No limits? Non-party costs against insurers

The jurisdiction to make a costs order against a non-party was first recognised in 1986.[1] There is now a wealth of guidance on non-party costs, including well-developed principles in relation to insurers, although the appellate courts continue to emphasise that this is an exercise of discretion, and that these are guidelines, ‘not… a rule-book[2]. This means that, as new factual scenarios arise such as the PIP breast implant litigation[3], the boundaries of non-party costs orders against insurers continue to be tested.

Basic ingredients

The basic ingredients for a non-party costs application are that:

  • An insurer has funded litigation
  • The litigation has been lost and an adverse costs order has been made against the insured
  • The insured is unable to pay the costs


The jurisdiction to make an order for costs against a non-party costs arises under s 51of the Senior Courts Act 1981. This provides that costs in civil proceedings are ‘in the discretion of the court’, and that the court shall have full power to determine by whom and to what extent the costs are to be paid.[4]


Balcombe LJ said in Symphony Group plc v Hodgson[5] that an order for payment of costs by a non-party will always be ‘exceptional’, but this is liable to mislead in the context of applications for non-party costs orders against insurers. As Phillips LJ said in T G A Chapman Ltd v Christopher[6] in relation to liability insurers, it must be rare for litigation to be funded, controlled and directed by a third party motivated entirely by its own interests, and although this is not extraordinary in the context of the insurance industry, that is not the test. The test is whether these features are extraordinary in the context of the entire range of litigation which come before the courts – and Phillips LJ said that he had no doubt that they were.

Solely or predominantly

Phillips LJ referred in Chapman v Christopher to a third party motivated ‘entirely’ by its own interests, reflecting the facts of that case. But in most cases, a liability insurer can credibly say that it is acting at least partly in the interests of its insured, and in Cormack v Excess Insurance Co Ltd[7], the Court of Appeal made clear that that is not sufficient to prevent a non-party costs order being made. In Cormack v Excess, Auld LJ referred[8] to what Sir Wilfrid Greene MR said in Groom v Crocker[9] about insurers being entitled to decide how to conduct the litigation provided that they did so in what they bona fide – in good faith – considered to be the common interest of themselves and their insured, and said that it might be sufficient for a finding of exceptionality that an insurer’s self-interest, though not its exclusive motivation (or effect) in its conduct of litigation, predominated over that of the insured to such an extent and in such circumstances that it strayed beyond this.

Auld LJ said that it was relevant to ask whether the insurer, when its insured was approaching or at risk of exceeding the limit of his indemnity cover, behaved solely in its own interest as if it were the defendant to the proceedings. It is important to recognise that, as Auld LJ said, this issue is distinct from the question of the reasonableness or justification of a tactical decision in litigation, such as whether to pursue or maintain a defence to an action. Once the defence has failed and an application for non-party costs is being considered, those are irrelevant considerations.

‘The’ real party or ‘a’ real party

The non-party need not be ‘the’ real party to the litigation; it is enough that they be ‘a’ real party. In Dymocks Franchise Systems (NSW) Pty Ltd v Todd[10], Lord Brown said[11]:

Where, however, the non-party not merely funds the proceedings but substantially also controls or at any rate is to benefit from them, justice will ordinarily require that, if the proceedings fail, he will pay the successful party’s costs. The non-party in these cases is not so much facilitating access to justice by the party funded as himself gaining access to justice for his own purposes. He himself is “the real party” to the litigation… Consistently with this approach, Phillips LJ described the non-party underwriters in T G A Chapman Ltd v Christopher… as “the defendants in all but name”. Nor, indeed, is it necessary that the non-party be “the only real party” to the litigation…, provided that he is “a real party in … very important and critical respects”’.

Liability insurance

Liability insurers invariably fund and take over the defence of the proceedings, and benefit from them, but something more than this is required for a finding on an application for non-party costs that the non-party liability insurer has not merely funded the proceedings, but substantially also controlled or at any rate was to benefit from them so that it is appropriate for an order to be made.

(1) Controlling the litigation

In Citibank NA v Excess Insurance Co Ltd[12], Thomas J said[13] that there were two aspects to the conduct of litigation:

  • First, obtaining information to draft the defence, instructing an expert, considering the statements of case, disclosure of documents, preparing witness statements, providing information to counsel, etc.
  • Secondly, the giving and receiving of advice and the taking of the decisions as to whether to defend the case through to trial or to attempt to settle the litigation.

