Robots to Cut 200k U.S. Bank Jobs in Next Decade, Study Says
bloomberg.com
bloomberg.com
One day I try go to the bank and the counter was ripped out and drywalled over, and in its place is a giant ATM machine. It's like an ATM machine on steroids that can do anything with your account. And if you had issues there was a screen on it that would connect you to a call center in Bangladesh or wherever.
Some corporate bank employee was posted next to the machine to explain the change to people who had questions. When I walked up a group of people were standing around listening to him talk, they looked ready to pull out the pitchforks. He assured us that all of the tellers were still employed, they were just working "behind the scenes". I didn't buy that at all, and neither did anyone else.
I continued to use the machine to deposit my cash, but it really sucked. I mean the machine worked absolutely fine. But it sucked to imagine how those tellers felt to be replaced by it.
Now that I don't get paid in cash, I don't even go to a physical bank at all. I deposit checks on my phone in the rare occasion I need to do that.
I'm still not sure how to feel about this. I want to hate it but it feels a little irrational. Maybe all of those tellers went to a coding bootcamp and now make triple their salary designing the ATM machines? Seems like that's how the story is supposed to go. But probably not. Nothing I can do about it anyway.
I think we can mourn our losses while also looking forward to the new opportunities created. It doesn't have to be all good or all bad.
Anyway, I'm not sure I completely understand your comment. Are human bank tellers impeding people from using "more complex financial instruments"?
Yes, but this value is clearly less than the value provided by automated convenience. Isn't this why you don't go to a physical bank at all?
It seems like the problem is that the people who are "automated away" should be able to get by somehow, even if that means entering retraining or adding a safety net for those unable to transition somehow.
I'm not sure how many viable solutions of substance we'll come us with, considering our day job is programming computers, not crunching intentionally complex economic policy.
Of course, that is where all the unemployed people go.
Some of them will have gone into other customer service roles. It's not like they had to be an expert in banking.
A few may have gone on to more general office jobs with titles like, "Marketing Associate" where they make PowerPoint slides all day.
Some might be in worse jobs now. But, given low unemployment figures, some might well be in better jobs. Not many people aim to spend their life working as a bank teller anyway.
Very good point. But as low-skill customer service jobs go, it's a great stepping stone for someone who is yet to figure out their career path. What happens when all those stepping stones are automated away and we only have high-level jobs left? I think it doesn't bode well for economic mobility.
Technology progresses and we don’t need experts to cash our checks anymore. If we want to give make-work jobs to people I would prefer they put on puppet shows and read books to kids and the elderly rather than paperwork nonsense at a grocery store bank.
People find new jobs. Losing a job doesn’t mean they are destitute - we have a vibrant jobs market in America. If there are enough people like you who would specifically go to a bank in person because you like talking to people, a bank may do that as a competitive advantage. But it isn’t a necessary thing anymore.
They also partner with some private companies like AAA so members can do some of the basic services in person. I was able to pay my registration and get proof in under a minute at AAA.
I do use AAA for some services. They are limited in what they can do with the DMV. It is risky for people with my personality to interact with state government workers.
The drivers are the AAA customers and AAA feels they can bring value to their customers by providing this service. Hopefully DMV uses this money to improve their own service times.
(This information is abundantly available from the Bureau of Labor Statistics or you can Google the obvious keywords if anyone wants to keep me honest.)
What is needed to really reduce costs is a payment system that eliminates error and irregularities, or which uses AI to handle them.
Perhaps some kind of... clearing house, which could be automated?
In the US we have a dozen startups for you to send money to your friends. In most of the world you just do a bank transfer.
I pay my rent and utilities with ACH, but I know people who still don't have that option and are stuck in the dark ages of checks or credit cards with processing fees.
I have accounts with multiple credit unions, and one doesn't permit online ACH transfers and another takes 3+ days to do it. I'm currently trying to consolidate those accounts but even that's a huge hassle. It's nearly 2020 and the form the payroll company my employer uses wants a voided check from me to change where my direct deposits go! I don't have checks and haven't written a personal check in probably 15 years.
We keep scratching our heads looking for that magic future tech solution to our problems (AI! AI for everything!) when we actually just need to fix adoption of the freaking 45 year old standardized solutions right in front of us.
Checks have always been an almost entirely American aberration and a bad solution on all sides. It took us almost 15 years after Europe to adopt chip cards, and we're still swiping in places and almost nobody requires a pin for security. NFC adoption isn't great either outside affluent areas. We're not lacking new financial solutions, we're lacking proper adoption of the solutions we already have and the rest of the world is happy to use.
Not everyone wants a change. One person's costs are another person's revenues. There are also a lot of managers managing organizations that process fraud, errors, and other irregularities. Costs don't matter if you can charge enough in fees and interest.
