Sir Olly Robbins spent much of Tuesday morning in front of the Foreign Affairs Committee describing a Downing Street that wanted a particular outcome quickly. “Constant pressure” to get Lord Mandelson in post before inauguration. A “dismissive” attitude to vetting. A Cabinet Office pressing for the appointment to proceed without the usual process. By the time Robbins arrived in the building, in his own words, the prime minister’s nominee “had been put out there to the public, announced, blessed by the king, agreed by the US government”. The space for independent official judgment had been closed down before the process could begin.
Commentary this week will treat the affair as a personnel story. Lord O’Donnell’s intervention in The Times is welcome, and his warning about the relationship between ministers and officials is right. But the personnel frame misses the larger point. The Robbins affair is a strategic state failure, and until we consider it as such we will keep producing more of them.
What the strategic state actually means
We have argued throughout this series that the central governance question of our era is whether Britain can build a genuinely strategic state: one capable of making consequential decisions well, under conditions of technological acceleration, geopolitical volatility, and fiscal constraint. A strategic state is not one with more strategy documents. It is one whose political and administrative leadership share an account of the tasks the state must perform, the decisions that properly belong to each side of the house, and the disciplines that protect the quality of those decisions when pressure rises.
The Fifth Industrial Revolution intensifies that demand. An age in which AI reshapes the cognitive base of the economy, supply chains and alliances are being reconfigured in real time, and the social contract is being renegotiated across every advanced democracy requires a state that can hold complex trade-offs together, act at speed without cutting corners, and learn from its own decisions. This is a materially different job from the one the British state was designed to do, and it demands materially different capabilities from both ministers and officials.
Measured against this standard, the Robbins affair is diagnostic. A consequential appointment moved from political intent to formal announcement without the machinery of the state ever stopping the clock to test the assumptions. A senior official took a binary decision on the basis of an oral briefing, in part because custom forbade wider consultation, in part because the political expectation had been set long before the process began. When the underlying judgment proved contestable, the centre looked for someone to carry the weight of the exposure. This is not a strategic state in operation. It is the opposite.
The capability gap the 5IR exposes
One of the clearest implications of the 5IR framing is that capability, long treated in Whitehall as a discretionary investment to be trimmed in tight spending rounds, is now the core of the job. Ministers who cannot read a technical risk assessment or interrogate a model output will make worse decisions in an environment where policy, delivery, and technology are increasingly intertwined. Officials who have not been trained in systems thinking, complex risk, or the governance of emerging technologies will struggle to give those ministers the advice they need. A centre of government without the analytic, delivery, and foresight functions to hold the whole picture together will default, under pressure, to managing the politics of exposure rather than the substance of decision.
The international contrast is uncomfortable. France retains a functioning Secrétariat Général du Gouvernement and a Direction Interministérielle de la Transformation Publique configured for delivery from the centre. Singapore invests systematically in the capability of its senior political and administrative classes, treating it as core infrastructure. Britain, which once led in this territory through the Delivery Unit and Strategy Unit of the Blair and early Brown years, has allowed the infrastructure to erode across successive reorganisations and has not replaced it. The AI Security Institute is a genuine exception, and Anthropic’s decision to use AISI as its sole independent assessor for its most powerful model to date is a quiet validation of what a capable British state can still do. The problem is that AISI is the exception rather than the rule.
Why personnel remedies will not reach this
The instinctive response to a week like this is to reach for personnel remedies. Better spads. A firmer prime minister. A more assertive cabinet secretary. Dame Antonia Romeo is, as O’Donnell rightly notes, a serious figure who deserves the leeway to rebuild trust with ministers. But personnel remedies alone cannot fix a problem that is institutional. If the machinery does not stop the clock on consequential decisions, no individual will do so reliably. If ministers and officials have not been prepared for the substantive demands of governing in the 5IR, no amount of goodwill will close the gap. The Robbins pattern will repeat.
The handling of Sue Gray, the ousting of Sir Chris Wormald, and now the dismissal of Robbins describe a governing posture in which the political centre has lost the habit of using the state’s own disciplines to test its own impulses. Each episode has been treated as a discrete personnel story. Read together, they describe a strategic state deficit. The prime minister acknowledged in his Commons statement on Sunday that the exclusion of ministers from vetting recommendations is not a legal requirement but an administrative convention, and one he wants changed. The willingness to say so is welcome. The question is whether the recognition extends to the deeper architecture of which the vetting convention is only one small part.
What a strategic state reset requires
A serious reset would treat the machinery of the state as itself a strategic asset. That means rebuilding a delivery-from-the-centre function with the authority and capability to stress-test consequential decisions before they become political facts; a sustained programme of capability investment for ministers and officials alike, explicitly designed around the demands of the 5IR; a centre of government configured to hold the long view alongside the short cycle; and a public reaffirmation of the ministerial-official partnership as a single leadership team with a clear division of labour, in which each side backs the other in difficulty rather than distancing themselves at the first sign of exposure.
None of this is the work of a single Budget cycle or a single reshuffle. It is the work of a government that has decided, as a matter of strategic choice, that the quality of its own machinery is a precondition for everything else it wants to do. The decisions ahead on fiscal consolidation, industrial strategy, health and social care, AI governance, defence, and the green transition will not be managed well by a state that cannot protect its own senior officials from the consequences of decisions they did not take.
The prime minister, as the first to have served as a permanent secretary, is uniquely placed to understand what the strategic state now requires. Dame Antonia Romeo is uniquely placed to help him build it. The Robbins affair is not the obstacle to that project. It is the evidence that the project can no longer be postponed.