Failings in LATCos and the Lessons to be Learnt

Executive summary

Local authority trading companies are now a structural feature of English local government. There were 743 of them across 59.2% of councils by September 2018, and more than 850, owned by 62% of councils, by January 2025. They employ roughly one in twenty local government workers. They were not a fringe experiment; they became the delivery model.

A significant minority of them have failed, and the failures have been expensive. Four examples, drawn from the nine cases examined in section three: Nottingham wrote off around £38.1m on Robin Hood Energy. Croydon lent Brick by Brick £200m, never received a dividend or a penny of interest, and in April 2026 wrote off £68m as the company entered liquidation having delivered three council homes in eleven years. Bristol capped its losses on Bristol Energy at £43.8m. Portsmouth wrote off £3.5m on an energy company that never supplied a single customer.

The central finding of this paper is that these were not a series of separate disasters. They were the same disaster, repeated. Read the public interest reports, the Best Value inspections and the independent reviews side by side and the same nine defects recur with a consistency that borders on the mechanical:

1. An untested business case. Built once to secure approval, then never re-run as assumptions moved.
2. A shareholder function that was absent or nominal. No one unambiguously responsible for protecting the council’s interest as owner.
3. Directors in unmanaged conflict. Duties owed to the company under the Companies Act, colliding with duties to the authority.
4. Management information that could not support a judgement. Late, aggregated, out of date, or simply absent.
5. Professional advice ignored or quietly suppressed. Commissioned, delivered, and never reaching the decision-maker.
6. Circular and related-party lending. The council lending money the company used to repay the council.
7. No exit strategy. No evaluated wind-down, sale or insourcing option, at any price.
8. Accounting used to defer the reckoning. Minimum revenue provision set implausibly low; impairments never assessed.
9. Scrutiny structurally disabled. Members given too little time, information or training to challenge.

Section three examines the nine cases in which these defects appeared. Section four sets out all nine failings in full, with the named report evidencing each.

Government’s own analysis reaches the same conclusion about timing. Reviewing 24 councils in Best Value intervention, the Ministry of Housing, Communities and Local Government recorded governance as a driver in 18 cases, financial failure in 17, and leadership in 16 — and reported “widespread agreement across sector bodies, best value experts, and MHCLG that all interventions happen too late.”

That is the gap this paper is written into. Every one of the failures examined here was visible in the company’s own affairs well before it was visible in the council’s accounts. What was missing was not information. It was a structured, repeatable way of asking the right questions early enough for the answers to still be useful. That is what the Nullam Method exists to provide, and section five sets out precisely which of the twelve modules would have surfaced which failing, and when.

1. Why this matters now

The commercial turn in local government was not a fashion. It was a response to a funding shock of a scale that has no modern precedent. The National Audit Office found that central government funding to English local authorities fell 49.1% in real terms between 2010-11 and 2017-18, with authorities’ overall spending power down 28.6% across the same period. Over a slightly longer series, the Institute for Government puts the real-terms cut to grant funding at 40.1% between 2009/10 and 2019/20, from £51.8bn to £31.0bn, and notes that the most deprived fifth of authorities lost 33.3% of spending power against 15.2% for the least deprived.

Councils facing that arithmetic had three options: cut services, raise what little tax they controlled, or generate income. Most tried all three. Trading companies were the vehicle for the third, and the sector grew accordingly. Localis found in March 2015 that 58% of surveyed councils already operated a trading company, against a 2010 projection of 15% adoption. Grant Thornton’s 2018 census counted 743 companies. UNISON’s FOI-based Trading Places research, published in January 2025, counted more than 850.

The political weather has since changed completely. The Public Works Loan Board rewrote its lending terms in November 2020 to stop authorities borrowing primarily for yield, requiring the section 151 officer to certify no such intention for three years ahead. CIPFA’s revised Prudential Code, issued on 20 December 2021, states plainly that “an authority must not borrow to invest primarily for financial return.” And statutory guidance on local government investments already treats loans to a council’s own wholly-owned company as investments in their own right, requiring proportionate exposure limits, expected credit loss modelling and formally agreed lending caps.

