Oversight Is Not the Same as Accountability

Oversight Is Not the Same as Accountability
What Northern Ireland’s arm’s-length body review reveals about a fault most governments have

In March 2026, the Northern Ireland Audit Office published a review of how government departments work with their arm’s-length bodies. The headline finding was that 22 of the 81 bodies requiring a Partnership Agreement – the document that establishes who decides what, and who answers for it – still did not have one, more than five years after the requirement was introduced.

That is a compliance failure, and it will be read as one. But two other findings in the same report are more interesting, and together they describe something much harder to fix than a backlog of unsigned documents.

The first: nobody knew how many bodies there were

When the auditors began, they discovered that neither the Department of Finance nor any other department maintained a definitive list of arm’s-length bodies. They had to write to each Permanent Secretary and ask. The returns produced a figure of 101. A separate list, supplied by the Department of Finance to the Assembly’s Finance Committee two months later, produced 140 public bodies, of which the department considered 95 to be arm’s-length bodies, and the two lists did not reconcile. Some bodies that departments said were ALBs were not counted as such by Finance; some that Finance counted were not recognised as such by their own sponsoring department.

The reason given was that Northern Ireland has no formal definition of an arm’s-length body.

These organisations spend roughly 77 percent of the Executive’s budget, £14.76 billion of a £19.20 billion departmental expenditure limit in 2024-25. The category through which three-quarters of public money flows has no agreed definition, and no one holds a list.

The second: oversight bears no relationship to what it oversees

The report sets out how many staff each sponsor team employs. The pattern is worth reading slowly.

Five point two staff in the Department of Health oversee the five health and social care trusts, with combined expenditure of £6.91 billion. Six point six one staff in the same department oversee the Northern Ireland Fire and Rescue Service, which spends £130 million.

Eleven staff in the Department for Communities oversee Sport NI, with a budget of £8.3 million. Thirteen oversee the Housing Executive, at £369.6 million.

Nine staff in the Department of Education cover the Education Authority, the Council for Catholic Maintained Schools and CCEA between them – £3.38 billion.

Within the Department for the Economy, Invest NI (£113.9 million) has 12.8 sponsorship staff; Northern Ireland Screen (£21 million) has 9.5.

There is no version of these numbers in which oversight capacity tracks scale, risk, or consequence. The auditors note as much, and observe that the differences are too large to be explained by particular governance issues.

And the third finding, which explains the other two

In 2019 the Department of Finance introduced “proportionate autonomy,” guidance explicitly designed to reduce unnecessary oversight of bodies that had demonstrated they could be trusted to run themselves. Between 2019 and 2025, staffing in sponsor teams increased by 10.9 percent.

A reform intended to reduce oversight produced more of it.

The pattern underneath

It would be easy to read all this as administrative failure: slow implementation, competing priorities, a pandemic, and long periods without ministers. The report notes all of those, fairly.

But look at what the three findings have in common. Nobody holds the list, because holding it was nobody’s job and nothing followed from its absence. Sponsor teams are sized by history and local circumstance rather than by risk, because no mechanism prices the mismatch. And proportionate autonomy failed to reduce oversight because the guidance changed what departments were permitted to do without changing anything about what they were rewarded for doing.

That last point is the load-bearing one. A sponsor team that reduces its scrutiny of a body and is later criticised when something goes wrong carries the consequence personally and immediately. A sponsor team that maintains excessive oversight carries no consequence at all: the cost of over-scrutiny is borne by the arm’s-length body, in paperwork and delay, and it never appears as a finding against anyone.

Given that asymmetry, more oversight is the rational response to a guidance note recommending less. The people running sponsor teams are not defying the guidance. They are reading their exposure correctly.

Which is why the Partnership Agreements matter, and why signing them will not be enough

A Partnership Agreement defines the governance framework, the respective roles and responsibilities, the delegated authorities, and what requires departmental approval. In substance, it is a decision-rights document. Where one is absent, an accounting officer is personally answerable for outcomes whose boundaries have never been settled with their department.

The Department of Finance was asked why nobody had been appointed to oversee implementation. It said it does not assume a policing role, and that accounting officers are responsible for ensuring guidance is followed. Which is a coherent position, and it also explains the outcome: a requirement with no owner, no deadline – the guidance deliberately set none – and no consequence for non-compliance produced 27 percent non-compliance over five years.

So the twenty-two missing agreements will now be progressed, and the recommendation to progress them is the right one. But the stakeholder evidence in the same report suggests what happens next. Of those bodies that do have agreements, the auditors heard: “no relationship with the department”, “sponsor branch does not understand the ALB’s business”, “parent/child relationship”, and “creation of partnership agreement has not happened in the spirit of how it was supposed to happen.”

The NIAO’s conclusion is blunt: Partnership Agreements are not working as intended.

Signing the remaining twenty-two closes the compliance gap. It does not address what produced it.

What would

Three questions, none of which require new guidance.

For any arm’s-length body: can the accounting officer alter the things they answer for? If the department retains approval over decisions the chief executive is accountable for, the agreement has assigned exposure rather than ownership, and no amount of partnership language changes that.

For any sponsor team: what happens to us if we oversee too much? At present, nothing. Until reducing unnecessary oversight carries as much protection as maintaining it, proportionate autonomy will remain a document rather than a practice.

And for the centre: who is worse off if the list doesn’t exist? Nobody, currently. Which is why it didn’t.

None of this is peculiar to Northern Ireland. Every government with an arm’s-length estate has a version of it, and the Cabinet Office’s own sponsorship code – which the NIAO recommends Northern Ireland consider – exists because the same pattern was found in Whitehall. What is unusual here is the clarity of the evidence, and the fact that a review of the whole landscape is already under way.

The opportunity in that review is not to write better guidance. It is to notice that the last guidance was good, was widely agreed, and produced the opposite of its intention, and to ask what would need to be true for the next one to do better.