United Kingdom Supreme Court

LORD REED: (with whom Lord Hodge, Lord Lloyd-Jones, Lady Arden and Lord Leggatt agree)

1.                  This appeal raises a question concerning the role of this court in relation to the principles governing the award of costs in lower courts.

2.                  The material facts can be shortly stated. The appellant, Mr Gourlay, is a life prisoner. The respondent is the Parole Board. In 2014 the Board met to consider whether it would be appropriate to direct the appellant’s release. If it did not direct his release, it was required to consider whether to recommend his transfer to open conditions. On 10 March 2014 the Board issued its decision. It did not direct his release or recommend his transfer to open conditions. The appellant then brought proceedings for judicial review of the decision. In the claim form, both aspects of the decision were challenged. The Board did not take part in the proceedings. In its acknowledgment of service, it ticked the box stating that “the defendant … is a court or tribunal and does not intend to make a submission”. At the hearing, counsel for the appellant confined his challenge to the decision not to recommend a transfer to open conditions. The challenge was successful, and that aspect of the decision was quashed: [2014] EWHC 4763 (Admin).

The treatment of costs at first instance

3.                  The appellant then applied for an order finding the Board liable for his costs. His submissions in support of that application reflected the established practice, described in R (Davies) v Birmingham Deputy Coroner [2004] EWCA Civ 207; [2004] 1 WLR 2739, of not making an award of costs against a court or tribunal which adopted a neutral stance in proceedings in which its decision was challenged, in the absence of exceptional circumstances, such as flagrantly improper behaviour. In Davies, Brooke LJ, with whom Longmore LJ and Sir Martin Nourse agreed, identified four issues as arising for consideration, at para 3:

“(1)     What is the established practice of the courts when considering whether to make an order for costs against an inferior court or tribunal which takes no part in the proceedings, except, in the case of justices, to exercise their statutory right to file an affidavit with the court in response to the application? (2) What is the established practice of the courts when considering whether to make an order for costs against, or in favour of, an inferior court or tribunal which resists an application actively by way of argument in the proceedings in such a way that it makes itself an active party to the litigation? (3) Did the courts adopt an alternative established practice in those cases in which the inferior court or tribunal appeared in the proceedings in order to assist the court neutrally on questions of jurisdiction and procedure and such like but did not make itself an active party to the litigation? (4) Whatever the answers to the first three questions, are there any contemporary considerations, including the coming into force of the Civil Procedure Rules 1998, which should tend to make the courts exercise their discretion as to costs in these cases in a different way from the way in which it was regularly exercised in the past?”

4.                  Brooke LJ set out his conclusions on those issues at para 47:

5.                  Against that background, the appellant argued in the present proceedings first, that the Board was not a court or tribunal falling within the scope of the practice described in Davies, and secondly, that there were in any event exceptional circumstances, arising from the Board’s behaviour in relation to the proceedings.

6.                  His submissions were recorded by the judge, King J, in a reasoned order dated 2 February 2015:

5.         The claimant relies on the Court of Appeal judgments in Bahta [R (Bahta) v Secretary of State for the Home Department] [2011] EWCA Civ 895, that even against government departments the normal provisions as to costs contained in the CPR apply, including the general rule starting point·(rule 44.3(2)) that a successful claimant is entitled to his costs and the importance of complying with Pre-action Protocols (an aspect of conduct identified in rule 44.3(5)). The point is made in this context that here the defendant did not respond to the letter before the claim.

6.         There are, says the claimant, no circumstances apparent in these proceedings which should lead the court when exercising its discretion as to costs, to depart from the usual position in judicial review proceedings that the unsuccessful defendant public body whose decision has been under challenge will normally be ordered to pay costs to the successful claimant. The fact that the claimant is publicly funded is not a good reason to decline an award of costs which would otherwise be appropriate (again see Bahta).

7.         In any event as a fallback position the claimant submits that even if the defendant is to be treated differently, there are ‘exceptional circumstances’ here to justify an order of costs, in that the defendant did not respond to the letter before claim, was late in filing its AOS [acknowledgment of service], and that which is described as a ‘failure to compromise the claim’ (as invited by the claimant’s letter of 29 September 2014) is only to be interpreted as a defended claim and puts the defendant in an adversarial role.”