In the case of chip cards, the US wasn't about to pay for the French smartcard patents. In general, financial technology is a patent thicket and its best to wait 20 years before implementing new ideas unless you have really positive financial case and can be sure you have the IP locked down.
I suspect there's also the old argument that standardisation takes regulation, and regulation is evil. Government should stay out of this (oh, btw, Government - can you just pay the next tranche of subsidies directly into the executive bonus account. Thanks!).
https://en.wikipedia.org/wiki/Interac_e-Transfer
https://en.wikipedia.org/wiki/Interac
It works quite well.
Other times, I've had e-transfers delayed by a few hours. This is fine if I'm sending off rent, but what if I'm trying to buy something off craigslist? I can't exactly hang out with the seller for a few hours waiting for Interac to get their shit together. This isn't an email issue: Interac acknowledges that transfers may take "up to a day" on their site.
Further, e-transfers cost money to both send and receive at most/some institutions, while cheques are free. E-transfers also can't be scheduled in advance like post-dated cheques can.
Cheques may be antiquated tech but they're still very much useful for certain applications.
I do use them to make transfers between banks because they’re virtually immediate rather than several days long pretty regularly these days without fault.
Checks are down to a third of what they were in 2000. The fastest growing type of non-cash payment is debit card, and a lot of these are chip and PIN.
Robots don't take jobs. Managers replace workers.
Illness in herd animals is healthy for wolves by making them easier to kill without injury.
The industrial revolution saw major growing pains and destitution but raised standards of living greatly.
The conflict with the goal of people living well is that we don't want to take wealth from those that have it and give it to those that don't, without making the ones that don't have wealth "earn it".
Otherwise, there is nothing stopping societies from gathering up all the extra profits from automation, and using it to invest in education, healthcare, reducing number of work hours, increasing vacation time, i.e. living well. But we would rather not have that.
In a 'healthy' society, automation should be employed to make humans work less hours and with less risks. Somehow many people think this statement is absurd.
The true issue is adaptability or lack of it in an economy. The replaced should be doing something actually useful but that process is less than straightforward.
Somewhere down the road some kind of intervention will be necessary, as too much will be on the line.
Software agent aka bot is not a robot. Robot is physical device. ATM is more robot-like than chatbot.
I imagine it's hitting the jobs a step or two above the tellers as well.
I don't know if you use ATM's a ton, but when it's the only option you have for interacting with your bank, it's pretty terrible, because your options for if something goes wrong is to call a number and get redirected to some general call center in probably some other country and wait while some minimally paid chap who couldn't care less about your problems follows a script and then when they can't figure it out, tell you they will file a ticket and get back to you in 3-7 business days.
I agree software and robots can do the job better with the caveat that you are using the well trodden majority use case. The moment something goes wrong (software bug, does the wrong thing, etc), or you are trying to do something slightly different, good luck.
My bank (First Direct) picks up only barely after the ringing tone starts, and the person you're talking to is an expert. They have no branches, so everything must be solved on a phone call (or you can use their web site of course). Nobody needs or wants branches, especially branches that are only open when you've got better things to do.
Their support is really good and always human.
I also bank with Citi, BA, and Chase. They have tellers and suck so much.
Wells Fargo employees had to generate more and more service fees to justify their employment. That was what came of the investigation a few years ago. If they weren't opening X new accounts every month they would be terminated.
I foresee a future where every major bank just gives their clients (hedge funds, asset managers) an API that they can call to electronically trade securities, rather than a phone number they can call. This is already in progress (see Tradeweb)
At the end of the day, when you’re competing with 6 other banks to offer a better bid-offer spread, you need to cut costs any way you can, and that means cutting humans.
Moreover, as sales and trading becomes more commoditized, all the salespeople get the axe. No need to take clients out to expensive lunches, etc.
Both are true, but this time the scale and speed is the difference. Even in software which is considered a safe job, I am much more productive today than just a few years ago because of better tools, third party services, libraries etc. We are able to do a lot more with a lot less with every passing year.
This time it is different and it is not going to be pretty, unless we take a step back and look at ideas like Universal Basic Income etc.
One other mindset that is quite difficult to change (at least in the U.S) is tying a person's self-worth to the amount of hard work they put in.
They do move to different jobs, but most likely not better ones. An observable consequence of automation is to put pressure on the 'labor market', leaving fewer and fewer opportunities with increasing requirements for job seekers.
There are a lot of assumptions in there. Getting everybody to work 168 hour weeks (24x7) seems like it should be even better, but it turns out that as you get close to that workload your employees are less productive. Fatigue will ensure your useful productivity is less than someone working somewhat lesser hours who is allowed to sleep. (I'm going to assume you require your people eat at their desk, otherwise starvation will kill them) You also lose because if there is even a rumor that other companies allow their people who have a life they will quit, and replacing people cost productivity.
