Don’t blame Fujitsu

The Post Office scandal exposes Britain’s broken state.

On 29 December – amid the quiet of the holiday break, two days before ITV broadcast the first episode of Mr Bates vs the Post Office and brought the Horizon scandal to national attention – a modification notice added to a government website revealed that the company at the centre of the scandal, Fujitsu Services Limited, had been granted an extension to one of its government contracts (for its work on the Future Flood Warning System), bringing the total cost of the contract to £19.5m.

Fujitsu has been awarded £6.8bn in public sector contracts since 2012. MPs on both sides of the House of Commons are now calling for its status as a “strategic supplier” to the UK government to be revised. David Davis told the Today programme on 8 January that the company should be “frozen out” of future government contracts. Kate Osborne, the Labour MP for Jarrow, used a Commons debate on the same day to call for the government to “pause existing contracts” with Fujitsu, and to “stop awarding them multimillion-pound contracts”. More outrage will doubtless follow when the company posts its yearly accounts for 2022-23.

But new analysis shared with the New Statesman shows the reach of Fujitsu across Britain’s outsourced state, and the difficulty of removing any such company. Fujitsu built the Horizon system that wrecked the lives of hundreds of sub-postmasters, but the disaster could have been prevented if officials in government, or at the publicly owned Post Office, had contracted it properly or held the company to account. The wider and more important problem is that this is something the government routinely fails to do.

Data collected by government contract analysts Tussell shows 197 public sector contracts have been awarded to Fujitsu since 2012, and it is hard to find an arm of the British state in which it is not involved: the Home Office, the Ministry of Defence, the Foreign and Commonwealth Office, HM Revenue & Customs, Transport for London, HS2, Scottish Water, Thames Valley Police, the Financial Conduct Authority, the Land Registry, NHS England, local authorities across the country – the list goes on and on. The Conservatives, for all their talk of a small state, have overseen a historic increase in spending on private companies, from £64bn in 2010 to £222bn last year.

The outsourced state is supposed to deliver value for money through competition, but these contracts show this rarely happens: only 17 of the Fujitsu contracts are recorded as having been awarded through an “open tender” system, meaning other companies were able to bid for the contract. Many more – 61 in total – are “call-offs” from framework agreements, meaning Fujitsu was pre-selected as one of a number of likely suppliers for a wider framework agreement. But in most cases – 100 of the Fujitsu contracts, and many thousands more across the public sector – officials simply have not recorded whether they used a competitive process to achieve value for money when spending public funds.

Public sector contracts are not just opaque, they are often hopelessly optimistic. The Future Food Warning System is a pertinent example: the first estimate of the total value of the contract, in 2016, was £4.6m, but the final value of the framework will be around four times that.

Companies know this, and routinely find their way into public contracts by offering a service for free that quickly comes to be seen as indispensable. A contract will be awarded or extended without competition, because there is only one company that can offer the service needed (especially if it’s a service that requires a proprietary technology). This is the route that the controversial American technology company Palantir used in 2020, when it agreed a call-off contract to supply services to the NHS for just £1; it would later win a contract (again, published just before Christmas and heavily redacted) to provide similar services to the NHS at a rather more chunky £182m – although the government admits that this could rise to £330m, or beyond.

The same is true of the Horizon contract, which rumbles on at a cost of £2.38bn, having been extended again in November.

Like many government contracts, Horizon began as a political ambition. In the early 1990s, the Conservative government was concerned about benefits fraud; at that time, most people claimed benefits from the Post Office by cashing a “giro” cheque and the government decided that by computerising this system, it could be made faster and more secure. John Major called the plan “the biggest anti-fraud drive in the history of the social security system”. A new means to pay for the system, private finance initiative, was used to keep the spending off the Treasury’s books.

The Post Office Horizon inquiry, which began in 2021, has heard that officials had “severe reservations” about the solution offered by Pathway (a consortium led by Fujitsu’s subsidiary), and considered it the weakest of the possible options. Years before the first sub-postmaster was wrongly accused, officials knew that Pathway would need “proactive management”, but it was chosen anyway. Deadlines and costs became elastic, but – as is often the case – enough good money had been thrown at the system that cancellation would have been embarrassing; such was the government’s concern that another Fujitsu fiasco, the failed NHS IT system, would cause “reputational damage”, that the legal proceedings took place in secret.

Horizon was eventually scaled down, the benefits-payment part of the system shelved, but as in many large institutions there existed a collective instinct to deny failure, and the system began rolling out in 1999. The prosecutions began less than a year later.

A reflexive condemnation of Fujitsu now is all but useless. The real problem was a collective failure on the part of the Post Office, the civil service, MPs and ministers to properly oversee a system to which they had committed huge amounts of public money.

This article was amended on 10 January to reflect that there were 61 call-offs from framework agreements, not 67.