7.                  The judge rejected those submissions. He stated that the first and critical issue was whether, for the purpose of the exercise of the court’s discretion as to costs, the Board was to be treated as a court or tribunal, or at least a judicial decision-making body, when making the decision under challenge. There was no doubt that it was acting in a judicial capacity when making its decision whether or not to direct a prisoner’s release. The second, consequent, decision was whether to recommend a transfer to open conditions: a recommendation which was not binding on the Secretary of State. It was that aspect which had been challenged. Although a theoretical distinction might be drawn between these two matters, the judge concluded:

8.                  Approaching the application for costs on that basis, and following the practice described in Davies, the judge concluded that he should make no order as to costs:

9.                  The appellant appealed against that decision to the Court of Appeal: [2017] EWCA Civ 1003; [2017] 1 WLR 4107. Senior counsel was instructed and presented a more elaborate argument than King J had heard. The appeal was, however, presented on a limited basis. The court noted, at para 33 of the judgment delivered by Hickinbottom LJ, with whom Gloster and David Richards LJJ agreed:

“The sole ground relied upon by Mr Southey [senior counsel for the appellant] is that the principles set out in Davies as applicable to courts and tribunals do not apply to the Board.”

Hickinbottom LJ also noted at para 44:

“Mr Southey - in my view, properly - conceded that this court is bound by Davies. Although it seemed to me that at times he came painfully close to doing so, he unconditionally accepted that he could not argue before this court that Davies was not still good law. The only ground upon which he relied was that the principles and practice set out in that case, and expressly applied by King J below, simply do not apply to the Board when it performs its function of making recommendations to the Secretary of State in respect of the transfer of a prisoner to open conditions.”

10.             In support of that proposition, counsel for the appellant made a number of submissions, which can be summarised as follows:

(3)              Davies should now be read in the light of R (M) v Croydon London Borough Council [2012] EWCA Civ 595; [2012] 1 WLR 2607, which signalled a new approach to costs in public law cases. That new approach was designed to encourage the compromise and settlement of claims against public bodies. (It should be explained that M was a more recent case following and elaborating upon the reasoning in the earlier case of R (Bahta) v Secretary of State for the Home Department [2011] EWCA Civ 895; [2011] 5 Costs LR 857, which had been cited to the judge.)

(4)              It would be helpful for the Board to be a party to a challenge to one of its decisions. The historic concern over the possibility of the decision-maker having to pay costs out of his own pocket had been alleviated in the case of the Board, which had been constituted as a body corporate.

(5)              Davies proceeded on the basis that it was irrelevant whether a claimant judicially reviewing a court or tribunal was legally aided, with the consequence that the fees paid to his legal representatives were below the rates paid for privately funded work. However, following In re appeals by Governing Body of JFS [2009] UKSC 1; [2009] 1 WLR 2353, that was or might be a relevant factor. Lord Hope of Craighead said in that case at para 25 that “[i]t is, of course, true that legally aided litigants should not be treated differently from those who are not. But the consequences for solicitors who do publicly funded work is a factor which must be taken into account. A court should be very slow to impose an order that each side must be liable for its own costs in a high costs case where either or both sides are publicly funded …”.

(6)              The scope of Davies should therefore be restricted, and it should not be applied where the Board was a defendant to a judicial review.

(7)              Even if Davies applied to the Board when deciding questions of release, it should not apply when the Board was considering whether to recommend a transfer to open conditions.

11.             Those submissions were carefully considered by Hickinbottom LJ, and each of them was rejected, for reasons which, taking each point in turn, can be summarised as follows:

(3)              The argument that M modified Davies had been considered, and rejected, in the case of R (Gudanaviciene) v First-tier Tribunal (Immigration and Asylum Chamber) [2017] EWCA Civ 352; [2017] 1 WLR 4095. In that case Longmore LJ, with whom David Richards and Moylan LJJ agreed, held that M did not constitute a new approach at all, but merely decided that to make no order for costs in a public law claim might not be appropriate where, as a result of some settlement or compromise, there had been some degree of success. It was only in that limited situation that M broke new ground. However, that had no relevance to the different question of whether a party that was wholly successful should obtain an order for costs against the other party. Hickinbottom LJ also cited para 36 of Longmore LJ’s judgment, where it was said:

“It would be a serious step to say that in any undefended appeal or judicial review, the tribunal would be at risk as to costs and any such conclusion cannot be implied into the decision of M. If such a step is to be taken, it cannot be by a court of coordinate jurisdiction with the court which decided Davies.”

Hickinbottom LJ observed at para 57 that Gudanaviciene was binding upon the court, and that he would in any event adopt Longmore LJ’s analysis and conclusions:

“M did not materially affect the approach of the courts to costs orders in public law cases involving a court or tribunal, ie the principles and application of Davies; and, in particular, the encouragement given by M to the early settlement of public law cases by the threat of costs sanctions has no application to judicial review claims against courts or tribunals.”