We can take the opposite extreme: you work 1 hour per week. At this point your productivity is very low because you can never get "in the zone", and in fact your entire hour at work is spend just trying to remember what you do.
Someplace in the middle is the sweet spot. 20 hour weeks may be worth it - perhaps you can pay your people half as much and still give them as much income as they want, and still get the productivity you need to compete. Perhaps you can justify paying as much for 20 hours as your competitors pay for 40 hours because you employees enjoy that extra time and so they don't quit - you get better productivity because you don't have to teach them new things.
Each company needs to figure it out. in general society has settled on 40 hour weeks as a good number. There is no legal law making it best, but considering the compromises it isn't a bad number.
Would a lot of people prefer to work 4 days a week all other things being equal? Sure. But a lot fewer would be interested in earning 80% (or maybe less when the cost of benefits are taken into account) of their salary especially given that their partner/friends/kids/etc. would probably not have that time off.
And working 5 days a week most weeks just seems pretty normal. As you say, Western society seems to have settled on that as a fairly reasonable number with the main variation in how much vacation people get on top of that.
It might be a bad number going forward, considering the widening income/wealth divides.
We are waiting for regulation of automation to spread the benefit.
We could fund a UBI or better unemployment by taxing some % of the total benefit of automation.
As long as most peoples' basic needs are met and there's a carrot in front of them promising someday they'll have it better/easier, people will keep on running towards that carrot and maintain the status quo.
[1] https://www.timeslive.co.za/news/south-africa/2019-09-26-pro...
[2] https://ewn.co.za/2019/09/25/busa-banking-sector-mass-action...
The first wave of labor downsizing in banking came from the closure of retail branches. The second wave of downsizing, I suspect will come projects like Libra or CBDCs [1] where issuers bypass banks (now reduced to a redundant middleman) and reach users themselves.
[1] https://www.bloomberg.com/news/articles/2019-09-10/why-china...
No, the 'non-clang' robots are an indication of the software engineers' gross failure to properly integrate and standardize data exchange between solutions.
Not surprisingly, only the last of these stand the test of time, and most organizations that jump on the bandwagon will realize that they need to hire more software engineers and automate their business processes at the root.
For some, this is a long and painful realization. I spent six months trying to convince the appointed leader of a "robotization" initiative that what they were doing is actually software engineering, and that involving developers at an early point would be a better use of their money. They eventually made the mistake themselves, and got their own developers 18 months later.
Maybe 20 years from now, we will have technologists on the board and in executive positions at these companies, and these decisions can be made in an adult manner.
Read this somewhere: in 1999 risk meant bumping into someone and finding out he/she makes more money daytrading stocks. In 2002 it meant owning stocks.
These are often easier to handle when you can walk into a branch in person. I just had something semi-complicated I needed to do with my brokerage a few months ago and it was very nice to walk into an office and get it taken care of in 10 minutes.
(That said, a lot of the real estate is rather unnecessary. I do have other financial folks who I deal with regularly who I have never met in person.)
I get it to some degree. I very rarely need a physical bank/brokerage branch but it's handy now and then so my choice of vendor for a largely commodity service is at least somewhat biased by retail availability.
Partly to upsell more complex products like investments and mortgages, but mainly because surveys have shown that wealthier demographics will only use online banking if there is a branch nearby “in case”. The cost of a branch is dwarfed by the additional revenue from those customers.
Source: used to work for a bank that did this, and those were the reasons.
Then maybe all those branches will start to die back for fewer rare services that truly need in person presence.
Plus I have a safety deposit box at my bank for important documents.
Did you see this article and discussion about safe deposit boxes? https://news.ycombinator.com/item?id=20545276
I don't know who these bank branches serve, as they are not ever open during hours that anyone with a job could get in to them.
In Australia you are lucky to find a bank with real people working in it, they have mostly been replaced by Automatic Teller Machines (ATM's) and if you do, they will normally be working at selling you insurance or investments.
I still feel like there are many human-contact roles where an AI isn't going to cut it.
Source: https://www.weforum.org/agenda/2019/04/skills-jobs-investing...
Banking as a profession relies on a fairly strict set of rules , most of which can be automated. Thus, the leftover role of the banker is to navigate the regulatory requirements within the the scope of the automated task.
As well as the programmers to automate it.
https://yourstory.com/2019/03/faas-agritech-startup-farmers-...
I will tell you what I see today. Salesmen selling 'culturally aware fuzzy logic' that promises to automate just about anything and let leadership just read beautiful dashboards. Depressing stuff. And it sells too.
And 200k over a decade? Wake me up when AI replaces CEOs.