Carillion Plc v KPMG LLP & Anor

The Claimant (“Carillion”) applies for pre-action disclosure pursuant to CPR 31.16. The application is made against KPMG LLP and KPMG Audit Plc (together, “KPMG”). KPMG were the auditors of Carillion and a large number of its subsidiaries (collectively “the Carillion Group”).

The basis of the intended action is that, as auditors, KPMG did not detect that the financial statements of Carillion and the consolidated financial statements of the Carillion Group failed to reflect their true position. The intended action arises from the well-publicised collapse of the Carillion Group. This has led to investigations by the Parliamentary Select Committee and the Financial Reporting Council. The Select

Carillion now asks the Court to make an order for pre-action disclosure in relation to these two aspects of the financial reporting of Carillion and its subsidiaries. Carillion contends that it has grounds to believe that KPMG’s auditing was negligent in relation to these matters. The documents are sought in relation to the audits for the years ended 31 December 2014-2016 only, and (as regards contract accounting) in relation to only 9 of the many construction contracts current in those years. These 9 contracts represent a subset of 58 contracts whose accounting had been reviewed by Carillion’s management between May and July 2017, prior to the Carillion Group’s collapse. Write-downs were then made by Carillion on at least 45 of the Carillion Group’s contracts, including the 9 in question. 16 of these contracts were specifically identified in the pre-application correspondence referred to below, but the application now concerns only 9. Some months after the application was issued, and as a result of evidence filed by KPMG, the application was significantly narrowed so that it now focuses exclusively on KPMG’s “eAudIT” electronic audit files described below.

Carillion contends that the disclosure is necessary because KPMG has refused to provide any of its working papers voluntarily, despite requests dating back to 1 October

KPMG resists the application on various grounds. It contends that the jurisdictional threshold in CPR 31.16 (3) (c) is not satisfied because Carillion cannot show on the balance of probabilities that standard disclosure would extend to the documents or classes of documents now sought; that the jurisdictional threshold in CPR 31.16 (3) (d) is not satisfied, because there is no real prospect that any of the matters there set out (disposing fairly of the anticipated proceedings, or assisting in settlement, or saving costs) applies in the present case; and that in any event discretionary considerations favour dismissing the application.

The application has generated a significant amount of correspondence and witness statement evidence, as well as cost. This includes a 38-page letter sent by Carillion’s solicitors in November 2019, and a 67-page witness statement in support of the application in February 2020. KPMG’s costs of resisting the application amount to in excess of £ 500,000. I do not have figures for Carillion’s costs, but it is certainly possible that they are in the same order of magnitude. The application has on any view given rise to “elaborate and expensive pre-action procedures” which are discouraged by paragraph B3.2 of the current edition of the Commercial Court Guide. However, Carillion contends that these have been necessitated by KPMG’s failure to comply with the Pre-Action Protocol for Professional Negligence (“the Protocol”), which is expressly referred to in that paragraph.

The relevant correspondence between the parties started in October 2019. It was relied upon by the parties in the course of argument, in particular because each side alleged that (in different respects) the other party had failed to comply with the Protocol. Carillion argued that KPMG’s failure to comply with that Protocol, in relation to its refusal to provide documents, was an unusual feature of the present case and was an important reason why pre-action disclosure should be ordered. By contrast, KPMG contends that there has been no such failure, and that Carillion has been focused on preaction disclosure rather than compliance with the Protocol. Carillion’s non-compliance with the Protocol is one reason why, on KPMG’s case, the present application is premature.

In considering these respective arguments as to breach, which are important to the resolution of the present application, it is important to pay regard to the “spirit” of the Protocol: see the valuable introduction to Pre-Action Conduct and Protocols in Section C1A of the White Book. Thus, as stated in paragraph C1A-005:

This is reflected in paragraph 13 of the Practice Direction on Pre-Action Conduct and Protocols:

The aims of the Protocol (paragraph 2.1) include enabling the parties to prospective claims to “(a) understand and properly identify the issues in dispute in the proposed claim and share information and relevant documents”, and “(b) make informed decisions as to whether and how to proceed”. Paragraph 3.1 says that the courts “will treat the standards set out in this protocol as the normal reasonable approach for parties to a professional negligence claim”. The scheme of the Protocol is as follows.

Paragraph 5 provides for a claimant to serve a “Preliminary Notice” which will set out a brief outline of the claimant’s grievance against the professional. Paragraph 6 provides for the sending of a “Letter of Claim” by the claimant, as soon as the claimant decides there are grounds for a claim against the professional. This Letter will normally be an open letter, and it should include the information set out in Paragraph 6.2. This includes:

Any reasonable requests which the claimant needs to make for documents relevant to the dispute which are held by the professional.

The allegations against the professional. What has been done wrong or not been done? What should the professional have done acting correctly?

An explanation of how the alleged error has caused the loss claimed. This should include details of what happened as a result of the claimant relying upon what the professional did wrong or omitted to do, and what might have happened if the professional had acted correctly.

An estimate of the financial loss suffered by the claimant and how it is calculated. Supporting documents should be identified, copied and enclosed. If details of the financial loss cannot be supplied, the claimant should explain why and should state when he will be in a position to provide the details. This information should be sent to the professional as soon as reasonably possible.

Confirmation whether or not an expert has been appointed. If so, providing the identity and discipline of the expert, together with the date upon which the expert was appointed.”

The Letter of Claim is to be acknowledged in writing within 21 days of receipt (Paragraph 7). If the professional considers that the Letter of Claim does not comply with section 6, then the professional should inform the claimant as soon as reasonably practicable why, and identify the further information which the professional reasonably required. (Paragraph 8.1). Paragraph 8.2 provides for a 3-month period for the professional to investigate and respond to the Letter of Claim by providing a Letter of Response. Paragraph 8.3 provides for the possibility of the professional requesting further time for the provision of the response. Paragraph 8.4 provides:

Paragraph 9.2 provides for the submission of the Letter of Response; it is to be an open letter and a reasoned answer to the claimant’s allegations. Paragraph 9.2.1 provides:

Section 10, headed Documents, provides:

Applications for pre-action disclosure in the Commercial Court are relatively rare, and the authorities to which I was referred contain no recent examples of successful applications. The 2018 edition of the White Book (paragraph C1A-010) said that pre-

The current (2017) edition of the Commercial Court Guide states:

In this section, I describe and comment on the relevant pre-action correspondence.

