Ho v Adelekun

This appeal is about the mechanics of Qualified One-way Costs Shifting (“QOCS”). There always has been, and probably always will be, an inherent inequality of arms between claimants and defendants in personal injuries (“PI”) cases. This is because the defendants in most cases have the benefit of insurance or, in the case of the NHS, large resources, whereas claimants are in general ordinary members of the public, only a few of whom have the benefit of legal expenses insurance or other sources for the funding of litigation. English procedural rules have for many years sought to ameliorate this imbalance, in particular by rules about costs.

QOCS may be described as the third generation of ameliorating procedural schemes. The first was Legal Aid, under which state funding of meritorious claims was (partly to protect the public purse) accompanied by a virtual prohibition on the recovery of costs by defendants against legally-aided claimants. The second was a combination of Conditional Fee Agreements (“CFAs”) between claimants and their solicitors and the use of After the Event (“ATE”) insurance which would cover the unsuccessful claimant’s liability to pay the defendant’s costs and so insulate claimants from costs risk, with both success fees under the CFAs and ATE premiums recoverable in a successful case from defendants as part of the claimant’s costs. Legal Aid had largely been withdrawn by the end of the 20th century, and the burden on defendants of having to pay the claimants’ solicitors success fees and the claimants’ ATE premiums was found to have tilted the playing field too far in favour of claimants, with a politically unacceptable knock-on effect on motor and other insurance premiums. What became QOCS was originally proposed by Sir Rupert Jackson in his ground-breaking report “Review of Civil Litigation Costs: Final Report” (“Final Report”) in December 2009. Following extensive public consultation it was implemented, with important changes to the mechanism originally proposed by Sir Rupert, by changes to the Civil Procedure Rules (“CPR”) on the initiative of the Ministry of Justice, debated in detail within the Civil Procedure Rule Committee (“CPRC”), approved by Parliament by the negative resolution procedure and came into force on 1 April 2013. A challenge to the validity of the QOCS regime was rejected by the Court of Appeal in Wagenaar v Weekend Travel Ltd (trading as Ski Weekend)[2014] EWCA Civ 1105; [2015] 1 WLR 1968. In that case it was argued that the rules impeded the full power of the court granted by section 51(3) of the Senior Courts Act 1981 to determine by whom and to what extent the costs are to be paid. The Court of Appeal held that the proviso in subsection (1) of that section that the power conferred was subject to rules of court applied to the power in subsection (3). Vos LJ (with whom Floyd and Laws LJJ agreed) noted that QOCS was “just one of a raft of interconnected changes” bringing about the wholesale reform of the funding of PI litigation (para 26).

The central rationale behind QOCS was that the burden falling on defendants and their insurers would be less if they were to forego costs recovery from claimants when the claim was dismissed than the burden they were forced to bear when they had to pay claimants not only their costs but also recoverable success fees and ATE premiums when the claimants were successful. The effect of success fees on defendants was replaced by a 10% uplift in certain categories of recoverable damages, see Simmons v Castle & Ors [2012] EWCA Civ 1288; [2013] 1 WLR 1239, para 50. Removing the risk of the claimant becoming liable to pay costs if they lost the claim was expected to enable claimants to do without ATE insurance, at least for covering defendants’ costs. But costs recovery by defendants was not to be removed entirely. Responses to the Government’s consultation expressed concern that adopting such an inflexible stance would mean that there would be no constraints on claimants pursuing dishonest or hopeless claims, and little incentive on claimants to settle. Hence the inclusion of “Qualified” in the title.

Contrary to Sir Rupert’s proposals, nothing in the QOCS scheme affects in any way (directly at least) the orders which a court may make in favour of defendants in PI cases, applying the general rules in CPR Part 44, either at trial, at pre-trial interim hearings, at the conclusion of contested costs assessment proceedings post-trial or later still on appeal. The scheme focuses entirely upon what a defendant can do by way of enforcement of a costs order in its favour once obtained. The qualifications to the ban on enforcement inherent in the phrase “One-way” are of two types, one general and the other specific.