The consequence is that a LATCo formed in 2016 is now operating under a materially different regulatory settlement from the one it was designed for. Meanwhile the intervention regime has accelerated sharply: of the 29 councils that entered Best Value intervention between 2014 and 2025, 22 did so in the five years from 2020. The local audit backlog has meant that in some cases the accounts which would have shown the problem were years from being signed. Spelthorne’s accounts for 2018/19 to 2022/23 remained unaudited when its Best Value inspection reported.

For a board or an audit committee, that combination — a changed rulebook, a delayed audit function and an intervention regime that everyone agrees arrives too late — makes self-assessment less of an administrative nicety and more of the only early warning system reliably available.

2. How we got here

2.1 The legal architecture

Two statutory routes underpin nearly every LATCo. Section 95 of the Local Government Act 2003, activated in England by the Local Government (Best Value Authorities) (Power to Trade) (England) Order 2009, permits a best value authority to do for a commercial purpose anything it may do for its ordinary functions. Alternatively, section 1 of the Localism Act 2011 gives a general power of competence to “do anything that individuals generally may do.”

Two features of that architecture deserve more attention than they usually receive, because both create risk rather than merely permitting activity.

First, the company form is mandatory, not optional. Section 4(2) of the Localism Act provides that where an authority does things for a commercial purpose under the general power, “the authority must do them through a company.” Section 95 of the 2003 Act carries the same constraint. Councils did not choose to place commercial activity at arm’s length in a separate legal entity with its own board, its own duties and its own information boundary. Parliament required it. The governance distance that so many reports later criticise is a feature of the statute, and it therefore has to be managed deliberately — it will not close on its own.

Second, a director’s duties are owed to the company, not to the council. Sections 171 to 177 of the Companies Act 2006 apply in full to councillor-directors and officer-directors: act within powers, promote the success of the company for the benefit of its members, exercise independent judgment, exercise reasonable care and skill, avoid conflicts of interest, declare interests in proposed transactions. A cabinet member sitting on the board of a company their own authority owns and funds is subject to a genuine, structural and continuing conflict. That is not a criticism of any individual. It is a description of the statutory position, and it is the single most frequently cited governance failing in the reports examined in section three.

Alongside these sit the Teckal arrangements permitting a council to award contracts to its own company without competition. Codified in regulation 12 of the Public Contracts Regulations 2015 — the control test, the 80% activity test, and the bar on private capital — these were replaced for new procurements from 24 February 2025 by the “vertical arrangements” provisions at Schedule 2, paragraph 2 of the Procurement Act 2023. The wording on private capital is materially stricter: a person is not regarded as controlled “if any person that is not a public authority holds shares.” Any legacy structure carrying even a token private shareholding warrants fresh advice.

Finally, the older control regime in Part V of the Local Government and Housing Act 1989, with the Local Authorities (Companies) Order 1995, remains live and is frequently overlooked. It classifies companies as controlled, influenced or arm’s length, and non-compliance with an applicable order renders expenditure unlawful for audit purposes.

2.2 The 2009 Order already required what most councils did not do

It is worth pausing on article 2(2) of the 2009 Order, because it is the quiet centre of this whole story. Before exercising the trading power, an authority must “prepare a business case in support of the proposed exercise of that power” and have it approved. Article 2(3) requires the authority to recover the costs of any accommodation, goods, services or staff it supplies to the trading operation.

The obligation to build a real business case, and to charge the company properly for what the council gives it, has therefore existed in English law since 1 October 2009. Section three is, to a very large degree, a record of what happens when that requirement is treated as a formality at formation and never revisited afterwards.

3. What actually went wrong

What follows is drawn entirely from published official sources: statutory public interest reports issued by external auditors, Best Value inspection reports commissioned by the Secretary of State, independent reviews published by the councils themselves, and parliamentary committee reports. Where figures are contested between sources, the more conservative and better-evidenced number is used and the basis is stated.

3.1 The scale, at a glance

Figures are confirmed write-offs or the council’s own stated ceiling, not gross exposure. Woking and Thurrock are excluded from this table because their losses are treasury and investment losses rather than trading company losses, though both were transacted substantially through council-owned vehicles.