On 1 October 2019, Carillion’s solicitors (Quinn Emanuel Urquhart & Sullivan UK LLP or “QE”) wrote a 7 page letter to KPMG’s solicitors (Orrick, Herrington &

The letter went on to request documents in a number of areas. Paragraph 12, under the heading “Accounting for Contract Revenue”, sought documents concerning 16 contracts. This list included the 9 contracts in relation to which (together with goodwill) the present application is made. Paragraphs 13 and 14 sought documents concerning goodwill. Paragraphs 15 – 17 sought documents concerning the going concern status of the Carillion Group as at 31 December 2016 and 30 June 2017. Paragraph 18, under the heading “Other aspects of KPMG’s work”, asked for various documents, including (i) those relating to the accounting treatment of transactions between the Carillion Group and a company called Wipro Ltd., and (ii) the accounting treatment of certain “reverse factoring” arrangements entered into from 2012 onwards. The letter concluded by indicating Carillion’s intention to apply for the requested documents under CPR 31.16, unless KPMG was willing to provide the material requested voluntarily.

The letter contained no reference to the Protocol, and clearly shows that Carillion’s focus, at this point, was on the application that I am presently considering, albeit that it was then in a much more expansive form than the application ultimately issued and subsequently narrowed.

On 16 October 2019, Orrick responded in detail. One point raised was that Carillion had made no attempt to comply with the Protocol. The letter also contained a central argument that KPMG continues now to advance, namely that Carillion had sufficient material to enable Carillion to ascertain whether there is any basis to criticise KPMG.

On 14 November 2019, QE responded at considerable length in a 38-page letter. This was headed “First Letter of Claim and Request for Pre-Action Disclosure”. After a comparatively short introduction, Section B of the letter set out the Carillion Group’s potential claims against KPMG. The letter stated that KPMG had failed to give proper consideration to two key areas of the Carillion Group’s accounts, namely (i) the reporting of contract revenue and profits/losses on construction contracts and (ii) the carrying value of goodwill.

Some 14 single-spaced pages then covered the first issue, concerning contract revenue and profits and losses. The analysis included a detailed discussion of three particular contracts (included in the original 16 requested, and the 9 now pursued) where KPMG had previously, in 2017, provided certain working papers to Carillion as part of the review which was carried out at that time. In the summary of this part of the letter, QE stated:

Paragraph 121 went on to identify three matters which a competent audit would have identified, including the consequences that losses were required to be provided for immediately in financial statements for 2014, 2015 and 2016.

The letter went on to state, however, that:

Pages 31 to 35 of the letter then analysed the position in relation to goodwill. The analysis here was less detailed, but the essential conclusion was the same. Paragraph 132 identified the carrying value of goodwill as a key audit risk, and identified the approach which KPMG had said that they would adopt to this risk. Paragraph 133 then said:

The letter went on to identify a number of specific shortcomings in KPMG’s approach, which were to be inferred on the basis of the materials which Carillion had. These included a failure sufficiently to scrutinise and challenge management’s base case; a possible failure to respond to “red flags” raised by the performance of certain business; and the impact on goodwill of KPMG’s failures in relation to the audit of construction contracts (i.e. the subject matter of the first issue discussed in the letter).

The letter concluded with a request for pre-action disclosure. This request was much reduced from the request made on 1 October. The request was now focused on 9 of the 16 contracts, and goodwill, rather than other aspects of the audit. It was in materially the same terms to the application subsequently issued in February 2020, although much broader than the application which is now pursued.

There are a number of features of this letter which are, in my view, relevant in the context of the present application.

As a preliminary matter, I agree with Mr. Gaisman QC (for KPMG) that this letter was clearly directed towards the present application for pre-action disclosure, and was not intended to serve the function of a “Letter of Claim” as described in the Protocol. For example, the letter attached no supporting documents. Nor did it confirm whether an expert had been appointed or identify him. This conclusion is confirmed by the subsequent correspondence. On 21 November, Orrick wrote saying that it was unreasonable to expect a response within 14 days (as the letter had requested). QE’s response was that QE was only expecting a response on the question of whether KPMG intended to provide the documents requested, or whether a pre-action disclosure application was required. On 6 December, Orrick wrote in more detail and said that Carillion had “not yet produced a Letter of Claim”. QE’s reply on 16 December was that a formal Letter of Claim was not a pre-requisite for an application under CPR 31.16, and that they intended to provide “a full Letter of Claim, addressing all the requirements of the Protocol, following receipt and review of the material we are seeking from your client”.

At a later stage, on 4 February 2020, QE did write saying that:

Ms. Sabben-Clare QC, on behalf of Carillion, submitted that the 4 February 2020 letter made it clear that the earlier 14 November 2019 First Letter of Claim “was intended to be, and was, a Letter of Claim under the Protocol.” I was not persuaded by this argument. The point being made in paragraph 2 of the 4 February letter was that, as

However, for reasons already given, it is important to consider the question of compliance with the Protocol as a matter of substance not form, to pay regard to the spirit of the Protocol, and to disregard minor or technical infringements (as indeed paragraph 3.1 of the Protocol makes clear). Even taking these matters into account, however, I do not consider that this First Letter of Claim complied with the spirit of the Protocol, or that KPMG was in breach of the Protocol or its spirit by declining to produce the documents which had been requested.

First, it is an important feature of the 14 November 2019 letter that it did significantly narrow the request for documents from the request made on 1 October. The request reduced from 16 to 9 contracts. No requests were made in relation to the going concern, WIPRO and “reverse factoring” matters which had featured in the 1 October letter, and which had then been described as “necessary” areas for the production of working papers. In so narrowing its request, Carillion had no doubt taken on board some of the points fairly made in Orrick’s 16 October response, and had recognised the need (as the case-law shows) for a 31.16 application to be tightly focused.

