UK banks are still using software code that was written in the 1960s and 1970s, with only a handful of employees who understand them.
According to a survey of 200 UK banks, 16% rely on software from the 1960s and almost 40% maintain 1970s code.
The survey, from management consultancy Baringa, questioned senior managers who had technology expertise. It found that 50% of banks admitted to relying on software that only one or two staff members, who are at or near retirement age, understand. Another 31.5% said they relied on one or two people, not yet at retirement age, to understand legacy systems.
Some 38 banks responded that they are still using code that was designed to run on physical systems such as punch cards, while 15% run code written for room-sized mainframes.
“Pockets of old technology is an unavoidable situation in complex technology estates,” said Paul Mihajlovic, who leads banking and markets technology at Baringa. “Banks are huge organisations, serving millions of customers across entire countries, and it would be impossible to demand they restart their infrastructure from scratch each time a tech innovation appears.
“However, the sheer age and criticality of some of the systems is eye-opening – if you have a current account, your money could rely on code written when Harold Macmillian was prime minister.”
One respondent to the survey said Microsoft software from 1993 is still used: “The ATM network of a bank depends on antiquated Windows NT servers that have been patched.”
Another said major banking core systems were built in the 1970s and still uses Cobol programming language. Cobol was the go-to technology for reliable financial and administrative systems, such as those used by the tax administration, banks, insurers and mortgage lenders.
One senior IT professional in the UK banking sector, who wished to remain anonymous, said he has worked on many systems dating from the 1960s, 1970s and 1980s.
“There is a lot of the old stuff out there and the reason it lasted is because it was very simple and it worked properly and it was high volume, simple transactions,” he added. “The banks are moving away from these systems because the people who understand them are leaving, and no young professionals want to learn languages like Cobol.”
Mihajlovic said the continued use of decades-old code causes two main risks for banks: “First, code initially written for long-shuttered systems, and maintained by a small handful of aging experts, is a significant risk to critical infrastructure. It might go wrong, and if it does, it will be difficult to fix.
“Second, old technology is seldom agile. If you have to employ specialists simply to keep something running, that thing is unlikely to respond with lightning pace to changing customer needs, and this becomes increasingly and disproportionately expensive.”