3.2 Robin Hood Energy, Nottingham City Council

Launched in September 2015 as a not-for-profit supplier, Robin Hood Energy also ran white-label supply for around ten other councils. By 31 March 2019 its cumulative losses stood at £34.4m, with a worst single year above £23m. The council had invested £43m and provided £16.5m of guarantees. The customer book was sold to Centrica in September 2020 for approximately £26m, leaving an estimated £38.1m write-off. In liquidation, creditors claimed £67.1m and received £13.7m — roughly twenty pence in the pound. Ofgem revoked the electricity supply licence with effect from 12 April 2021.

Grant Thornton’s statutory Report in the Public Interest, issued in August 2020, produced the phrase that has followed the sector ever since. Governance, it found, was “overshadowed by the council’s determination that the company should be a success”, creating what the auditors called institutional blindness to escalating risk.

The thirteen recommendations map almost directly onto assessable control weaknesses:

• “insufficient sector (or general commercial) expertise at non-executive Board level” — a board composition failure, testable at any point from incorporation onwards.
• “insufficient understanding within the Council of RHE’s financial position, partly due to delays in the provision of information” — a management information failure. The council was not being told, and had no mechanism to compel.
• “lack of clarity in relation to roles within the governance structure” — an undefined shareholder function.
• Company risks sat outside the council’s corporate risk management and escalation processes entirely.
• Recommendation 1 required the council to “urgently determine the future of Robin Hood Energy, with options properly evaluated” — which is to say that at the point of failure, no evaluated exit option existed.

Nottingham’s difficulties were not confined to the energy company. Ring-fenced Housing Revenue Account money was spent unlawfully on general fund services from 2014/15, a figure that escalated from £15.86m at the first section 114 notice in December 2021 to more than £51m by January 2023. Its district heating company, Enviroenergy, was taken back in-house in January 2022 with more than £11m of council loans outstanding. A second section 114 notice followed in November 2023 and commissioners were appointed in February 2024.

3.3 Brick by Brick, London Borough of Croydon

Incorporated in 2015 to build homes on council land and return dividends and capital receipts, Brick by Brick received £200m of council lending between 2015 and 2020 under loan agreements totalling £221m. It never paid a dividend and never paid loan interest. By March 2020, £14.4m of interest was outstanding and £110m of loans due for repayment had not been repaid. Croydon’s 2020/21 budget nonetheless assumed a £3m dividend from a company that had never paid one. The company passed a resolution to wind up on 2 April 2026; the council’s write-off is £68m. Over eleven years and £200m of borrowing it delivered three council homes.

Grant Thornton’s public interest report of October 2020 and the independent Penn Report, published by the council in February 2023, together produce the most complete anatomy of LATCo failure available in the United Kingdom. Among the findings:

• Loan covenants were unmonitored and breached. The 90-day accounts covenant was unmet at 153 days; the interest covenant was unmet; the repayment covenant was unmet, with no evidence that written consent was ever formally requested.
• Management information did not exist in usable form. The Penn Report found no recognisable profit and loss statements or balance sheets, and forecasts aggregated without the detail needed to assess performance.
• The company had not registered as a Shared Ownership Provider. This was discovered only in January 2020, when buyers could not obtain mortgages. The auditors described it as indicating “a lack of understanding.”
• Financing became circular. A July 2020 decision saw the council borrow a further £30m to buy properties from its own company; the auditors found the decision “did not address the circular nature” of the transaction.
• The group structure was opaque. London Borough of Croydon Holdings LLP was dissolved by compulsory strike-off in December 2019 for failure to file accounts — £55.1m of loans sat behind that vehicle — and this was never reported to Cabinet or the Audit Committee.
• Governance was replaced by relationships. The Penn Report found the arrangement “over reliant on relationships rather than formal structure,” with no complete repository of loan agreements, and recorded that the company’s chief executive had “direct access to previous Cabinet members” who “instructed others to make happen what they wanted.”
• Reporting was manipulated. The report describes “officers being asked to re-word Cabinet reports to present the most favourable picture.”

“collective corporate blindness to both the seriousness of the financial position and the urgency with which actions needed to be taken”

Auditor recommendations from 2017/18 and 2018/19 had not been implemented. Croydon issued three section 114 notices; Brick by Brick accounted for over half of the overspend cited in the first.

