The case of Smailes -v- McNally [2015] EWHC 1755 (Ch) has appeared in the reports before.  In his judgment today His Honour Judge Pelling QC refused relief from sanctions after the claimant had failed to give adequate disclosure in compliance with a peremptory order. It contains important observations on the care needed in giving disclosure in addition to the principles relating to relief from sanctions.

The reality is that proceedings brought on behalf of the Crown and funded by the tax payer ought to be conducted throughout to best practice standards. These proceedings have not been”


  • A judge hearing an application for relief from sanctions was not confined to considering those breaches that had been found by the Court of Appeal on appeal. The judge was entitled to look at the matter fully.
  • Certain breaches of a peremptory order to give disclosure would not, in themselves, have led to relief from sanctions being refused.
  • However there had been fundamental, and inexcusable, failures by the claimants. These had led to unnecessary hearings and expense. It was inappropriate for relief from sanctions to be granted.


The court made a peremptory order that the liquidators conduct a search for documents falling within CPR r.31.6. Originally the judge found that there had been compliance with the order. However this decision was overturned by the Court of Appeal who found that there had not been any compliance.

1. The liquidators failed to make any, or any reasonable, search for the documents.

2. The judge had erred in considering that the methodology of the search had been agreed between the parties.

3. The question of whether there had been compliance with the unless order was not a matter of discretion.

4.  The liquidators had failed to disclose relevant documents that were in their possession, whose disclosure had been promised.

5. The judge found that the liquidators had acted in good faith but this did not warrant an inference that a reasonable search had been made.

6. There had also been a failure to make a proper disclosure list.

The Court of Appeal refused an application for a stay whilst an application for relief from sanctions was pending.  The liquidators made an application for relief from sanctions following the Court of Appeal finding.


The judge reviewed the rules and practice directions relating to disclosure and the three stages of the Denton test.


The First Stage

The Scope of the Enquiry

  1. As I indicated at the start of this judgment there is a difference of principle between the parties concerning how I should approach the first stage exercise. It is submitted by Mr Steinfeld QC on behalf of the Liquidators that the only category of breach that I am entitled to consider is that which was identified by the Court of Appeal as the basis on which the appeal from the Order of Birss J should be allowed – that is the omission from the list served in purported compliance with the Default Order of any reference to the scripts. Mr Steinfeld argued that the effect of the judgment of Birss J was that the Default Order had not been breached in any of the ways alleged before him and that in consequence the only breach the respondents were entitled to rely on was that relied on by the Court of Appeal. Mr Crow argues that this approach is wrong in law, that I should consider each of the breaches that were being relied on before Birss J, and the OCR Issue. I am not able to accept Mr Steinfeld’s submission for the following reasons.

  2. The only authority that each of the parties relied on in relation to this issue is P & O Nedlloyd BV v. Arab Metals Co and others (No.2) [2006] EWCA Civ 1717 [2007] 1 WLR 2288 (“Nedlloyd”). In summary in that case Colman J had refused permission to amend Particulars of Claim so as to allege an alternative contract to that which had been pleaded originally on the basis that the new claim was statute barred. The claimants appealed that decision to the Court of Appeal. That court allowed the appeal on the basis that even if the new claim was statute barred it arose out of the same facts and matters as the original claim and so the court had jurisdiction to give permission to amend and permission was given to amend the Particulars of Claim. Subsequently, the Defendants met a claim for specific performance of the alternative contract with a defence of Laches. The Claimant applied for summary judgment on its specific performance claim and failed. The issue on appeal was whether it was open to a defendant to rely on Laches in relation to a claim that was (as Colman J had held) the subject of a statutory limitation period – a conclusion that had not expressly been overturned by the Court of Appeal, when it had set aside Colman J’s Order.

