There used to be a lot of telephone operators in America. And now hardly any. When there were hundreds of thousands rather than just a couple of thousand, it was the work of women — and demanding work it was. As one historian describes it (quoted in this Richmond Fed analysis), the life of an operator was a real juggling act that meant “While … keeping an eye open for lights indicating new calls, and sweeping the board of old connections, operators had to complete several hundred calls an hour during peak times. Months of practice were required before they mastered the ‘overlaps,’ or the knack of performing multiple tasks simultaneously.”
If that sounds like a job better done by a machine, it was. Almost all were mechanized (and then eventually digitized) as people started doing their own dialing, with each turn of the rotary dial signaling switching equipment at a telephone exchange. So what happened to all those operators — almost entirely women, many young — when their jobs were automated? And what about future female workers denied a common entry path in the job market? Here are the key results from the new working paper “Automation and the Fate of Young Workers: Evidence from Telephone Operation in the Early 20th Century” by James Feigenbaum and Daniel P. Gross, which looked at the period of 1920 to 1940 when more than half of exchanges serving more than half of the US telephone network were mechanized:
Using panel variation in the local adoption of mechanical switching and population outcomes from complete count census data from 1910 to 1940, we show that dial cutovers presented a large negative shock to local labor demand for young, white, American-born women, with the number of young operators dropping by upwards of 80%—a near-total collapse in entry-level hiring in one of the country’s largest occupations for young women—and accordingly around 2% of jobs for this group being permanently replaced by machines, essentially at the flip of a switch.
These 2% of jobs represented entry-level opportunities for several times as many young women, and the fear was that its automation might choke off future generations from the labor force. We find that this shock did not reduce future cohorts’ employment rates. It appears that comparable middle-skill office jobs and some lower-skill service sector jobs absorbed future generations of young workers workers [sic], and did so fairly quickly, with women of only the youngest ages on average ending up in lower-paying occupations than they would have been in otherwise. The adverse consequences of automation were concentrated in incumbent telephone operators, who were subsequently less likely to be working, and conditional on working, more likely to be in lower-paying occupations—but even then, the magnitudes of these impacts were relatively modest.
So not terrible, which might seem encouraging for all those “cashiers, office clerks, or customer service workers” of today who seem at great automation risk. But the researchers are cautious drawing that conclusion given the unique economic factors of the time such as “rapid growth in women’s educational attainment, labor demand, and labor force participation.” Today, for instance, female labor force participation has been declining since 2000.
But one thing I found especially interesting is that the operators typically had plenty of warning of the “cutover.” While moving to rotary dial ultimately happened at a “flip of the switch,” that flip only occurred after a “multi-year planning and installation process.” This advance notice, the researchers speculate, “may have potentially allowed young women nearing working age to adjust their educational investments and early career choices (or for their parents and high schools to do the same), mitigating the impact of the shock.”
One wonders if the same will be true of one of the most commonly mentioned and fretted about automation threats, autonomous trucks. While there has been great concern about the potential impact on truck drivers, it won’t happen overnight — if at all — giving the labor market time to adjust. As technology analyst Benedict Evans has explained, the transition might take 20 or 30 years. By that time, “effectively all current truck drivers will have quit anyway — you won’t replace them, but you won’t necessarily put anyone directly out of work,” Evans says. Already, he added, the average trucker is 49 years old, and truckers are leaving the industry faster than they can be replaced.