Most of what comes out of Deutsche Bank’s cardboard cut-out CEO John Cryan’s mouth is so ridiculous that it’s best ignored. Yes, that Deutsche Bank (DB), the one which is having a perpetual near-death experience. But when Yves highlighted this story in Business Insider about (yet another) clever scheme to drag the bank into something hopefully approaching a going concern using robotics technology, it was so annoying I decided that enough was enough.
And as – bad – luck would have it, other industry illuminati started jumping on the same bandwagon and even my lumpen Too Big To Fail (TBTF) is trying to get in on the same act (well, someone’s created an internal website about it and a supernumerary exec has been appointed to head a “disrupter” team as he whiles away a year or two until his exit package is agreed, which is as close as it gets to being on the curve).
So what’s got everyone in TBTF-land so excited? Robotics. Let’s turn the floor over to our new best friend John (from the Business Insider article):
John Cryan, the chief executive of Deutsche Bank has warned that a “big number” of staff at the company will ultimately be replaced by robots and other forms of technology as the firm embraces a “revolutionary spirit” going forwards.
[Clive, interjecting briefly to apologise to readers for not warning about an imminent management babble speak overload and to reassure that, having got through that paragraph, the worst is over; let’s return to the quote.]
Cryan told the audience that the era of accountants and bankers acting like “abacusses” is coming to an end.
“In our banks we have people behaving like robots doing mechanical things, tomorrow we’re going to have robots behaving like people,” he said, according to a report from the Financial Times.
[That’s enough of that; Clive back again]
Complete and utter rubbish. Every business process automation project I’ve ever been involved with has failed and ended up a graveyard for multiples of millions of investment. Even my dumb, inept and hopelessly gullible TBTF looked at (actually, is still looking at) robotics but can’t see how to integrate it anywhere meaningfully.
Here’s the science bit. Business processes are divisible into three broad classes:
- “routines” (a multitude of predictable ostensibly simple stuff that is easy to detail, document and calculate volumes)
- “runners” (still predictable but more complex processes the are less numerous, typically lower volume and take longer to complete) and
- “rarities” (uncommon, ad-hoc variable length and variable complexity tasks that just crop up from time to time).
The “routines” have already been automated as far as possible. The “runners” have been assessed for automation and automated where a business case can be made to do so but some are highly complex and don’t lend themselves to be automated in a cost-effective way no matter what the technology applied is. And “rarities” are too unpredictable to get a coherent business process defined for.
Even if DB haven’t automated a lot of “routines” already, and still have some low hanging fruit to pick, the numbers talked about aren’t credulous. They would have to have little to none in the way of automation in place already for them to be even half true. Even DB aren’t *that* bad.
It isn’t as if any of this is really new. Information Technology didn’t just come on the scene once Apple, Goole and Amazon’s stock prices started heading for the stratosphere. There have been over the past couple of generations some significant leaps.
Computerised ledger processing beginning with IBM’s RAMAC system in the mid 1950’s, which was the first truly usable accounting system for banks to manage their customers accounts other than by paper-based bookkeeping because of its ability to scale and perform overnight batch runs to update up to a million accounts, was one of those kinds of game-changer. The ATM was similarly a big efficiency and productivity enabler.
Then there’s been incremental improvements such as bar-coding of forms and Optical Character Recognition (OCR) to read what customers or other bank operations staff have entered into them, then route the paperwork around to more specialised (as opposed to general multi-skilled) teams.
Processing check clearing without having to move paper vouchers around, the migration of card payments to a fully electronic system of authorisation and payment processing probably fell somewhere between the major systemic overhauls and the incremental enhancement ends of the productivity-enhancing spectrum.
But to reiterate, the big banks have been working on this for 60 years or more. A big claim, such as there is a significant but as-yet untapped potential to further reduce human-based back-office operations, needs a big proof. Thus far, while the announcement from DB made a lot of promises, it was very light on specific details.
Yves ran my comment with the link to Cryan’s speech in the day’s “Links” section and I was happy to leave it at that. But then when former Citigroup CEO Vikram Pandit popped up, with all the depth and worth of a worn dime, to weigh in on the subject on Bloomberg, I suspected there was something else afoot.
From the Blooming piece, here’s what Pandit had to say (don’t worry, it’s not quite as bad as Cryan’s waffle):
Artificial intelligence and robotics reduce the need for staff in roles such as back-office functions, Pandit, 60, said […]
“Everything that happens with artificial intelligence, robotics and natural language — all of that is going to make processes easier,” said Pandit, who was Citigroup’s chief executive officer from 2007 to 2012. “It’s going to change the back office.”
