Neutral Citation Number: [2006] EWCA Civ 1134
Case No: A2/2005/2783, A2/2005/2783(A), A2/2005/2783(B)
 

IN THE SUPREME COURT OF JUDICATURE
COURT OF APPEAL (CIVIL DIVISION)
ON APPEAL FROM CARDIFF COUNTY COURT
His Honour Graham Jones, sitting as a Deputy Circuit Judge
Deputy District Judge Sandecott

&nbdp; Royal Courts of Justice
Strand, London, WC2 A 2LL

31 July 2006

  LORD JUSTICE BROOKE
Vice-President, Court of Appeal (Civil Division)
LORD JUSTICE LAWS
and
LADY JUSTICE SMITH

  JONATHAN LUKE ROGERS
Claimant/Appellant

and

MERTHYR TYDFIL COUNTY BOROUGH COUNCIL
Defendants/Respondents

  Andrew Bartlett QC and Nicholas Bacon (instructed by Messrs Coleman Coyle LLP) for the Appellant
Benjamin Williams (instructed by Messrs Dolmans) for the Respondents
Nigel Cooksley QC (instructed by the Law Society) for the Law Society (as interveners)
Richard Drabble QC (instructed by Coleman -- ctts) for Temple Legal Protection Ltd & Keystone Legal Benefits Ltd (as interveners)


HTML VERSION OF JUDGMENT
Crown Copyright ©

 

  
INDEX
  Part Para
1. The proceedings in the courts below 1
2. The preliminary proceedings in this court.. 6
3. The evidence before Master Hurst: Appellant's witnesses: (1) Christopher Cater 16
4. The evidence before Master Hurst: Appellant's witnesses: (2) Phillip Bellamy 30
5. The evidence before Master Hurst: The DAS 80e Product 39
6. The evidence before Master Hurst: The Respondents' witnesses 58
7. The Law Society's submissions to Master Hurst 60
8. The experiences of Temple and Keystone 61
9. The Law Society's evidence 74
10. Allianz Cornhill's experience 83
11. The evidence from LAMP 84
12. Brit's submissions 89
13. Abbey's observations 92
14. The Liability Insurers' Group's submission 93
15. Discussion 96
16. Proportionality 102
17. The legitimacy of a staged premium model 107
18. The size of the premium in the present case 108
19. The Respondents' submissions 113
20. The procedure to be followed in future 116
21. Evidence justifying the ATE premium claimed 117
22. Self-insurance of the premium in the event of failure or only partial success 118
23. An issue we have not decided 119
24. Conclusion 120
  Annex by Lady Justice Smith 123
    
   Lord Justice Brooke: This is the judgment of the court.
   1. The proceedings in the courts below
1.   This is an appeal by the claimant, but in reality by DAS Legal Expenses Insurance Co Ltd ("DAS"), who acted as the claimant's "after the event" ("ATE") legal expenses insurers, on appeal from a decision of His Honour Graham Jones sitting as a deputy circuit judge on 23 November 2005 when he allowed an appeal by the defendants from an assessment of costs by Deputy District Judge Sandecott in the Pontypridd County Court on 5 March 2005. This case is concerned with an injury to the claimant on the evening of 23rd August 2001 when he was 11 years old. He fell over on the perimeter of a play area in a local park in Merthyr Tydfil, and was cut by shards of glass embedded in the timber edging of the play area. There were no witnesses.
2.   The damages were agreed in the sum of £3,105 plus interest. We were told that one of the reasons why the defendants resisted liability was that there was a hospital note to the effect that the boy had suffered his injuries after falling from a swing. It was not until the trial that it became clear that it was his mother who was the source of this misinformation, and that she had simply told the hospital that night how she understood the accident had occurred. We have not seen the parties' statements of case or the witness statements disclosed before the trial. We therefore do not know whether the costs of this very expensive trial might have been avoided if the defence had been settled in such a way (pursuant to CPR 16.5(2)(a)) that this explanation would inevitably have surfaced in the mother's witness statement before the parties prepared for trial. However that may be, the matter did go to trial on liability and the deputy district judge held that the defendants were liable. He then assessed the costs summarily in the sum of £16,821.30, which was a reduction from the sum of £18,632.73 on the schedule produced to him. The sums he allowed included an ATE premium of £5,103, being £4,860 premium plus insurance premium tax ("IPT"). He upheld a 100 per cent success fee.
3.   The defendants appealed against that order. There was no new evidence adduced for the purposes of the appeal, apart from some material from the April 2002 edition of a magazine called Litigation Funding (a date chosen because the defendants wrongly believed that the ATE insurance had been taken out in July 2002). This material was sent by the defendants to the claimant's advisers five months before the appeal was heard. It was included in the appeal bundle without being formally put in evidence. The approach of the judge was influenced very considerably by the contents of that document. He heard the appeal on his own without an assessor, and he reduced the costs to £12,628.30, including VAT. In particular, he reduced the recoverable ATE premium to £900. The base profit costs of £4,000, the 100 per cent success fee of £4,000, counsel's fees of £1,100, and court fees, disbursements and VAT were all left undisturbed.
4.   The judge was influenced by the fact that DAS was using an ATE product with a three-stage premium. £450 was payable at the outset, a further £900 when proceedings were issued, and a further £3,510.60 days before the trial. This third stage premium was individually rated, and IPT of 5% was added at each stage. The judge observed that according to Litigation Funding, ATE insurance was available at a cost of between £450 and £1,350 from a range of companies, and he could not see that the staged premium justified the wholly excessive charge that was made, given that other companies in the market were charging much less than £1,800 for the whole proceedings.
5.   He said that nobody conducting litigation reasonably would have agreed to pay a premium of this order if the case did not settle, when he could obtain an insurance policy that covered him for the whole of the conduct of the proceedings for much less than was payable at Stage Two of the DAS policy, and indeed for not very much more than was payable at Stage One. He decided that the evidence before him clearly indicated that products were available from a range of companies for cheaper premiums providing the same amount of cover, including the self-insurance aspects of it. That was why he reduced the recoverable premium to £900, which he said was in the middle of the range of £450 to £1,350 quoted in Litigation Funding.
   2. The preliminary proceedings in this court
6.   In the grounds of appeal for a second appeal to this court, the claimant's insurers asserted that the judge had been wrong to reduce the premium in this way. He had failed to pay any or any sufficient regard to the risks being insured under the policy. If he had done so he would have allowed the premium in the sum claimed. They said that the £900 did not reasonably reflect the risks insured, and that the judge had failed to apply paragraph 11.10 of the Costs Practice Directionpdp-43. He had been wrong to take any account of the material in Litigation Funding. Reference was made in this context to what Master Hurst had said about this in his judgment in re RSA Pursuit Test Cases [2005] EWHC Costs 9003 (Costs) (27 May 2005).
7.   They asserted that the sum of £900 was unreasonable and therefore wrong in the context of CPR 44.4, and that the judge was wrong not to have ordered a detailed assessment as opposed to attempting a summary assessment on the limited evidence and information available. They accepted that the arguments they wished to advance on the appeal were significantly different from those advanced by the different solicitors and counsel instructed before the deputy judge. In particular, DAS wished to adduce a good deal of evidence about the way they set about their task in fixing a premium, none of which was placed before the judge.
8.   With their notice of appeal they filed a witness statement from Mr Bellamy, who is DAS's Operations Manager, with particular responsibility for the operation of a product called the DAS 80e policy. He said the appeal was of extreme public importance and of critical importance to the insurance industry in general. The judge's arbitrary reduction of the premium was so severe that if it should be used as a persuasive precedent (with the result that all ATE trial premiums were reduced to this level), some ATE insurers would stop writing this business overnight. With an already very limited UK-based ATE insurance market, this result would be at odds with the Government's Access to Justice Act requirements. The judge's decision, if unappealed, would impact on the entire ATE market, and the question of a second appeal had to be considered from this global position. He said that, in many hundreds, if not thousands, of settled cases throughout the market, defendants were refusing to pay an ATE insurance premium and were awaiting a decisive ruling from the Court of Appeal.
9.   He said that another hindrance for the fledgling ATE market was that insurers not only had to predict future claim numbers and claims costs on losing cases, but they also had to be aware they might not collect the premiums due to them on the cases they won for many months or years. This made accurate underwriting of this market almost impossible. Numerous high profile insurers had departed from the ATE market over the last two years, and a judgment of this magnitude was exactly what this sector of the insurance industry needed in order to provide certainty and clarity in an otherwise uncertain market.
10.   On the basis of this evidence, Lord Justice Brooke directed that a second appeal should be permitted, and on his instructions the Civil Appeals Office wrote to a number of interested parties, including other ATE insurers, to inquire whether they wished to make submissions or to intervene in the appeal. As a result of that invitation, submissions of one kind or another were made by Allianz Cornhill ("Allianz"), Abbey Legal Protection ("Abbey"), Brit Insurance Ltd ("Brit"), LAMP Insurance Co Ltd ("LAMP"), Temple Legal Protection Limited ("Temple"), Keystone Legal Benefits Ltd ("Keystone"), and the Law Society. Lord Justice Brooke then convened a case management conference on 23rd May.
11.   On that occasion he decided that this matter should proceed as a second appeal as a test case, and he permitted the parties and the interveners to adduce such evidence and submissions as they thought fit. The appeal was to be by way of re-hearing. In view of the fact that the defendant council had little interest in any wider issues, and were self-funding so far as their liability to the claimant was concerned, he directed that the appeal should go ahead on condition that the orders as to costs in the courts below would stand in any event, and that each side should bear its own costs in the Court of Appeal (save that if the appeal was dismissed or if the defendants wished to assert that the appellant had secured a less advantageous result than any relevant offer, he made special provision in relation to their ability to apply for costs (up to a capped limit) in the Court of Appeal). He granted permission to the Law Society and to all the companies listed in para 10 above to intervene in the appeal. He also permitted intervention from any liability insurers who wished to intervene and who were willing to comply with the terms he set for their intervention. He adjourned further directions to the senior costs judge, who has acted as our assessor throughout.
12.   At the request of Lord Justice Brooke Master Hurst held a case management conference on 12th June, and a further hearing on 3rd July. He then furnished a report to the court which provides an admirable foundation of fact on which we can decide the issues of law that arise on this appeal.
13.   The evidence showed that after the accident the claimant's father instructed a local firm of solicitors called Spicketts. His son's case was conducted by a partner in the firm, Mr Christopher Cater, who was and is a member of the Law Society's Personal Injuries Panel. On 18th July 2002 a conditional fee agreement ("CFA") was entered into with a 100% success fee. No "after the event" ("ATE") legal expenses insurance was taken out at that time. On 26th January 2004 Spicketts completed a proposal form for a DAS 80e policy in which they stated the likelihood of winning the claim to be 60-70%. The policy was backdated to cover the period from the inception of the CFA. Although DAS maintains panels of solicitors for their BTE insurance product, they do not maintain a panel for their 80e insurance.
14.   At the hearing on 3rd July 2006 Master Hurst received written and oral evidence from two witnesses for the appellant. These were Mr Cater and Mr Bellamy (for whom see para 8 above). The respondents, for their part, had adduced written evidence from three witnesses, none of whom was able to attend for various reasons before Master Hurst on 3rd July. These were Mr Mitchell (an assistant solicitor in the firm acting for the Respondents), Mr Boobier (a partner in that firm), and Mr Hudman (a claims executive with Marsh Insurance Brokers). Master Hurst ruled that he would admit their evidence, but he would bear in mind that evidence given by live witnesses and tested by cross-examination was likely to carry far more weight than the evidence of witnesses which had not been so tested.
15.   The evidence before Master Hurst ran along the following lines.
   3. The evidence before Master Hurst: Appellant's witnesses: (1) Christopher Cater
16.   As we have said, Mr Cater acted for the claimant under a CFA, and it was on his recommendation that the ATE insurance was obtained under a DAS 80e policy. This policy protected the claimant against any liability for his own disbursements and/or against liability for the defendants' costs and disbursements.
17.   Mr Cater's firm had originally put all its ATE insurance through a company I will call X. He was on a panel of solicitors with this company and they gave him delegated authority to place cover for his clients. The cost of this insurance was £250 plus IPT. At about the end of 2002 he started looking around for an alternative ATE provider because he had experienced difficulties with X after one of his cases failed at trial. X had been reluctant to pay the defendant's costs, and it had required the production of the claimant's own file because it wished to review the conduct of the case. It had also indicated on more than one occasion that it might look to Mr Cater's firm to recoup any outlay. Although this had never happened, Mr Cater decided to seek a different insurer who would not cause him these difficulties.
18.   At the material time his firm was relatively small, and it placed a relatively small amount of personal injury work with ATE insurers. More recently the firm has amalgamated and the volume of work has increased. At the relevant time Mr Cater believed he had limited purchasing power with ATE insurers. Although DAS suggested that the average was 50 cases per year, he thought that the figure might be higher than that. He agreed that the figure of 50 was not a 2002 figure.
19.   In 2002-3 his firm had run a number of cases through The Accident Group. His normal procedure was to wait for a response to the pre-action letter before deciding whether it was appropriate to take out ATE insurance. He mainly obtained insurance for tripping and slipping cases.
20.   When he was looking for an alternative ATE provider he decided to approach three reputable providers. These were Law Assist, Legal Protection Ltd, and DAS 80e. The first two sent him literature to read, while DAS sent a representative, and he was impressed by the presentation he was given about their product and the service which they gave.
21.   He said that there were at least five features in any insurance that he would want to recommend to clients. These were:
  
