-tellers-rise-ro...) we see the case of "teller employment is going up AND ATMs are going up, ergo ATM adoption does not steal teller jobs". Except we don't know that, all we know is that both are going up. There is no way of knowing what the teller growth rate would have been absent ATM technology. Looking at the graphs shown, it is fairly clear that tellers were increasing from 1970-1980 at a strong rate and then it more or less levelled off. ATMs grew at a very similar rate from 1990-2000. It seems reasonable that if ATMs were not as successful as they were, the teller growth would have been higher and therefore ATMs have "stolen" at least some teller jobs growth potential. Absent a technological efficiency increase, the base demand of US population growth would have created more need for human provided services. In the long term the truth is laid bare: in the AEI article (2016) it says "employment of tellers is projected to decline 8% over the next decade". That would have left us with ~543k tellers -- instead we have 378k, a drop of 35%. When the tailwind stops, we will often see capital choose technology over humans and often at higher rates than projected. A parting concern: the rate of technological change and capability of industries to swap inputs was impressive in 1970-2000 but I hope you will agree it is nothing compared to the pace of change now. There is every reason to believe the trends, short and long term, we saw with ATM:tellers is going to repeat faster in the future.