Loose Threads: The UK “Delivery Teams”

19/05/2026

There is a very specific discourse about delivery and implementation that keeps happening in the same way. A government announces that the state is broken. A minister commissions new teams, new advisers, and new task forces. The centre tightens its grip. And then, after a year or so, the same problems return, wearing slightly different clothes. The new thread never quite gets picked up. More often than not, it gets rewound.

Last week, Darren Jones, the Chief Secretary to the Prime Minister and Chancellor of the Duchy of Lancaster, announced delivery teams for every Whitehall department, delivery advisers in ministerial private offices, and taskforces modelled on the Vaccine Taskforce for major implementation projects (Jones, 2026).

Then the King’s Speech picked up the theme, promising stronger accountability and productivity (Prime Minister’s Office, 2026). The instinct, on my read, was to add capacity, tighten control, and make the centre feel in command. That instinct is entirely understandable. I have watched it play out in three jurisdictions over four decades, and the pattern is remarkably stable. The results, however, are always partial. These moves treat failure as a shortfall of capability and control. My thinking, reading, and writing across this Substack, my empirical work and in my doctoral research point to a different shortage: the quality of the space that translates policy into delivery.

I have been calling this the interpretive hinge. Readers of the Te Rā Whakamana series will recognise the concept. It is the slow, relational, sense-making work that ensures ministerial intent is legible to those who must deliver it. It is the capacity to surface what implementation will actually cost, to keep policy design and operational reality in the same room long enough to rehearse failure, share risk, and adjust. It is not a unit. It is not a dashboard. It is not a reporting line. It is practice: face-to-face translation, delegated judgement, proportionate information demands, and the calibrated sharing of responsibility. It is the small, often invisible labour that stops a policy from collapsing once it leaves headquarters.

So when the centre announces delivery reform, I find it is worth reading the announcement against the interpretive hinge.

The same structural problem

Jones’s diagnosis is blunt. Capacity is inconsistent, dependence on consultancies is rising, and the state is broken (Jones, 2026). The prescription follows the diagnosis logically enough: if the problem is a lack of capability, add more capability. Delivery teams in every department. Advisers in private offices. Task forces modelled on the one that worked for vaccines. Each measure assumes that what the system lacks is people and coordination at the centre.

But the question worth pausing on is whether any of this, however well resourced, can actually reach the place where implementation happens. Because if the problem is not capability but the institutional spaces between policy and delivery systems, then more capability at the centre simply produces more centre. It does not improve or necessarily lift the hinge work.

How we got here

Some historical context is worth establishing before turning to the present announcement. Westminster systems do not inherently separate policy and delivery. New public management reforms did that. In the United Kingdom, the Ibbs Report of 1988 proposed hiving off the executive functions of government into semi-autonomous agencies, and the Thatcher government moved quickly: 155 Next Steps agencies were created over the following decade, until three quarters of the civil service worked at arm’s length from policy departments (Efficiency Unit, 1988; Elston, 2013).

That reform disaggregated delivery from policy by design, splitting departments into small policy cores close to ministers and larger delivery agencies farther away. Blair’s “joined-up government” from 1997 was an explicit attempt to repair the fragmentation, but in practice it layered coordination mechanisms atop the separation rather than reversing it (Ling, 2002). The essential agency model, as Elston (2013) concludes, remains largely intact.

In Aotearoa, the State Sector Act 1988 and the accompanying reforms went further. Operational and service delivery functions were hived off into Crown entities and state-owned enterprises, leaving most departments as policy ministries advising their portfolio ministers. A small number of departments retained both policy and delivery functions, but they were the exception. The separation of policy from delivery was not merely an accountability arrangement. It was bedded into the institutional landscape through legislation, and despite the problems of fragmentation it created, it proved extraordinarily durable.

The Review of the Centre in 2001 to 2002 gave the fragmentation a formal diagnosis and recommended lateral, problem-driven approaches to reconnect what the reforms had pulled apart (State Services Commission, 2001).

But governments had already been trying to solve the problem before it was officially named. As I traced in a recent Loose Thread on the Prime Minister’s role, each government furnished the space between politics and administration with its own coherence architecture: Bolger and Shipley with the Strategic Result Areas and Key Result Areas through the 1990s, Clark with whole-of-government governance narratives managed through exceptional Cabinet discipline, Key and English with the Better Public Services result areas, the Performance Improvement Framework and input and output costing tools against four and ten year plans, and Ardern with the child poverty targets and the Wellbeing Budget.