Interesting, neither Cryan nor Pandit themselves specified that there is also a subtler factor in play: which jobs done where? The Bloomberg piece relied on an earlier report by Citigroup to get some numbers:
In a March 2016 report, the lender estimated a 30 percent reduction between 2015 and 2025, mainly due to automation in retail banking. That would see full-time jobs drop by 770,000 in the U.S. and by about 1 million in Europe, Citigroup said.
[Clive again] … but that merely reflects widely acknowledged significant overcapacity (Section 4 refers) in banking, especially retail banking. As we’ve so often noted here at Naked Capitalism, there is simply too much finance — much of which is merely extractive and rent-seeking — for the economy to be able to cope with and remain healthy. The “automation” referred to by Citigroup includes nudging, coercing or otherwise deterring customers from using high-cost channels such as branch or telephone banking and instead foisting internet-based self-service onto them. It’s long been an industry business model to offer white glove service to high or, increasingly only ultra-high, net worth clients. But that level of service was what used to be the norm even for the regular living-on-a-pay-check customer base.
Some of what Cryan and Pandit are up to is merely a demonstration of the usual elite class warfare — scare your employees into thinking that if they have the temerity to demand higher wages and better treatment, the robots are coming to get you. Yes guys, we’ve all seen Terminator at the movies.
But there is another possibility.
About a decade ago, when the offshoring boom really took off, I made a mental note to look for signs that what was a low cost base (axing onshore people in favour of low cost countries like India and the Philippines) was starting to turn sour. It was inevitable that — unless the low-cost countries got badly stuck in a middle-income trap — sooner or later they’d not only cease to be low cost countries but they’d present a rapidly rising cost base which would start to bite the bottom line quite badly.
Business Process Outsourcing (BPO) destination countries are starting to tire of providing corporate welfare and are increasingly implementing specific measures to share some of the gains. It’s no longer sufficient for transnational giants to simply turn up and say “we’ll employ some locals” while omitting to add “so long as you pay for all the infrastructure (power grids, office space, transport, logistics, housing) we’ll rely on being there”.
Let’s say a typical TBTF had an operational cost base of $500m p.a. for labour for business processes which it could offshore. That cost base, again let’s say for the sake of argument, a decade ago, fell to $250m a year in year one as a result of the offshoring of operations. But the offshore labour costs were increasing at a fairly rapid lick, let’s assume 5% in real terms as a minimum. So now, 10 years on, we have a labour cost base of $400m p.a. And it’s increasing in real terms by $20m every year.
In less than five years, they’ll be right back where they stated in terms of costs — and locked into an escalating cost base. There may be a few true low-cost countries to relocate to but few are now without significant political and country-specific risks (bribery is probably the greatest; there are stiff penalties for companies which engage in bribing officials or other enterprises and in countries where the rule of law is, erm, rather hazy, this is often essential for getting anything done).
Bank management is so fixated into thinking that offshoring is always and everywhere cheap. They expect something for nothing. It’s coming as a nasty shock that their cost base is creeping up and up, even though it’s offshore.
Another issue is, having offshored a business process which wasn’t especially efficient, it becomes harder to implement process improvements because the operational activity is now in a different country and so is at the end of a long supply line. The bank whose business process it is therefore has poor visibilty of how the process is actually being executed. And, of course, the business process outsourcer has little incentive to improve process efficiency and their own productivity. Just the opposite, in fact.
A frequently worked scam in the BPO sector is to act as a consultant for the process owner to, supposedly, help optimise the business process. But because the business process is entirely outsourced, the business whose process it is supposed to be now lacks key operational data about the nuts-and-bolts of what it takes to perform the process. When, as often happens, the outsourcer claims that they are operating as efficiently as possible, the process owner is poorly positioned to say any differently.
Hence the sudden genuflecting at robotics. But it’s just another mirage. Bank management — and bank shareholders — had better get used to a cost base which is not only increasing but now unstable. Formerly low-cost countries of operation cannot continue to subsidise BPO-based employment indefinitely. And the investment track record for trying to squeeze out the last vestiges of business process automation are pretty woeful. Will Cryan and Pandit’s faith in robotics technology to come to their rescue be yet another triumph of hope over reality?