Part 26A Restructuring Plans at Five: Cross-Class Cram-Down From Virgin Active to Thames Water

On 26 June 2020 the Corporate Insolvency and Governance Act dropped a new Part 26A into the Companies Act 2006 and, in doing so, gave the English courts something they had never quite had before: a statutory power to bind dissenting classes of creditors and shareholders to a compromise none of them had voted through. Five years on, the restructuring plan has eaten most of the work the company voluntary arrangement (CVA) used to do, has produced its first appellate authority, and is starting to look less like a rescue tool and more like a court-supervised valuation contest.

This is the editorial five-year audit: what Part 26A is, how the data compares with its predecessors, where the case law has actually landed, and the question now sitting on practitioners' desks — whether the procedure is becoming too expensive and too contested for the mid-market it was partly meant to serve.

What Part 26A actually changed

The core innovation sits in two sections. Section 901F gives the court the same broad sanctioning discretion it has under a Part 26 scheme of arrangement, requiring approval by 75% in value of each class voting at the relevant meeting. Section 901G then adds the cross-class cram-down: even if one or more classes vote the plan down, the court may still sanction it provided two conditions are met. First, the dissenting class members must be no worse off under the plan than they would be under the "relevant alternative" — almost always administration. Second, at least one class that has a genuine economic interest in the relevant alternative must have approved the plan by the requisite majority.

The drafting is short. The litigation has not been.

For any reader who has only worked with the older toolkit, the relationship between the three procedures is the cleanest place to start.

Scheme, plan and CVA compared

The practical answer to "why a plan and not a scheme" is almost always the cram-down. The practical answer to "why a plan and not a CVA" is almost always the ability to bind secured creditors and landlords with leasehold rights that no CVA can touch — a lesson the retail sector learned the hard way after the Lazari Properties and New Look challenges to landlord CVAs in 2020 and 2021.

The data: 2020 to 2026

The Insolvency Service does not publish a clean monthly series for Part 26A plans the way it does for administrations or CVAs, so the public count comes from sanction judgments. As of the May 2026 register, the cumulative numbers look like this:

1. Plans convened, mid-2020 to May 2026: approximately 50. That number includes a small handful of withdrawn or abandoned proposals between convening and sanction hearings.
2. Plans sanctioned: approximately 40. The 2024 calendar year was the busiest on record, with around a dozen sanctioned in twelve months.
3. Plans refused or not sanctioned in current form: at least five. These include Hurricane Energy (2021), Project Lietzenburger Strasse / Aggregate Holdings (2023), and the well-publicised refusal in Great Annual Savings (2023).
4. Cross-class cram-down deployed: in roughly two-thirds of sanctioned plans. The mid-market is now using cram-down as a starting assumption, not a fallback.
5. Year-on-year CVAs sanctioned: down from 116 in 2020 to fewer than 35 in 2024 (Insolvency Service statistics). Some of that decline is cyclical; a meaningful share has migrated to Part 26A.
6. Companies House filings tagged "order of court — restructuring plan": now appearing routinely on registers of FTSE 250 and AIM-listed issuers, where a decade ago the same companies would have used a scheme.

Those volumes are still small in absolute terms — by comparison, Companies House processed roughly 4,400 strike-off applications a week in 2025 — but the value of the liabilities being compromised is not. Adler Group alone restructured around €3.2 billion of bond debt on its first attempt.

The doctrinal arc: five judgments that matter

Five plans have done most of the work of building the case law. They are worth holding in mind in this order, because the appellate corrections only make sense against the first-instance gloss they were correcting.

1. Re Virgin Active Holdings Ltd [2021] EWHC 1246 (Ch). Snowden J's sanction of the Virgin Active plan in May 2021 was the first contested cross-class cram-down. The judgment laid out the now-standard two-stage approach: first determine the relevant alternative, then test "no worse off" against it. Landlords were crammed down to a fraction of contractual rent.

2. Re Hurricane Energy plc [2021] EWHC 1759 (Ch). Sanctioned conditionally and then refused. Zacaroli J held that cram-down requires a meaningful showing that the dissenting class would actually be worse off under the alternative — speculative future upside is not enough. Hurricane is the early reminder that s.901G is not a rubber stamp.

3. Re Smile Telecoms Holdings Ltd [2022] EWHC 740 (Ch). First plan to disenfranchise out-of-the-money creditors entirely under s.901C(4) — a procedural shortcut that has since been used selectively.

4. Re AGPS BondCo Plc (Adler) [2024] EWCA Civ 24. The Court of Appeal's January 2024 judgment, handed down by Snowden LJ — now sitting upstairs — is the single most important development in five years. The Court overturned the first-instance sanction, holding that fairness under s.901G(5) requires a benchmark of how value would have been distributed in the relevant alternative, not how the plan proposers chose to distribute it. "Departure from the pari passu principle" is permissible but must be justified.

5. Re Thames Water Utilities Holdings Ltd [2025] EWHC 357 (Ch). Leech J's February 2025 sanction of the first emergency-liquidity plan for a regulated water company tested whether Part 26A could reach a quasi-public utility. It can; whether it should is a different question, and the appeal is expected to push the boundary again.

The direction of travel is unmistakable. Sanction hearings are now full-blown valuation trials, with competing expert evidence on what the "relevant alternative" would actually look like. Adler in particular put plan proposers on notice: the court will not simply accept their distribution waterfall just because in-the-money classes approved it.

Where the procedure now sits

Five findings should sit in any sober assessment of Part 26A on its fifth birthday.

The mid-market has been priced out. Sanction hearings now routinely run for four to seven days, with leading silks on both sides. The Houst plan in 2022 showed the procedure could work for an SME, but very few have followed. For a company with under £25 million of debt to compromise, Part 26A legal costs alone can eat 5–10% of that figure.

Landlord challenges have migrated upstream. The contested terrain of 2018–2021 — CVAs being challenged on "unfair prejudice" grounds — has moved into Part 26A convening and sanction hearings. Landlords now appear with valuation experts at convening, not merely as an afterthought.

Cross-border use is settled, not contested. The English courts have sanctioned plans for German, Spanish, Romanian and Dutch holding companies. Sufficient connection — typically a COMI shift or English-law-governed debt — is being established with relatively light evidence.

The PRA / Bank of England carve-out is real. Banks and insurers cannot use Part 26A. That leaves a meaningful gap in the toolkit, addressed only partially by the special administration regimes.

The Companies House register now carries a parallel record. Court orders sanctioning plans are filed under s.901L; the explanatory statements under s.901D give the registered file a level of detail no CVA chairman's report has ever matched. For corporate due diligence purposes this is, quietly, one of the most useful disclosure developments since the PSC register.

The editorial view

Part 26A has done what CIGA 2020 set out to do: it has given England a true cram-down tool and kept London competitive with Chapter 11 and the Dutch WHOA. It has also done something Parliament did not quite legislate for: it has turned the sanction hearing into a forensic valuation contest that prices most distressed UK SMEs out of the procedure entirely.

The next two years will decide whether that is a feature or a bug. If Court of Appeal authority continues to converge around Adler-style fairness scrutiny, the procedure becomes more predictable and the costs may settle. If first-instance judges keep using their s.901F discretion to second-guess proposers' commercial judgement, the cost premium will keep rising. Either way, the CVA's quiet retreat is now structural rather than cyclical, and the Companies House filing pack of a serious UK restructuring will, for the foreseeable future, carry a court order under s.901L on top.