However, the 14 November letter was careful to indicate that the matters set out therein are not a complete statement of the claim which Carillion was contemplating at that stage. The letter said that it was focused on “those aspects of KPMG’s audits which we anticipate will be the subject of the core allegations in the intended proceedings”. Paragraph 11 of the letter indicates that KPMG’s alleged failure to give proper consideration to a number of key areas included “at least” contract revenue and the carrying value of goodwill. That paragraph also said that:

This reservation of other potential claims for other areas of the audit, and indeed other contracts, is significant in the context of compliance with the Protocol. That Protocol envisages, in my view, that there will only be one Letter of Claim, to be sent when the claimant decides there are grounds for a claim against the professional. It must then include the intended allegations against the professional, and this must mean all of those which are intended at that stage. The professional’s response is to include specific comments on the allegations made, and this cannot be done (at least completely) if the professional has not been told all of the intended allegations. In my view, the Protocol does not envisage that there will be a sequence of Letters of Claim, with the initial letter being confined to certain particular aspects of the claimant’s case, and then subsequent letters being sent in relation to other aspects. I cannot see how such an approach would be consistent with the overriding objective or indeed the spirit of the Protocol.

Secondly, the documents requested in the Schedule to the letter go well beyond, in my view, a reasonable request for documentation in the context of the Protocol. The Protocol envisages and indeed requires both parties to provide “key” documents: see paragraph 10. Whilst there is obviously a certain amount of elasticity within that concept, it would be surprising if in most cases the “key” documents could not fit very comfortably within one lever arch file. Indeed, when eventually (on 24 April 2020) Carillion did provide its “key” documents, there were 12 such documents or groups of documents. KPMG submitted that this volume was “the right order of magnitude”, and I agree.

Here, however, the “Requested Documents” sought by Carillion were very extensive. The request was not confined to KPMG’s “Working Papers”. It included a further 11 categories of documents, including for example: “Internal emails relating to the areas of KPMG’s work referred to and the working papers requested in this letter” and “Emails between KPMG and the Carillion Group relating to the areas of KPMG’s work referred to and the working papers requested in this letter”. These very broad categories were themselves expanded upon by the terms of the final paragraph of the Schedule:

I could not see any realistic basis upon which it could be contended that a request of this magnitude was consistent with either the letter or spirit of the Protocol. Leaving aside the question of whether Carillion “needed” all of documents requested (see paragraph 6.2 (c) of the Protocol), I do not consider that this could fairly be regarded as request for key documents under the Protocol. Indeed, it must be doubtful whether all of the documents requested would even fall within the scope of standard disclosure: see paragraph 10.2 of the Protocol. Mr. Simon Willis, a partner at Orrick, later explained in his evidence the magnitude of the task presented. He had applied a range of search terms to a subset of documents relating to the Carillion Group audit which had already been collated by KPMG. The total number of documents which were responsive, after de-duplication, was over 190,000. But he said that further documents would need to be collected in order to meet Carillion’s requests. I do not consider that this is the sort of exercise contemplated by the Protocol. I therefore do not think that KPMG could be said to be in breach of the Protocol by failing to produce the documents requested, or indeed to engage with this request pursuant to the Protocol.

As matters developed, Carillion’s request was very substantially narrowed in late April 2020, some 3 weeks before the hearing. Ms Sabben-Clare’s submissions therefore ultimately focused on the narrowed application, rather than the original request. However, in my view the terms of the original request are relevant to the question of whether KPMG had breached the letter or spirit of the Protocol in the months after the request was made. For the above reasons, I do not consider that they did.

There are a number of other features of the 14 November 2019 letter which are relevant to the ultimate exercise of my discretion in this case.

First, in relation to the claim concerning contract revenue, the letter stated that, on the basis of the current material, it was likely that KPMG must have breached their contractual duties and been negligent “in relation to some or all of the other contracts for which provisions were made in July and September 2017”. Carillion’s intended case therefore appears to extend not only to the 9 contracts where documents are sought, and the additional 7 where documents were originally sought, but also to some or all of the other contracts within the 45 for which provision was made in July 2017.

Secondly, the letter made clear that the Claimants had a “settled intention to bring claims in respect of the matters discussed below”. This is not therefore a case where there is any doubt as to whether or not proceedings will be started. The basis of the request was that they were required “in order to properly particularise those claims”.

Thirdly, the letter dealt with causation and loss, albeit briefly and in a manner criticised by Mr. Gaisman. The size of the intended claim is clearly very substantial. The letter identified dividends of £ 234.2 million paid out in respect of the 2014-2016 years alone, but claims for other types of losses were referred to in general terms. The litigation therefore contemplates very substantial claims indeed.

Fourth, a fair reading of the letter to my mind leads readily to the conclusion that a considerable amount of work had already been carried out in investigating the potential case against KPMG, and that Carillion could provide adequate particulars of negligence in support of that case. Indeed, given the 25-page limit for pleadings in the Commercial Court, the analysis contained in the letter would require some considerable degree of compression in order to fit within a pleading of reasonable length.

After the exchange of e-mails concerning timing described above, Orrick replied on 6 December. They said that whilst efforts had been made in the direction of complying with the Protocol, Carillion had not yet produced a Letter of Claim which met the requirements thereof. They suggested that compliance with the Protocol should now be the focus of Carillion’s efforts. The correct order of events was for Carillion to engage properly with the Protocol, for KPMG to respond, and then for Carillion to make any application under CPR 31.16 if “they conclude that additional documents are required by way of pre-action disclosure, absent agreement”. The letter identified five respects in which there was alleged non-compliance by Carillion with the Protocol. Orrick also said that the 14 November letter demonstrated that Carillion already had sufficient information to articulate its case on breach of duty in accordance with the Protocol, should it choose to do so. Points were also made as to the width of the disclosure sought and that this would require the review of tens of thousands of documents. Orrick made it clear that KPMG was not to be taken as declining to provide any pre-action disclosure, but that should be “within an appropriate compass, take place at the appropriate time and be provided in the context of proper compliance with the Protocol”. The letter concluded by asking for a proper Letter of Claim, setting out the case not only on duty and breach but also causation and loss, and enclosing key documents.

QE provided a short response on 16 December. They considered that KPMG’s response had been misconceived, and that a continuation of the current dialogue would be ineffective. There was therefore no alternative but to issue an application under CPR 31.16. The letter engaged in a limited way with KPMG’s argument that no Letter of Claim under the Protocol had been provided. QE said that they had provided the best particulars of negligence that they were currently able to provide; that a formal Letter

On 10 January 2020, Orrick replied. They said that KPMG had complied with the Protocol by informing QE why their 14 November 2019 letter did not comply with the Protocol. They reiterated that Carillion already had the necessary documents to particularise their case on breach under the Protocol. They said that none of the documents sought related to matters of causation and loss where Carillion’s case remained “almost entirely unformulated”. It was not “appropriate for KPMG to be put to the burden of engaging in a substantial disclosure exercise” in circumstances where Carillion had “not provided a single document themselves (nor even identified the categories of document to which they already have access) and where they have not articulated any case on causation and loss”.

