From London: As Australians woke this Tuesday morning, one of the quieter but more consequential stories in British public administration was playing out before the House of Commons Public Accounts Committee. The Bank of England informed MPs that it had formally ended its relationship with consulting giant Accenture on the £431 million renewal of the Real-Time Gross Settlement (RTGS) system — the plumbing behind nearly every significant financial transaction in the United Kingdom.
The RTGS system manages £790 billion in transactions every day, making it one of the most consequential pieces of digital infrastructure in Britain. It underpins the stability of the UK's payment systems, including CHAPS, the high-value payment system, and retail systems like Bacs. Its failure, even briefly, would ripple through mortgage settlements, corporate payrolls, and interbank transfers across the country.
The Bank announced its intention to modernise the RTGS in 2016, contracting Accenture as its technical delivery partner in 2020. The previous system relied on mainframe technology and was difficult to maintain because of its specialist hardware and a shortage of skills; the new platform is still internally hosted but relies on cloud-native technologies, making it more flexible. The renewed RTGS service went live on 28 April 2025.
Appearing before the Public Accounts Committee on Monday, Deputy Governor Dave Ramsden, who holds responsibility for Markets and Banking, gave MPs the assurance they were seeking. "It was a key part of the contract with Accenture that they would pass everything back to us to run, so we are not contracting out. In fact, Friday (February 27) was the last involvement of Accenture," he told the committee.
The committee's concern was well-founded. Whitehall has a long and inglorious history of becoming captive to the consultants it hires, particularly in complex technology programmes where institutional knowledge erodes and the only people who understand the code sit on the contractor's payroll. In January last year, the UK tax collector awarded Accenture an additional £35.2 million without competition to run the National Insurance and PAYE System because only Accenture could manage the technical risk, age, and intricate interdependencies of the solution. That cautionary tale was squarely in the committee's mind.
Committee member Rupert Lowe gave voice to the scepticism plainly. "In my experience, when you get involved in writing bespoke software, particularly with people like Accenture, if you're not careful, you end up running up a very big downstream bill," he said.
The Bank's response was that it had deliberately structured the programme to avoid precisely that trap. Chief information officer Nathan Monk told the committee that Accenture and Bank of England staff had co-located from the beginning of the contract, while bank technical staff were placed in Accenture teams during development. The bank also owns the intellectual property for the system. Ramsden added that in-house engineers could now read the code, diagnose problems, and build new features independently.
On the question of cost, the picture is mixed but not alarming. The programme's estimated cost rose to £431 million, 15 per cent above the original £375 million budget set in the 2020 business case, which included only 11 per cent contingency — lower than industry norms. Around £23 million of the increase stemmed from the change in the European Central Bank migration date. Still, the costs remained lower than the industry standard for similar projects, and the National Audit Office said "the increase in cost was reasonable given the size and complexity of the programme, and the number of uncertainties and risks that the bank had to manage."
Annual running costs tell a slightly more complicated story. The NAO noted that annual operating and maintenance costs had increased to around £41 million, compared with £21 million previously, with Ramsden explaining that the increase reflected the cost of running the system coming in-house, along with the technical expertise. In other words, what was previously billed by a contractor is now on the Bank's own ledger. Whether that represents a genuine saving or simply a reclassification of expenditure is a fair question, though the Bank's argument is that retaining capability in-house reduces long-term risk.
For Canberra, the implications are more than academic. The Reserve Bank of Australia is managing its own payments infrastructure modernisation, and the British experience offers a useful case study in how a central bank can structure a major technology programme to avoid vendor lock-in. The deliberate embedding of internal staff within the contractor's team, and the contractual requirement to repatriate intellectual property, are lessons that could travel well. The broader question of how governments avoid dependency on large consulting firms while still accessing the technical expertise those firms carry is one that Australian federal and state agencies continue to wrestle with.
The RTGS story is, in the end, a relatively encouraging one: a complex, nine-year infrastructure programme delivered over budget but within a defensible range, with the institution emerging owning its own technology rather than renting access to it. That is not the norm in public-sector technology. When it happens, it is worth paying attention to how.