(4)              The principle that, unless it had acted improperly or had actively participated in the challenge, a court or tribunal was not required to pay the costs of rectifying one of its orders - whether by way of judicial review or appeal - was therefore well-settled. It was a principle supported by many strands of public policy, and it had been well-established since Davies. It could not be undermined in the case of a particular court or tribunal simply because one of the historic concerns that it sought to address was absent, for example where, as in the present case, the relevant judicial decision-maker had an indemnity against costs. Nor was it undermined by the fact that the active intervention of the court or tribunal in the proceedings might be of assistance. As Brooke LJ indicated in Davies, it was open to a court or tribunal to make a neutral submission without incurring a costs order. The Administrative Court could also ask for the assistance from the tribunal. The fact that specialist courts or tribunals might assist in this way was no reason for overriding the principle in Davies.

(5)              Lord Hope’s comments in JFS were made in an entirely different context, in which they were consistent with the proposition that legally aided litigants should not be treated differently from those who are not. He was not dealing with a retrospective application for costs, or with an application made against a court or tribunal. Nor were the present proceedings a high costs case. Following Davies, the source of funding was immaterial to the principle that, where a court or tribunal had not acted improperly and had taken no active part in the claim, the appropriate course was to make no order for costs.

(6)              In respect of decisions concerning the release of prisoners, the Board was clearly acting as a court or tribunal for the purposes of Davies. The principles of Davies thus applied to challenges to such decisions.

(7)              Although the recommendation of the Board in relation to transfer was only advisory, a similar approach to the costs of challenging decisions of the Board was appropriate. Whether dealing with decisions concerning release or decisions concerning transfer, the Board performed a similar function, in that it had to obtain relevant material from the National Offender Management Service and the offender himself, and evaluate that material in making an assessment of the risk posed by the offender, and of whether that risk was at an appropriate level for him to progress by way of transfer to a category D prison or release on licence, as the case might be. In respect of release, it had to reach its own objective judicial decision, in order to comply with the requirements of article 5(4) of the European Convention on Human Rights. In respect of transfer, it reached its decision in the same way, and to the same procedural standards. It had to use the same procedures for practical reasons: it was often the case that a panel was considering both transfer and release at the same time. However, it was also required to adopt the same procedural standards, not as a result of article 5(4), but by the common law. Therefore, in considering transfer decisions, the Board both in practice acted, and in principle was required to act, as if it were a court or tribunal.

12.             Hickinbottom LJ added that it was likely to be open to the prisoner to make the Secretary of State a party to any challenge to a decision in relation to transfer, so that in appropriate cases the prisoner would be able to obtain a costs order against the Secretary of State.

13.             The issues arising on this appeal were agreed by the parties to be, first, whether Davies continues to represent the approach that should be adopted to costs orders against courts or tribunals, and secondly, if Davies continues to apply, whether the Board should be treated as a court or tribunal for that purpose.

14.             In relation to the first of those issues, counsel for the appellant submitted that “Davies is wrong”: a submission which, it was said, was not available to him below. In support of his submission, counsel argued that the rationale for the practice described in Davies had not been explained or justified. To the extent that reasons were given, many if not all of them had fallen away. In that regard, counsel referred to a number of changes which were said to have occurred since Davies was decided: in particular, the decline of legal aid, a diminution (in real terms) in legal aid payment rates, an increased reliance on conditional fees, and the introduction of the general rule that costs follow success, in CPR rule 44.2(2). Davies was argued to be incompatible with CPR rule 44.2(2), since it created an exception to that general rule. The Board must be viewed as the unsuccessful party even if it played no active part in the proceedings, since its decision had been quashed. If the Board did not concede a challenge, then in substance it opposed it.

15.             In relation to the second issue, counsel submitted, as in the courts below, that the Board should not be treated as a court or tribunal, at least when deciding whether to recommend a transfer to open conditions. In support of that submission, counsel referred to a number of factors, placing particular emphasis on the fact that the Board’s recommendation was not binding on the Secretary of State, and on the fact that the Board had published a “litigation strategy”, which noted that in any form of litigation, it could decide to concede the case.