 i)   The insurance provider must be reputable and one in whom he and his clients could have confidence;
 ii)   If a claim failed at trial the insurer must be relied on to pay out under the policy without argument and without seeking to decline cover. When asked why he thought DAS would be less difficult to deal with than X, Mr Cater said that he was aware that DAS was a major underwriter in the BTE market, and that when he had asked about their 80e procedure he was satisfied that they would not take the same negative approach as X.
 iii)   There had to be an adequate level of indemnity. DAS 80e provided a limit of indemnity of £100,000 which he considered ample for the cases with which he would be dealing. He agreed in cross-examination, however, that £25,000 would have been sufficient for a fast track case.
 iv)   He wished to ensure that the premiums were deferred and that the client was not required to take out any loan or to incur any interest. He was also keen that it should not be necessary for the client to enter into any agreements to which the Consumer Credit Act ("CCA") would apply, since he thought that the signing of CCA forms might cause a client undue concern.
 v)   He was seeking an efficient and helpful service from the ATE provider, and also a flexible approach, particularly when decisions had to be made during the currency of a case. In this context he was very impressed by the DAS representative, both at that meeting and later, because he always offered a friendly and helpful service which Mr Cater thought ideal for his clients.
22.   He was also attracted to the three-stage premium He thought that this could operate as an incentive to defendants and their solicitors to settle claims early, once they realised that the first stage premium was going to be low. He also hoped that it would dissuade insurers from denying liability in cases where, in reality, his client's claim should be accepted. He did have concerns that the third stage premium would inevitably be "quite high", but he countered this with the thought that relatively few cases go to trial. In the event rather more of his slipping and tripping cases went to trial than he had originally anticipated. He put the failure rate of such cases at trial as being in excess of 50%.
23.   Given the type of work undertaken by his firm, he said it was neither proportionate nor practical to seek an individual policy in every case. He accepted that ATE providers would require him to insure all his available cases through them to avoid the difficulties caused by adverse selection.
24.   In the light of these considerations he concluded that DAS 80e was a good product which fulfilled all his requirements. His firm therefore entered into an agreement with DAS on 1st July 2003. They do not receive any commission or other payment from DAS, nor do they receive any work from that source.
25.   When asked if he had considered cheaper products, he said that in his experience cheaper products were not as good as that offered by DAS. The X policy with which he had had such difficulties had been one of the cheapest on the market at £250. His concern was to protect his client.
26.   He later considered other products in the light of challenges to the level of premium in the DAS 80e policy, but he concluded that this product was better than most. In this context he mentioned one product which involved an off-shore insurance company and the payment of commissions. He also knew that other products on the market might require a solicitor's firm to be able to commit itself to much greater volumes of work than his firm could provide. There was evidence before Master Hurst that Allianz say in their brochure (apparently dated July 2005) that they require a minimum of 100 policies a year from the solicitors who insure with them because "this is due firstly to reduce the risk of adverse selection, and secondly the cost of setting up a firm to issue policies on delegated authority."
27.   Once he had decided that ATE insurance was required in any particular case, Mr Cater would recommend it to a client in accordance with the terms of his agreement with DAS 80e. Para 1.4.4 of that agreement provides:
  
   "The Firm agrees to sell 80e products as their preferred provider and agrees not to supply any after the event product using another provider, unless the Claim has first been offered to and refused by 80e."
28.   He said that if he felt that the 80e policy was inappropriate for a client he would not recommend it. In those circumstances, given the terms of his agreement with DAS 80e, his only option would have been to approach another insurer on a one-off basis. Since in the present case a 100% success fee had been applied in the CFA, he said that any insurer approached on a one-off basis would either have refused insurance or would have demanded a substantial premium. If the case had then settled early his client would have been confronted with an argument to the effect that there had been no need to take out such an expensive policy when many other policies (including 80e) offered much cheaper options if settlement occurred at an early stage.
29.   He said that another factor he had to take into account was that if an insurer declined cover or quoted an unacceptably high premium, he would have to disclose this fact to any subsequent insurer, with inevitable consequences. He added that it would be disproportionate if he had to review the market in every individual case, in contrast to using a reputable product which, at least in the first two stages, involved very modest premiums.
   4. The evidence before Master Hurst: Appellant's witnesses: (2) Phillip Bellamy
30.   Because of the attention the deputy judge paid to Litigation Funding, Mr Bellamy produced a table based on the table from that publication which showed the main providers of ATE insurance for CFA cases in April 2002. He also produced a similar table for April 2006. He pointed out that the tables did not give full details of the policy cover, or any indication as to the conditions that related to their availability and usage. The publication made it clear that the figures stated are indicative only.
31.   In his judgment in re RSA Pursuit Test Cases Master Hurst said (at para 235):
  