3.4 Bristol Energy, Bristol City Council

Established in 2015 by cross-party cabinet decision, Bristol Energy received an initial commitment of £15.7m, raised to a maximum of £37.7m in 2018. Operating losses ran at £8.4m, £11.2m, £12.1m and £14.8m across the four years to 2019/20. The domestic and business books were sold in September 2020 for a combined £15.3m. The council stated its total loss “will not exceed £43.8m.”

Grant Thornton’s value-for-money work found the business plan “represented an overly unrealistic view of how BE might perform.” But the finding that should concern every shareholder representative is narrower and sharper than optimism:

“The independent shareholder advisor was unable to support the business plan — and this was not formally communicated to Cabinet.”

The advice existed. It was commissioned, it was delivered, and it did not reach the decision-makers. The auditors further found that papers at the January 2020 Cabinet “did not clearly state the risks faced by BE,” that information was out of date, and that risk reports were “buried in the appendices.” The flow of information from shareholder to Cabinet was described as inadequate, and decisions made by the shareholder representative were not routinely published.

At formation, the options appraisal “lacked a detailed and robust methodology on which the preferred option and the alternatives could be considered,” and there is no record of why an energy services company — a materially lower-risk model — was not established instead. And in October 2017 the council considered selling, then declined because a sale would not recover its investment. That single sunk-cost decision preceded roughly £27m of further losses.

One footnote captures how interconnected the sector’s risk had become: Bristol Energy’s domestic customers were sold to Together Energy, half-owned by Warrington Borough Council, which collapsed four months later.

3.5 Woking Borough Council and the ThamesWey group

Woking operated 23 wholly-owned companies, principally the ThamesWey energy and regeneration group and Victoria Square Woking Limited. Borrowing reached £1.9bn, projected to £2.4bn by 2025/26, against a net revenue budget of £24m — debt of roughly £19,000 per resident, with annual capital financing costs of £62m. Victoria Square was approved at around £460m in November 2016 and expanded to £700m by February 2021; its asset value was subsequently assessed at £300–350m against approximately £700m of debt. A section 114 notice was issued on 7 June 2023.

The external assurance review commissioned by the Secretary of State found that no shareholder function existed until October 2022 — after the council had lent well over £1bn. It found investment decisions “made without appropriate business cases and records of robust land valuation,” affordability assessed on cash flow forecasts rather than robust financial appraisal, and “little capacity and capability put in place to manage the scale of commercial activity.” Shares held by the council had never been revalued since acquisition.

Most instructive is the lending pattern. The council lent to companies which then used the money to repay the council — companies “making operational losses but still extending their borrowing to cover principal and interest repayment costs.” A revolving facility of that kind produces an income line in the council’s accounts that looks like commercial return and is in fact the council’s own capital coming back to it.

“It was the belief that the council’s commercial income and its ability to take margins on loans to companies would cover any financial eventuality. This has proved not to be the case.”

3.6 Thurrock Council

Thurrock is properly a treasury and investment failure rather than a trading company failure, but it belongs here because it demonstrates the same control weaknesses operating at maximum amplitude. The council borrowed short-term from other authorities to invest in solar bonds issued by Rockfire Capital vehicles, including £268m into Miramar in December 2017. The investment portfolio grew from £446m at the end of 2017/18 to £847m a year later and approached £1bn by early 2020. A £470m in-year deficit was reported at the second quarter of 2022/23; total borrowing reached an estimated £1.3bn. A section 114 notice followed in December 2022, and the Serious Fraud Office opened an investigation.

The Best Value inspection led by Tony McArdle OBE, published on 19 May 2023, found that the section 151 officer made investments totalling over £500m “without meaningful reference to elected members”, by treating transactions as falling below the £10m threshold or as continuations of existing arrangements, and expanded his own delegated authority for non-specified investments from £200m to £550m.