Pausing at this point in the chronology, there is no doubt in my view that, as a matter of form, Carillion had not yet provided a Letter of Claim in accordance with the Protocol, for at least some of the reasons articulated in Orrick’s letter of 6 December. On any view, key documents had not been copied and enclosed, and there was no provision of documents to support the estimated calculation of financial loss. There had also been no confirmation relating to Carillion’s proposed expert. More importantly, however, I consider that there had been non-compliance with the spirit of the Protocol, because Carillion had not set out the full scope of its intended allegations against KPMG.

These conclusions are not surprising, since the correspondence indicates that Carillion’s focus at this stage was not upon sending a Protocol-compliant Letter of Claim. Rather, as KPMG submitted, it was upon obtaining documents and laying the ground for the present application.

I did not understand Carillion’s case (that KPMG was in breach of the Protocol) to be that a fully compliant Protocol letter had been provided by this time, so that the 3-month period for KPMG to respond had started to run and that KPMG were in breach in failing to respond. Indeed QE’s letter of 16 December 2019 stated that they intended to provide a “full Letter of Claim, addressing all of the requirements of the Protocol, following receipt and review of the material we are seeking from you client”. That seems to me to acknowledge (as I would in any event have held) that a Protocol compliant Letter of Claim had not yet been sent.

However, Carillion’s argument, as articulated in QE’s letter of 16 December 2019, and in subsequent correspondence, is that a request for documents (pursuant to the Protocol) can be made at any time, including prior to the sending of a compliant Letter of Claim. Such a request is itself made pursuant to the Protocol. Here, a request was made and KPMG’s failure to provide the documents was itself a breach of the requirements of the Protocol. I will return to this argument below. However, even accepting (as I do) that a reasonable request for key documents can be made under the Protocol prior to sending a Letter of Claim, I do not consider that this was a reasonable request for key documents. KPMG were not therefore in breach by failing to respond to it.

QE wrote on 4 February 2020 indicating their intention to apply under CPR 31.16 forthwith. They provided, in draft, the documents which they intended to serve in support of the application. KPMG was asked to admit liability in respect of the intended allegations described in the 14 November 2019 letter, or to agree to provide the documents requested in the Draft Order voluntarily.

As previously stated, the documents sought in the draft application were materially identical to those sought in the schedule to the 14 November letter. There was a small change in the wording of the final paragraph, but this did not alter the substance. The application is in wider terms to the application that is now pursued, principally because the result of the exchange of evidence within this application has been that Carillion has been able to be more precise in relation to the documents sought.

Orrick responded on 17 February 2020. Their letter largely repeated points which had previously been made. It complained of the unfocused and wide-ranging nature of the application. It said that it also appeared from the draft that there is “the threat of further, similarly wide, requests being made at an unspecified later stage in respect of other audit areas (and possibly other audit years)”. This was said to exacerbate the deficiencies in the application as presently mooted. The proper order of events under the Protocol was for Carillion to prepare a Letter of Claim “that sets out its claim fully and identifies and encloses key documents in support”.

The present application was then filed and served on 19 February 2020. The application was supported by the lengthy 67-page witness statement of Mr. Bunting, a partner in QE with responsibility for the case. The issue of construction contract revenue and profits/losses was addressed in detail over some 24 pages. A further 9 pages were devoted to the issue of goodwill. These pages largely repeated, with some development, the analysis in the 14 November 2019 letter.

The reasons for the application were essentially those set out in the prior correspondence described above. Mr. Bunting said that there was a “strong prima facie case against KPMG, based on the information known to Carillion plc at present”:

A central point made in the witness statement was that the documents were necessary in order to enable Carillion to reach a “concluded and fully informed view” on whether KPMG in fact breached its duties and if so, in what respects and which specific years, and in order to formulate a “properly particularised claim”. The reason why the present case fell outside the “usual run” of cases (an expression used in the authorities referred to below) was KPMG’s prior refusal to provide any documents at all in response to Carillion’s requests.

Evidence in reply was served in a witness statement of Mr. Simon Willis, a partner at Orrick. As described above, he had applied a range of search terms to a subset of documents already been collated. The number of responsive documents was 190,000. Further documents would need to be collected in order to meet the proposed order.

Mr. Willis’ evidence did provide information about how KPMG’s working papers and associated relevant documents were structured and categorised. The subset of documents, to which the search terms had been applied, included KPMG’s “eAudIT files”. At the relevant time, KPMG used an electronic audit file known as “eAudIT”. KPMG had collected documents from those files, including all documents on the 2014, 2015 and 2016 year-ends. He explained that the eAudIT file comprised a combination of:

'screens' (that is, documents created within the eAudIT architecture which are completed by audit team members to record the audit work carried out);

working papers and other documents created outside of eAudIT (say, in the form of Excel and Word documents) uploaded to eAudIT;

evidence provided by the audit client supporting the conclusions reached and which is saved within eAudIT; and

'sign-off histories' which record the name of the individuals who prepared and reviewed each of the audit documents saved to eAudIT (and the relevant date).

This information in Mr. Willis’ statement then led to a request by QE on 15 April 2020 for, amongst other things, an index of the eAudIT. Orrick’s response on 22 April 2020, whilst reiterating criticisms of the application, provided an index showing the structure of the eAudIT files in each year.

On 24 April, QE sent a further letter in a final attempt to reach agreement on pre-action disclosure. The letter was sent along with a short form letter before claim, described as “the Second Letter of Claim”. This Second Letter of Claim was intended to meet the points previously raised by KPMG, and “unequivocally complies with the requirements of the Protocol”. This Second Letter of Claim cross-referred to the first letter sent in November. It identified and addressed each relevant paragraph of the Protocol. The Appendix to the letter contained a list of documents requested. This list was an adaptation of the request originally sent on 14 November 2019. But it was very significantly narrower in that the key definition of ‘Requested Documents” were now “defined as documents in the following categories contained on KPMG’s eAudIT files for the 2014, 2015 and 2016 financial year end audits of the consolidated financial statements of the Carillion Group and of Carillion Construction Limited”. There then followed 9 separate numbered sub-paragraphs. Previously, the ‘Requested Documents’ had contained no such definition, and there were 11 sub-paragraphs. The Second Letter of Claim requested a response by 29 May 2020, and production of the documents by the same date.