   "As to the information contained in Litigation Funding and The Judge website, this is no more than an indication of policies which might be available in certain circumstances. As [counsel] points out, the premiums on his website are 'indicative only' and the website contains further warnings. Litigation Funding has similar warnings and reservations. I can derive no firm data from these sources."
32.   Master Hurst said that he accepted what was said in Mr Bellamy's statements in so far as they contained factual evidence, as opposed to submissions. His evidence was along the following lines.
33.   For all but the more serious or specialised personal injury claims, ATE insurance is block rated rather than individually assessed. For block rating to work the insurer needs to be sure that it is receiving a full and fair selection of cases, ranging from those where liability is unlikely to be in doubt to those where it is contested. In order to avoid adverse selection it is standard practice for ATE insurers to require solicitors to insure all available cases with the ATE provider. In practice, therefore, claimants' solicitors cannot simply pick and choose from a variety of products and offer different policies to different clients. This approach is, in any event, incompatible with block rating.
34.   The level of premium is not generally, nor in this particular case, predicated by or related to the amount of damages. It is set by reference to the risk to the insurer and his exposure to costs and disbursements in a failed case.
35.   In setting a premium ATE insurers adopt two different approaches. They either adopt a single fixed premium that applies throughout the case whenever it is concluded, or a staged, or stepped, premium (as in this case), where the amount payable increases as the case progresses. Although fast track cases can be insured by single fixed premiums, most providers require multi-track cases to be separately rated.
36.   Mr Bellamy suggested that care must be exercised when one looked at the available market, since there were a number of products offered by off-shore providers "which have serious drawbacks". These drawbacks included the fact that in some off-shore locations the level of regulation was lower than it is here. He accepted, however, that the regulation of off-shore providers was not his speciality.
37.   He then referred to "ghost policies", where a certificate of insurance is issued, but there is also an agreement with the solicitor that no claims, or only a limited number of claims, will be made against the insurer. In effect the solicitor insures, and the premiums are shared between the insurer and the solicitor. The insurer may thus offer what are apparently attractive premium rates, but will bear little, if any, loss. When asked what difference this would make to the client, Mr Bellamy accepted that the client would still have the policy even if his solicitor went out of business. He would therefore not be affected by the "hold harmless" clause.
38.   The April 2002 edition of Litigation Funding had identified 22 ATE products on offer through eight different underwriters. Since that time a number have left the market (including NIG, Wren, Lloyds, Goshawk, Catlin, Atrium, Lexington, St Paul and Europ Assistance). Some of those who have remained have changed the products they offer. Because of this volatility in the market solicitors find it difficult to keep abreast of what is available and on what terms. Those who left the market are thought to have suffered significant losses, particularly as a result of their experiences with Claims Direct and The Accident Group. Some off-shore insurers, however, have entered the market since 2002, as has Allianz, which is also a liability insurer.
   5. The evidence before Master Hurst: The DAS 80e Product
39.   The 80e policy is called "Justice Solutions". It has a three-stage premium. It is available for all non-motor claims, subject to a number of exclusions which are not covered under delegated authority. The policy may be cancelled within 90 days with no charge. The premium structure is set out in para 4 above, starting with the low Stage One premium of £450 plus IPT. The DAS database shows that 63% of 80e cases settle at Stage One, of which 21 will be lost or discontinued. This figure excludes cases where cover is cancelled during the first 90 days: in practice DAS incurs no liability within this period.
40.   The larger Stage Two premium (a further £900 plus IPT) becomes payable if proceedings are issued. This sum is said to reflect the greater risk in relation to adverse costs. 95% of all 80e cases do not reach Stage Three. This figure includes the 63% which settle (or are otherwise disposed of) without the need to issue proceedings at all.
41.   The Stage Three premium becomes payable in the remaining 5% of cases as the case approaches trial. In the present case, the liability was triggered 60 days before trial. This trigger-point has been subsequently reduced to 14 days before trial for a fast track case, and 21 days before trial for a multi-track case, in order to bring things into line with the new predictable costs regime under CPR Part 45.
42.   At Stage Three the risks involved vary significantly, and it was felt better to rate this element of the premium individually. DAS's aim is to make sure that the trial premium is directly proportional to the risk involved, so that each case is individually underwritten, taking into account the merits and the estimated maximum loss figure (EML). In order to calculate the EML the claimant's solicitor is asked to provide details of the disbursements he has already incurred and an estimate of his own side's disbursements up to and including trial. He is also asked for an estimate of the opponent's total costs and disbursements up to and including trial. The estimates provided in the allocation and pre-trial questionnaires are used when they are available.
43.   In the present case the figures passed to the senior underwriter were:
  
 i)   Claimant's disbursements estimated at £2,000;
 ii)   Opponent's profit costs estimated at £3,000;
 iii)   Opponent's disbursements estimated at £1,500;
 iv)   Total EML £6,500.
44.   The underwriter is then required to assess the risk and to apply a percentage in order to calculate the premium. In this case liability had been denied and there was no Part 36 offer. The prospects of success had been assessed by case handlers as "acceptable", which in effect meant 51%. Mr Bellamy would not expect prospects of success to be rated much higher than this in a case about to go to trial where liability was still denied. Based on that information the underwriter applied a rate of 54% to the EML, producing a third stage premium of £3,510 plus IPT. The insurers expect to lose about half the cases which go to trial. Defendant insurers will therefore be called upon to pay the Stage Three premium in about 2½% of all cases.
45.   DAS have carried out research every month for the last 12 years in relation to their claims experience. The latest figures show that the company loses 7 out of 10 slipping and tripping cases.
46.   It was put to Mr Bellamy in cross-examination that the fact that the client had already paid the first two stages of premium was not taken into account, and the Stage Three premium was rated against the whole of the loss. Counsel for the defendants suggested that if the earlier premium were added back in, the rating would become almost 75%.
47.   Mr Bellamy produced a comparative table showing the premiums payable under the 80e policy against those of Brit (Abbey) and Allianz (National Accident Helpline). These figures were necessarily based on average premiums. The total average premium per case was as follows:
  
 i)   DAS 80e: £1,033 plus IPT;
 ii)   Brit: £1,270 plus IPT;
 iii)   Allianz: £950 plus IPT.
48.   Mr Bellamy said that in his experience the vast majority of cases are settled without the need for a trial. Although a trial was always a possibility, it was therefore far more likely that the case would in fact be settled, either by the defendant accepting liability or possibly by the claimant abandoning his claim.
49.   Mr Bellamy accepted that the Abbey policy, with a premium of £813.75, would have been cheaper than the 80e policy in this particular case, but he pointed out that it would have been more expensive in the majority of cases which settle at Stage One. The initial premium for the Abbey policy is £775 (as compared with DAS's £450). Similarly, the Temple policy has a lower trial premium than the DAS 80e policy, but it has higher premiums at Stages One and Two: their Stage One premium is £850, and Stage Two £1,500 (see the table in para 66 below).
50.   Mr Bellamy exhibited to his witness statement a table analysing "claims disposed of" by stage and income against costs. This table sets out the total premium recoverable, from which is deducted the total paid out in respect of lost or discontinued cases, and also the premiums rendered uncollectible because the case has been lost or discontinued. It demonstrates that in 100 average cases
  
 i)   The total premiums due would be £101,377;
 ii)   The total paid out in respect of lost and discontinued cases would be £57,779;
 iii)   Total uncollectible premiums would be £28,718;
 iv)   There would therefore be a gross profit (prior to expenses) of £14,880.
51.   This figure breaks down to a gross profit of £2,351 for Stage One and £21,802 for Stage Two, but a loss of £9,273 for Stage Three. Mr Bellamy points out that, in arriving at his average figures, he is dealing with a very immature book which is currently only five years old, and there are still a large number of cases to run off. Until these cases close DAS cannot calculate its losses accurately.
52.   Mr Bellamy went on to suggest that a reasonable allowance for overheads would be 25% of net premiums. Using his figures he suggests that DAS should expect to recover £24,137 for overheads. This would in fact produce an operating loss overall of £9,257 for every 100 policies written. His figures are only on the basis of overheads and running costs, and they make no allowance for any operating profit. Nor is there any allowance for commission to intermediaries. (There is no commission element in this particular case in any event). When the DAS 80e product was originally conceived the idea was that the Stage Three premium should break even. The calculations were based on an assumption that about one third of cases would fail at trial. On this basis a premium of 54% (see para 44 above) would have meant roughly a break even point. Experience has shown, however, that in fact some 50% of cases fail at trial. 54% is therefore inadequate.
53.   In calculating a Stage Three premium, an assumption is made that cases do not have a better than 51% chance of success: were it otherwise DAS would have expected the defendant to have conceded. On this basis the underwriter does not consider counsel's advice at that time.
54.   In response to a suggestion that the DAS 80e premium is unreasonable in the way it is staged (since it does not reflect the way that the risk of liability for costs is incurred throughout the claim process), Mr Bellamy produced a schedule that showed six recently lost cases for each of the three stages of premium, giving the figures paid out in each case in respect of "own side disbursements", "other side's costs" and premium. The premium is self insuring if the claim is lost, and the loss of premium is therefore included in the figures.
  