Three findings deserve particular attention:

• Professional advice was rejected and then removed. Treasury advisers Arlingclose warned in March 2018 that the strategy was “extreme” and had “moved the Council well beyond the bounds of what we consider to be prudential risk management boundaries.” The advice was rejected; Arlingclose resigned in December 2018; and the council then operated with no treasury management adviser at all for 33 months.
• The risk was recorded as an opportunity. The investment programme appeared on the strategic risk register only as an “opportunity” — never as a material risk. Internal audit was never instructed to examine the investment strategy and did not self-identify it as a priority.
• External challenge was systematically deflected. The LGA Corporate Peer Challenge of September 2018 identified significant risks; the resulting action plan appeared to dismiss the recommendations and the report was never circulated to Cabinet or public meetings. A £14m loss on Chip Chip Ltd was kept out of full Council records, and members’ questions were refused as “commercially sensitive” or “operational.”

“the Council’s lack of openness and transparency has given rise to a culture of insularity and complacency. Internal challenge has been discouraged, and external criticism and challenge have been routinely dismissed.”

3.7 Knowle Green Estates, Spelthorne Borough Council

Spelthorne’s debt rose from £4.0m in March 2016 to £1,087m in March 2024 — over £10,000 per resident. Its housing company, Knowle Green Estates, had incurred £16.0m of revenue expenditure by 31 March 2024 and, as at January 2025, had not developed a single new home. Site values had declined significantly and the £16m remained capitalised on the balance sheet. Minimum revenue provision stood at £12.4m, approximately 1% of the capital financing requirement against a prudent range of 2–3%.

The Best Value inspection led by Lesley Seary, published in March 2025, contains the single most useful sentence in this literature for anyone thinking about assurance:

“We have not seen evidence of internal audits conducted on the Council’s housing company, Knowle Green Estates, or its regeneration sites over the past three years, despite the high level of associated risks.”

The inspectors also found that sites were approved using “indicative appraisals and optimistic assumptions,” without the Red Book valuations standard practice requires; that reports were issued to Cabinet in some cases only one or two days ahead of the decision, with access “restricted to a select group of Councillors,” preventing call-in; and that senior officers displayed “optimism bias” and lacked “awareness of the true situation.” The council was found to be failing its best value duty across five areas.

3.8 Three shorter cases that make the same point

Victory Energy, Portsmouth City Council. Final business case approved by Cabinet in October 2017; scrapped in 2018 before it ever traded; £3.5m written off. Internal audit rated three areas high risk. Nearly £400,000 was spent after Cabinet approval without following procurement rules. The council failed to verify the CVs of its lead consultants for more than a year. There was no evidence that consultant day rates of £600–750 were cost effective. A company that never supplied a customer still cost the equivalent of a small service.

Foundations Homes, Liverpool City Council.Established in 2018 to deliver a promised 10,000 homes; delivered 18; liquidation approved October 2022. The council’s own Lessons Learnt report found it was created “without a clear policy strategy and without robust delivery and business plans,” that the 10,000-home promise was “communicated before funding was secured,” and — decisively — records the “absence of a shareholder agreement with reserved matters.”

Sandwell Land and Property Company.Established in January 2011 to protect the freehold of 99 education assets. Grant Thornton’s December 2021 governance review found that “on taking external legal advice after establishing SLaP it was confirmed the freehold of education assets would not pass to academies” — the company’s entire rationale was invalid, and the advice that would have revealed it was taken after incorporation rather than before. The same review found no council-owned company demonstrating effective shareholder oversight during the period examined.

4. The nine recurring failings

Set the reports beside one another and the pattern is unmistakable. Nine defects account for substantially all of the documented failure, and every one of them is observable from inside the company before it becomes visible in the council’s accounts.

Every entry is drawn from a named public interest report, Best Value inspection, independent review or council-published lessons-learnt document.

4.1 What the pattern actually tells us

Three observations follow, and they are more important than the individual scandals.

First, none of these failings is exotic. Not one requires forensic accounting or specialist market knowledge to detect. “Is there a shareholder agreement with reserved matters?” is a question anyone can ask and anyone can answer. So is “when was the business case last re-run?” and “what is our evaluated exit option?” These failures were not hidden. They were unasked.