Thomas J said that although the involvement of the actual defendant in the first of these activities was a factor, ‘the decisive factor as to who has the conduct of litigation must be the control and direction exercised through the giving and receiving of advice and the taking of the decisions’.

(2) Benefitting from the litigation

Where a defence is run sensibly and reasonably for the benefit of both the insured and insurers, and the court is unlikely to view the insurer as benefiting from the litigation to the extent required for a non-party costs order, and is unlikely to exercise its discretion to make an order in that situation. Simply benefiting from the litigation is not enough: this is why the issue of whether insurers have allowed their interests to predominate over that of the insured is so important in this context. Consider for example the facts of two of the cases referred to above:

  • Chapman v Christopher: The defendant, Mr Christopher, lived at home with his mother and had no assets. He did though have liability insurance with a limit of indemnity of £1m under his mother’s household insurance policy. The claimants’ warehouse and factory were extensively damaged in a fire which was caused by Mr Christopher’s negligence: he had thrown a lighted match which landed in an open tin of beeswax which immediately caught fire. The insurers were the defendants in all but name: by reason of the Third Parties Rights Against Insurers Act 1930, they were contingently liable to the claimants up to the policy limit of £1m. They took the decision to contest the litigation and subsequently conducted the defence, in an attempt to avoid or reduce their liability to the claimants. Phillips LJ described this as a ‘paradigm case for the exercise of the court’s discretion under s 51 to make a costs order against a non-party’.
  • Cormack v Excess: The insured were civil engineers with a reputation to protect, and the Court of Appeal accepted that the decision to defend was not taken entirely by the insurer, and that, until liability was determined, the proceedings were not defended solely for the benefit of the insurer. However, once liability was determined, and the judgment exceeded the limit of indemnity, the proceedings in relation to quantum were defended entirely for the benefit of the insurer. The insurer was ordered to pay the claimant’s costs of the quantum stage of the action.

(3) Conflict of interest

Where solicitors are on the record for the insured in proceedings, and are acting under a joint or dual retainer for the insured and for the liability insurer, it seems that a failure to recognise or deal appropriately with a conflict of interest, and as a result to allow the interests of the insurer to predominate over those of the insured, will strongly incline a court to exercise its discretion in favour of making a non-party costs order.

XYZ v Travelers

In XYZ v Travelers, the PIP breast implant litigation, the insured went into insolvent administration. A mixture of insured and uninsured claims were being pursued in group litigation, and the effect of the Group Litigation Order (‘GLO’) was that the more uninsured claims which were pursued, the lower the proportion which the insurer would have to contribute to the insured’s liability for common costs. The insured wanted to disclose to the claimants whose claims were uninsured that that was the case, but were advised by the legal team appointed by insurers not to do so. There was a conflict of interest, which no one recognised at the time. The Court of Appeal said that this was not decisive, but they did agree with the judge that a non-party costs order should be made.

(1) ‘Asymmetry’

One particular feature which made the case ‘exceptional’ was the ‘obvious asymmetry’ in insurers’ position: if the insured had succeeded on the preliminary issues then all claimants (whether insured or uninsured) would have been liable equally to contribute towards the insured’s costs which, ultimately, would have been to the insurer’s advantage; but failure on those very same issues meant that, unless a non-party costs order was made, insurers were ultimately liable for only approximately 32 per cent of the claimants’ costs. This asymmetry arose because of the terms of the GLO.

(2) Non-disclosure of the insurance position

In Cormack v Excess, Auld LJ said[14] that as there was no obligation in litigation to disclose the limit of indemnity, it was difficult to see why an insurer should be penalised in costs for not doing so, and that a court should be cautious before regarding a failure to disclose the extent of cover as sufficiently exceptional to justify the making of a non-party costs order. In Travelers v XYZ, while they based their decision on the asymmetry they identified between the position of the claimants and the insurers referred to above, the Court of Appeal identified a number of reasons why Auld LJ’s observation did not apply:[15]

  • It was not alleged in Cormack v Excess that the failure to disclose the cover limit had any causative effect on costs. In Travelers v XYZ, by contrast, the judge was satisfied that, if the lack of insurance had been disclosed, costs would not have been incurred.
  • The non-disclosure in Cormack v Excess was the cover limit. The non-disclosure in Travelers v XYZ was the non-existence of any insurance at all.
  • The policy itself and the pre-action protocols in Travelers v XYZ required any response to a letter of claim to include details of the insurance policy.