These changes to the documents requested by Carillion were also mirrored in a revised draft order sought in the context of the present CPR 31.16 application. Mr. Bunting’s second witness statement, served on 24 April, therefore enclosed a revised draft order.

On 1 May 2020, Mr. Willis responded to Mr. Bunting’s second statement and the amended draft order. He addressed, first, the practical implications of the amended application. The statement acknowledged that the exercise now contemplated was reduced. But this would still require the review of around 8,500 documents. This figure of 8,500 did not reflect the total number of pages, since many documents would have multiple pages. Even if KPMG approached the exercise on the basis that certain areas of the audit files were irrelevant (as Mr. Bunting had suggested), there were still in the region of 6,000 documents to review. He also said that there was a strong likelihood that KPMG would have to perform a disclosure review of the documents in question twice, if the order was granted. A second review would be required once the issues in the action were known, since they would be very unlikely to be exactly the same, and in the same scope, as now presented.

The final relevant letter is Orrick’s reply, on 6 May 2020, to the Second Letter of Claim. They gave reasons why KPMG would not be responding to the Second Letter of Claim by 29 May 2020. The letter raised a number of points. One issue raised was whether the First and Second Letters of Claim encompassed the full scope of the anticipated claim against KPMG. Orrick requested confirmation that the intended claim would be confined to the matters addressed in the existing Letters of Claim: i.e. 9 contracts, goodwill on the basis identified and three audit years. Orrick said that the usual period was 3 months to respond to a Letter of Claim, but that this was highly unlikely to be an adequate period in the present case, even ignoring their various criticisms of the letters sent. The letter concluded:

The relevant legal principles are conveniently summarised by Blair J. in paragraph 17 of Assetco. CPR 31.16 provides that the court may make an order for pre-action disclosure only if certain conditions are satisfied:

The respondent and applicant must both be likely to be parties to subsequent proceedings. It is not however necessary to show in addition that the initiation of such proceedings is itself likely: Black v Sumitomo Corp [2002] 1 WLR 1562 at [71 – 72], Rix LJ, which is the leading case on the rule.

The documents sought must fall within the scope of the standard disclosure which the respondent would have to give in the anticipated proceedings. It follows that at the time of the application, the issues must be sufficiently clear to enable this requirement to be properly addressed.

Disclosure before proceedings have started must be desirable (i) to dispose fairly of the anticipated proceedings, (ii) to assist the dispute to be resolved without proceedings, or (iii) to save costs: CPR 31.16 (3) (d).

In considering whether to make an order, among the important considerations are the nature of the loss complained of, the clarity and identification of the issues raised by the complaint, the nature of the documents requested, the relevance of any protocol or pre-action inquiries, and the opportunity which the complainant has to make his case without pre-action disclosure (Black v Sumitomo Corp at [88]).

The anticipated claim must have a real prospect of success.

In the commercial context, a pre-action disclosure order, even if not exceptional, is unusual.

The request must be “highly focussed” and confined to what is “strictly necessary” for the purposes for which pre-action disclosure may be ordered: Hutchinson 3G UK Ltd. v O2 (UK) Ltd. [2008] EWHC 55 (Comm) at [40] (Steel J.); Snowstar Shipping v Graig Shipping [2003] EWHC 367 (Comm) at [35] (Morison J).

Two other judicial observations are pertinent. In Total E&P Soudan SA v Edmonds [2007] EWCA Civ 50, the Court of Appeal described applications under CPR 31.16 as being “in the nature of case management decisions requiring the judge to take a “big picture” view of the application in question”. This was said in the context of discouraging detailed investigation of legally complex and debateable issues of law, but in my view it is true as a general proposition. This means that a judge should look at the case in the round, and take into account how matters are likely to proceed in the litigation in the event that disclosure is ordered or refused. In Hands v Morrison Construction Services Ltd., [2006] EWHC 2018 (Ch) Mr. Michael Briggs QC (as he then was) said that the court must also stand back at some point and look at the matter in the round – in other words, not fail to see the wood for the trees. The question at that level “may include the general question: does the request for pre-action disclosure further the over-riding objective in this case, or not” (paragraph [30]).

In order to make an order under CPR 31.16, the court must be satisfied that, if proceedings had started, the respondent’s duty by way of standard disclosure, set out in rule 31.6, would extend to the documents or classes of documents of which the applicant seeks disclosure. In Hutchinson 3G UK Ltd. v O2 (UK) Ltd., Steel J. said that it was necessary for all the documents sought within a class or category to be within the scope of standard disclosure. This means that the application must “clearly” not encompass categories of documents which will simply prove to be relevant (if at all) as part of the background, let alone documents which might merely lead to a train of inquiry. It is therefore necessary for the applicant to show that it is more probable than not that the documents are within the scope of standard disclosure in regard to the issues which are likely to arise. The authorities (such as Total E&P discussed below) show that if a request is too broad in a particular respect, the court is not required to dismiss the application but can modify the order appropriately.

The material parts of the present application, as amended, request:

Aberdeen Western Peripheral Route.

Royal Liverpool Hospital.

Battersea Power Station.

Midland Metropolitan Hospital (this request is made in relation to the period from December 2015, when the relevant contract was entered into, onwards).

Southmead.

TTC Vaughan (Canada).

Union Station (Canada).

Msheireb (MENA).

Al Dara (MENA).

The allocation of businesses and intangible / goodwill balances to the different Cash Generating Units within the “Group” (i.e. the group of companies comprised of Carillion plc and its subsidiaries).

KPMG’s audit of the annual impairment reviews carried out by management in accordance with IAS 36, including details of any sensitivity analysis performed by KPMG.

KPMG's review of the Group's disclosures regarding the sensitivity of the outcome of the impairment assessment.

Documents in the categories described in Willis 1 paragraphs 110(a)-(d), ie ‘screens’ created within eAudIT, working papers and other documents created outside eAudIT and uploaded to eAudIT, audit evidence saved within eAudIT and sign-off histories).

Audit planning documents.

Notes of or prepared for year-end clearance meetings and calls.

Technical consultation documents.

Documents recording consideration of relevant financial statement disclosure matters.

Internal emails.

