When is it too late to revise an approved costs budget? For most claimant firms the honest answer is earlier than you think — but Car-Wizard Ltd v Vixen Surface Treatments Ltd [2026] EWHC 2177 (Ch), a decision of HHJ Paul Matthews sitting as a judge of the High Court on 17 August 2026, shows that a late Precedent T is not automatically a dead one.
A draft judgment had already been circulated. The trial had nevertheless not finished, because the judge had decided he needed further written submissions on damages. That work had been foreseen by nobody. The claimant applied to increase its trial phase before doing the work, and the court allowed it.
The claimant then beat its Part 36 offer, obtained 90% of its costs with indemnity assessment running across both periods, and took £214,000 on account. It is an unusually complete costs judgment for one case.
The variation, and the numbers
The underlying dispute concerned a commercial lathe sold to a small vehicle repair business. The defendant was found liable for supplying a machine that was not new and could not cut tapered edges as specified.
Liability and quantum had always been intended to be dealt with together. The judge abandoned that only after circulating a draft judgment, at which point further written submissions on the assessment of damages became necessary.
The claimant applied to increase the trial phase:
| Budget | Amount |
|---|---|
| Approved before variation | £275,201.79 |
| Increase sought | £19,710 |
| Revised total | £294,911.79 |
| Overall uplift | c. 7% |
The size was not really the issue. The question was whether it was still procedurally possible to revise the budget at all at that stage.
The objection: you cannot vary a budget after trial
The defendant argued that post-trial budget variations are impermissible, relying on the line of authority warning against retrospective attempts to repair an exhausted budget — Elvanite Full Circle Ltd v AMEC Earth & Environmental (UK) Ltd [2013] EWHC 1643 (TCC), where a retrospective variation was described as a contradiction in terms, and Queensgate Place Ltd v Solid Star Ltd (No. 3) [2024] EWHC 2139 (Ch).
That principle is sound and the judgment does not disturb it. Costs management exists to control future expenditure, and it would lose most of its function if parties could spend what they liked and ask the court afterwards to make the budget match.
How the judge distinguished them
The distinction drawn was unexplained delay. Elvanite and Queensgate were cases where the application came late and without explanation. Set against them were Cranstoun v Notta [2021] Costs LR 47 and Barry v Barry [2025] 4 WLR 56, where variations were permitted.
The line is therefore not simply chronological. What matters is:
- whether the relevant stage of the litigation has genuinely concluded;
- whether the additional costs have already been incurred; and
- whether the application followed promptly on the development.
On all three Car-Wizard was on the right side of it. The trial was still on foot, the work had not been done, and the application followed the event that caused it.
The development was the court’s own direction
This is the part worth remembering. The significant development relied on was not that the trial had proved expensive. It was the judge’s decision to require further written submissions on damages — something nobody had contemplated when judgment was reserved.
That is a genuine change in what the litigation required, arising from the court rather than from either party. Compare it with the alternative framing, which fails:
| Not a significant development | Capable of being one |
|---|---|
| The phase cost more than expected | The court directs further substantive work nobody had foreseen |
| Work took longer than assumed | An additional expert discipline becomes necessary |
| Foreseeable litigation steps | An amended case changes the issues |
| An overrun discovered at bill stage | A material change in the shape of the trial |
The counterpoint: Bassey v Whittaker
Car-Wizard should be read next to Bassey v Whittaker and the significant-development threshold, decided by Cavanagh J on 10 August 2026, a week earlier.
There the appeal was allowed because the judge below had permitted a variation despite finding that there had been no significant development. The High Court held that the court has no general discretion to vary a budget: the power lives in CPR 3.15A and its preconditions have to be met. Extra work by medical experts, a modest extension of the trial window and further disclosure did not qualify, because they could have been anticipated when the budgets were drawn.
Put the two together and the position is unusually clear for costs budgeting:
- Bassey — no significant development, no power to vary. More money needed is not a gateway.
- Car-Wizard — a real development, identified promptly, with the application made before the resulting work. Lateness alone did not defeat it.
The rule those cases apply, and how Precedent T fits into the wider budgeting cycle, is set out in our guide to costs budgets, the CCMC and the rules that decide recovery.
Beating the Part 36 offer
The claimant had offered to accept £65,000 on 24 October 2024, with the relevant period expiring on 14 November 2024. It recovered approximately £86,000.
The defendant tried to argue the comparison the other way: strip out one head of loss and the award fell to £62,276.50, below the offer. That failed twice over. The arithmetic did not work, because interest accrued would have carried the figure past the offer even on the reduced damages. And the losses had been pleaded generally, the specific sourcing being a matter of evidence rather than a pleading requirement.
The practical point survives the facts: when testing whether a judgment beats an offer, compare the right figures on the right basis. Interest can decide it, and looking only at the bare damages number can produce the wrong answer.
Indemnity costs across both periods, on two different bases
This is the part most likely to be misreported, so it is worth stating precisely.
| Period | Basis | Source of the entitlement |
|---|---|---|
| To 14 November 2024 | Indemnity | General costs discretion — conduct out of the norm |
| From 14 November 2024 | Indemnity | CPR 36.17(4)(b) — judgment at least as advantageous as the claimant’s offer |
On the earlier period the judge concluded at [41] that, taking the matters cumulatively, he was quite satisfied the defendant’s conduct had been sufficiently out of the norm to justify indemnity costs to 14 November 2024. The conduct relied on included defending the “not new” allegation without evidence to support it, defending the “cannot cut tapered edges” allegation without expert rebuttal, destruction of documents, and apparent interference with expert evidence.