16.             In response, counsel for the Board submitted that the Court of Appeal was correct to hold that Davies was good law, and that the approach which it laid down applied to the Board. That approach, it was argued, accorded with the settled and long-established approach of the High Court and the Court of Appeal, and ensured the continuing impartiality of judicial decision-making bodies where their decisions were challenged by way of judicial review. Any decision which swept away their long-established ability to maintain neutrality in proceedings challenging their decisions would have a very significant impact on the way in which litigation was conducted. An approach which declined to award costs against a neutral party was compatible with CPR rule 44.2(2). Such a party was not “unsuccessful”, and in any event rule 44.2(2) provided that “the court may make a different order”.

17.             At an early stage in the hearing of the appeal, the court raised with the parties the question whether the issues which were said to arise on the appeal (para 13 above) raised any question of law, as distinct from questions of practice in relation to the award of costs, which it might be appropriate for this court to leave to the courts below. The parties were invited to file written submissions on the question after the hearing.

18.             In the post-hearing submissions filed on behalf of the appellant, counsel submitted that the Court of Appeal regarded itself as bound by its decision in Davies, as was apparent from para 36 of Gudanaviciene (para 11(3) above) and para 57 of Hickinbottom LJ’s judgment in the present case (ibid), and from its acceptance in the present case of the appellant’s concession that Davies was binding upon it (para 9 above). That understanding was also apparent from the judgment of Hickinbottom LJ, with which Davis and Ryder LJJ agreed, in R (Faqiri) v Upper Tribunal (Immigration and Asylum Chamber) [2019] EWCA Civ 151; [2019] 1 WLR 4497, para 24, where Gudanaviciene was said to have “confirmed that Davies was still good law, and binding on this court”. It followed that only the Supreme Court could decide whether Davies was “good law”.

19.             If, however, the question was one of practice, then it followed that the Court of Appeal had been mistaken in regarding itself as bound by the decision in Davies. This court should therefore find in favour of the appellant, on the ground that the courts below were mistaken in regarding Davies as imposing a strict rule, when their discretion was in reality much broader.

20.             The submissions on behalf of the respondent identified the relevant question of law as being whether King J had exercised his discretion correctly, or whether his approach was erroneous in law. In response to the submission of the appellant summarised at para 19 above, counsel observed that, given that counsel for the appellant had conceded before the Court of Appeal that it was bound by Davies, the implication of his argument was that the Court of Appeal should, of its own motion, have decided the appeal on a basis for which the appellant was not contending.

21.             Section 51(1) of the Senior Courts Act 1981 provides, so far as material:

22.             Section 51 is concerned with the jurisdiction of the court to make orders as to costs. That jurisdiction is expressed to be subject to rules of court. It is therefore open to the rule-making authority (now the Civil Procedure Rule Committee established under section 2 of the Civil Procedure Act 1997) to make rules which control the exercise of the court’s jurisdiction under section 51(1). It is also open to appellate courts to provide guidance to lower courts as to how their discretion should be exercised. As Lord Goff of Chieveley remarked in Aiden Shipping Co Ltd v Interbulk Ltd [1986] AC 965, 975, section 51(1) “is consistent with a policy under which jurisdiction to exercise the relevant discretionary power is expressed in wide terms, thus ensuring that the court has, so far as possible, freedom of action, leaving it to the rule-making authority to control the exercise of discretion (if it thinks it right to do so) by the making of rules of court, and to the appellate courts to establish principles upon which the discretionary power may, within the framework of the statute and the applicable rules of court, be exercised”.

23.             The rules of court relating to the court’s discretion as to costs set out a number of general principles. Some are contained in CPR rule 44.2. Under paragraph (1), the court has discretion as to whether costs are payable by one party to another, the amount of those costs, and when they are to be paid. Under paragraph (2), if the court decides to make an order about costs, the general rule is that the unsuccessful party will be ordered to pay the costs of the successful party, but the court may make a different order. Certain exceptions to the general rule are set out in paragraph (3), but do not apply in the present case. Under paragraph (4), in deciding what order (if any) to make about costs, the court is to have regard to all the circumstances, including the conduct of the parties. Other principles are set out in other rules, notably CPR rule 44.3, and apply to circumstances falling within the scope of those rules.

24.             The rules of court do not, however, set out a comprehensive code. It is also important, as Lord Goff indicated, “that the appellate courts … establish principles upon which the discretionary power may, within the framework of the statute and the applicable rules of court, be exercised”. This court, like the House of Lords before it, proceeds on the basis that responsibility for developing those principles falls principally upon the Court of Appeal. In proceeding on that basis, the court recognises that such principles are generally matters of practice, rather than matters of law, and follows its general approach to matters of practice.