 i)   At Stage One the totals paid out ranged from £482.50 to £1,776.52;
 ii)   At Stage Two they ranged from £2,121.50 to £11,224.59;
 iii)   At Stage Three they ranged from £8,119 to £13,348.13.
55.   A medical report has to be obtained prior to issue in a number of cases (as well as medical records), and it may also be necessary to incur other disbursements. The average cost to DAS pre-issue, exclusive of premium, is £833. At Stage Two the average cost, exclusive of premium, is £4,472. (A substantial part of this represents the costs paid to the successful defendant). At Stage Three the average cost, exclusive of premium, is £8,950. (Again, most of this figure is made up of defendant's costs).
56.   Mr Bellamy also exhibited a copy of the summary of the work which DAS undertake in relation to a new agent application, case management and review. All that this shows is that DAS have a fully structured system. Mr Bellamy believes that DAS's administrative costs have increased significantly due to the large number of challenges to premium recovery made by defendants. The company has employed a senior controller in order to handle these issues. In turn, he regularly has to instruct costs draftsmen to deal with them on detailed assessment.
57.   Finally Mr Bellamy said that the company expects to write 7,000 to 8,000 non-motor policies every year.
   6. The evidence before Master Hurst: The Respondents' witnesses
58.   Mr Boobier (see para 14 above) produced a schedule of recent cases in which ATE insurance had been used by claimants in connection with claims brought against his clients and other local authorities. For the reasons Master Hurst set out in his report he was unable to find in this evidence anything which would assist the court, and it is unnecessary to say anything more about it.
59.   Mr Hudman (see para 14 above) had conducted some market research within the ATE insurance market in about December 2004 to determine, if possible, the extent to which DAS 80e premiums measured up with premiums offered by competing insurers. For the reasons he gave in his report, Master Hurst was able to derive little value from this evidence save to say that there might possibly be products available on the market which were cheaper than the DAS 80e policy if a case went to trial, but that the situation might well be reversed if a case settled early.
   7. The Law Society's submissions to Master Hurst
60.   Master Hurst also recorded some submissions made by Mr Cooksley QC, who attended the hearing on 3rd July on behalf of the Law Society. He made the point that there were two approaches to ATE insurance. Some solicitors insured cases from the outset, an approach approved by this court in Callery v Gray [2001] EWCA Civ 1117 at [100]; [2001] 1 WLR 1112. This court, however, did not disapprove of those who delayed taking out ATE policies, since this could result in a saving. If it is necessary to take out a policy at a later stage, the premium is likely to be significantly higher than that required for a policy taken out earlier. In this case Mr Cater, as was his usual practice, waited until January 2004 before taking out the 80e policy, i.e. some 18 months after the date of the CFA. This method of insurance would inevitably be more expensive in the individual case, but when this cost is measured against other cases, which are all insured from the outset, although the individual premiums will be cheaper, the overall cost is likely to be greater. Mr Cater did not insure unless he felt he had to.
   8. The experiences of Temple and Keystone
61.   Mr Drabble QC appeared for Temple Legal Protection Ltd ("Temple" and Keystone Legal Benefits Ltd ("Keystone"). These two companies gave the court details of their experiences as ATE providers and their practice in relation to stepped premiums.
62.   Temple was formed as an underwriting agency in 1997, and they provide underwriting and claims management functions for insurance carriers in the UK. Underwriting and claims management services are carried out under a binding authority. Unlike many of their competitors, Temple decides on premium rates and terms and conditions for all risks. They have established themselves as a market leader since 2000, and they are among the top five ATE providers for personal injury claims.  They have been involved in many of the leading cases concerned with CFA and ATE funding.
63.   The company has insured over 30,000 claims (both personal injury and non--personal injury) using a stepped premium model. Of these claims, about 10,000 are still on risk. These include certificates of insurance written under delegated authority schemes with firms of solicitors as well as certificates based on a one -- off quotation.
64.   The current rates for their stepped premiums (exclusive of IPT at 5%) for delegated authority/block rated personal injury claims are shown in the table below. The figures are not cumulative.
    
  
  Premium (a) -- prior to issue of proceedings Premium (b) -- after issue of proceedings but more than 45 days before trial Premium (c) -- case settled 45 days or less before trial or at trial itself.
Fast track road traffic accident £350 £750 £1800
Fast track Accident at work £750 £1200 £2,700
Fast track Public liability £850 £1,500 £ 2,700
Multi-track Road Traffic accident £650 £1,100 £3,500
Multi-track accident at work £980 £1,900 £4,200
Multi-track Public Liability £1,150 £2,800 £5,500
    
65.   Keystone was incorporated in 1988. At first they administered uninsured loss recovery schemes for insurance brokers and insurance departments of national companies, but they then evolved to become an established provider of ATE insurance. They launched their own ATE scheme several months before the Law Society's "Accident Line Protect" product.
66.   As binder holders, Keystone underwrote policies for leading Lloyds syndicates and a major European Composite Insurer. Given the dramatic shrinkage in ATE capacity in the UK, Keystone took the decision to form their own insurer, and Bastion Insurance Co Ltd was duly authorised as an insurer on 14th December 2005.
67.   Keystone has issued about 15,000 stepped premium policies in their various forms. They generally underwrite all categories of personal injury litigation, principally specialising in Road Traffic Act ("RTA"), Employers' Liability ("EL") and Public Liability ("PL") matters. They have a limited exposure to clinical negligence and product liability cases. Keystone will also provide cover to insure appropriate cases in the Court of Appeal. They have developed a significant market position in underwriting cases that are perceived as difficult, or that have been rejected by other ATE providers. For many policyholders Keystone are effectively the final option. Most of the policies they issue are individually underwritten (rather than block rated) and Keystone have a high degree of technical expertise in its underwriting staff.
68.   In Keystone's stepped premium model Stage 1 covers the period between the inception of the policy and service of the defence (or judgment in default). The premium effectively covers the disbursement exposure on all cases (and own costs exposure on a Both Sides policy), and for the litigated cases the court fees and the initial stages of adverse costs. For CFA cases (not being cases in which both sides' costs are covered), although all cases are individually underwritten, the Stage 1 premium is generally within the following ranges:-
  