Generally, defendants’ costs may be enforced up to an amount equivalent to the aggregate of court orders for damages and interest in favour of the claimant. This is, as we shall later explain, a form of monetary cap on the amount of the costs orders made in the defendant’s favour which the defendant may enforce. The specific type of qualification consists of defined circumstances where there is no limit on enforcement, namely where the claimant’s claim has been struck out as disclosing no cause of action, as an abuse of process or on account of obstructive conduct of the claim, where it has been found to have been fundamentally dishonest, or where it has been pursued for the benefit of a third party. Costs incurred in the same proceedings in the pursuit of claims other than for personal injuries (such as replacement car hire) may also be enforced without limit.

The present appeal concerns the question whether QOCS constrains in any way the defendant’s liberty to seek, or the court’s discretionary power to permit, a set-off between opposing costs orders, that is orders in favour of the claimant and the defendant respectively. Opposing costs orders may occur in a single set of proceedings in numerous situations. At trial the judge may make opposing costs orders to reflect the varying success of the parties on specific issues, although this is generally discouraged. Opposing orders for different periods of the litigation may be made when the claimant fails at trial to beat the defendant’s Part 36 offer. In that situation the defendant is commonly ordered to pay the claimant’s costs up to the date specified by the claimant in the Part 36 offer for this purpose and the claimant is ordered to pay the defendant’s costs thereafter: see CPR rule 36.17. More typically a party may lose at trial, but win on an interim application, or on a costs assessment or on appeal, with costs following each event.

In cases where the aggregate of the damages and interest that the defendant is ordered to pay a successful claimant is large, this question may not matter. Provided that the defendant’s costs order (or the aggregate of more than one costs order) is less than the aggregate of the orders for damages and interest, there is no constraint on the ability of the defendant to enforce its costs orders, whether by set-off (if that is a species of enforcement) or by any other process of enforcement known to the law. But there are at least three types of case where it may be critical to any use which can be made by the defendant of a costs order in its favour. The first is where the claimant fails at trial and is ordered to pay the defendant’s costs, but is successful (with an order for costs in its favour) at an earlier interim stage, such as in fending off an application for summary judgment by the defendant, or later in winning on a costs assessment. The second is where the claimant succeeds, but by way of settlement rather than at trial. In such a case there is no court order for damages or interest, even if the settlement agreement is annexed to a Tomlin order, and therefore no headroom below the cap available under QOCS for the defendant’s costs enforcement: see Cartwright v Venduct Engineering Ltd[2018] EWCA Civ 1654; [2018] 1 WLR 6137 (“Cartwright”). The third type is where the aggregate of the costs that the claimant is ordered to pay the defendant substantially exceeds the aggregate of the orders for damages and interest which the defendant is ordered to pay the claimant. This is by no means a rarity; a disproportionality between the damages and the costs is all too frequent in modest to medium-sized PI claims which do not settle within the Pre-action Protocols.

This case comes to the Supreme Court because of a disagreement between two differently constituted Courts of Appeal about the question outlined above. In Howe v Motor Insurers’ Bureau [2020] Costs LR 297 (“Howe”) the Court of Appeal (Sir James Munby P, McFarlane and Lewison LJJ) decided in 2017 that set-off of opposing costs orders was not affected by QOCS, essentially because set-off is not a type of enforcement. In the present case the Court of Appeal (Sir Geoffrey Vos C, Newey and Males LJJ) were inclined to take the opposite view in April 2020, but considered themselves bound to follow Howe.

We should say at the outset that we doubt the appropriateness of a procedural question of this kind being referred to this court for determination. The very fact that two eminently constituted Courts of Appeal have differed profoundly over the interpretation of a provision of the CPR suggests that there must be an ambiguity which practitioners need to have sorted out. The CPRC exists for the purpose of keeping the CPR under constant review. It is better constituted and equipped than is this court to put right such ambiguities, all the more so where, as here, the outcome is suggested by both parties and by the Association of Personal Injury Lawyers (“APIL”), intervening, to have potentially profound policy consequences for the maintenance of a reasonably fair and level playing field in PI litigation, something which this court is much less well equipped than is the CPRC to assess. Nonetheless, permission having been given, this court must decide the question of construction, leaving it to the CPRC to consider whether our interpretation best reflects the purposes of QOCS and the Overriding Objective, and to amend the relevant rule if, in their view, it does not.