Each mechanism was an attempt to make a vertically designed system work laterally. Each was discontinued by the subsequent government. The vertical architecture proved more durable than any of the instruments layered atop it. In both countries, three decades of attempted correction have not undone the original unbundling of policy from delivery. Policy still sits in the headquarters, close to the ministers. Delivery still happens in agencies, regions, local authorities and contracted providers, where citizens and communities actually are. The two rarely meet laterally. Where they connect at all, it is through vertical lines, often through poorly drafted contracts: upward to ministers through reporting, downward to frontline workers through targets. What is missing is the lateral space where policy is translated so it makes sense in practice and is tested before it fails in public.

The right instinct, the wrong architecture?

Jones’s most recent announcement suggests he may understand part of the problem. He has proposed delivery teams led by senior civil servants in every department, experts with delivery experience embedded in ministerial private offices, the task force model used by the Vaccine Taskforce with direct lines to the top of government, a reduction in the layers of approval that slow decision-making and secondments from experts from outside the public service (Jones, 2026).

The intention is not wrong. Placing people with practical, frontline experience closer to ministerial decision-making is precisely the kind of move that is required. In my doctoral research, I found that practitioners across the advisory system share a common floor beneath their otherwise divergent postures on what free and frank advice is: an expectation that advice will always be practical. Not political. Not abstract. Not theoretical. Practical: what I describe as practical statecraft, the knowledge of what can actually be done, what implementation will cost, and what will break if the design is wrong (Te Kawa, 2026).

It seems to me that knowledge is exactly what Jones is trying to get into ministers’ private offices and into the delivery teams. But the question is whether his proposal is sufficient.

An adviser with delivery experience who sits in a private office is closer to the minister, not to the frontline or the work at the hinge. A delivery team reporting to a permanent secretary still operates within a vertical line. A task force with a direct line to the top of government is, by definition, a vertical instrument.

So, while the practical statecraft these people carry may be exactly what is needed, it is being inserted at the top of a system when the problem lies in the middle: in the lateral, often ambiguous spaces between headquarters and the places where citizens and communities encounter the state.

Practical statecraft does not only need to reach the minister. It needs to flow in both directions, connecting ministerial intent to operational reality and feeding operational reality back into design. If you don’t resource and support the hinge work – the slower, relational work of translation and sense making – even well-placed and the best practical statecraft risks being absorbed into the reporting architecture it was meant to challenge.

The purple zone, again

The evidence for this is not merely theoretical. My Whitehall’s Woes pieces set it out at length, and the Institute for Government’s Whitehall Monitor 2026 points to dysfunction originating in what I have called the purple zone: the contested institutional space where political judgement, ministerial ambition, official advice, and operational reality are forced to meet (Institute for Government, 2026).

Political incoherence degrades administrative capacity in ways that reorganisation, new units, or fresh public appointments simply do not cure. Leadership matters, but its effect is conditional on the coherence of the strategic frame within which officials are asked to act.

That conditionality shows up in incentives. Ministers are told to encourage risk-taking. Ministers and officials then learn the lesson of scapegoats when embarrassment strikes. The Mandelson vetting fallout and the departures it produced, which I explored in last week’s “Dear Colleagues” post, made that dynamic painfully visible. The constitutional conventions that should protect free and frank advice too often create a regime where candour is simultaneously expected and penalised. More staff in private offices will spread the pressure and amplify the noise. They will not necessarily dissolve the underlying asymmetry.

What Aotearoa tried, and what happened next

We have seen a different organising logic attempted in Aotearoa. After the Review of the Centre in 2001 to 2002, Wellington piloted circuit-breaker teams: locally constituted, multi-agency groups assembled around concrete, intractable problems (State Services Commission, 2001). They were bottom-up, problem-driven, and grounded in relationships. These are precisely the conditions the hinge requires.

Over time, however, the centre reasserted control. Better Public Services targets and Wellington-managed results took over, and the circuit-breakers receded. Today’s Regional Public Service Commissioners are centrally appointed and dual-hatted as MSD Regional Commissioners (Te Kawa Mataaho Public Service Commission, n.d.). Their language is collaborative, but their architecture remains vertical. The difference matters. One model starts from the problems a community or regulated party experiences, and what they think the solution might be. The other starts from the centre’s need for coherence.