Meeting notes.

Internal notes of site visits.

Internal notes of reviews of position papers on contracts.

Accordingly, the application is now focused only on documents on KPMG’s eAudIT files, and only on those which specifically concern the audit work on the 9 contracts and goodwill. Leaving aside the final paragraph (“For the avoidance of doubt”), I consider it more likely than not that the documents sought by this focused request are within the scope of standard disclosure in regard to the issues that are likely to arise.

Mr. Gaisman submitted that, by way of example, no specific criticisms had been advanced in relation to KPMG’s planning of the audit in relation to accounting for contract revenue. The court could not therefore be satisfied that standard disclosure was satisfied in that regard. I disagree. Given that the request is now narrowly focused on the eAudIT files, it is in my view inherently probable that one side or the other will seek to place reliance on the planning for the audit. If there were sound planning, then it is probable that KPMG will place reliance upon that aspect of the audit. If there was not, then this will adversely affect KPMG’s case. Either way, the documents probably fall within standard disclosure. The decision of the Court of Appeal in Bermuda International Securities Ltd. v KPMG [2001] Lloyd’s Rep P.N. 392 shows that a focused application for specific audit working papers, in the context of a potential claim for audit negligence, will likely fall within the scope of standard disclosure. Of course all depends upon the nature of the case, and the wording of the order sought. But here (leaving aside the final paragraph) I consider that the focus and drafting is now sufficiently narrow, and the scope of the issues in the potential case sufficiently wide, that this jurisdictional requirement is satisfied.

That conclusion does not, however, apply to the final “for the avoidance of doubt” paragraph. The motivation for including these words in the order is that Carillion’s expert has indicated that general documents of this kind are likely to be on the audit file. No doubt that is so, but it does not follow that this entire class of documents fall within the scope of standard disclosure. I agree with Mr. Gaisman that Paragraph 29.3 of Carillion’s skeleton, fairly read, acknowledges that, certainly in relation to these

Mr. Gaisman suggested that the order would require KPMG to have to work out for itself what documents were within standard disclosure, and that this is not permissible. I disagree. The order defines the documents to be produced, and does so sufficiently narrowly. The exercise required would be for KPMG to review the eAudIT files in order to identify the documents falling within the scope of the order. There would be no requirement to carry out additional work in order to decide whether standard disclosure extends to those documents. If I were minded to make the order, any debate on that score would have been resolved.

The case-law indicates that this is a relatively low jurisdictional threshold which, as Rix LJ said in Black v Sumitomo, may well be crossed in “very many if not most cases”. The requirement is to show a “real prospect” of one of the matters set out in those subparagraphs being satisfied.

There are many authorities where judges have been persuaded that the jurisdictional threshold is crossed because it will enable the pleadings to be more focused, and avoid the need and cost of later amendments: see Bermuda v. KPMG at [27]; Black v. Sumitomo at [83]; Marshall v. Allotts [2005] P.N.L.R. 11 at [61]; Hands v Morrison Construction Services Ltd. [2006] EWHC 2018 (Ch) at [31]; Total E&P & Soudan S.A.

I consider, with some hesitation, that this threshold is crossed in the present case. If the requested documents are made available, Carillion will be able to prepare a more focused pleading at least in relation to the 6 (out of 9 contracts) where Carillion has not previously seen any relevant audit working papers, and also in relation to the issue of goodwill. This will enable both parties to know each other’s fully particularised case on those matters.

My hesitation derives, in part, from the fact that I consider that this is a case where there will inevitably in due course be amendment of the pleadings, following disclosure, even if documents were now produced pursuant to the order sought.

This is one consequence of a dilemma which Carillion has, to my mind, faced in advancing the present application. In order to make the application attractive, and to fit within the case-law which underlines the importance of a focused approach, Carillion has been forced to limit its case to a relatively small number of contracts and goodwill. But this means that other aspects of the case will not, on any view, be particularised at the present stage and that amendment is likely in the future; at least unless the present application is only the first in a series of similar applications for further documents. It is tolerably clear from the correspondence described above that Carillion’s case in relation to contracts will not be confined to the 9 where pre-action disclosure is currently sought. It will almost certainly extend to the other 7 where disclosure was originally requested, and in all likelihood to other contracts as well. Indeed, QE’s

This conclusion (which is relevant to the exercise of my discretion, below) is borne out by the lack of confirmation provided in response to a pertinent point raised by Orrick in their letter of 6 May 2020:

Moreover, even in relation to the 9 contracts and goodwill, standard disclosure is likely to extend to documents other than those on the eAudIT file. Carillion will certainly so contend, and indeed the present application originally sought documents well beyond the eAudIT file. The consequence will inevitably be that even in relation to the 9 contracts and goodwill, amendments are reasonably in prospect.

There is therefore no real prospect of amendment being avoided. Indeed, it is even now obvious that the proposed litigation, arising out of such a high profile collapse, is likely

A further consequence of these considerations is that whilst there may be some saving of costs because the pleadings could be better focused in relation to the 6 or 9 contracts and goodwill, there is likely to be consequent increased cost because of the need for KPMG to re-review the eAudIT file once the case has been pleaded out and the full extent of Carillion’s claims are known. This duplication in cost is a point made by Mr. Willis in his second statement, and I consider it well-founded. In Assetco, a similar point was made by Grant Thornton, and this was one reason why Blair J. exercised his discretion against granting the order sought: see paragraphs [35] and [36].

I will return to these matters in the context of the exercise of my discretion. Notwithstanding my hesitation, as described above, I consider that the jurisdictional requirement is satisfied in this case; because, as in other authorities, disclosure would assist in sharpening the initial rounds of pleadings, albeit only in relation to 6 or 9 contracts and goodwill and even then with the prospect of later amendment on those and other issues. Whilst the jurisdictional threshold is crossed, this says very little about how my discretion should be exercised.

I was unpersuaded that there were any other reasons for saying that CPR 31.16 (3) (d) was satisfied. In a case of this potential magnitude, and where Carillion has yet to identify the extent of its alleged loss, I do not consider that there is any realistic prospect that disclosure will assist the dispute to be resolved without proceedings.

For the following reasons, I do not consider that this is an appropriate case for the court to exercise its discretion in favour of ordering pre-action disclosure.