  3. The Court of Appeal held that the order of a first instance judge ceased to have any effect once it was set aside by the appeal court, regardless of the grounds on which it was set aside, and, therefore, the first instance judgment was incapable of giving rise to an issue estoppel once it had been set aside. The Court of Appeal’s judgment was given by Moore-Bick LJ who at paragraphs 26 to 29 said:

“When the matter came before it on appeal this court approached the question from a different direction. At the outset of the hearing, as is reflected in para 11 of the judgment of Thomas LJ, the court directed that it would hear argument first on the question whether the new claims arose out of the same, or substantially the same, facts as the existing claims and on the exercise of its discretion. In the event the court was satisfied that the new claims did arise out of the same facts as the existing claims and exercised its discretion in favour of permitting the amendment. Thomas LJ, with whom Carnwath and Tuckey LJJ agreed, concluded his judgment with these words [2007] 2 Lloyd’s Rep 148 , para 25:

“For these reasons, therefore, I consider that, on the assumption that the claims are time-barred, they arise out of the same facts as originally pleaded and the court should exercise its discretion to allow the amendments. In the circumstances, therefore, the issues on limitation so carefully and clearly considered by the judge do not arise. We have heard no argument on the issues and I express no view upon them.”

27 There can be no doubt, in my view, that the court was concerned to make it clear that it was neither endorsing nor disapproving the decision of Colman J on the limitation question. It did not need to do so because, having concluded that the new claims arose out of the same facts as the existing claims, it was satisfied that it had jurisdiction to permit the amendment, whether the claims were time-barred or not. In the event, it set aside Colman J’s order and substituted its own.

28 Mr Rainey was quite right in saying that this court did not overturn the judge’s decision on limitation, but despite that I am unable to accept that his judgment is any longer capable of giving rise to an estoppel in relation to that issue. The effect of the order made on appeal is to avoid entirely the order made by the court below. In Spencer Bower, Turner & Handley, at para 60, the matter is put as follows:

“When a tribunal with original jurisdiction has granted, or refused, the relief claimed and an appellate tribunal reverses the judgment or order at first instance, the former decision, until then conclusive, is avoided ab initio and replaced by the appellate decision, which becomes the res judicata between the parties.”

29 … As a matter of principle, when an appellate court sets aside the order of a lower court that order ceases to have any effect and the decision of the appellate court alone is determinative of the issue between the parties. That is sufficient to determine the present case. Although the decision of Colman J was originally capable of giving rise to an issue estoppel, it could no longer do so once it had been set aside on appeal, regardless of the grounds on which this court made its order… .”

  1. In my judgment, the reasoning in Nedlloyd is of limited relevance to the question between the parties before me namely whether the decision of the Court of Appeal limits the respondents to reliance on the failure by reference to which the Court of Appeal concluded that the obligation to carry out a reasonable search contained in the Default Order had been breached. The issue before Birss J and the Court of Appeal was different from the issue I have to decide. The Court of Appeal and Birss J were concerned only to determine whether the Default Order had been breached, not to determine the seriousness or significance of the breach of that Order – an issue that is material only to the relief from sanctions application. That application was not determined by either Birss J or the Court of Appeal. That is undoubtedly why the Court of Appeal approached the appeal in the manner that it did.

  2. The effect of the reasoning in Nedlloyd on the facts of this case is only to preclude the Liquidators from denying that they have breached the default order in at least the manner identified by the Court of Appeal and also from relying on the ground that Lewison LJ expressly stated he had not based his judgment on. Birss J’s Order could only be relevant if it created some form of issue estoppel in relation to the issues considered by him and not considered by the Court of Appeal. It does not because Birss J’s Order ceased to have effect for all purposes once it was set aside. Thus no estoppel could arise in relation to either the issue concerning the form of the list, which Birss J had resolved against the respondents, or the OCR Issue, which could not have been considered by Birss J and was not considered by the Court of Appeal.

  3. Although Mr Steinfeld argued that Nedlloyd could be distinguished on the ground that Lewison LJ did not express himself in the way that Thomas LJ had in Nedlloyd as described in the quotation set out above, in my judgment that submission is mistaken. The presence or absence of express wording such as that used by Thomas LJ is not determinative of the point I am now considering, which follows as a matter of law for the reasons given by Moore-Bick LJ in Nedlloyd. Mr Crow also suggested that this was also the case because the Court of Appeal’s Order expressly records that the appeal had been determined only by reference to Ground 1 and not any of the other grounds. In my judgment that point assists in part on the issue I am now considering. Ground 1 concerned whether there had been a reasonable search and thus Mr Crow’s point begs the question I am now considering – that is whether the respondents are confined to the ground on which the Court of Appeal concluded there had not been a reasonable search. I accept however that this point is a valid one in relation to the alleged failure to comply with CPR r.31.10.