Second, the failings cluster. No company in this paper failed on one of the nine. Croydon exhibits at least eight. A weak shareholder function permits poor management information, which disables scrutiny, which allows an untested business case to survive, which removes the trigger for an exit review, which is then concealed by accounting treatment. This is why point-in-time reviews of single topics tend to miss the problem: they examine one link in a chain that only fails as a chain.

Third, and most consequentially, the lag between detectable and detected was measured in years. Croydon’s auditors made recommendations in 2017/18 and 2018/19 that were still unimplemented when the public interest report landed in October 2020. Woking’s auditors recommended an MRP policy review in 2018/19. Thurrock was warned by its treasury advisers in March 2018 and issued its section 114 notice in December 2022 — a gap of four years and nine months. Spelthorne’s company went three years without an internal audit. MHCLG’s conclusion that all interventions happen too late is not a criticism of the intervention regime. It is a description of an assurance gap that opens years earlier, inside the company, where nobody was looking.

5. Where the Nullam Method fits

The Nullam Method is a structured self-assessment for local authority trading companies: twelve modules arranged across five layers — strategic, governance, financial, operational, and transition and exit — producing a red, amber and green score, a section-by-section breakdown and a tailored action plan delivered as a branded report. It is designed for boards, audit committees, executives and shareholder representatives, and it is built to be re-run so that movement can be evidenced over time.

It is worth being precise about what that is and is not. A self-assessment is not independent assurance, and no board should present it as such to an audit committee. External audit, internal audit and peer challenge provide independent assurance; a scored self-assessment prepares for them, feeds them, and answers their questions before they are asked. That is a narrower claim than “assurance,” and a considerably more defensible one.

What the Method does provide is the thing the case record shows was missing: a repeatable, structured instrument that asks every one of the nine questions, on a cycle, in a form a board can act on and a shareholder can read. The mapping below sets out which module addresses which failing, and — more usefully — what it would have caught, and when.

5.1 Why a repeatable score, and not another review

There is no shortage of guidance in this sector. Local Partnerships publishes company review guidance, CIPFA publishes a good practice guide and a governance checklist, the LGA maintains an improvement and assurance framework, and the major firms will conduct a governance review for a fee. A capable monitoring officer with time can assemble a serviceable review from the free material alone — and it would be dishonest to pretend otherwise.

What none of that material produces is a score you can run again. Guidance is qualitative and narrative. Reviews are point-in-time and expensive. Neither gives a board the one thing the case record shows it needs: a baseline this year, the same instrument next year, and evidence of whether the gap closed or widened.

That matters because the expectation of periodic review already exists. The Best Value duty guidance expects self-monitoring, self-assessment and early assurance. The LGA framework requires that councils undertake regular reviews of the governance of their trading companies, with outcomes reported in the Annual Governance Statement. There is, in other words, a stated expectation of recurring review with no recurring instrument to meet it. Spelthorne went three years without one, and the inspectors said so in terms.

The second gap is at the other end of the life cycle. Every framework surveyed covers formation, governance and monitoring. Almost none has a genuine transition and exit dimension — which is precisely the failing that turned recoverable losses into unrecoverable ones at Bristol, at Nottingham and at Croydon. Two of the twelve Nullam modules sit in that layer for exactly this reason.

5.2 What we would say to a board

The companies in this paper were not run by fools. They were run by capable people, under real financial pressure, doing something their council had asked them to do, inside a legal structure Parliament had required them to use. What they lacked was not intelligence or integrity. It was a habit of structured, periodic, honest self-examination, and a shareholder relationship robust enough to survive an uncomfortable answer.

Three questions are worth asking at your next board meeting, and none of them requires a consultant to answer:

If any of those three is difficult to answer, the company is already carrying one of the nine failings. Every organisation in section three could have answered them badly, in public, years before the write-off. Nobody asked.

The Nullam Method asks all three, and the ninety-odd others that follow from them, in a form that produces a score, a gap analysis and an action plan a board can act on and a shareholder can hold. The cost of asking is trivial against the cost, repeatedly demonstrated, of not asking.

Sources

All figures and quotations in this paper are drawn from published documents. Principal sources are listed below.

This paper is published by Nullam. The Nullam Method is a self-assessment framework for local authority trading companies. It is not a substitute for independent audit, legal advice or professional financial advice.