First, Carillion can in my view satisfactorily plead its case in relation to the 9 contracts and goodwill without the materials now sought. Those materials concern the alleged negligent conduct of the audit. Carillion has been able to articulate its claim as to negligence on those issues in a very lengthy letter and even longer witness statement. Ms. Sabben-Clare acknowledged that if the application were refused, the claim against KPMG would not go away. Carillion would then, as she said, plead a case that it was to be inferred that the audits were negligent in reliance upon their expert’s prima facie views. The level of detail into which Carillion has been able to descend, both in 38page letter and Mr. Bunting’s witness statement, in reliance upon expert evidence already obtained, shows that the present case is a very long way from the circumstances of Pantelli Associates ltd. v Corporate City Developments Number Two Ltd. [2011] PNLR 12, upon which Carillion relied. That case concerned a wholly inadequate pleading of negligence in circumstances where no expert evidence had been obtained. Even though the present case has not yet been pleaded out, Carillion will in my view have no real difficulty in articulating a pleadable case on negligence.

I accept that Carillion’s ability to plead a case on negligence is not determinative of the present application. In Total E&P, the Court of Appeal accepted that Total would be able to plead a case along the lines indicated by the material which it already had, but nevertheless upheld the judge’s decision to order disclosure. However, the ability satisfactorily to plead a case is a relevant factor when considering the big picture. Rix

In Assetco (which was an audit negligence case), a significant reason for the refusal of the order sought was that, as Blair J. said:

Both sides referred to the facts and decision in Assetco as providing some support for their respective positions. I do not consider that it is helpful in the present context to compare the facts of one case with another, and I note that the above passage (upon which Mr. Gaisman particularly relied) was directed towards whether documents fell within the scope of standard disclosure. However, the decision in Assetco does indicate (unsurprisingly, and consistently with Black v Sumitomo) that it is important for the court to consider the range of materials already available to the party who intends to advance a case of negligence against an auditor in respect of alleged misstatements in financial statements. Here, as Mr. Gaisman submitted, the question of whether there were misstatements by Carillion or its subsidiaries in their financial statements is a matter which depends, principally if not exclusively, upon the company’s own records. The audit working papers are relevant to the separate question of whether, on the basis of the materials available to the auditors, the latter acted negligently. I consider that, just as in Assetco, Carillion’s own documents should enable it to say why, on the material available to KPMG, it acted negligently in relation to the company’s figures. Indeed, Carillion has already articulated its case, in some detail, as to why that was so.

Secondly, against that background, I agree with Mr. Gaisman’s submission that, in seeking further disclosure in order to seek to obtain a view on negligence which is “concluded” and “fully informed”, Carillion is seeking a level of assurance and certainty which is inappropriate and does not justify the application which is made. It would be possible to say, in every professional negligence case in the Commercial Court, that pre-action disclosure would assist the claimant’s expert in coming to a fully informed or concluded view on the issue of negligence. However, pre-action disclosure is not the norm, even in audit cases, as shown by Assetco and the absence of any recent authority where such disclosure has been ordered. The nature of civil litigation is that it is uncertain, and the process of providing information in order to enable experts and others to give concluded or “fully informed” views is necessarily a lengthy one. The provision of disclosure is certainly a key part of that process, but the norm is for that to be provided once the litigation has started, and the Commercial Court has taken steps (reflected in the Disclosure Pilot contained in CPR PD 51U) to control that disclosure. The process of enabling concluded or fully informed views to be reached by experts does not, however, stop with disclosure. Experts need to consider what the fact witnesses say in their statements. Hence, expert reports are invariably exchanged or served only some time after factual witness statements, so that the reports can be informed by relevant witness evidence. The experts themselves will invariably only give their evidence, and hence their final views, after the factual witnesses have given evidence at trial. Accordingly, even if the disclosure now sought were to be provided, it would only provide an additional step on the road to an expert reaching a concluded or fully informed view. However, as the 38-page letter and Mr. Bunting’s statement show, Carillion and its expert have been able on the materials available to take significant steps down that road already, and have been able to reach a view which is clearly strongly held. Whilst this is said to be based on “inference”, there is nothing inherently unsatisfactory about that: inference is, generally speaking, a conclusion which a person can sensibly reach on the basis of existing information.

An expert engaged for the purposes of litigation would, of course, always like as much information as possible as early as possible. However, I do not think that the desire of an audit expert to have more information so as to convert his “prima facie” views into more certain “concluded” or “fully informed” views, is a matter which generally speaking should lead to an order under CPR 31.16. In the present case, where it is clear that the expert has already formed a prima facie view which Carillion has been able to articulate in detail, the case for doing so is far from compelling. I do not consider that ‘out of the norm’ expensive pre-trial procedures are appropriate in order to provide a level of perfectionism or comfort for an expert’s views, in circumstances where the expert has material on which he has been able to express a view which is sufficient to enable Carillion to start proceedings and plead its case.

Thirdly, the case for ordering the production of documents in order to assist an expert to reach a concluded or fully informed view, at the pre-action stage, must take into account the scale of the exercise which is contemplated. The burden of complying with the order sought is an important factor to be considered in the exercise of the court’s discretion. It is also important to have in mind case management considerations in this context. If, as here, a very large piece of litigation will be commenced in any event (as Ms. Sabben-Clare said, the claim against KPMG will not go away), giving rise to a significant disclosure exercise likely to stretch well beyond the documents sought in the CPR 31.16 application, then a court may conclude that disclosure in the case should simply follow the ordinary “normal run” course. It is also relevant in this context that the current disclosure pilot in the Commercial Court provides for a more focused and controlled approach to disclosure, and that the Commercial Court Guide encourages restraint and discourages elaborate and extensive pre-action procedures.

In the present case, the application has now been substantially narrowed, albeit only some months after the application was mooted and then served. Nevertheless, this remains an application which, if granted, would require KPMG to review a very large number of documents. Even after confining its requests to documents saved on

Fourth, as indicated in Section F above, this is a case which will develop further, even if the disclosure were to be ordered. Even in relation to the 9 contracts and goodwill, there will inevitably be requests for further standard disclosure of documents outside those saved to the eAudIT system. The pleadings are likely to encompass at least 7 more contracts, and possibly some 45. There is the possibility that Carillion will plead a case on the other aspects of the audit which formed part of the original request in October 2019, but which were then not pursued – no doubt, as KPMG submitted, in order to make the present application more palatable. Carillion was asked (in Orrick’s letter of 6 May 2020) to confirm that the scope of the intended claim would be confined to the matters addressed in the existing Letters of Claim. No such confirmation has been provided, and it is obvious from the terms of the correspondence that Carillion is contemplating a wider claim.