25.             In modern times, that approach was articulated by Lord Hope in Girvan v Inverness Farmers Dairy 1998 SC (HL) 1, an appeal from the Inner House of the Court of Session concerned with the level of damages for personal injuries which had been awarded by a jury. In the course of his speech, with which the other members of the Appellate Committee agreed, Lord Hope considered a submission that a change in practice should be introduced in the Court of Session, so that juries would be given guidance by the judge as to the appropriate level of damages. He observed at p 21:

“In the first place it would, I believe, be inappropriate for your Lordships to recommend changes in the practice which is followed by the Court of Session in the conduct of jury trials in its own court. The Court of Session is, in a very real sense, the master of its own procedure. It has been said several times in the Inner House, on motions for leave to appeal in interlocutory matters, that it is not appropriate to refer matters of practice for decision by the House of Lords … The basis for this view is that the Court of Session is far better placed than your Lordships can ever be to assess what changes in procedure or practice can appropriately be made and, if they were to be made, what would be their consequences.”

26.             Lord Hope went on to explain that the Court of Session had a statutory power to make rules of court regulating its procedure, and that the Court of Session Rules Council, established under the same legislation, had the function of considering what changes ought to be made from time to time. Both branches of the legal profession were represented on the Rules Council, so that it provided an appropriate forum for consultation when changes to the rules were being proposed. In those circumstances, Lord Hope concluded:

27.             That case was concerned with a different question from the one which arises in the present appeal. The emphasis on rules of court, in particular, distinguishes it from a case concerned with the principles upon which a discretionary power should be exercised, as set out in the judgments of appellate courts. Nevertheless, Lord Hope’s remarks about the role of the House of Lords in relation to matters of practice, and about the greater ability of the Court of Session to assess what changes can appropriately be made, and to modify practice flexibly and without delay by means of judicial decisions, are of wider scope. They are relevant, mutatis mutandis, to the relationship between this court and the Court of Appeal of England and Wales, and the Court of Appeal of Northern Ireland, and have been cited in a number of different contexts in more recent decisions of the House of Lords and of this court.

28.             The case of Callery v Gray (Nos 1 and 2) [2002] UKHL 28; [2002] 1 WLR 2000, is more directly in point, as it concerned appeals against an order for costs. Questions arose, where the claimant had entered into a conditional fee agreement with his solicitors, as to whether the success fee, and an after the event insurance premium, were recoverable from the defendant as part of an award of costs; whether the amount of the success fee was reasonable; and whether the insurance premium was reasonable. The House of Lords decided not to intervene in the Court of Appeal’s determination of those questions.

29.             Lord Bingham of Cornhill, with whom Lord Nicholls of Birkenhead and Lord Hope agreed, stated at para 8:

30.             Mention should also be made of the case of Global Torch Ltd v Apex Global Management Ltd (No 2) [2014] UKSC 64; [2014] 1 WLR 4495, which concerned an order striking out a defence as a result of the defendant’s non-compliance with a case management decision. Lord Neuberger of Abbotsbury stated at para 39, in a judgment with which Lord Sumption, Lord Hughes and Lord Hodge agreed:

“[I]ssues such as those raised by this appeal are primarily for the Court of Appeal to resolve. It would, of course, be wrong in principle for this court to refuse to entertain an appeal against a decision simply because it involved case management and the application of the CPR. However, when it comes to case management and application of the CPR, just as the Court of Appeal is generally reluctant to interfere with trial judges’ decisions so should the Supreme Court be very diffident about interfering with the guidance given or principles laid down by the Court of Appeal.”

31.             Similar remarks have been made in other decisions of this court concerned with analogous matters. For example, in the case of BPP Holdings Ltd v Revenue and Customs Comrs [2017] UKSC 55; [2017] 1 WLR 2945, which concerned guidance given by the Court of Appeal and the Upper Tribunal (UT) to the First-tier Tribunal (F-tT), Lord Neuberger, with whom the other members of the court agreed, stated at para 26:

“It is not for this Court to interfere with the guidance given by the UT and the Court of Appeal as to the proper approach to be adopted by the F-tT in relation to the lifting or imposing of sanctions for failure to comply with time limits (save in the very unlikely event of such guidance being wrong in law).”

32.             The position is different where an appeal on costs raises a question of law. Appeals to the House of Lords, or in more recent times to this court, which are purely on costs have long been discouraged, as a general rule, and will rarely meet the court’s central criterion for the grant of permission to appeal, namely that the appeal must “raise an arguable point of law of general public importance”: Practice Direction 3.3.3. Nevertheless, where permission to appeal has been granted, the court will intervene if an error of law is established. The Aiden Shipping case is itself an example. Three more recent examples can be given.