The cumulative approach mirrors the reasoning in Lawrence v Associated Newspapers, where an accumulation of features rather than any single episode took the case out of the norm — and where the consequence for a budgeted case was that the approved budget stopped operating as the CPR 3.18 ceiling.
Keep the distinction on the file, because it has to survive into the bill. Two periods, two legal bases, one basis of assessment. A bill that does not show where the entitlement changes and why is harder to defend than it needs to be.
90% of costs, despite recovering less than pleaded
The claimant did not recover everything it had originally claimed, and the defendant argued for a substantial percentage reduction on that basis. The court awarded 90%.
That is a useful corrective. Costs do not track a simple ratio between the sum claimed and the sum recovered. The court can weigh what actually drove the litigation, which issues generated the work, overall success and conduct.
It is not a rule that shortfall never matters — plainly it can, and an exaggerated claim or a distinct issue lost can reduce recovery. The narrower point is that the damages percentage does not mechanically set the costs percentage. Proportionality falls away on the indemnity basis in any event, which is the subject of our note on indemnity costs and the proportionality argument.
The payment on account, split by category
The £214,000 is the most directly reusable part of the judgment, because of how it was built.
| Category | Amount | Allowed | Contribution |
|---|---|---|---|
| Incurred costs | £148,244 | 65% | c. £96,359 |
| Approved budgeted costs | £146,668 | 80% | c. £117,334 |
| Ordered on account | £214,000 | ||
The reasoning is the useful bit: approved budgeted costs have already been through costs management, so they are more robust and carry a higher percentage. Incurred costs were only ever recorded at the costs management hearing, not approved, so they carry a wider margin.
That is now a visible pattern across recent judgments. Lawrence allowed 90% of budgeted and 60% of pre-budget incurred costs. Car-Wizard allowed 80% and 65%. The percentages differ; the structure does not — and a single blended percentage across the whole bill is the weaker application. On how the court arrives at the figure at all, see our note on how a payment on account is calculated.
What this means in practice
1. Monitor the budget, do not file and forget it
Compare what is actually happening against the approved assumptions as the case moves. The budgeting file should be reviewed before every major stage and after any material procedural development.
2. Identify the development, not the overspend
A Precedent T has to say what changed in the litigation. “More work was required” is the consequence. Bassey is the authority a paying party will cite if the application cannot name the development.
3. Apply before doing the work
This is what saved Car-Wizard. An application made before the additional work is a prospective variation; one made afterwards asks the court to approve money already spent, which is the thing Elvanite and Queensgate warn against.
4. Record why it was not foreseeable
The file should show what the original budget assumed, what later happened, when it became apparent, why it changed the work required and which phases are affected. Bassey turned on exactly that: the developments relied on could have been anticipated.
5. Keep Part 36 under review as the case develops
A well-pitched claimant offer changes the economics. It is worth revisiting as the evidence and valuation position move, rather than treating it as a one-off decision — and note that the consequences follow a judgment, not an acceptance, as our note on late acceptance and the CPR 36.17 uplift explains.
6. Split the payment on account by category
Separate incurred from approved budgeted costs and propose a percentage for each, with the reason. Both Car-Wizard and Lawrence show courts engaging with exactly that structure.
What a strong Precedent T has to answer
Five questions, quickly:
- The assumption — what did the approved budget contemplate?
- The development — what specifically changed in the litigation?
- Significance — why is that material rather than foreseeable?
- Causation — what additional costs does it cause, and in which phases?
- Timing — why now, and has the work already been done?
If those cannot be answered cleanly, the problem is usually that the case has cost more than expected rather than developed. That is a Bassey problem, and no amount of drafting fixes it.
Key takeaway
A late Precedent T is not necessarily a dead Precedent T — but Car-Wizard is not a licence to repair overspent budgets after the event.
Its value is in the chronology. An unexpected development arose from the court itself, the relevant stage was still alive, and the claimant applied before doing the work. Change any one of those and the outcome likely changes with it.
We prepare Precedent H budgets, assumptions, Precedent R reports and Precedent T variations for claimant solicitor firms, and deal with the bill and detailed assessment that follow — see our costs budgeting service, or send us the file and we will confirm the fixed fee before any work begins.
Sources: Car-Wizard Ltd v Vixen Surface Treatments Ltd [2026] EWHC 2177 (Ch), HHJ Paul Matthews sitting as a judge of the High Court, 17 August 2026, from the approved judgment published by the National Archives, the indemnity conclusion at [41]; Bassey v Whittaker [2026] EWHC 2126 (KB), Cavanagh J, 10 August 2026; Elvanite Full Circle Ltd v AMEC Earth & Environmental (UK) Ltd [2013] EWHC 1643 (TCC); Queensgate Place Ltd v Solid Star Ltd (No. 3) [2024] EWHC 2139 (Ch); Cranstoun v Notta [2021] Costs LR 47; and Barry v Barry [2025] 4 WLR 56. Rule references are to CPR 3.15A, CPR 36.17 and CPR 44.