In my view, these matters give rise to a number of consequences relevant to the exercise of the court’s discretion, and which point away from ordering the disclosure at this stage. They reinforce the conclusion that KPMG will need to repeat the exercise contemplated by the present application in order to identify documents presently falling outside the scope of the intended order. They indicate that the concept of Carillion’s expert forming a concluded or fully informed view is questionable; because he will not be able to form such a view in relation to potential areas of the claim which fall outside the scope of the present request. They indicate that this is a case where amendment of the pleadings after disclosure is inevitable.

They also give rise to the possibility that there will be further serial applications for further pre-action disclosure. KPMG submitted in its skeleton argument for the hearing that Carillion had not said that it was unable to articulate its case in relation to the further contracts and areas of audit work previously identified; that it seemed likely that Carillion had excluded those other contracts and areas in the hope of making its present application slightly less unattractive; that it was unclear (and Carillion had not said) how many further requests there would be, or how they would fit with the further and more detailed letter of claim Carillion promised. It concluded that “Carillion does not intend the cost and delay of this application to be the end of its pre-action manoeuvres”.

Ms. Sabben-Clare’s answer to this point was that this was all for another day; that the court should simply decide the present application; and that if any further applications were made, KPMG would be able to say that Carillion should not have a second bite at the cherry.

I did not consider that this was a persuasive response. In considering the exercise of discretion, it is important to look at the “big picture”, to take into account overall case management considerations, and not to lose sight of the wood for the trees. Here, Carillion has provided no indication that its case will be limited to the 9 contracts and goodwill which form the subject-matter of the present application, and no indication or assurance that it is not contemplating further serial applications. I consider that the prospect of serial applications is another factor which should lead the court to decline to exercise its discretion in favour of making the order sought. I will return below to the way in which, as a matter of case management, I consider that matters should proceed.

Fifth, I have not been persuaded that there has been any breach by KPMG of the letter or spirit of the Protocol. I consider that the only substantial argument as to why this case would fall “outside the usual run” would be if Carillion could demonstrate that KPMG had failed to act in accordance with the Protocol. In Black v Sumitomo, Rix LJ said “the relevance of any protocol or pre-action inquiries” was a relevant matter to consider.

For reasons given in Section C above, I do not consider that KPMG had acted in breach of the Protocol at the time that the present application was launched. At that time, Carillion had not, as a matter of form, purported to serve a Letter of Claim compliant with the Protocol. More importantly, neither the letter nor the spirit of the Protocol had been complied with because, for tactical reasons relating to the need to keep Carillion’s proposed application within bounds, Carillion had confined its 15 November 2019 letter to 9 contracts and goodwill, and had not set out the full scope of the allegations that it intended to pursue. I have also explained why I consider that the request for documents made at that stage was far too wide, and beyond anything that in my view is contemplated by the Protocol.

I accept Carillion’s point that the Protocol does contemplate that a reasonable request for documents can be made “at any time”, at least if the claimant “needs to make” such a reasonable request. However, the usual course – as Assetco indicates – is that such request will be made at the time when the claimant has formulated his Letter of Claim. There may of course be cases where a request is justified at an earlier stage, for example where a claimant knows that something has gone wrong very badly, but has little idea as to why, and needs documents in order to formulate a Letter of Claim. But that does not apply in the present case, and I was unpersuaded that Carillion needed to make any request for documents, let alone the request that was made, prior to setting out its case in a compliant Letter of Claim.

I have so far considered the position at the time when the application was launched. However, the landscape did change on 24 April 2020, when Carillion substantially narrowed its document requests, and when it also sent its “Second Letter of Claim” to KPMG. This letter did address (for the first time) each relevant paragraph in the Protocol. The question therefore arises as to whether the case now falls outside the usual run, because KPMG has acted in breach of the Protocol since 24 April 2020 by failing to produce the documents identified in this narrower request?

Orrick’s response to this letter was, first, that various “deficiencies” in the first Letter of Claim had not been cured. It seemed to me, however, that some of these arguments (for example that there had been insufficient identification of the misstatements in the accounts) did not go so much to the question of whether a compliant Letter of Claim had been served, but rather whether the proposed claim was sufficiently well-founded. Those are arguments which I would expect KPMG to develop in its Letter of Response, but they do not in my view make the Letter of Claim non-compliant.

Orrick’s second point, however, is in my view more substantial. This is the point that Carillion had yet to set out all of its intended allegations, and that it was inappropriate for KPMG to be required to respond to what appear to be partial Letters of Claim. As already indicated, I do not consider that a partial Letter of Claim is within the spirit or letter of the Protocol.

This does not quite dispose of Carillion’s argument, however, because I have accepted their argument that reasonable requests for key documents can be made prior to the service of a Letter of Claim, and hence could be made in circumstances where only a “partial” Letter of Claim has been served. However, the usual course is, as I have said, for a claimant to seek documents at the same time as a Letter of Claim (i.e. a full Letter of Claim) is sent. I do not consider that there is any “need” in the present case for a request to be made in advance of sending the full Letter of Claim, or that it is reasonable to do so. Furthermore, I consider that the request made, even though now more narrow, still goes well beyond a request for “key” documents, and would require a substantial disclosure exercise.

Finally, I have endeavoured to stand back from the detail and give consideration to how matters should proceed as a matter of sensible case management. I do not think that the overriding objective or the efficient case management of this litigation would be furthered by granting the application. Rather, I consider that what should now happen is that Carillion should either provide the confirmation sought as to the scope of the intended claim (see paragraph 2 (b) (iii) of Orrick’s letter of 6 May 2020), or send a further (compliant) Letter of Claim setting out the scope of the intended allegations. Carillion can of course make a request for documents pursuant to the provisions of the

If Carillion’s allegations do not extend substantially beyond those set out in the Second Letter of Claim, I would expect KPMG to be able to respond well within the 3 month period, given the history of this case, and the fact that the substance of the case was set out in detail in the 38-page letter in November 2019. Even if they do so extend, I doubt whether a period of significantly more than 3 months would be required.

However, I would not wish to be seen to be encouraging a further application under

Accordingly, Carillion’s application is dismissed.