33.             First, in R (Hunt) v North Somerset Council [2015] UKSC 51; [2015] 1 WLR 3575, the appellant successfully challenged the legality of a decision taken by the respondent council, but the Court of Appeal declined to grant a quashing order because of the practical problems which that would cause, given the time which had elapsed before the appeal came on for hearing. Declaratory relief was not sought. In those circumstances, the Court of Appeal dealt with costs on the basis that the council was the successful party to the appeal. Its order was set aside by this court, Lord Toulson stating at para 15:

“The discretion of a court in a matter of costs is wide and it is highly unusual for this court to entertain an appeal on an issue of costs alone. But the Court of Appeal said that it reached its decision as a matter of principle, treating the respondent as the ‘successful party’. In adopting that approach, I consider that the court fell into error.”

34.             Secondly, in Cartier International AG v British Telecommunications plc [2018] UKSC 28; [2018] 1 WLR 3259 the question arose whether the costs to internet service providers (ISPs) of implementing website-blocking orders should be borne by the ISPs or by the persons who had obtained the orders in order to protect their intellectual property rights. The courts below had considered that the costs should in principle be borne by the ISPs, first for reasons of commercial equity, and secondly on the view that that was required by EU directives. This court intervened, on the basis that the ordinary rule in equity, in situations where a person was ordered to prevent the use of his facilities by a third party to commit or facilitate a wrong, was that the applicant bore the costs of compliance by an innocent intermediary, and the EU directives contained no contrary requirement.

35.             A third example of intervention is the case of XYZ v Travelers Insurance Co Ltd [2019] UKSC 48; [2019] 1 WLR 6075, where this court reversed the decisions of the courts below in relation to the award of a non-party costs order against a liability insurer. Earlier decisions of the Court of Appeal had laid down the applicable general principles, which in the context in question implied that the relevant question was whether the insurer had become the real defendant in relation to an insured claim, or had intermeddled in an uninsured claim, thereby causing additional costs to be incurred. The judge had erred in the exercise of her discretion by basing her order on conduct by the insurer, some of which could not reasonably be regarded as intermeddling, and the remainder of which had not in any event resulted in the incurring of additional costs. The Court of Appeal had also erred, by relying on an irrelevant consideration.

36.             In summary, therefore, this court will ordinarily be slow to intervene in matters of practice, including guidance given by the Court of Appeal as to the practice to be followed by lower courts in relation to the award of costs. The court recognises that responsibility for monitoring and controlling developments in practice generally lies with the Court of Appeal, which hears a far larger number of cases. This court is generally less well placed to assess what changes in practice can appropriately be made. It cannot respond to developments with the speed, sensitivity and flexibility of the Court of Appeal. Nevertheless, it can intervene where there has been an error of law, and has done so where a question of law arose which was of general public importance. Bearing in mind, however, the discretionary nature of decisions on costs, and the rarity of their raising any question of law of general public importance, appeals solely on costs are not ordinarily appropriate.

37.             The counterpart of this restraint on the part of the Supreme Court is that the Court of Appeal must fulfil its primary responsibility for monitoring and controlling developments in practice, including developments in relation to costs. It cannot do so, however, unless it is able to keep its decisions laying down principles of practice as to how lower courts should exercise their discretion in relation to costs, such as Davies, under review. That entails that its decisions on such matters cannot be treated as binding precedents, in the sense in which that expression is generally understood: that is to say, precedents which the Court of Appeal is required to follow in accordance with the principles laid down in authorities such as Young v Bristol Aeroplane Co Ltd [1944] KB 718 and Davis v Johnson [1979] AC 264. Were the position otherwise, the Court of Appeal would be severely restricted in its ability to introduce changes in practice, since any departure from its previous decisions could only be brought about by appeals to this court.

38.             However, the appellant’s submission that such decisions are treated by the Court of Appeal as binding precedents, in the same sense as decisions on questions of law, appears to be a misleading over-simplification of the position. In the first place, the principles of practice laid down by the Court of Appeal to guide judges in the exercise of their discretion as to the award of costs are not strictly binding even upon those judges, in the way in which a decision of the Court of Appeal on a point of law is binding upon them. There is always a residual discretion as to costs. Since “the discretion is to be judicially exercised” (Pepys v London Transport Executive [1975] 1 WLR 234, 237), the application of the principles laid down by appellate courts must be tempered by an ability to respond flexibly to unusual situations, and to reach a just result in the individual case. As was said long ago in relation to the discretion to order a jury trial, “the Court cannot be bound by a previous decision, to exercise its discretion in a particular way, because that would be in effect putting an end to the discretion” (Jenkins v Bushby [1891] 1 Ch 484, 495 per Kay LJ). Brooke LJ’s judgment in Davies itself recognised that there was scope for judges to exercise their discretion: see para 47(4), cited at para 4 above.