The appellant, Ms Adelekun, was injured in a road traffic accident on 26 June 2012 for which she alleged the respondent, Ms Ho, was liable. Ms Adelekun instructed solicitors and on 15 January 2014 they notified Ms Ho’s insurers of the claim in accordance with the Pre-Action Protocol for Low Value Personal Injury Claims in Road Traffic Accidents (the “RTA Protocol”). Ms Ho did not admit liability so the claim left the RTA Protocol and Ms Adelekun issued proceedings on 7 January 2015.

On 19 April 2017 Ms Ho’s solicitors offered to pay Ms Adelekun £30,000 in settlement of her claim in what was described as a “Part 36 Offer Letter”. In that letter, Ms Ho also offered to pay Ms Adelekun’s costs “in accordance with Part 36 rule 13”, such costs to be subject to detailed assessment if not agreed, if the offer was accepted within 21 days. Ms Adelekun decided to accept the offer and a Tomlin order was subsequently made by consent on 24 April 2017. The parties agreed at the time that the effect of that was that Ms Ho was liable to pay Ms Adelekun’s costs of the claim.

A dispute, however, arose between the parties as to the basis of assessment for those costs. Ms Ho argued that Ms Adelekun’s costs were limited to the fixed costs recoverable in accordance with the terms of CPR Part 45 Section IIIA. Ms Adelekun argued that she was entitled to recover her costs of the claim assessed on the conventional standard costs basis because that was what had been offered in the letter and accepted by her. The difference between the fixed costs and the standard costs was considerable. The fixed recoverable costs to which Ms Adelekun would have been entitled were about £16,700 whereas her costs of bringing the claim if assessed on the standard basis would be about £42,000.

In February 2018, Deputy District Judge Harvey decided that only fixed recoverable costs were payable by Ms Ho to Ms Adelekun. That decision was overturned by His Honour Judge Wulwik in October 2018 but the Court of Appeal allowed the further appeal and decided that only fixed recoverable costs were payable: see [2019] EWCA Civ 1988; [2019] Costs LR 1963. Ms Ho was awarded her costs of the appeal to the Court of Appeal, the appeal to His Honour Judge Wulwik and the hearing before Deputy District Judge Harvey. We shall refer to that stage of the proceedings as the “assessment dispute”. At the conclusion of the assessment dispute and the award of costs in her favour, Ms Ho asked the Court of Appeal to direct that she could set off her obligation to pay Ms Adelekun the fixed recoverable costs for the claim against the much larger costs liability that Ms Adelekun owed her for the assessment dispute. The costs that Ms Ho has incurred in succeeding in the assessment dispute and which have been awarded in her favour (including the hearing giving rise to the judgment under appeal) are said to be about £48,600.

Because Ms Adelekun’s claim included a claim for damages for personal injury it was accepted by the parties that the QOCS scheme applies. It is further agreed that since Ms Adelekun’s personal injury claim concluded by way of an acceptance of a CPR Part 36 offer, there was no “order for damages” made here within the meaning of the QOCS regime. This, as we have said, follows from the decision in Cartwright. Ms Adelekun contends that she is entitled to be paid the damages and interest pursuant to the settlement plus her fixed recoverable costs of the claim and that although she has been ordered to pay Ms Ho’s costs of the assessment dispute, that order cannot be enforced against her, not even by treating the £16,700 fixed costs she is owed as being absorbed by the £48,600 costs she owes to Ms Ho.

Ms Ho accepts that she must pay the damages and interest to Ms Adelekun. But she contends that she should not be required to pay over the fixed recoverable costs of £16,700 to Ms Adelekun because they should be set off and absorbed by the £48,600 costs that Ms Adelekun owes her. Ms Ho accepts that she cannot enforce her costs order for the assessment dispute against Ms Adelekun beyond that so she must bear the remaining £31,900 herself. But she contends that the QOCS regime does not require her to pay Ms Adelekun’s fixed recoverable costs in addition to the damages and interest, and bear the whole cost of the assessment dispute herself.

This case is therefore an example of the second kind of situation we described earlier. It is one of a number of situations where a defendant might be required to put their hand in their pocket to pay out money to the claimant even though they have been awarded costs against the claimant in an amount which, if they could set it off, would cancel out that liability.

