When IBM rolled out its IBM System/360 back in 1964, these big metal cabinets housing primitive circuitry and endless spools of magnetic tape were revolutionary pieces of technology, hurtling the banks into the future.
50 years later, the magnetic tape is long gone but 92 of the world’s top 100 banks still rely on IBM mainframes, according to the company. The majority of them are in retail banks, but a “significant number” of investment banks are also still using the technology, according to Keith Bear, vice president, financial markets at IBM. Their clients’ ranks include Morgan Stanley, HSBC and UBS.
Mainframes are just one example of legacy IT systems banks would love to eliminate. Frank Sanchez runs Finxact, a Florida-based fintech that develops cloud-based banking software. In one example, he said: “I’ve had a banker telling me: I don’t know how to change it because I don’t know what it does. But it must do something because I am spending $100m annually on maintenance.”
So what is stopping them? Some in the tech community say banks do not even know where to start.
Banks typically spend 80% of their IT budgets on legacy technology maintenance and a tier one bank could easily spend up to $300m a year on existing software which constantly needs expensive updates in order to meet regulatory requirements, experts say. While it is hard to put a figure on the precise level of savings for banks to move data and applications to the cloud, some technology experts predict it could cut IT costs by 75%.
HSBC’s chief architect David Knott said depending on the area where it is used, cost savings could be 50%, but added that he has seen instances where HSBC has been able to eliminate up to 90% of costs.
Taking a cost-cutting approach to deal with the growing price tag for maintaining legacy systems appears attractive. But making the case to budget for a major spend that will ultimately save banks money is harder in these relatively lean times. As Bear put it, “There’s a limit to the ‘change the bank’ funding.”
‘University graduates often don’t know this code and don’t want to learn it because there’s a limited career path’
Keith Saxton, who spent more than 16 years in IBM’s financial services research team said that banks can no longer delay major reform. He added that the banks often frame legacy-related problems as individual issues, “as opposed to saying, hang on a minute, we have a big problem because our systems don’t really look as if they are futureproof. We can’t get all this done without spending a lot of money”.
Those who urge banks to completely overhaul these systems would have a stronger case were there to be a crisis, but so far this has not occurred. Instead, IT departments have plugged holes one by one.
They have also often turned to short term fixes such as middleware software, which can translate between different programmes. Middleware can “knit together different parts of the organisation and allow it to interface with your legacy systems”, according to James O’Neil, a senior analyst at Celent.
Add the challenge of mergers among banks, where middleware was often employed to integrate the various systems, and bank IT departments are left to manage a raft of complicated structures and systems that have often been customised and changed over time. In short, it is a mess.
Andra Sonea, head of strategy and architecture at fintech group 11FS, used to work as an architect at both IBM’s financial services business lab and at Lloyds bank. She compares updating legacy systems with a box of tangled up jewellery. “You want to get your ring out of the box. You take it, but then you also have earrings and necklaces hanging on it. It’s the same with moving banking systems; you just want the ring, but there are a lot of other things hanging around it.”
HSBC’s Knott admitted these “spaghetti connections” are challenging for banks. “Everything is connected to everything else, pull on one thread and everything comes with it.”
Mazy Dar, chief executive of fintech business OpenFin, said: “The problem is that with every new feature that you are adding to the existing system, you are making migrating off that system harder. You are continuing to dig that hole.”
‘The banks aren’t dealing with it properly — they blame old stories like the mainframe’
Despite the existential threats from challenger banks, some argue that executives at the incumbents see little upside to their career in initiating a big, expensive overhaul when incremental fixes will do. Sanchez said: “If I am a banker and I make a good salary and have three years until retirement and I know my bank will live for at least another five years, why take that risk?”
Even if all the other hurdles can be overcome and everyone is on board with replacing a complex IT system by rewriting code or updating software — which experts say can take as long as 18 months — there is the issue of keeping the bank in continuous operation.
Craig Butterworth, managing director and global head of client eco-system at Nomura, compares such IT upgrades to renovating a major train station. “We can’t just say ‘London Bridge is shut for a month’. We need to try and upgrade systems while keeping the bank up and running, which is really tough.”
This is also a key issue for Mike Dargan, the chief information officer of UBS. “Clearly we make all of our money from the systems we have today, not from systems of the future. [Upgrades] can’t put any of the systems of today at risk,” he explains.
Though banks may stick their heads in the sand, time is running out. The coding language Cobol that was primarily used in banking on IBM mainframes is completely outdated and no longer competitive, according to Sanchez: “No Silicon Valley company would ever consider using that [language] now.”
The pool of developers in this coding language is also shrinking. Oliver Bussmann, ex-chief information officer at UBS and president of Crypto Valley Association, a Swiss blockchain network, said: “University graduates often don’t know this code and don’t want to learn it because there’s a limited career path for them while the developers that know this software are close to retirement now.”
A long-awaited solution
The growth of cloud computing may offer a solution. Operating banking systems in the cloud would allow developers to make continued updates, minimising the need for major overhauls in the future. But the main benefit according to IBM’s Bear is that banks would not have to “feed and water” their mainframes or other old systems.
HSBC’s Knott sees moving to the cloud as an important way of cutting its IT budget and is currently working with Google, Amazon and Microsoft to migrate some of its data.
He pointed out that cloud providers have become more “credible” in recent years and have bolstered their security, a major concern in the past around adopting a cloud-based system. He said: “We can move significant workloads to the cloud and be confident that we will still be as secure and resilient as anybody expects us to be including ourselves”.
‘With every new feature that you are adding to the existing system, you are making migrating off that system harder’
Turning what was a capital expenditure into an operational cost is “probably one of the major shifts,” Knott said. “Previously we would have to buy significant [bits] of infrastructure and we would [pay for] licenses upfront. Obviously in the cloud model we pay for consumption.”
For now, most investment banks are trying to unbundle some of their legacy systems, so they run separately, Bear said. However that can create further problems — especially when it comes to access to data, according to Saxton. “The big problem with legacy is stranded data. You can’t get the right data in the right format.”
Saxton said banks should stop blaming legacy systems and face the problem head on. “The banks aren’t dealing with it properly — they blame old stories like the mainframe.” However, he added that “the smart guys like Goldman [Sachs]” are seriously looking at cloud solutions, “so it’s starting to happen.”
Sanchez also believes leading banks will make the jump first and, as soon as that happens, others will have to follow.
“I think banks will save a lot more than 50% of their infrastructure costs, maybe 75% or even more. Within five years, the banks that are maintaining their legacy will be so disadvantaged on the cost they will either go out of business or be acquired.”