39.             Secondly, since a decision such as Davies establishes principles which should generally be applied as a matter of practice, as Brooke LJ repeatedly made clear (see, for example, para 47, cited at para 4 above), rather than deciding a question of law, it falls outside the scope of the rules of precedent laid down in authorities such as Young v Bristol Aeroplane Co Ltd, which are concerned with the effect of a “decision on a question of law” (p 729). It also falls outside the scope of the rationale of those rules, namely to promote legal certainty. It is therefore appropriate for a decision on a matter of practice to be reviewed where sufficient reason, such as a material change of circumstances, is put forward. Indeed, the fourth of the issues which Brooke LJ identified in para 3 of his judgment in Davies (cited at para 3 above) was whether there were “any contemporary considerations … which should tend to make the courts exercise their discretion as to costs in these cases in a different way from the way in which it was regularly exercised in the past”: a question to which he gave an affirmative answer in para 47. Previous decisions on matters of practice are not, therefore, binding in the ordinary sense of that term.

40.             At the same time, Brooke LJ observed at para 4 of his judgment in Davies that some of the issues in that case had recently been considered by the Court of Appeal in another case, R (Touche) v Inner London North Coroner [2001] EWCA Civ 383; [2001] QB 1206. He added (ibid):

“Needless to say, the latter judgment is binding on us unless we were satisfied that the court in Touche’s case overlooked relevant case law of a material nature or that its decision could be treated as not binding on us on other grounds.”

That approach, indicating that the court would only review its recent decision on the same issue of practice if persuaded that the previous case had been decided per incuriam, or that there was some other sufficient reason to do so, makes evident sense, if the Court of Appeal is to avoid repeated arguments about the principles to be adopted in costs cases, potentially divergent decisions, and the attendant risk of inconsistency and incoherence. Accordingly, although a previous decision of the Court of Appeal on a question of practice is not binding upon it in the sense in which its precedents on questions of law are binding, it is appropriate that such a decision should only be reviewed where there is sufficient reason to do so: as, for example, where there has been a material change of circumstances, or where the previous case was decided per incuriam. Following that approach, the Court of Appeal can ensure a consistent approach, and prevent the repeated re-arguing of the same points without justification, while also avoiding a situation in which its previous decisions become a potential barrier to the development of practice.

41.             Placed in that context, the dictum of Longmore LJ in Gudanaviciene at para 36 (cited at para 11(3) above) takes on a different complexion. The central issue in that case was whether the approach to costs in cases of judicial review which had been adopted in the M case, and the observations of Lord Hope in the JFS case, required the guidance given in Davies to be revisited. The Court of Appeal did not close its mind to those arguments, on the basis that Davies was a binding precedent, but examined the issues fully. As has been explained, Longmore LJ concluded that M did not represent a new approach, as had been argued, but reflected the approach generally adopted to challenges to the decisions of public bodies. That approach did not impinge upon the distinct approach taken to cases involving courts and tribunals which adopted a neutral stance towards challenges, exemplified by Davies. The case of Gudanaviciene itself was a case of the latter kind. In addition, the observations in the JFS case had been made in very different circumstances and were not in point. It was in those circumstances, where the arguments in favour of a new approach had been considered and rejected by the Court of Appeal, that Longmore LJ observed at para 36 that, if such an approach were to be adopted, that step would have to be taken by this court.

42.             Similarly, in the present case, Hickinbottom LJ carefully examined the arguments which were advanced in support of a modification of the approach adopted in Davies. Since the arguments were in substance the same as those which had recently been rejected by the Court of Appeal in Gudanaviciene, and no good reason had been put forward for reviewing the decision in that case, Hickinbottom LJ was correct to conclude that that decision should be followed. His description of it as binding was perhaps open to misinterpretation, but did not in reality involve any error of law: he plainly did not treat Gudanaviciene as a binding precedent in the Bristol Aeroplane sense. The same is true of his description of the decision in Davies as binding, and of his similar remark in Faqiri, cited at para 18 above.