This is not abstract. It changes who pays to interact with the state. The administrative burden framework, which I have been developing across a separate series here, shows that learning costs, compliance costs, and psychological costs fall hardest on those least able to bear them. The Māori Communities COVID-19 Fund succeeded in part because officials deliberately lowered those costs: outcomes-based contracts, proportionate reporting, with experienced staff who helped newer providers navigate the system. The interpretive hinge held around the MCCF because it was place-based, relational, and judged success by whether whānau and communities felt ownership of the delivery, not whether central metrics checked out.

A task force with faster procurement authority does not automatically mitigate those burdens, nor does it sit at the interpretive hinge where its work occurs between the policy and delivery systems.

The questions nobody is asking

I think Jones is right to say the state is broken in important respects. But both the Waitangi Tribunal archive and Practical State case studies I have been highlighting suggest a narrower diagnosis.

The state was built around particular distributions of authority, legitimacy, and recognition, and it reproduces those distributions irrespective of outcomes. So the question is not merely how to speed up the machinery. It is whose questions the state is prepared to answer, and whose it is not.

Delivery teams might be a beginning. But they will not fix the hinge unless they start by asking different questions. Three questions, in particular, seem worth posing early and deliberately, because the shape of delivery changes depending on whether anyone asks them at all.

The first concerns cost, and specifically whose costs rise. Every new compliance requirement, every reporting template, every approval layer carries learning costs, compliance costs, and psychological costs. These fall hardest on those least able to bear them. The question for any delivery design is therefore not whether it improves central visibility, but whether it lowers the burden on the people and communities at the other end of the chain.

The second concerns translation. Policy does not arrive at the frontline in the form it left the minister’s office. It is interpreted, adapted, and made legible through local relationships and delegated judgement. The question is which relationships are essential to that translation, and who will share the risk if implementation falters, rather than reaching for the scapegoat.

The third concerns timing. Too often, ministerial intent meets operational reality only after a policy decision has been taken, and by then the room for adjustment has narrowed to the point of embarrassment. The question is how that encounter can be staged before a proposal reaches Cabinet, not after failure forces a retreat.

These are operational questions, not rhetorical ones. They require lateral forums, place-based translation roles, proportionate information regimes, and explicit risk-sharing between ministers and officials. Without that architecture, faster procurement, shinier private offices, and more reporting will make central government feel better while doing little for the citizens whose lives policy is meant to change.

The instinct to fix the machinery is honourable, and the attention to practical statecraft is, in itself, a significant development. But the conversation about delivery keeps stalling at the same point, and the reason it stalls is that the question being asked is “how do we make the centre stronger so we have more control?” when the question that matters is “how do we build the sort of public institutions Aotearoa New Zealand needs: the sort that translate intent into practice, closest to the people whose lives depend on it?”

References

Efficiency Unit. (1988). Improving management in government: The Next Steps. HMSO. https://www.civilservant.org.uk/library/1988_improving_management_in_government_the%20_next_steps.pdf

Elston, T. (2013). Developments in UK executive agencies: Re-examining the ‘disaggregation-reaggregation’ thesis. Public Policy and Administration, 28(1), 66-89. https://doi.org/10.1177/0952076711432580

Institute for Government. (2026). Whitehall Monitor 2026. https://www.instituteforgovernment.org.uk/publication/whitehall-monitor-2026

Jones, D. (2026a). Move fast. Fix things. Modernising Whitehall to deliver for Britain [Speech]. GOV.UK. https://www.gov.uk/government/speeches/move-fast-fix-things

Jones, D. (2026b). All departments ‘will have a delivery team led by a top civil servant’. Civil Service World. https://www.civilserviceworld.com/professions/article/darren-jones-all-departments-will-have-a-delivery-team-led-by-a-top-civil-servant

Ling, T. (2002). Delivering joined-up government in the UK: Dimensions, issues and problems. Public Administration, 80(4), 615-642. https://doi.org/10.1111/1467-9299.00321

Prime Minister’s Office. (2026). The King’s Speech 2026. GOV.UK. https://www.gov.uk/government/speeches/the-kings-speech-2026

State Services Commission. (2001). Report of the Advisory Group on the Review of the Centre. New Zealand Government.

Te Kawa, D. (2026). The bounded plurality of free and frank advice: Practitioner postures, tensions and the practical state in Aotearoa [Doctoral thesis, University of Canterbury].

Te Kawa Mataaho Public Service Commission. (n.d.). Our work: Regional Public Service Commissioners. New Zealand Government. https://www.publicservice.govt.nz/publications/strategic-intentions/our-work