Hi everyone! Some context on this question: I've been working on making some predictions about the economy of the future, and recently wrote a piece laying out the most aggressive possible case for AI-accelerated automation and job loss.
It's often held as true that new jobs will always replace old ones, though some research by Daron Acemoglu has suggested that technological unemployment can at times outstrip re-employment. I'm interested in tracking both trends over time, and have been thinking of building a tool that that pulls detailed occupational data from the Census' Current Population Survey (specifically to PTIO1OCD variable), and highlights any significant trends (e.g. where might technological unemployment be increasing the most, and where is it being offset by increases in hiring in other occupations?). I'm also curious to compare current trends in occupational change, and how they compare to previous ones. Inspired by this analysis from the Cleveland Fed, I did a more detailed analysis about the distribution of occupational change from 1870 to 2019.
I know that the Current Population Survey data is quite noisy, and detailed occupational categories shift and may not always be comparable over time. But I have two questions:
- Is there another organization or person that is already tracking these trends (or similar ones) in real-time and publishing the data?
- Is the approach I'm proposing a potentially fruitful one, or are there other sources of data (e.g. JOLTS, scraping LinkedIn data) that might provide a better answer to my questions?