   RTA   £300-500
    EL/PL   £550-£750
    Industrial Disease   £750-£950
69.   Stage 2 covers the period between service of the defence and 28 days before Trial. The amount of the premium will, to a great extent, be dependent upon the nature of the defence and the issues in the case. These premiums are individually underwritten in each case, but in most cases the additional premium is likely to be between £1,000 and £1,500.
70.   Stage 3 covers the period from 28 days up to and including the trial itself. The additional premium will be variable and will be significantly dependent upon all the issues and risks in the particular case, the costs at risk, and the prospect of success. In all cases the Stage 3 premium is rebated by 75% if the matter settles prior to the commencement of the hearing.
71.   Temple believes that the reason why Mr Bellamy's product is making a loss may be because their Stage 1 and Stage 2 premiums are higher than DAS's. In contrast to Mr Bellamy's experience, Temple's experience has been that 20-30% of the cases they insure are abandoned or discontinued (excluding those that fail at trial or where a Part 36 offer is not bettered), and for Keystone the equivalent figure is 40%.
72.   With these caveats, Temple and Keystone broadly support Mr Bellamy's figures. They stress that some policies they issued as long ago as 2000 are still at risk, and they assume that this will be the experience of other ATE providers (including DAS). It follows that any data given now will not reflect the true picture as to how an account will run off. All the risks that are still live exist in cases progressing to trial. The risks in these cases are likely to be at or close to 50%, since the defendant obviously believes that it has a realistic prospect of success.
73.   Temple and Keystone take the view that 25% (of net written premiums) is a minimum allowance for overheads. A more realistic allowance would be closer to 50%. The cost of dealing with premium challenges and dealing with reductions in premiums is an important factor here. They consider that Mr Bellamy's estimates of defendants' costs are broadly consistent with their experience.
   9. The Law Society's evidence
74.   The Law Society adduced evidence in the form of a statement by Helena Yearwood, who is the Funding Manager for Russell Jones & Walker. In that capacity she is responsible for the administration of ATE cover in their personal injury department on a national basis. Her responsibilities extend to dealing with insurance brokers, reviewing the market, and presenting to the Board and other senior members of the firm information that is relevant to the choice of appropriate ATE insurance. The firm has been insuring cases since 2000, but in high volume since 2003. This increase in cover followed an extensive review of the market when the various products and schemes on offer were all considered. The firm has insured all types and values of case (including those referred to it by trade unions acting under a Collective Conditional Fee Agreement ("CCFA")). They review the market periodically to ensure that their clients are offered the best combination of cover for a reasonable premium.
75.   She says that there are a relatively small number of suitable products available on the open market that would meet a prospective firm's "wish list". This is because UK-based insurers have little interest in entering the market (because of constant challenges to premium levels and the huge losses experienced by other insurers in the past). The information available in publications like Litigation Funding or on the Internet is fairly limited. Detailed information on products, however, can be obtained through independent brokers. They can access all the insurers on the market and try to find schemes which match the firm's requirements. They can also provide ongoing support to solicitors by negotiating more flexible terms with insurers. This service is more likely to be provided in high volume schemes.
76.   She thought that many firms of solicitors would encounter problems over lack of choice (if they limited their search to UK-based insurers), their obligations under a proposed scheme (because insurers demand exclusivity), restrictions on cover, rigorous reporting controls, a capped limit of indemnity, the absence of any facility to unwind premiums and cancel the policy "when the file is turned down", the fees payable on application (in relation to some products), and a general lack of flexibility (for instance, in negotiating individual exceptions in difficult cases). Small firms that lack negotiating power (because they have less work to offer) may find their choice particularly limited.
77.   In her view an ideal product would include full delegated authority with minimal reporting and administration requirements; deferred premiums; guaranteed recoverable premiums; maximum flexibility; no requirement for the client to sign a CCA agreement to cover the premium; prompt payments in respect of valid claims; and added value to the firm's business (such as links with work providers and advice on costs recovery).
78.   She thought that the idea that solicitors should shop around for the best premium in every case would involve an extremely protracted process. When cases were presented individually, insurers would be reluctant to entertain any negotiation of their premium and their policy terms. The offer of a sizable book of business (where premiums can be calculated on the basis of the firm's previous claims record and insurers' expected profit under the scheme), is a much more satisfactory way of proceeding. If there were numerous quotes that would never be taken up, ATE providers would have to factor the cost of providing them into the cost of their policies.
79.   She said that one of the main difficulties solicitors faced was that they require flexibility in a product so that they can run their cases without the fear that insurers will not support their position in the later stages of the case. They wish to offer their clients full access to justice: for example, a policy that would provide full protection for their client in the event that they failed to better a Part 36 payment that was rejected on reasonable grounds. This type of cover did not come cheaply, since the potential for an adverse costs order was obviously increased. For some firms litigating without insurance could be commercially disastrous.
80.   A general review of the market might reveal what appeared to be reasonable premiums, but a cheap premium did not necessarily mean a better product. Those policies often did not provide for matters on a solicitor's standard "wish list". They might require the payment of the premium immediately, or the client might have to sign a CCA agreement to finance the cost of the premium, and interest would be deducted from the client's damages to finance the interest payments on the loan. A more expensive product might offer additional benefits such as reasonable practice-based funding packages; ongoing case management service through the life of the claim; support for premium challenges; and cover for Part 8 "costs only" proceedings.
81.   She said that there was a wide range of premium structures available on the market, including the stepped or staged premiums with which we are concerned on this appeal. The stepped pricing of the premium could be used as a negotiating tool to encourage early settlement, but other insurers did not use stepped premiums, and charged one rate to cover all risks.
82.   All in all, she said that in practice it is difficult for solicitors to select a product based on any one factor, such as lower premiums. Solicitors can and will choose a product on the basis of what is available to them in the market, the overall suitability of the product, and a scheme or policy that meets their firm's business requirements.
   10. Allianz Cornhill's experience
83.   The court received written submissions from Allianz Cornhill Insurance plc ("Allianz"). They used a model for ATE insurance for fast track non-RTA claims whereby a single premium was set at the outset. They said that both their model and a model based on a variable premium worked on the insurance policy of pooling risk. Both models constituted valid methods of assessing and qualifying risk. In essence the underwriter should expect the same overall outcomes from each, since the average premium in the second model was probably close to the single premium in the first. The second model would be undermined if the premium for the post-issue and trial stages were reduced on assessment, because the overall average would be affected. They, too, considered that it was essential to look behind the details given in a publication like Litigation Funding to establish the experience and long term financial stability of the particular insurer, the risk assessment and premium model being used, and the cover being provided. They added:
  
   "Only then can any true judgment be made. With respect, it is questionable whether a judge can make an adequate assessment given the complexities.
   The case serves to highlight the difficulty faced by ATE insurers and helps to explain why in actuality there are few underwriters of [ATE] insurance. The uncertainty of securing a recovery of the full premium some years down the line is very difficult to actuarially model. The judiciary applying hindsight underwriting is likely to arrive at a very different view of the suitability of a given premium. These decisions can drastically alter the actuarial model and make the difference between underwriting at an acceptable profit and making a loss. Additionally this uncertainty increases the amount of capital an underwriter needs from its shareholder (the more risk equates to a greater need for capital). This leads to margins being harder to achieve, risk appetite diminishes, and underwriters become more selective."
11. The evidence from LAMP

84.   The court also received valuable evidence from Mr Strange, who is the Underwriting Director at LAMP Services Ltd and LAMP Group Ltd. Although LAMP started trading in April 2005, Mr Strange had had four years' experience in the ATE market before joining LAMP, whose ATE insurance is underwritten by a Gibraltar-based company. LAMP has provided ATE policies in this country for over 7.000 claimants. The vast majority of these policies relate to personal injury cases, and his evidence focused on this class of case.
85.   He, too described a market in which a number of different products are now on offer. LAMP offer four premium structures for ATE insurance: a single rate for all cases of the same broad classification; a two stage/rebated premium; a three-stage premium; and a periodic insurance premium, whereby the premium increases by a predetermined rate at predetermined dates or intervals. LAMP do not insure a shortfall in costs recovery, and only the single or two-staged premium models result in an upfront payment of the premium. In paras 20-29 of his statement Mr Strange describes all the factors that contribute to the rating of risk on an ATE policy. The starting point is to assess the estimated maximum loss which may arise for any particular claimant, and Mr Strange makes it his business to form an estimate of what the costs are likely to be in relation to particular types of case, such as fast track RTA claims or fast track EL claims.
86.   He also described all the other costs associated with his business. These costs, together with a profit element, have to be taken into account when establishing the company's necessary premium income. Risk analysis may also have to take into account the different success rates of the different solicitors LAMP insures; the attitude of defendants to a given solicitor's claims and their conduct of these cases; the incidence of Part 36 offers; the different levels of assessment of damages in different parts of the country; and so on. A further key component in risk analysis is the community of risk, and the quality of that community. For instance, if a solicitor insures all his claimants, there will be a higher success rate than if he only insures those claimants who are more likely to lose their case. LAMP currently operates over 40 different schemes, all of which are slightly different because of these different factors.
87.   Mr Strange provided a detailed exposition of his company's approach to staged premiums. LAMP sets its third stage premium at the outset. His experience has been the same as Mr Bellamy's inasmuch as two thirds of claims probably settle pre-proceedings, and a little less than one third prior to trial. Where premium is assessed on a bespoke basis it will be the most expensive methodology since it requires an individual assessment of a particular claim and the risk may -- and is highly likely to -- operate outside a community which is able to share its risk (as happened in Mr Cater's case, where he selected the cases he wished to insure, as opposed to insuring all his personal injury cases, regardless of how likely they were to succeed).
88.   Mr Strange considered that the insurers' premium must be based on the costs risk he faces, not the damages recoverable. Because it is a matter of public policy that ATE insurance shall be available to fund personal injury claims, ATE insurance must be self-financing, and in order to be self-financing the premium must be reasonable. If it is reasonable, the courts should consider it to be proportionate. It has long been known that litigation costs frequently exceed the sums claimed, and Mr Strange considered that there was no reason why ATE insurers should be penalised on an assessment as compared with the legal costs awarded to a party. He added that, in his view, a staged premium, or a periodic insurance premium, offer advantages in funding access to justice in many circumstances by spreading the cost of the insurance. The fact that they would be more expensive in the later stages of a case than a single premium did not make them unreasonable or disproportionate. The methodology of staged premiums recognised the overriding objective of encouraging resolution.
   12. Brit's submissions
89.   The court also received written submissions from Brit Insurance Co Ltd ("Brit"). Their insurance business ranges more widely than ATE insurance, but they do provide ATE insurance to claimants in personal injury cases. In this context they underwrite the Accident Line Protect ("ALP") product issued by Abbey and the Solus product provided by Amicus Legal. Both their ATE policies support CFA retainers.
90.   Brit underwrites staged premiums through their ALP product. In their view, staged or discounted premiums offer a defendant's insurer the opportunity to settle litigation earlier, thereby giving him the benefit of a reduced ATE insurance liability if he loses. Brit observed that no two schemes within the ATE market were identical.
91.   Although the ATE market has developed over the last six years, Brit observed that the number of UK-based underwriters has shrunk. Since the demise of Claims Direct and The Accident Group, however, the market has started to mature, and results were starting to mature. It is still young, however (in that its data is not mature), and care had to be taken to ensure that any decision in a case like this did not adversely affect the sustainability of the market with consequent knock-on effects for access to justice. If "tariffs" setting guidance for reasonable premiums were introduced, this might mean that ATE providers would be keen to insure only the very low risk cases. This would defeat access to justice and deny such access to many people who were entitled to a legal remedy through the courts.
   13. Abbey's observations
92.   Abbey is a specialised Legal Expenses insurance underwriting agency. It was responsible, with the Law Society, for developing the Society's Accident Line Protect ATE policy in 1995, and it has been involved in the ATE market ever since then. In October 2001 they introduced their first staged premium policy, and they have described to the court how they have expanded their range of staged premium ATE products over the years since then. This strategy has enabled them to introduce significant reductions in their initial premiums, so that their premiums in RTA cases concluded prior to allocation, for instance, are significantly below the level of the premiums approved by this court in Callery v Gray. Their comments echoed the comments the court received from other providers. They said, among other things:
  
