If crises bring with them new opportunities to think afresh, then the combined impact of Brexit and COVID has been to focus attention on the capacity and structure of the British state. This rethinking is increasingly framed in terms of “smart government”. Dominic Cummings, the prime minister’s now departing chief adviser, has been at the centre of a drive to “harness the power of data and technology” at every turn. Technology has a role to play in modern government, but what Cummings seems intent on fashioning appears something closer to a populist technocracy based on a belief in algorithmic governance.
Our argument is simply that this logic, and these ideas, should be dropped. Indeed, a succession of recent failures and fiascoes has only underlined the paucity of the intellectual thinking behind this agenda as well as its lack of emotional intelligence.
Cummings appears to see British government as defined by institutionalised failure and dysfunctionality – nothing more, nothing less. What’s needed is strong “flexible, adaptive and empirical” leadership and the mastery of technology in order to control uncertainty. More than just “reform” is needed – only “transformative reform” will do. Cummings has described this in his blog in the following terms:
Most of our politics is still conducted with the morality and the language of the simple primitive hunter-gatherer tribe … Our ‘chimp politics’ has an evolutionary logic: our powerful evolved instinct to conform to a group view is a flip-side of our evolved in-group solidarity and hostility to out-groups … This partly explains the persistent popularity of collectivist policies … and why ‘groupthink’ is a recurring disaster.
Cummings has sat at the centre of a powerful hub-and-spoke model of governance that promotes a strident data-driven model of technocratic depoliticised governance. For every problem there is, in his worldview, a metric. For every social challenge, there is an algorithm. Data and technology are, as might be expected, the twin pillars of this (latest) revolution in government which, in turn, creates a need to recruit a new technological elite.
In July, Michael Gove delivered a lecture in which he enthused about these ideas. Gove said he wants more government decision makers “to feel comfortable discussing the Monte Carlo method or Bayesian statistics”. This is the scientisation of politics; the belief in a pure, structured, depoliticised, technocratic and highly mechanical view of decision-making.
His view offers a glimpse of a rather unattractive model of hybrid populist technocracy that is devoid of emotional content and lacking political understanding. It deifies a rather pure model of brutal governing efficiency that is more nightmare than vision. It is a form of sacrificial statecraft because it sacrifices any understanding of why feelings so often trump facts in politics.
When algorithms go wrong
The blatant failure of algorithmic governance during the A-levels and GCSE fiasco of August 2020 is a prime example of where all this goes wrong. Computers are good at crunching big data, but their digits and dashes will at some point be translated into a real-world impact on someone’s life. To pretend that algorithms provide a somehow neutral, technical or depoliticised way of taking difficult decisions is torpedoed by the wealth of evidence on embedded biases and how these tend to mirror preexisting structural inequalities. Prime Minister Boris Johnson blamed a “mutant algorithm” for the exams crisis. There is no sign that the broader lesson has been learned.
The government’s “algorithm for homes” could be its next policy fiasco. Housing Secretary Robert Jenrick’s newly announced plans to build more homes propose replacing the current method for determining housing need by introducing an algorithm which will determine targets for every English region based on relative affordability and the extent of development in those areas. Sounds great until you dig into the data and discover that instead of “levelling-up” the new algorithm will continue to concentrate growth in wealthier regions.
Technology, evidence and big data may well have a role to play in informing government policy, but let’s not pretend that it offers simple answers to complex problems.
We must also recognise that this is all part of a pattern of centralisation around Johnson – and to a large extent around Cummings. From the Nasa-style mission control “hub” in 70 Whitehall complete with floor-to-ceiling screens, real-time data and rolling news coverage, through to the decision to centralise government communications and hold Whitehouse-style televised daily press briefings. A new machine is being built almost by stealth.
And the metaphor of a machine is really quite apt. As a more centralised, presidential and technology-driven “hub” takes shape, then, so the capacity of local MPs or regional leaders to question the system declines.
Meanwhile, it’s not surprising that public officials have been left unable to understand the rules of the game. Where they used to be expected to keep their heads down, gain experience and expertise and remain politically neutral, times appear to be changing. When problems occur, it is officials that are sacrificed and several permanent secretaries have decided this is a game they don’t want to play. Political office-holders are landing top public roles, which raises serious concerns. Mary “Dido” Harding’s appointment to lead the new National Institute for Health Protection, for example, is not only questionable because she sits as a Tory backbench peer but also because her performance leading NHS track and trace has been less than impressive – certainly not “world class”. Even the health secretary’s defence of her appointment was far from convincing.
But maybe that’s the problem. The UK has, as Peter Hennessy has eloquently warned, a constitution that relies on the “good chaps theory of government”. That structure now looks incredibly vulnerable when faced with a prime minister and key advisers who reject the rules, lack self-restraint and engage in populist posturing. The result is sacrificial statecraft wrapped around a naive vision of populist technology. Ditch it now.