Task automation and the reshaping of job value
MIT Economist David Autor –one of the world’s most-cited labour economists– and Neil Thompson, director of MIT’s FutureTech group, have just released a report that analyses how different kinds of task automation affect wages and head-counts.
Looking into 40 years of task data across 303 US occupations (1980–2018) the findings highlight that automation affects workers extremely differently depending on the expertise needed for the role after automation occurs, even within similar occupations.
Between 1980 and 2018, accounting clerks and inventory clerks both faced computerisation of routine tasks. Yet their outcomes diverged dramatically:
- Accounting clerks: Wages rose substantially, employment fell
- Inventory clerks: Wages fell, employment rose
Computers eliminated routine bookkeeping from accounting work, leaving behind expert problem solving and decision making tasks requiring specialised knowledge. This raised the occupation's expertise requirements. Fewer workers could qualify for the remaining work, driving wages up even as total employment declined.
For inventory clerks, automation removed their most skilled work –things like checking items against government price supports and technical analysis. What remained was basic counting, stocking and weighing. Lower expertise barriers meant more workers could enter, pushing wages down while employment expanded.
The researchers found this pattern repeated across occupations. When automation eliminates inexpert tasks, it makes remaining workers more valuable by concentrating the work around higher-skill activities. When automation eliminates expert tasks, it democratises access to jobs but reduces their value.
The Space of Tasks and Occupations: Expert, Automated, and Generic Tasks Performed in each Occupation at Current State of Automation
A one-standard-deviation decline in the complexity of an occupation’s tasks –equivalent to the decline in expertise required by a telephone operator from 1980 to 2018– is linked to an 18% wage decline and a 40% increase in employment per decade. The opposite occurs when expertise required rises (wages increase, employment demand decreases).
This explains why simple "automation exposure" predictions often miss the mark. The same technology can simultaneously replace experts in some occupations while augmenting expertise in others.
For professionals navigating technological change, instead of asking "how much of my job can be automated?", ask "what expertise will remain valuable after automation removes routine work?"