   "A … weakness with regard to recoverability has arisen because of the practice of some courts of assessing proportionality of premiums to the value of the damages in a case. Whilst this concept is clearly familiar in the assessment of costs, it disregards the whole basis of insurance which calculates the risk in terms of both the chances of loss and the cost of loss. The damages play no part in that calculation."
   14. The Liability Insurers' Group's submission
93.   At Lord Justice Brooke's invitation the Liability Insurers' Group (a group of major liability insurance companies that have vast experience as defendants in the personal injuries market) filed submissions with the court, although they preferred not to intervene with oral submissions once the appeal was limited to the particular issues that arose as between the present parties. They observed that there were clearly difficult issues in the ATE market context, such as the level of premium, risk, proportionality, competition, and payment to intermediaries. These were contentious issues and they continued to cause uncertainty in the area as a whole. However, they differed from case to case and from provider to provider. The group observed that although they were generally most exposed to premium recovery, there would always be a residue of cases in which ATE premiums were sought from individuals, businesses and other organisations (such as the present defendants) who might for a variety of reasons be uninsured and thus unable to redistribute the cost of the claimant's premium payment.
94.   Hitherto they had had no objection in principle to the concept of stepped insurance premiums. However, the existence at one and the same time of both stepped premium products and block rated premium products introduced a tension into an immature market which, they accepted, was difficult to resolve. If straightforward cases were habitually allocated to block premium products, and only the more complex cases were dealt with by stepped premiums, there would be no saving to anyone. The block premiums would tend to be higher than any of the stepped premiums, but almost by definition it was unlikely that the stepped premiums would ever be disposed of at the first step. Everyone accepted that the fact that a particular product should encompass both good and bad risks was essential to the need to spread insurance risk. Otherwise, so far as the premium on a particular product, was concerned, the many could not pay for the few.
95.   Liability insurers were concerned about the question of payment to intermediaries (an issue with which we are not directly concerned on this appeal). They also said that difficulties rose because there was no real commercial pressure in the present arrangements. Most models now deferred the payment of premium by the claimant in some way or other, so that he did not feel any pressure in terms of price, and his solicitors anticipated that whatever the outcome their client would not in fact be paying the premium. There was therefore, in the group's view, no pressure to reduce premium rates at the point of sale, and when a court came to consider whether or not a premium was properly paid, there was effectively no clear competitive material against which its reasonableness could be tested. Issues related to the selection of a particular product therefore tended to be a simple question of preference and relationship as between a particular solicitor and a particular provider, and this was unlikely to maintain any pressure on premium levels.
   15. Discussion
96.   This is the first occasion on which the reasonableness of an ATE premium has been considered authoritatively by this court since its landmark decisions in Callery v Gray (No 1) [2001] EWCA Civ 1117; [2001] 1 WLR 2113 and Callery v Gray (No 2) [2001] EWCA Civ 1246;, [2001] 1 WLR 2112. In those two judgments the court was principally concerned to decide whether an ATE premium was reasonable when ATE cover was taken out by a claimant before the first letter of claim was made, and when the defendant's reaction to the claim was still unknown. On the present appeal, three main issues arise for decision:
  
 i)   What is the proper approach to proportionality in a small personal injury case where the ATE premium may appear large in comparison with the amount of damages reasonably claimed?
 ii)   What is the proper approach to evidence of reasonableness of the choice and of the amount of the ATE premium in such cases?
 iii)   Are both staged (or stepped) premiums and single premiums for ATE insurance legitimate for the purposes of the recoverability of an ATE premium by a successful claimant, and is it reasonable that such premiums should be wholly or partially block-rated?
97.   During the exchanges that preceded the hearing of this appeal it was suggested that a fourth issue might arise for decision, namely whether it was permissible for a recoverable premium to include an element for the cost of self-insuring the premium in the event of failure, or only partial success. It was rightly conceded, however, that an affirmative answer to this question had already been given by this court in Callery v Gray (No 2), so that it will only be necessary to deal with this issue quite briefly in the present judgment (see para 118 below).
98.   Although an ATE insurance market existed in a fledgling state before s 29 of the Access to Justice Act 1999 came into force, that section transformed things in the sense that it was now possible for a successful claimant to recover the ATE premium from the defendant instead of having to fund it out of the damages he or she received. Section 29 provided that:
  
   "Where in any proceedings a costs order is made in favour of any party who has taken out an insurance policy against the risk of incurring a liability in those proceedings, the costs payable to him may, subject in the case of court proceedings to rules of court, include costs in respect of the premium of the policy."
99.   The following provisions of the Civil Procedure Rules are relevant in the present case:
  
 43.2    
 (1)  In Parts 44 to 48, unless the context otherwise requires --
  
 (a)   'costs' includes … any additional liability incurred under a funding arrangement …
 (k)   'funding arrangement' means an arrangement where a person has --
  
 (ii)   taken out an insurance policy to which section 29 of the Access to Justice Act 1999 (recovery of insurance premiums by way of costs) applies; or
 (m)   'insurance premium' means a sum of money paid or payable for insurance against the risk of incurring a costs liability in the proceedings, taken out after the event that is the subject matter of the claim;
 (o)   'additional liability' means … the insurance premium ...
 44.3B  
  
 (1)   A party may not recover as an additional liability --

  
 (c)   any additional liability for any period in the proceedings during which he failed to provide information about a funding arrangement in accordance with a rule, practice direction or court order;
 44.4  
  
 (1)   Where the court is to assess the amount of costs (whether by summary or detailed assessment) it will assess those costs --

  
 (a)   on the standard basis; or
 (b)   on the indemnity basis,
   but the court will not in either case allow costs which have been unreasonably incurred or are unreasonable in amount.
 (2)   Where the amount of costs is to be assessed on the standard basis, the court will --
  
 (a)   only allow costs which are proportionate to the matters in issue; and
 (b)   resolve any doubt which it may have as to whether costs were reasonably incurred or reasonable and proportionate in amount in favour of the paying party.
 44.5  
  
 (1)   The court is to have regard to all the circumstances in deciding whether costs were --
  
 (a)   if it is assessing costs on the standard basis --
  
 (i)   proportionately and reasonably incurred; or
 (ii)   were proportionate and reasonable in amount, or

 (3)   The court must also have regard to --
  
 (b)   the amount or value of any money or property involved; …
100.   Section 11 of the Costs Practice Directionpdp-43 (which is entitled "Factors to be taken into account in deciding the amount of costs") contains further assistance:
  
 11.   In applying the test of proportionality the court will have regard to rule 1.1(2)(c). The relationship between the total of the costs incurred and the financial value of the claim may not be a reliable guide. A fixed percentage cannot be applied in all cases to the value of the claim in order to ascertain whether or not the costs are proportionate.
 11.2   In any proceedings there will be costs which will inevitably be incurred and which are necessary for the successful conduct of the case. Solicitors are not required to conduct litigation at rates which are uneconomic. Thus in a modest claim the proportion of costs is likely to be higher than in a large claim, and may even equal or possibly exceed the amount in dispute.
 11.3   Where a trial takes place, the time taken by the court in dealing with a particular issue may not be an accurate guide to the amount of time properly spent by the legal or other representatives in preparation for the trial of that issue.
 11.5   In deciding whether the costs claimed are reasonable and (on a standard basis assessment) proportionate, the court will consider the amount of any additional liability separately from the base costs.
 11.6   In deciding whether the base costs are reasonable and (if relevant) proportionate the court will consider the factors set out in rule 44.5.
 11.7   Subject to paragraph 17.8(2), when the court is considering the factors to be taken into account in assessing an additional liability, it will have regard to the facts and circumstances as they reasonably appeared to the solicitor or counsel when the funding arrangement was entered into and at the time of any variation of the arrangement.
 11.10   In deciding whether the cost of insurance cover is reasonable, relevant factors to be taken into account include:
  