  4. The conclusion that Birss J’s Order ceased to have effect for all purposes once it was set aside also enables me to resolve the related submission made by Mr Steinfeld, namely that it is not open to me as a matter of judicial comity to reach a different conclusion from that reached by Birss J concerning the applicability of Practice Direction 31B. In my judgment the effect of the Court of Appeal’s Order in this case is that Birss J’s Order was avoided ab initio and with it the reasoning that led to that Order.

  5. In the result, in my judgment the respondents are fully entitled to contend that the Default Order had been breached in each of the ways they assert and are not confined to the single ground on which the Court of Appeal allowed the appeal from Birss J’s Order.

The Alleged Breaches of the Default Order

  1. The Scripts Issue

The respondents allege that the failure to disclose the scripts was illustrative of a failure to carry out a reasonable search (the basis on which the Court of Appeal reached its conclusion) but was also a failure to disclose documents that were required to be disclosed on standard disclosure. Whether what occurred is characterised as a breach of both the obligation to carry out a reasonable search and/or as a failure to disclose relevant documents does not add materially to the seriousness or significance of the omission, which in reality depends on the seriousness and significance of the omission of the scripts from the list served in purported compliance with the Default Order.

  1. The primary facts relating to this issue are set out above in the sections of the various judgments that I have quoted from. Mr Steinfeld’s argument was that this breach was technical and one that could not in any relevant way have prejudiced the respondents because the existence of these documents or at least this class of documents had been apparent to the respondents from a very early stage. He drew attention to the fact that the documents had been referred to in the evidence filed in support of the application for a freezing order made against the respondents at the commencement of these proceedings and thus were disclosed within the meaning of CPR r.31.2 at that stage. Mr Steinfeld submitted that the respondents had been entitled to call for copies of the scripts from that stage pursuant to CPR r.31.14 but had never done so, which he said was illustrative of the reality – that the documents were not considered to be of significance so far as the respondents’ own case was concerned. He also maintained that the documents (and the bank statements) had been fully disclosed in the 2012 List and thus inspection of the scripts (and/or the bank statements) could have been sought at any stage thereafter. He accepted that the scripts and bank statements should have been supplied to Unified for inclusion in the disclosure database, and should have been included in the list served in purported compliance with the Default Order but maintains that the failure to do so was human error and the omission was neither serious nor significant in all the circumstances.

  2. The scripts were plainly important documents. They were used as the evidential basis for obtaining freezing relief against the respondents. That of itself is not sufficient to make this failure either serious or significant given that they had been disclosed before, referred to in the evidence filed in support of the application for a freezing order and that the respondents could have but did not seek inspection of these documents at any stage.

  3. What makes the failure to include the scripts serious is that, as the Court of Appeal pointed out, the scripts were secondary documents and the decision to omit from the disclosure process all the documents that supported the figures within the scripts was taken (and agreed to by the respondents’ solicitors) only on the basis that the scripts would be uploaded on to the database and disclosed in compliance with the Default Order. As Lewison LJ put it at Paragraph 38 of his judgment:

“Mr Gibbs knew about the importance of the scripts and knew that their disclosure had been promised to Mischon de Reya, yet he failed to check that they had been delivered to Unified. That they had not been uploaded was apparent on checking the database; and that should have alerted him to the need to find them. In my judgment he made no search for these critical documents let alone a reasonable search.”

Given this analysis, it is difficult to avoid the conclusion that the omission of the scripts from the list served in purported compliance with the Default Order was a serious breach of that order and of the obligation to carry out a reasonable search imposed by CPR Part 31.

  1. Although the scripts had been referred to in Schedule A to the 2012 List, Mr Crow maintained that that schedule was inaccurate and in truth was no more than an inventory of documents located at a particular place. Mr Crow says, and it is not disputed, that what is listed in Schedule A are 104 scripts not the 102 items that in the end were listed in the supplemental list served by the Liquidators in part to correct the omission of the scripts from the list they had served pursuant to the Default Order. That is a discrepancy that is not explained but in my judgment has no real impact on the issue that I am now concerned with. It is a discrepancy that could have been dealt with reasonably and proportionately by a short exchange of correspondence between solicitors. It is the sort of omission that ought to lead readily to relief from sanctions because it is neither serious nor significant.