The Court of Appeal referred to the arguments put forward by Mr Mallalieu on behalf of Ms Adelekun. The first was that to allow a defendant to set off costs owed to him against costs owed by him would undermine the QOCS regime and so impair access to justice. Underlying the regime was the recognition that the claimant is likely to have entered into a CFA with her solicitor and will be liable to pay the solicitor’s fees if the claim succeeds. The absence of set-off helps ensure that there will be a fund available from the defendant from which the claimant can pay those fees, even if the claimant has incurred a liability to the defendant for costs along the way. The court did not consider this argument had much weight since there were other circumstances in which a claimant might end up owing more to her solicitor than she received from the defendant: see para 14.

The court saw more force in Mr Mallalieu’s other argument that rule 44.14 was intended to operate as a complete code and to bar any enforcement of costs orders in excess of damages and interest unless an exception in rule 44.15 or rule 44.16 applied. Newey LJ considered that there were “compelling reasons” for concluding that set-off was to be regarded as a form of enforcement for this purpose: para 16. He said that had there been no authority on the issue, he would have been inclined to hold that, where QOCS applies, the court has no jurisdiction to order costs liabilities to be set off against each other. However, the Court of Appeal had decided in Howe that set-off was not a species of enforcement and that the court did have power to order set-off. The court held that it was bound by Howe to proceed on the basis that it did have jurisdiction to order set-off: para 25. Howe could not be treated as having been decided per incuriam by reason of the Court of Appeal not having been referred to Darini v Markerstudy Group (County Court at Central London 24 April 2017 (unreported)) where His Honour Judge Dight came to the opposite conclusion to the court in Howe. The court went on to exercise its discretion in Ms Ho’s favour adding in conclusion that the CPRC may wish to consider whether costs set-off should be possible in a QOCS case. Males LJ delivered a concurring judgment also expressing the view there was considerable force in the submission that the reference in rule 44.14 to the costs order being “enforced” should be understood as extending to the exercise of a right of set-off with the consequence that set-off of costs orders against each other is precluded: para 39. Sir Geoffrey Vos C agreed with both judgments. The Court of Appeal gave permission to appeal to this court on the issue of the availability of set-off. There is no appeal against the exercise by the Court of Appeal of its discretion to order set-off if such an order can be made.

The QOCS scheme is set out with commendable brevity in Section II of CPR Part 44. Rule 44.13 provides that Section II applies to proceedings which include a claim (including a counterclaim) for damages for personal injuries or under the Fatal Accidents Act 1976, or which arises out of death or personal injury and survives for the benefit of an estate by virtue of the Law Reform (Miscellaneous Provisions) Act 1934.

The question on this appeal concerns rule 44.14, which is worth setting out in full:

Rules 44.15 and 44.16 contain those exceptions where defendants may enforce costs orders free from the cap in rule 44.14. No issue arises as to their interpretation. The exceptions have already been listed above. Rule 44.15 enables enforcement without limit (“to the full extent”) without the permission of the court. In these cases, where the claim has been struck out, the court order doing so will already make it plain whether one of the qualifying exceptions applies, so that no further application to court is necessary. Rule 44.16 contains those exceptions (dishonest claim, claim brought for the benefit of a third party etc) where enforcement without limit will need the court’s permission, not because any exercise of discretion is involved, but because the defendant will need to establish that the relevant exception applies.

Various extracts from Sir Rupert’s interim and final reports were cited to us by the parties, including paragraph 4.10 of Chapter 19 of his Final Report, in which it does appear that Sir Rupert thought that a favourable costs order for the claimant could count as part of the fund from which to meet an element of costs liability to the defendant arising from a costs order arising from Part 36. But Sir Rupert’s observations do not significantly assist with the issue before the court because QOCS as enacted in Section II of CPR Part 44 is significantly different in its essential mechanism from Sir Rupert’s recommendation. Whereas he contemplated a restraint on the types of costs order which the court could make in favour of defendants, QOCS as enacted imposes no such restraint, but merely (albeit severely) limits what defendants can do with costs orders in their favour, once obtained.

Of slightly greater assistance are two statements which immediately preceded the approval by Parliament of the QOCS rules, once formulated by the CPRC. The first is a statement to Parliament made on 10 July 2012 by Mr Jonathan Djanogly, Parliamentary Under Secretary of State for Justice as to the intent of QOCS. By contrast with Sir Rupert’s assumption in the passage cited above, Mr Djanogly said that whereas the principles set out in CPR Part 36 will override QOCS, that will be “only up to the level of damages recovered by the claimant”.