 (1)    where the insurance cover is not purchased in support of a conditional fee agreement with a success fee, how its cost compares with the likely cost of funding the case with a conditional fee agreement with a success fee and supporting insurance cover;
 (2)   the level and extent of the cover provided;
 (3)   the availability of any pre-existing insurance cover;
 (4)   whether any part of the premium would be rebated in the event of early settlement;
 (5)   the amount of commission payable to the receiving party or his legal representatives or other agents.
101.   Because of the wealth of information with which the court has now been provided, it is possible to answer the questions posed by this appeal quite shortly.
   16. Proportionality
102.   It is to be noted that proportionality's role in the conduct of civil litigation travels much further than the arena of costs issues. In addition to what is provided in CPR 44.4(2) and CPR 44.5 (1) and (3) (see above), CPR Part 1.1 sets out the "overriding objective", to deal with cases justly. Part 1.1(2) provides in part:
  
"(2)   Dealing with a case justly includes, so far as is practicable --
  
   …
 (c)   dealing with the case in ways which are proportionate --
  
 (i)   to the amount of money involved;
 (ii)   to the importance of the case;
 (iii)   to the complexity of the issues; and
 (iv)   to the financial position of each party."
103.   The idea that justice requires not only impartial judges and procedures that are fair as between the parties, but also a use of resources that is proportionate to what is at stake, is an important -- perhaps the most important -- dimension of the philosophy which underpins the civil procedure reforms carried into effect by the CPR. It is plainly expressed in Part 1.2 which we have just set out. Without doubt it applies in the costs context. In the ordinary way, the value of the claim is a principal element in the judgment of what is proportionate for the purpose of assessing costs on the standard basis under Part 44 (see CPR 44.5(3)(b) above). But in this case, on the approach taken by Mr Bartlett QC, who appeared for the appellants, and supported by Mr Drabble and Mr Cooksley, the fact that the damages were agreed at only £3,000 does not touch -- at least not directly -- the issue whether the costs claimed are proportionate. How can this be right?
104.   This court has had to consider proportionality in the costs context on a number of occasions. Giving the judgment of the court in Lownds v Home Office [2002] EWCA Civ 365;, [2002] 1 WLR 2450, Lord Woolf MR said this , at para [31]:
  
 "31.   [W]hat is required is a two-stage approach. There has to be a global approach and an item by item approach. The global approach will indicate whether the total sum claimed is or appears to be disproportionate having particular regard to the considerations which Rule 44.5(3) states are relevant. If the costs as a whole are not disproportionate according to that test then all that is normally required is that each item should have been reasonably incurred and the cost for that item should be reasonable. If on the other hand the costs as a whole appear disproportionate then the court will want to be satisfied that the work in relation to each item was necessary and, if necessary, that the cost of the item is reasonable. If, because of lack of planning or due to other causes, the global costs are disproportionately high, then the requirement that the costs should be proportionate means that no more should be payable than would have been payable if the litigation had been conducted in a proportionate manner. This is turn means that reasonable costs will only be recovered for the items which were necessary if the litigation had been conducted in a proportionate manner."
    In Simms v The Law Society [2005] EWCA Civ 849 at [28] Carnwath LJ stated that "the 'reasonableness' of the costs… in practice may overlap with issues of proportionality". With respect we think this observation needs to be treated with some care: reasonableness and proportionality are conceptually distinct. However we apprehend that Carnwath LJ was saying no more than that reasonableness may be a necessary condition of proportionality, and we are sure that is so.
105.   In this case it might be thought that all the considerations urged on the court by Mr Bartlett which favour the course taken by Mr Cater, the appellant's solicitor, might go to demonstrate the reasonableness of his bill of costs -- specifically, the ATE insurance staged premium -- but not its proportionality: precisely because they have nothing to do with the quantum of the claim. But we do not think that is right. If the court concludes that it was necessary to incur the staged premium, then as this court's judgment in Lownds shows, it should be adjudged a proportionate expense. Necessity here is, we think, not some absolute litmus test. It may be demonstrated by the application of strategic considerations which travel beyond the dictates of the particular case. Thus it may include, as we are persuaded it does, the unavoidable characteristics of the market in insurance of this kind. It does so because this very market is integral to the means of providing access to justice in civil disputes in what may be called the post-legal aid world.
106.   It is important to recognise that this conclusion runs with, not across, the grain of the procedural reforms expressed in the CPR. The very recognition that justice requires a use of resources that is proportionate to what is at stake implies the rightness of a strategic approach. There can be no touchstone of a proportionate use of resources so understood, without an eye to the context in which any such resources are expended. Once it is concluded that the ATE staged premium here was necessarily incurred, principle and pragmatism together compel the conclusion that it was a proportionate expense. We turn therefore to the question whether the ATE staged premium was necessarily incurred.
   17. The legitimacy of a staged premium model
107.   Nobody has suggested that a staged premium model is in itself an illegitimate way of rating the risk. Although this court has never previously had to address this issue, there is in principle no difference between a two-staged success fee (whose merits this court has consistently endorsed) and a staged ATE premium. The financial risk to which the ATE provider is exposed inevitably rises as a case proceeds towards trial. While defendants may be liable to pay a higher premium if they take a case to trial and lose, the situation is no different from that facing them in relation to their liability to pay a higher success fee when claims are resolved against them 14 days or less before the date fixed for the commencement of a trial in the cases covered by the new arrangements for fixed recoverable success fees in CPR Part 45. Exposure to this greater liability requires a defendant to think very seriously about the merits of his position before a trial takes place. This obligation, too, runs with the grain of the philosophy of the CPR.
   18. The size of the premium in the present case
108.   The evidence before this court is entirely different from the evidence before the deputy circuit judge, since DAS had at that time made no effort to place the information contained in Litigation Funding in its proper context. If more evidence had been available to this very experienced judge, he would no doubt have reminded himself that the fact that the ATE premium was large compared with the agreed damages of £3,000 did not necessarily mean that it was disproportionate. Under CPR 44.5(1) a court must take into account "all the circumstances". These include the financial risk faced by the insurer.
109.   In the present case the total EML of £6,500 compared favourably with the actual outturn of £6,875 revealed to the deputy district judge. Given the defendants' determination to go to trial, the handler's assessment of the risk at no greater than 51% was not unreasonable, particularly as DAS's experience was to the effect that more slipping and tripping cases taken to trial were lost than won. And on these figures it is impossible to say that a total premium of £4,680 was unreasonable. Put quite simply, if in two cases insurers face a 50% risk of having to pay out £6,500 on one of them, it is reasonable for them to charge a premium of £6,500 (not allowing for overheads or profit) on each. On the one they win, they will be able to get their premium paid by the defendant, and this will recompense for them having to pay out £6,500 on the one they lose.
110.   On these figures it is quite impossible to say that a total premium of £4,860 was unreasonable. Indeed, it was not fixed at a sufficiently high level to reflect the risk and to provide a contribution to the insurer's reasonable overheads and profit.
111.   On the evidence now before the court the judge's reliance on Litigation Funding as a source of dependable evidence was not well founded. We would endorse what Master Hurst said about this material in his judgment in Re RSA Pursuit Test Cases. It is not legitimate to compare the total premium payable at the third stage of a three-stage premium model with the single premium under a single premium model that is payable throughout the progress of a claim to trial. As the evidence in this case shows, DAS's first stage premium was significantly lower than Temple's single premium. This discrepancy would benefit defendants in 63% of the cases covered by a DAS 80e policy. Many might think that this demonstrated a preferable approach to the rating of risk.
112.   Master Hurst's report showed that the defendants had been unable to identify any cheaper alternative provider for Mr Cater, given the reasonable way he approached his client's insurance needs, and Mr Bellamy's evidence showed that DAS's average premium compared favourably with the average premium charged by their competitors.
   19. The Respondents' submissions
113.   Mr Williams, who appeared for the claimant, boldly asserted that in tying himself to DAS in the way he did, Mr Cater was in breach of Section 4(1) of the Solicitors' Introduction and Referral Code 1990, which reads:
  
