The robot revolution in banking is still a long way off

Register now

(Bloomberg Gadfly) -- For more than a year now, a sort of phony war has been going on in the financial-services industry.

Respected seers such as Deutsche Bank AG Chief Executive Officer John Cryan and Citigroup Inc.'s former chief Vikram Pandit have been prophesying a digital revolution that will see routine banking jobs replaced by automation, algorithms and artificial intelligence. To date, as Gadfly's Marcus Ashworth and Lionel Laurent have argued, there's been precious little sign of the mass layoffs that could spark the dawn of the robots.

Suddenly, that's changing. National Australia Bank Ltd. said Thursday it will cut 4,000 jobs, equivalent to about 12 percent of its workforce, in a shift toward digital banking. One week earlier, Sweden's Nordea Bank AB cited automation in announcing plans to lay off 6,000 workers. Mizuho Financial Group Inc. is looking to eliminate 19,000 jobs over the coming decade, equivalent to about one-third of its workforce, as artificial intelligence replaces clerical positions, the Yomiuri Shimbun newspaper reported last week, without saying where it got the information. Commerzbank AG, meanwhile, is grinding through plans to use greater automation to scrap 9,600 jobs over four years.

Should bankers be getting worried? Not quite yet.

Such announcements are the business-world equivalent of Donald Trump announcing that some new factory or other will create 2,000 jobs or so: accurate, but misleading compared with the bigger picture. Despite rising levels of automation and weak profitability creating the impression of a lost decade since the 2008 financial crisis, jobs in the sector appear to have kept rising until they peaked in 2015, according to Gadfly's analysis.

Finance employment kept climbing right into 2015.

While a plan to cut workforce numbers by 12 percent within a few years can sound scary, it's worth putting into context. It's about the same as the median annual employee turnover rate for the financial-services sector, and other industries such as retail and media see far higher rates of firings and resignations. Anyone who's worked at a company experiencing mass redundancies will be familiar with the way all the slashed jobs can soon come back as new channels of revenue are opened up. That's particularly likely because, as Lionel Laurent has argued, enthusiasm about the potential for automation to reduce costs and increase productivity in finance has run way ahead of evidence.

Machines can be great at many tasks. They can also be remarkably incompetent at times. One-shot generalization -- the ability that allows toddlers to identify pictures of (say) a dog after being shown a couple of examples -- remains one of the most difficult challenges in artificial intelligence. But the ability to spot patterns without a mound of data to mine remains crucial to banks' ability to carry out such basic tasks as generating M&A ideas for their clients or spotting when a customer is trying to launder money.

Jobs will go in banking, but they will largely be the soulless, repetitive ones and will mostly be replaced by more jobs elsewhere in a company. As one recent study pointed out, the introduction of automated teller machines didn't even reduce employment of human bank tellers.

That won't delight shareholders hoping to see falling costs and rising profits, but if you depend on financial services for your wage rather than investment income, it's reason to be cheerful. The robot revolution may come, but not quite yet.

For reprint and licensing requests for this article, click here.