Underlying and linking all these points seems to be a single factor, which is that the claims, both insured and uninsured, were subject to a GLO:

  • The GLO gave rise to the asymmetry in relation to the recoverability of costs
  • The GLO gave rise to the conflict of interest between insurers and the insured about the desirability of disclosing the fact that some of the claims were uninsured, and in turn to the flawed advice given to the insured by the legal team not to disclose that fact
  • The GLO formed the basis for Thirlwall J’s case management decision[16] that the insured should provide her with information about its insurance position
  • The GLO resulted in the common issues of both insured and uninsured claims being tried together, so that the insurers were funding the costs of defending all the claims, including the uninsured claims

Against this background, although the information about the insurance position provided to the judge was not disclosed to the parties, the Court of Appeal said that it must have been obvious to the insured and to insurers that the perception of the uninsured claimants was that all of the underlying claims were insured.[17] In these circumstances, it is unsurprising that the Court of Appeal decided that ‘on balance’ the flawed advice given by the legal team appointed by insurers in relation to the disclosure of the insurance position was relevant to the insurers’ liability for costs, although not decisive, and that it was not unjust for the insurers to bear some responsibility for the advice given under the joint retainer[18].

BTE insurance

The same principles apply in relation to before the event (‘BTE’) legal expenses insurers as to liability insurers. Unlike liability insurers, BTE insurers will not usually control the proceedings, or have an interest in the result of the litigation save insofar as it affects their liability to pay costs. They are therefore not generally balancing their interests with those of the insured, and will not usually be exposed to a risk of a non-party costs order.[19]

ATE insurance

After the event (‘ATE’) legal expenses insurance has features which are distinct from BTE insurance: ATE insurers decide which cases to insure after the cause of action has arisen, and the recoverability of the policy premium may depend on the outcome of the litigation.

In Herridge v Parker[20], Mr Recorder James Thom QC, sitting as a judge of the County Court, held[21] that ATE insurance was in the public interest as it facilitated access to justice and that, by analogy with the case-law in relation to solicitors acting under conditional fee agreements, an ATE insurer would not be ordered to pay costs as a non-party simply by virtue of issuing a policy which was insufficient, either because the policy was avoided, or because the limit of indemnity was exceeded.

The judge also observed that an ATE insurer who chose to prolong the proceedings for the purposes of seeking to negotiate a favourable exit might well be acting in a sufficiently self-interested way to become a ‘real party’. In those circumstances, a non-party costs order might be made against the insurer in respect of the additional costs. The same logic would apply if a BTE insurer acted in this way.

Subrogated claims

Where insurers have funded a subrogated recovery, they will usually pay the costs if the action is unsuccessful. But if they do not, an application for a non-party costs order may be made against them, and the same principles applied as on an application against liability insurers. ‘What has been sauce for the goose would have been sauce for the gander’, as Phillips LJ said in Chapman v Christopher, in which although the focus was on the position of the liability insurers who had funded the unsuccessful defence, both parties were ‘litigants in name only’, as the claim itself was brought by underwriters exercising subrogated rights.

Alison Padfield QC

  1. In Aiden Shipping Co Ltd v Interbulk [1986] AC 965, HL.
  2. Petromec Inc v Petroleo Brasileiro SA Petrobras [2006] EWCA Civ 1038, para 10 (Laws LJ); Travelers Insurance Co Ltd v XYZ [2018] EWCA Civ 1099, para 30 (Lewison LJ; Patten LJ agreeing).
  3. Travelers Insurance Co Ltd v XYZ [2018] EWCA Civ 1099.
  4. Section 51(1) and (3).
  5. Symphony Group plc v Hodgson [1994] QB 179, CA.
  6. [1998] 1 WLR 12, CA.
  7. [2002] Lloyd’s Rep IR 398, CA.
  8. At 406, cols 1 to 2.
  9. [1939] 1 KB 194, CA.
  10. [2004] UKPC 39, [2004] 1 WLR 2807.
  11. At para 25.
  12. [1999] Lloyd’s Rep IR 122.
  13. At 133, col 1.
  14. At 406, col 2.
  15. At para 44 (Lewison LJ; Patten LJ agreeing).
  16. XYZ v Various [2013] EWHC 3643 (QB).
  17. At para 42 (Lewison LJ; Patten LJ agreeing).
  18. At para 45 (Lewison LJ; Patten LJ agreeing).
  19. See Murphy v Young & Co’s Brewery [1997] 1 WLR 1591, CA.
  20. [2014] Lloyd’s Rep IR 177.
  21. At paras 89-96.