 "4(1)   If a solicitor recommends that a client use a particular firm, agency or business, the solicitor must do so in good faith, judging what is in the client's best interest. A solicitor should not enter into any agreement or association which would restrict the solicitor's freedom to recommend any particular firm, agency or business."
114.   This was a surprising submission, given that the success of ATE insurance has been dependent from the outset on arrangements like these. They are designed to prevent "cherry-picking" and to ensure that very many low risk cases are available as a counterweight to the few high risk cases. Mr Cooksley immediately disavowed this proposition on behalf of the Law Society. He told us that solicitors had been advised by the Law Society that they would not act in breach of the Code if they made reasonable contractual arrangements of this kind with ATE insurers. The use of the milder word "should", as opposed to the more prescriptive word "must", shows that this approach by the Law Society to the construction of this part of its own Code was not an unreasonable one.
115.   We mean no disrespect to Mr Williams's spirited submissions if we do not respond to them in detail. He was inevitably advancing the cause of clients who found themselves liable to pay a high third stage premium. Mr Cater's approach towards his clients' insurance needs means, however, that many other defendants would not have to pay any premium at all to his clients (because he did not insure their cases), or only a first stage premium that was lower than the single premium charged by other ATE insurers. We find it impossible to say that the approach Mr Cater adopted, which we have described in this judgment, was an unreasonable one.
   20. The procedure to be followed in future
116.   During the course of argument it was accepted on all sides that a party who has an ATE insurance policy incorporating two or more staged premiums should inform its opponent that the policy is staged, and should set out accurately the trigger moments at which the second or later stages will be reached. This obligation should be undertaken in addition to the obligations set out in CPR 44.15(1) and in paras 19.1(1) and 19.4 of the Costs Practice Directionpdp-43. If this is done, the opponent has been given fair notice of the staging, and unless there are features of the case that are out of the ordinary, his liability to pay at the second or third stage a higher premium than he would have had to pay if the claim had been settled at the first stage should not prove to be a contentious issue.
   21. Evidence justifying the ATE premium claimed
117.   If an issue arises about the size of a second or third stage premium, it will ordinarily be sufficient for a claimant's solicitor to write a brief note for the purposes of the costs assessment explaining how he came to choose the particular ATE product for his client, and the basis on which the premium is rated -- whether block rated or individually rated. District judges and costs judges do not, as Lord Hoffmann observed in Callery v Gray (Nos 1 & 2) [2002] UKHL 28 at [44]; [2002] 1 WLR 2000, have the expertise to judge the reasonableness of a premium except in very broad brush terms, and the viability of the ATE market will be imperilled if they regard themselves (without the assistance of expert evidence) as better qualified than the underwriter to rate the financial risk the insurer faces. Although the claimant very often does not have to pay the premium himself, this does not mean that there are no competitive or other pressures at all in the market. As the evidence before this court shows, it is not in an insurer's interest to fix a premium at a level which will attract frequent challenges.
   22. Self-insurance of the premium in the event of failure or only partial success
118.   Finally, we would confirm that as the law now stands it is permissible and reasonable for the premium itself to be insured by the policy. This was decided by this court in Callery v Gray (No 2) [2001] EWCA Civ 1246 at [63]; [2001] 1 WLR 2142.
   23. An issue we have not decided
119.   We have noted in this judgment that an ATE premium may "buy" certain benefits that are not directly linked with the financial risk faced by the client who takes out an ATE policy. It was common ground that this was not an issue we were equipped to handle on the present appeal. Whether that element of the premium, if identifiable, should properly be chargeable to the unsuccessful defendant must await the outcome of a wider inquiry than has been possible to carry out within the four corners of the present appeal.
   24. Conclusion
120.   For all these reasons, we will allow the appeal, set aside the decision of the deputy circuit judge, and restore the decision of the district judge in relation to the recoverability of the ATE premium as claimed. We should make it clear, once again, that on this re-hearing we have received far more evidence than was before the deputy circuit judge on the first appeal.
121.   In accordance with the directions given by Lord Justice Brooke on 23rd May 2006 (see para 11 above), the deputy circuit judge's costs order will stand and the parties and the interveners will each bear their own costs in this court, save that if the Respondents have any submissions to make pursuant to the reservation in their favour in Lord Justice Brooke's order, they should make them by way of written representations to the court.
122.   We are adding to this judgment an Annex written by Lady Justice Smith. Although the two other members of the court agree with what she says, we felt that it would be far better that she expressed her anxieties in her own words, in view of her immense experience of the practice of personal injuries litigation in this country.
   Annex by Lady Justice Smith
123.   I wish to add a few words of my own because it seems to me that information provided to the court for the purpose of this appeal has thrown useful light on the operation of the funding arrangements for personal injury litigation brought into effect pursuant to the Access to Justice Act 1999. At least, it has shed some light on how the system is operating in respect of public liability (mainly slipping and tripping) cases.  The picture that emerges is, to my mind, rather worrying. 
124.   First, the figures produced by DAS show that, of the 5% of slipping and tripping cases which proceed to trial, about 70% fail. Legal advisers are being insufficiently robust in the advice they give at the late stage when a decision has to be taken whether to abandon the case or go on to trial. 
125.   In the days before CFAs and ATE insurance, a small proportion of clients would fund the litigation themselves; a substantial proportion would be supported by a trade union and another substantial proportion would be in receipt of legal aid.  Lawyers advising a private client had to take a realistic view of the risks of failure and an adverse order on costs. It would not often be sensible to proceed if the prospects of success were only just over 50%.  Usually one would look for a chance of success in the region of 60% to 70%.  It would hardly ever be sensible to risk paying £10,000 in costs for the sake of recovering £5,000 in damages. Trades unions were often prepared to take a greater risk on costs than a private client could sensibly accept.   When giving advice for the purpose of legal aid, solicitors and counsel were supposed to apply the same degree of rigour as they would apply when advising a private client.  After becoming a judge, I became aware that that rule of practice was not always honoured as it should have been. I was shocked at how many obviously 'difficult' or even weak cases were proceeding to trial with the support of legal aid and were being lost.  The provision of civil legal aid was, I believe, more expensive to the public purse than it should have been.  
126.   When the government of the day abolished legal aid for most personal injury actions and brought in the provisions of the Access to Justice Act 1999, my understanding was that it was intended that all claimants would have as good a means of access to the courts as a litigant who could afford to fund his claim from his own resources. This was to be achieved through CFAs and ATE.  By the use of uplifts on base costs, solicitors would take the rough with the smooth and make about the same level of profit as they would if they were acting for private clients.  The other costs of litigation, particularly the risk of paying the defendants' costs, were to be insured by ATE and, in the end, passed to liability insurers and through them to the general premium paying public. 
127.   However, I do not think it was the intention of Parliament that would-be claimants should be able to litigate weak cases without any risk whatsoever to themselves.  But it seems to me that this is what is happening. ATE premiums are set on the basis of a high expected failure rate at trial.  Even cases that are assessed at a prospect of success of only 51% receive ATE insurance. Thus the premiums have to be significantly higher than they would be if a more rigorous standard were applied.  Often no premium has to be paid upfront. If the case is lost the premium is rarely paid. That practice inevitably increases the premiums even further.  If the case is won, the premium is in principle recoverable from the liability insurer and, as this court has held in the instant case, if it was necessary for the claimant to take out ATE insurance and the solicitor has acted reasonably, the whole premium will be recovered. 
128.   Two things concern me about this situation.  One is that there is very little incentive for solicitors to look for the best value in ATE insurance.  One can understand the position of someone like Mr Cater whose primary concern is to protect his client from the kind of problem that he had experienced with his previous provider.  He can quite sensibly justify opting for a more expensive product.  His client will never have to pay the premium regardless of the outcome.   As the judgement of the court acknowledges, there is a pressure on insurers to keep their premiums at a reasonable level in order to avoid challenges such as has occurred in this case. However, the decision in this case may well have the effect of reducing that pressure.
129.   As the judgment of court observes, the proper use of a three stage premium should require a defendant to take a second, more rigorous, look at the merits of the case before it proceeds to trial. What concerns me is that the ATE system does not provide any incentive for the claimant's side to have a second and more rigorous look at the merits, often even after a Part 36 offer has been made and refused. It appears that the claimant can just carry on.  The only assessment that is made is as to the size of the premium; the lower the chances of success, the higher the premium. I have the impression that the insurer does not ask whether the claim should be stopped.  It seems to me that if, at the third stage, when much more information should be available than at the time of issue, solicitors and ATE insurers were to make a careful assessment of the prospects of success, and were to stop the weak ones, the proportion of cases which fail at trial would be much reduced. All sensible private clients would insist on such an assessment. At present, the insured claimant can notionally pay the high premium which reflects his poor chances of success, secure in the knowledge that, if he wins, the premium will be recovered and, if he loses, he can walk away unscathed. I find it hard to believe that Parliament intended that claimants should be in so much better a position than a private litigant.
130.   I recognise of course that the CFA itself should provide a strong incentive for the solicitor to examine the merits of the case before allowing it to proceed to trial. However, the 70% failure rate suggests that that incentive is not enough to produce real rigour.  How solicitors are managing if they lose 70%, I know not. The figures we have seen suggest that the ATE insurance market is not managing at all well, at least on cases that go to trial.   
131.   It seems to me that this lack of rigour is resulting in the system being more expensive (eventually to the premium paying public) than it should be.  I do not know how these problems could be rectified. However, I felt it right to draw attention to them.