The second is the Explanatory Memorandum prepared by the Ministry of Justice for the Joint Committee on Statutory Instruments when the statutory instrument making the amendments to the CPR to introduce the QOCS scheme was laid before Parliament in The Civil Procedure (Amendment) Rules 2013 (SI 2013/262 (L1)). It contains the following passage:

This statement regards the claimant’s damages as the only fund against which the defendant might recover costs, and uses the Part 36 situation only as an example of a wider principle. It points with reasonable clarity away from a set-off of costs against costs as an additional resource from which a defendant might recover costs from a claimant.

A major plank in the respondent’s case is that there was by the time of the introduction of QOCS a well-settled general tendency of the court to allow set-off of opposing costs orders as a means of recovering costs, in particular under the Legal Aid regime which, much more than the intermediate ATE regime, may be seen as the true ancestor of QOCS. Although originally well within the doctrine of equitable set-off, the jurisdiction of the court to direct set-off between opposing costs orders had long been statutory by the time of the introduction of QOCS, being found in section 51 of the Senior Courts Act 1981. It is now repeated in CPR rule 44.12 which provides as follows:

In Lockley v National Blood Transfusion Service [1992] 1 WLR 492 the Court of Appeal had to decide whether allowing a costs against costs set-off against a legally aided claimant was contrary to section 17 of the Legal Aid Act 1988. In deciding that it was not, Scott LJ described the equitable origin of the jurisdiction to order set-off of costs against costs and continued, at p 497D:

Both parties before this court relied upon Lockley. Mr Mallalieu QC for Ms Adelekun submitted that this showed that, in the minds of the legal costs community, set-off of costs against costs was generally regarded as a form of enforcement. Mr Bacon QC for Ms Ho submitted that this case both demonstrated and fully justified an assumption that set-off of costs against costs was not regarded as a derogation from the form of one-way costs shifting constituted by the Legal Aid regime, from which QOCS was partly derived.

Lockley was followed and applied by the Court of Appeal in 2004 in R (Burkett) v Hammersmith and Fulham London Borough Council[2004] EWCA Civ 1342; [2005] 1 Costs LR 104, again in a Legal Aid context. Apart from emphasising the essentially discretionary nature of the jurisdiction to order set-off under section 51, it does not significantly add to the debate. But the two decisions appear to remain the leading cases on the jurisdiction to order set-off of costs against costs, outside the QOCS context.

Counsel for both parties, and the intervener in written submissions, placed great emphasis on what they characterised as the adverse policy consequences of the opposing answers to the question before the court, seeking to admit those policy issues under submissions about the purposes of the QOCS scheme. For the appellant and APIL it was said that the real effect of a costs against costs set-off in the QOCS context was to deprive the claimant’s solicitor of the means of payment for work done on credit in parts of the case in which the client had been successful and recovered costs. This would, they said, undermine the whole economic basis upon which PI litigation under QOCS could be undertaken by solicitors for deserving clients of modest or non-existent means. This would strike at the heart of what QOCS was seeking to achieve. For the respondent it was submitted that by depriving defendants of any fund against which (in a no court order for damages case) they might recover costs, the court would be giving a green light to the pursuit by claimants of weak interim applications and unmeritorious points. It would also remove any real incentive to settle before trial, if the adverse costs consequences of losing at trial (or failing to beat a Part 36 offer) led to a purely unenforceable costs sanction.

It is not necessary or appropriate to describe or examine those policy considerations in any detail. First, as already emphasised, this court is not well placed to assess them reliably. If the true construction of the QOCS scheme set out in Section II of CPR Part 44 has adverse policy consequences, that is a matter for the CPRC to put right. The purpose of QOCS is tolerably clear, to seek to rebalance an inherently tilted playing field. The question underlying this appeal is how far that levelling process was intended to go. The answer to that question will not affect that levelling process to any great degree. It is common ground that there can be no costs recovery at all against claimants who simply lose, and obtain no damages or costs order in their favour. A much larger effect on the levelling process was arrived at by the decision in Cartwright that damages and interest payable under a settlement did not count for the purposes of rule 44.14(1), since far more cases settle than go to trial. Where a claim does conclude with a court order, in many cases, the defendants’ costs orders will be less than the claimants’ damages and interest, and the defendant will undoubtedly be able to deduct the costs it is owed from the damages and interest it must otherwise pay over to the claimant.

