Dominic Cummings, the aide so important to the Prime Minister that he seems to be doing a fair chunk of his job, is known to be a reader. Those 12,000-word blogs he used to spew out into the ether with depressing regularity would be packed with learned references to history and political economy. In mid-summer he insisted staff read two books before attending a “weekend boot camp”, which can’t possibly have been as fun as it sounds: High Output Management (1983) by Andrew Grove, the former chief executive of Intel, and Philip Tetlock’s Superforecasting (2015), which sounds a lot like someone read about “psychohistory” in Isaac Asimov’s Foundation series without realising that the whole thing was made up.
[see also: The dangerous legacy of the Cummings affair]
But something Cummings appears not to have read is any news coverage of major government IT projects over the past 20 years, because he’s just proposed another one. From the Times:
“No 10 plans to create online ‘ID cards’ for British citizens as Dominic Cummings tries to revolutionise the use of data across government. Under proposals announced yesterday each person will be assigned a unique digital identity to help them with such tasks as registering with a new GP.”
At the risk of saying something nice about Classic Dom, this might not be an entirely terrible idea. Digital ID cards could, in theory, make life easier for anyone trying to engage with the tax or welfare systems online, for landlords checking tenant references, even for bar staff checking that they aren’t about to sell three pints of Stella and a bottle of WKD to an unusually tall chap who was, nonetheless, born in 2005. Making it easier for people to prove who they are when interacting with the government or other services is probably A Good Thing.
But the key words in that last paragraph are “in theory”. Let’s leave aside for now the privacy concerns inherent in big government databases, or the moral questions around whether forcing people to prove their immigration status might prevent vulnerable people from gaining access to the housing or healthcare they need (the problem there is less the entirely theoretical big scary database than years of hostile government policies towards migrants). Because the odds are this new database is never, ever going to happen.
So familiar is the tale that Computer World put together a slideshow as a sort of greatest hits of government IT failure – a phenomenon it summarises as “ambitious project, outsourcing, bloat, disaster, crash, spend, spend, spend”. Particularly exciting examples of major IT project failures include:
1. Universal Credit which, apart from being a terrible welfare policy, has also been a terrible IT project. From 2010, the Department for Work and Pensions tasked Accenture, IBM, Hewlett Packard and BT with building a tech platform for the new policy. It didn’t work, was decommissioned and its functions brought back in-house, but only after costing £837m in taxpayer money. And this, remember, on a policy that was supposed to cut costs.
2. The modernisation of the Home Office’s Disclosure & Barring Service. A 2018 report from the public accounts committee described it as a “masterclass in incompetency”, which was running four years late and £229m over budget.
3. The e-borders programme, when the UK Border Agency abandoned its supplier Raytheon after it had already paid it £188m, but before the company was awarded another £224m in compensation for the unlawful termination of its contract.
4. My personal favourite: the Emergency Services Mobile Communications Programme, intended to create new radio infrastructure for said emergency services, but which as of May 2019 was running an eye-watering £3.1bn over budget.
The list goes on, each story as depressingly predictable as the last. As a client, the government isn’t sure what it wants, keeps changing its mind, and is terrible at negotiating contracts. The inevitable results are delays, failure to deliver and, often, enormous compensation payments to contractors which, it turns out, are rather good at negotiating contracts.
One other government IT failure is worth noting: the “Verify” system, which sounds suspiciously similar to what No 10 wants to create now. The programme, which was launched by the Government Digital Service in 2016, was intended to give people a single system to prove their identity. By this year it was meant to have 25 million users; as of February 2019 it had managed only 3.6 million, and I’m guessing that, if Dom Cummings wasn’t aware of it, you probably weren’t either. The fact this new system was announced without even mentioning that the old one existed is very, very striking.
I would bet a not insignificant sum of money that what happens next will go roughly like this. The government will award a juicy contract to a tech company, which specifies the parameters of this project so vaguely as to make it unachievable, but which nonetheless locks ministers into paying tens, if not hundreds, of millions of pounds of public money regardless of delivery. A pilot scheme will be attempted, and fail; full rollout will be repeatedly delayed. And then, years from now, when Boris Johnson and Cummings are long gone and no one can remember what this thing was meant to do anyway, it will be quietly cancelled. Cummings’ latest big idea is nothing more than a promise to spend untold sums of public money on not doing the thing he says he wants to do.
All of this has happened before; all of this will happen again. Matt Warman, the digital minister, told the Times that he is looking forward to “working with partners in the private sector”. I’ll bet he is.