43.             The central question raised in an appeal on costs is generally whether the decision awarding costs was vitiated by a failure to exercise the statutory discretion rationally and in accordance with established principles. In this case, the relevant decision was taken by King J. The submissions before him took as their starting point the established practice described in Davies, and argued that the Board should not be treated as a court or tribunal for that purpose since (in particular) it could concede a challenge to one of its decisions; that it should be treated in the same way as the generality of public bodies, and as such should normally be liable for the costs of successful challenges to its decisions; and that there were in any event circumstances justifying an exception to the usual practice, since its conduct, in failing to respond to the appellant’s letter before action in accordance with the relevant protocol, and in filing its acknowledgment of service after the due date, should be penalised by an award of costs.

44.             The judge’s reasons for declining to make such an award took full account of the appellant’s submissions, and reflected the established practice. In particular, he identified the point of central importance as being that the Board had taken the decision under challenge while acting in a judicial or quasi-judicial capacity, and had not made itself an active party to the litigation. His conclusion that the Board consequently fell within the scope of the practice described in Davies did not involve any error of law. The question whether the Board should be treated as a court or tribunal for that purpose is itself a question of practice: it is not determined abstractly or on the basis of definitions used for other purposes, such as the meaning given to the expression “court or tribunal” in the European Convention on Human Rights. Nor was there any error of law involved in the judge’s conclusion that the Board’s failure to respond to the letter before action was regrettable, but did not amount to improper or unreasonable behaviour which would justify an award of costs against it. In short, there is nothing in the judge’s reasoning which was erroneous in law, or with which this court would consider it appropriate to interfere as a matter of practice. Given that conclusion, it is difficult to see any basis upon which this court could properly allow this appeal.

45.             The same arguments were presented to the Court of Appeal in a more fully elaborated form, together with an argument concerning the significance of the appellant’s being in receipt of legal aid. The appellant’s contention that the Court of Appeal erred in law by regarding itself as bound, as a matter of precedent, to follow its decision in Davies, has already been considered and rejected. Although the court might have proceeded on the basis that no adequate grounds had been advanced for it to review its recent decision in Gudanaviciene, in which the same arguments had been considered and rejected, it considered the appellant’s submissions fully, and gave detailed reasons for rejecting them on their merits. In doing so, it committed no error of law.

46.             In particular, there is nothing in CPR rule 44.2 which is inconsistent with the approach which was described in Davies and followed in the present case. A body which takes a decision in a judicial or quasi-judicial capacity, and then declines to defend it when it is challenged in court proceedings, choosing instead to maintain its impartiality and to let the reasons which it gave for its decision speak for themselves, acts in accordance with principles of judicial independence and impartiality which have long been recognised both in English law and at an international level: see, for example, the United Nations Commentary on the Bangalore Principles of Judicial Conduct, paras 72 and 74. A judicial or quasi-judicial body which acts in that way cannot be what the framers of the CPR rule had in mind when they referred to an unsuccessful party. Furthermore, CPR rule 44.2 was in force at the time of the decision in Davies, although differently numbered. As Brooke LJ observed in Davies at para 45(1), the rule “provides unequivocally that the general rule is that the unsuccessful party will be ordered to pay the costs of a successful party, but this does not throw any light on the position of a neutral party”.

48.             The dictum of Lord Hope in the JFS case which was cited in para 10(5) is clearly consistent with that approach in so far as he said that “legally aided litigants should not be treated differently from those who are not”. It may, however, appear at first sight to conflict with it in so far as he said that “the consequences for solicitors who do publicly funded work is a factor which must be taken into account”. However, as Hickinbottom LJ explained at para 60 of his judgment in the present case, there is no inconsistency when the dictum is placed in its context. In the relevant passage of his judgment, Lord Hope was considering, obiter, whether it would have been appropriate to make a protective costs order under which each party would be liable for its own costs whatever the outcome of the appeal, in the event that such an order had been sought. The result of making such an order would have been to disapply the principles which normally govern an award of costs, so as to prejudice the legally aided claimant’s lawyers by making it impossible for them to recover remuneration at the rates which would have applied if the claimant had been privately funded. The point made by Lord Hope, in explaining why such an order would not have been made, was therefore consistent with his endorsement of the principle that legally aided litigants should not be treated differently.

49.             In so far as the decision whether to award costs against the Board turns on matters of practice, it would not be appropriate for this court, for the reasons explained earlier, to impose on the Court of Appeal its own assessment of the merits of the parties’ arguments.

50.             For all these reasons, I would dismiss this appeal.