As both counsel were constrained to accept, the present question is one of construction of the language of the QOCS provisions in the CPR set in their context. For that purpose, it is necessary to return to study the language of rule 44.14 in more detail. Certain points were, or became, common ground. First, the QOCS regime does not in terms prevent a trial judge from making a single, one way, costs order which is the result of adjusting in the judge’s mind the appropriate costs order in the light of the parties’ respective successes or failures on different issues in the litigation. This is because QOCS does not seek to constrain the court from making costs orders, but merely the use which defendants can make of costs orders in their favour.

Secondly, the appellant submitted and the respondent did not seriously challenge that the QOCS regime is essentially mechanical rather than discretionary, so that the phrase in rule 44.14(1) “without the permission of the court” did not preserve a general discretionary power to permit a defendant costs enforcement beyond that expressly provided for by the permission process in rule 44.16. That process was necessitated only by the need for the court to see whether the qualifying facts existed, such as dishonesty.

Thirdly, and this only emerged during the hearing, rule 44.14 does not in terms operate as a total ban of set-off of opposing costs orders. It just imposes a monetary cap. It does so by requiring the monetary value of any set-off by the defendant to be brought into account against the monetary amount of the claimant’s orders for damages and interest. That will amount to a ban only if there are no orders for damages or interest (as in the present case) or if the aggregate amount of damages and interest has already been used up by other means of enforcement.

The appellant’s main submission on construction was that Section II of Part 44 was a complete costs code for the PI cases to which it applies, so that the express mention in rule 44.14(1) of set-off of costs against damages and interest necessarily ruled out (by silence) set-off of costs against costs.

For the respondent it was submitted that the QOCS rules were not a complete code. They did not expressly displace the court’s jurisdiction to direct set-off of costs against costs in rule 44.12. Such a deep-rooted jurisdiction, routinely applied in Legal Aid cases and founded on plain justice and equity, could not be displaced without clear words of exclusion. Cases like Burkett showed that the only costs which were subject to the cap in rule 44.14(1) were the net costs liability (if any) of the claimant to the defendant after all opposing costs orders had been netted off. If this was, by concession, what the judge could do by way of mentally adjusting the amount or percentage of a single costs order in one party’s favour at the end of a trial, why should QOCS be any more restrictive, merely because opposing costs orders were made at different times during the litigation? Finally, “enforcement” should be defined by its use elsewhere in the CPR. Part 70 provided a complete list of enforcement measures, which did not include set-off.

We would not accept Mr Mallalieu’s submission that QOCS is a complete costs code, or that it wholly excludes set-off of costs against costs under rule 44.12. But we would accept that QOCS is intended to be a complete code about what a defendant in a PI case can do with costs orders obtained against the claimant, ie about the use which the defendant can make of them. The defendant can recover the costs ordered, by any means available, including set-off against an opposing costs order, but only up to the monetary amount of the claimant’s orders for damages and interest. This is what the Explanatory Memorandum states in terms.

We consider that rule 44.14(1) works in the following way. First, it requires two comparators to be constructed. First, the aggregate amount in money terms of all costs orders in favour of the defendant. Secondly, the aggregate amount in money terms of all orders for damages and interest in favour of the claimant. We will call them A and B. If A is less than or equal to B, the defendant can enforce his costs orders without limit. If A is more than B, then the defendant can only enforce his costs orders up to the monetary limit of B. The effect of this cap, as we have called it, is to require the defendant to keep a running account in money terms of all costs recoveries which it makes against the claimant, and to cease enforcement when limit B is reached.

The question remains: does the defendant have to bring into account the benefit in money terms of the set-off of a costs order in his favour; in other words does the limit B only apply to the net amount of costs owed by the claimant, having set off any costs the defendant is ordered to pay to the claimant? Plainly the defendant must bring into account the monetary benefit of setting off costs against the claimant’s damages, despite the fact that this may not generate actual cash but only save the defendant from having to put his hand in his pocket to pay the damages and interest to that extent. That is what “money terms” means. For example, assume that the claimant is ordered an award of £20,000 in damages and interest, but that the defendant has costs orders for an aggregate amount of £30,000. If the defendant has not yet paid the damages, it can set off its damages liability against the claimant’s costs liability, but only up to £20,000. It must bring that £20,000 into account under rule 44.14(1) and cannot enforce the balance of its costs entitlement of £10,000, by any means of enforcement. If the defendant has already paid the damages before its costs are assessed, then it can enforce its costs orders by any other available means (set-off being in practice unavailable), but only up to £20,000. It cannot therefore be said that use of a set-off is not a means of enforcement, where costs are set off against damages.