  2. Whilst I am satisfied that the omission of the scripts was a serious breach of the default Order, it was not of itself significant. This is not a case of highly material documents that supported the respondents’ case or damaged the Liquidators’ case being concealed. The respondents and their solicitors were fully aware of the existence of the documents, their importance in the matrix of these proceedings and that other documents had been omitted on the basis that the scripts would be included within the list that was to be served in purported compliance with the Default Order. I conclude that the fact that the documents had not been called for by the respondents at an earlier stage is a relatively unimportant point. I say that because the documents were likely to be much more important to the steps that followed completion of discovery than to the steps in the proceedings that preceded it. What is more important is the point made by Lewison LJ in another context – namely that once the omission was pointed out it was the work of a few phone calls to establish what had happened and the omission was easily resolved, and was in the end resolved, by service of a supplemental list that was compliant so far as the scripts were concerned. Had this been the sole difficulty, then in my judgment whilst the omission was serious, it had limited if any real significance.

  3. Had this been the only breach complained of, and given that the omission was the result of human error alone, it is difficult to see why it would not have been appropriate in all the circumstances to grant conditional relief in relation to it – that is relief from sanction conditional upon the scripts (and bank statements) being uploaded onto the disclosure database and a supplemental list served by a short fixed future date coupled with a debarring sanction for default. However given my conclusion that I must also consider all other ways in which the Default Order was allegedly breached, I postpone reaching a final conclusion on that point until after reaching conclusions about the other alleged ways in which the Default Order was breached, not least because Mr Steinfeld conceded that if I conclude that the OCR Issue shows that there has been a serious and significant breach of the duty to make a reasonable search then the Liquidators could not seek relief in relation to that breach by reference to any findings I might make in relation to the second and third stages of the Denton enquiry.

  4. Non Compliance with CPR r.31.10

The essential facts relating to this issue are those set out in the judgment of Birss J that I have quoted from above. In summary, the respondents complain that the list served in purported compliance with the Default Order failed to comply with CPR r.31.10 by failing sufficiently to identify the documents disclosed because it described 44 documents as “Spreadsheet“, 60 documents as “Transcript“, 52 documents as “Accounts“, 29 documents as “Report” and 23 documents as “Table” in each case without further identification or description. It was submitted that some 555 of the documents disclosed had no file name, 421 had no author and 238 had no date. It was submitted that this breach was serious and significant because it meant that the list served in purported compliance with the Default Order was “… so vague as to be of no practical utility”.

  1. In response to this submission the Liquidators make two points. First, they say that they are entitled to rely on Practice Direction 31B either because on proper construction the exercise being undertaken is e disclosure because of the methodology adopted with at least the tacit agreement of the respondents’ solicitors and the approval of Henderson J or was so close to such an exercise that it ought to be considered by analogy with the requirements of Practice Direction 31B but secondly and in any event the problem has been grossly exaggerated because in practice access to all the documents can be obtained by the respondents’ solicitors at the click of a button on a mouse so as to enable a check to be undertaken of what the document consists of.

  2. Birss J rejected the submission that Practice Direction 31B was applicable to the exercise being undertaken by the Liquidators and I agree with him for the reasons that he gives at paragraph 56 of his judgment – that is:

“… the Practice Direction is concerned with electronic documents. A document which is held in the control of a party as a piece of paper is not an electronic document. Thus the Practice Direction would not ordinarily apply to the bulk of the disclosure in the Atrium case. “

That is so because as he said in the previous paragraph, when summarising a submission from leading counsel for the respondents:

“… the documents being disclosed were hard copy documents which were scanned into the database for the purposes of carrying out the disclosure exercise.”

