Banking Fintech

COBOL and the big tin bank

COBOL and the big tin bank
Written by Alex Letts

In 2017, Reuters published the following findings from a piece of research conducted by Celent, Accenture, IBM and others, into the technology supporting major US banking systems:

  • 43% of banking systems are built on COBOL
  • 80% of in-person transactions use COBOL
  • 95% of ATM swipes rely on COBOL
  • 220 billion lines of COBOL are in use today

For the less tech-savvy among us, COBOL is a computer programming language designed by an astonishing woman, Rear Admiral “Amazing” Grace Hopper, in 1959. And no, that’s not a typo. At a time when trillions of pounds are transacted every year, and with the UK economy depending on six banks to keep the show on the road, regulated banks are relying on a computer language that’s nearly 60 years old, designed for an age when computers as powerful as your smartphone filled entire rooms.

Against this backdrop, it’s easy to see why new fintechs get so excited:
“I’ve never seen a big bank do something cool – if they did, we wouldn’t exist.”
– Nikolay Storonsky, Founder of Revolut (Daily Telegraph, 10 April 2018).

Innovation among the big banks is slow at best, and it’s easy to see why – the technology that promised to liberate banking in the 1960s has today cemented the industry in analogue processes. But when asked the question “why do 80% of transactions and 95% of ATM swipes still rely on COBOL?”, the answers from those in banking have ranged from “if it ain’t broke, don’t fix it” to “it’s too big, too risky, and too expensive to change”. Imagine you’re the CEO of a major bank – would you bet on a migration project that could destroy your entire company if it went wrong? Perhaps it’s easier to keep the underlying ‘deadware’ on life support and pass the problem on to your successor.

So what happens when the expertise to support COBOL literally dies off? That’s a real scenario. The rumours of one bank even making calls to a nursing home seem credible.

The position of those of us in fintech is that the retail business of major banks is in terminal decline because of their outdated attitudes towards customers. But below the surface there is a greater truth – their 1950s attitudes survive largely because of the 1950s code designed by a person born in New York in 1906, just after the end of the Boer War.

About the author

Alex Letts

Alex Letts is founder and Chief Unbanking Officer of U. His disruptive career has brought digital technology into the advertising industry, electronic trading into the Lloyd's of London insurance market and, now at U, a transformation of the retail banking current account model.


  • They must start migrating few products in parallel with other systems gradually else time will come they would left behind the race of fintech. Every system takes time and obviously one has to change with time.

  • The challenge facing FS as well as other industries is the pace of change of the digital era and the unprecedented complexity of the IT options available both to the organization and the marketplace.

    The premise of this blog is that if something is an idea from the past, then it cannot be innovated. So the logic goes, if COBOL is from 1959 then it can’t possibly be involved in anything innovative.

    If we look to other technological innovations of the past, let’s test the logic of that argument. A trivial example is the telephone. A device that connects two parties separated by distance with a vocal communications facility. Well I think that one has just about stood up to the test of time and at the best part of 6 billions users of such a facility, it begs the question is this unique? How about something post-war? How about a particular brand? Let’s say the Porsche 911. Launched in ’63 and constantly modernized over the decades, its regular reinvention has helped keep it top of the heap in that market.

    The issue is simple. To assume that the original invention remains the sole basis of its development or modern incarnation is in most cases unsound. And any inference that because something is old means it cannot be evolved is equally dangerous.

    The telephone is largely unrecognizable today compared with its original instance. The Porsche probably shares a badge and little else with its 1960s forerunner, as it has modernized over time to support fuel, safety, crash protection, pedestrian safety, road legality, and any number of technological and engineering advances. People love the original, but they purchase the latest model. Same with the phone. Same with technology that helps build banking systems. Sure it might be something called COBOL in there, but it is the 2018 version: the version that supports Web Services, Cloud, Containers, Managed Code, the latest platforms and an unimaginable array of other contemporary technology.

    COBOL is often the scapegoat for inadequate funding of core IT systems. But all too often, it is the funding that is at fault – the budget is to blame, not the tech’. Indeed COBOL’s very raison d’etre is to support big business systems (that’s what the B stands for – business). For a view of what COBOL will help organizations do nowadays, there are plenty of vendors (e.g. Micro Focus) who can help.

Leave a Comment