If set-off of costs against damages is therefore a form of enforcement in this context, so as to make sense of rule 44.14, then why should set-off of costs against costs not equally be a means of enforcement? Both achieve a recovery measurable in money terms for the defendant on account of its costs entitlement, and by the same self-help means of appropriating an asset of the claimant (his damages entitlement) to the part satisfaction of the defendant’s entitlement against the claimant for costs. Strictly it might be said that set-off of costs against damages pursuant to rule 44.14 requires less assistance from the court than set-off of costs against costs, because the latter requires the court’s direction under rule 44.12. But that just makes it more like a form of enforcement.

Real assistance on the contextual meaning of enforcement is gained by reflecting on the language and structure of rule 44.14(1). The requirement is to calculate A by reference to the aggregate amount in money terms of all the defendant’s costs orders made against the claimant, not the net amount arrived at by netting off opposing costs orders and striking a net balance. Costs orders in favour of the claimant are not even mentioned in the formula, and the aggregate expressly referred to is a gross not a net amount.

Some slight further assistance may be available from rule 44.14(3), which prohibits the “unenforced” part of a costs order from being registered as an unsatisfied or outstanding judgment. That would act to the detriment of the claimant’s credit rating, such that the threat of it would otherwise be an incentive towards payment, and therefore a prohibited form of enforcement. This suggests a wide contextual meaning of enforcement. The concept of the judgment for costs being partly unsatisfied because of the impact of the cap in sub-rule (1) would be very odd if set-off were not a kind of enforcement for this purpose.

Returning to the respondent’s submissions, we do not consider that the well-established jurisdiction to direct set-off of costs against costs under rule 44.12 is displaced by the QOCS scheme, provided that there is an order for damages or interest and that the headroom provided by that order has not been exhausted by other means of enforcement. But for the reasons already given we do not accept the submission that it is only the net costs entitlement that has to be brought into account under rule 44.12(1).

We recognise that this conclusion may lead to results that at first blush look counterintuitive and unfair. Why should a defendant which has a substantial costs order in his favour have to pay out costs to a claimant under an order made against him when the two costs orders would net off against each other, leaving both sides to meet their own solicitor’s costs themselves? Whether or not the intervener in this appeal is right that such a result accords with the policy underlying QOCS, we hold that it is the result that follows from the true construction of the wording used in Part 44. Any apparent unfairness in an individual case such as this dispute between Ms Ho and Ms Adelekun is part and parcel of the overall QOCS scheme devised to protect claimants against liability for costs and to lift from defendants’ insurers the burden of paying success fees and ATE premiums in the many cases in which a claimant succeeds in her claim without incurring any cost liability towards the defendant.

We also recognise that this construction of rule 44.14 may lead to results that appear anomalous. We have already referred to the fact that a judge in making an order for costs might be invited to adjust the amount or percentage ordered to reflect the relative success of the parties though, as Mr Mallalieu pointed out, a judge might also take into account when so invited, that this would water down the protection that would be afforded to the claimant if the judge made cross costs orders instead. Mr Bacon also argued that in the PI cases where legal aid is still available, particularly certain birth defect claims, it would appear that costs against costs set-off would still be available following Lockley and Burkett. No one has claimed that the QOCS scheme is perfect. It is, however, the best solution so far that the opposing sides in the ongoing debate between claimant solicitors and defendant insurers have been able to devise. It works to achieve the aims for which it was introduced in the great majority of straightforward cases in which one side or the other is entirely successful.

Finally there is in our view nothing in the point that set-off is not mentioned in the list of the court’s enforcement powers in Part 70. If set-off of costs against damages must be a form of enforcement in order to make rule 44.14 work, then the QOCS context requires set-off to be treated as a form of enforcement even if not mentioned as such elsewhere in the CPR.

So for all those reasons we consider that the Court of Appeal was right in the present case to doubt whether Howe was correctly decided. We would allow the appeal.