The UK government’s consultancy spending target will probably be met. The consultancy spend won’t fall by as much. Both things can be true because the target and the spend are measuring different things.
The Public Accounts Committee’s March 2026 report found that government ‘does not have a grip on levels of consultancy spend’. HM Treasury estimated central government consultancy expenditure at £1.36 billion for 2022-23. Other data sources put it at up to £2.23 billion. That is a gap of nearly £900 million. The Cabinet Office, tasked with halving consultancy spend, chose to benchmark against the lower figure. The PAC’s characterisation of this decision was pointed: the Cabinet Office ‘appears unconcerned at the many and obvious inconsistencies’, and non-compliance with existing directives is ‘going unmonitored’.
Stated clearly in the report: ‘Departments often have large contracts that combine consultancy and professional services, making it difficult for departments to isolate the spend on consultancy services only.’ And it names the consequence: ‘The variability in definition and lack of rigour in reporting consultancy spend may also allow departments to reclassify spending and under-report rather than reduce spending.’
That is described in the report as a data integrity risk. It is actually the incentive the target creates.
But reclassification within a category label is the shallow version of the problem. The deeper version is substitution across the full external resourcing taxonomy.
Consultancy is one category in a continuum that runs from permanent employees through contingent workers, contractors, professional services firms, and management consultants. These categories are managed separately in government: different budget lines, different approval thresholds, different audit frameworks, different definitions. A department reducing consultancy spend while increasing its contingent workforce or expanding a managed services contract is not breaking any rules. It may not even recognise it is doing the same thing through a different door.
The firms on the selling side understand this. A major professional services firm does not have a consultancy relationship with a government department. It has an account. The work may flow through consulting, technology implementation, managed services or contingent labour placement depending on which vehicle is available, which budget is accessible, and which category is currently under political pressure. Cross-category account management is not a workaround. It is the commercial model. When consultancy spend is squeezed, the account manager’s job is to find the next vehicle. The relationship continues. The label changes. The buying organisation, tracking categories rather than relationships, sees a compliance win.
This is not new behaviour and it is not hidden. Selling organisations have structured their service lines and contract vehicles to maintain this flexibility for decades. Buying organisations have largely chosen not to see it because seeing it would require managing external resourcing as a single strategic question rather than a collection of procurement categories.
The evasion is not confined to the selling side. Internal stakeholders in the buying organisation are equally active participants. When procurement scrutinises consultancy spend, the same capability requirement reappears as a contractor engagement: visible to HR under a different framework, subject to different approval thresholds, outside procurement’s reporting scope. When HR starts watching contractor headcount, the engagement is restructured as a statement of work. A statement of work is particularly effective as an escape vehicle because it is not a person. It does not appear in headcount counts. It does not appear in contractor registers. A statement of work issued to provide ‘digital transformation advisory support’ is, in commercial substance, a consultant. In the data, it is a deliverable.
Governance fragmentation enables this. Procurement, HR and finance each own different segments of the external labour taxonomy. Each function’s reporting covers its own category. No single function has visibility of the whole. The stakeholder who needs the capability will find whichever door is currently unguarded. This is not circumvention. It is rational navigation of a governance architecture that was never designed to see the full picture. The buying organisation is not being misled by its suppliers. It is being navigated by its own people.
The PAC’s own findings contain the demand signal that makes substitution structurally inevitable. One of the committee’s recommendations is that the government publish a civil service workforce plan by May 2026, including an assessment of skills gaps — specifically to understand why it is so dependent on external consultants for digital capability. That dependency is not a spending preference. It is a capability gap. The internal workforce cannot meet the demand. External providers fill it. A category-specific spending target that reduces the consultancy headcount without closing the internal skills gap does not reduce demand for external capability. It creates a vacancy in the external resourcing mix that contingent workers or contractors fill, through a different budget line, under a different approval process, without appearing in the consultancy spend figures.
The PAC’s recommendations are accurate responses to the stated problem: better data, clearer definitions, a workforce plan. None of them changes the structure. Better definitions, once finalised, will produce cleaner data for the category that is currently targeted. The same bundled contracts will still exist. The same account relationships will still be managed cross-category. The same skills gaps will still generate demand. What changes is that the reclassification will be better documented.
The mechanism that would fix this already exists in another part of government. HMRC’s off-payroll working rules exist precisely because ‘contractor’ is a label that obscures an employment relationship when there is a financial incentive to apply it. The test is not ‘what does the contract say?’ It is ‘what is the nature of the engagement?’ Indicators of control, substitution, and mutuality of obligation. Substance determines routing. The label is irrelevant.
A spending control built on the same logic would apply an engagement substance test at the point of authorisation. Advisory work that shapes decisions is consultancy regardless of whether it arrives under a consulting agreement, a contractor rate card, or a statement of work. The test has to happen before the contract is signed, not after the invoice is coded. Documents produced after an engagement to justify a classification are not evidence of what the engagement was. They are evidence of what the parties agreed to call it. The artefacts are designed to clear a low bar. That is the point of them. Government has already learned what happens without upfront substance testing: IR35 reform in the public sector produced years of disputes and blanket status determinations because the financial incentive to label an employment relationship as something else did not disappear because the label was convenient. HMRC made the label irrelevant by testing the substance at source. The same financial incentive is operating in the consultancy spending context. The response in one case was a legislative substance test with enforcement consequences. The response in the other, as of March 2026, is a definition the Cabinet Office has been drafting for 16 months.
This is not an unsolved problem. Australia’s APS Strategic Commissioning Framework defines the three-way split operationally: a consultant produces an independent intellectual output under their own direction; a contractor works under Commonwealth direction to produce a Commonwealth product; labour hire supplies a worker whose day-to-day supervision belongs to the agency. The test is the nature of the engagement, who directs the work and who owns the output, not the label on the contract. New Zealand’s government procurement guidance applies the same logic: outcome responsibility and supervision determine the classification. Both frameworks are applied at the point of engagement authorisation, not retrospectively. Both have been in operation for years. The Cabinet Office is not navigating territory that comparable jurisdictions have not already mapped.
The PAC’s workforce plan recommendation addresses a different but related problem: why does the demand for external capability exist in the first place? Closing internal skills gaps is the only intervention that reduces the demand rather than redirecting it. But a workforce plan that asks why the dependency exists without asking where the demand goes when one category is restricted will answer one of three questions. The routing problem and the governance fragmentation problem remain.
The Chancellor’s pledge was announced in November 2024. The Cabinet Office was still working on revised definitions for consultancy in July 2025. At the time of the PAC report in March 2026, no revised definition was in place. The spend target is benchmarked against a figure that other sources contradict by nearly £900 million. None of this is surprising. A definition that takes more than 16 months to finalise is not technically difficult. It is a decision that reduces the flexibility of departments to reclassify, and that flexibility has value to the people who currently use it.
The target will be met on the chosen benchmark. The firms will still be in the building. They will just be on a different invoice.
Source: UK Public Accounts Committee, ‘Chancellor’s pledge to cut billions in govt’s consultancy spending in doubt — PAC report’, March 2026, https://committees.parliament.uk/committee/127/public-accounts-committee/news/212566/chancellors-pledge-to-cut-billions-in-govts-consultancy-spending-in-doubt-pac-report/