What ATMs and Bank Tellers Actually Teach Us About Automation
The most-cited "proof" that automation doesn't kill jobs is a bank teller story -- and it's true, right up until the part everyone leaves out.
Between 1980 and 2010, U.S. banks installed roughly 400,000 ATMs, a technology built specifically to do a bank teller's core job: dispense cash. Bank teller employment didn't fall over that period -- it grew, from around 500,000 to nearly 600,000, according to research by economist James Bessen. It's one of the most-cited data points in the automation-and-jobs debate.
The mechanism is worth understanding, not just the headline. ATMs did cut costs -- the number of tellers needed to run a typical branch fell from 20 to 13 between 1988 and 2004. But cheaper branches meant banks opened more of them to compete for customers, so total teller demand held up. The job itself changed too: tellers spent less time on routine cash handling and more on the things a machine still couldn't do -- explaining products, handling complex requests, building the relationship that keeps a customer at that bank.
That story has an ending most retellings skip. Starting around 2010, teller employment did enter a steep decline -- from about 332,000 full-time tellers to roughly 235,000 by 2016. The cause wasn't the ATM. It was mobile banking, which didn't just automate a task, it removed the need to visit a branch at all.
Drawing on economist James Bessen's research on ATMs and bank employment (IMF Finance & Development, 2015) and subsequent labor-market reporting on teller employment through 2016.
Get your own AI exposure score
Two minutes, free — see how exposed your specific role is and what to do about it.
Get your AI Vulnerability Score