Birss J considered that this position was altered by the agreement of the respondents’ solicitors to this methodology. On this point I find myself in respectful disagreement. In my judgment what was agreed was that disclosure would be given by scanning the hard copy documents and then identifying documents to be disclosed using a key word search methodology. They did not as far as I can see either expressly or impliedly agree to the application of Practice Direction 31B to the product of that activity. Birss J considered that such agreement was implicit from what the respondents’ solicitors say in their letter of 1 March 2013, that is:

“We understand and accept that it is your desired intention to upload the documents referred to in Schedules C & D (excluding the obviously irrelevant documents) onto an E-disclosure platform in order to perform such a search using the E-disclosure tools and presumably working within the terms of the PD 31B and treating these documents as documents in electronic form for that purpose.”

I am not convinced that this was a reference to anything other than the search methodology that was being adopted, particularly because the solicitors had said in an earlier part of the letter that:

” … although we welcome your approach to discuss with us the issues in relation to disclosure we cannot validate your clients’ disclosure or indeed any proposed approach to disclosure any more than the Court can, until your clients’ disclosure is complete and their list served.”

  1. In those circumstances, I am satisfied that the Liquidators failed strictly to comply with the obligations imposed by the Default Order – that is the obligation to provide to the respondents a list that was “… in compliance with the requirements set out in CPR 31.10 …“. In my judgment, for this requirement to be complied with there had to be provided a list that was prepared in conformity with the requirements imposed by paragraph 3 and in particular paragraph 3.2 of Practice Direction 31A. In my judgment the list served by the Liquidators did not comply with that requirement for the reasons given by the respondents.

  2. The question remains however whether this breach of the Default Order was serious and significant. Had the assertion made on behalf of the respondents that the list was of no practical utility been correct, then obviously the breach would have been both serious and significant. On this issue however I agree with Birss J’s comment that “… the argument that the lists are of no practical utility is seriously exaggerated …”. I am in agreement with that view for the following reasons.

  3. First, the impact of a failure of this sort will depend on how disclosure is given. If disclosure was given by list with inspection to follow either by the recipient of the list calling for photocopies or carrying out a physical inspection exercise at the premises of the disclosing party or its solicitors then the omission of the material required by paragraph 3 of Practice Direction 31A is potentially both serious and significant. Depending on the circumstances, it could affect very significantly the amount of time that had to be spent on inspection and thus the cost of that exercise for the inspecting party. It would mean that every document with a generic description would have to be examined to see whether it might be relevant or not, when, if properly described, at least some could safely have been ignored. Whilst this additional work might not take very long for each document it could easily become very significant if there were a large number of documents within each inadequately described category that had to be inspected for potential relevance. The time needed would be increased in the case of physical inspection by the need to locate each document. All of this could add significantly but unnecessarily to the costs incurred by the inspecting party. If inspection by provision of copies was being undertaken, it would be necessary to call for far more copy documents than might otherwise be the case (thereby adding significantly to the copying costs, which almost invariably is borne in the first instance by the party asking for the copies) as well has having to make time available to examine the copy documents concerned.

  4. In my judgment however, the impact of this sort of problem was significantly mitigated (though not eliminated) in this case by the fact that the inspection process was to be undertaken by giving the respondents’ solicitors access to a searchable database of scanned disclosed documents. Thus, for example, it would have been the work of a moment to pull up a transcript in order to identify who was being interviewed by or on behalf of the Liquidators and when. That sort of preliminary examination is all that we are here concerned with. The detailed review of any document thought to be relevant would be undertaken in any event. All that arises in this case is the effect of failing to provide the concise descriptive material in the list, which would enable a judgment to be made as to whether the document needed to be inspected at all. The use of a disclosure database eliminated much of the additional work that such a failure would have caused where inspection was to be conducted in the traditional manner.

  5. I conclude that the omission of an appropriate concise description in breach of the Order and CPR Part 31 was serious and significant because the failure to comply would have added significantly to the time needed to carry out, and therefore the cost of carrying out, a competent inspection exercise. The seriousness and significance of this omission was mitigated but not eliminated by the fact that inspection was to be carried out by access to a database of electronically scanned documents so that the task of initial examination of generically described documents could have been accomplished in less time than would have been the case if inspection had to be undertaken in a conventional manner. As with the Scripts Issue, I postpone considering whether relief ought to be granted until after I have considered the OCR Issue.

  6. The OCR Issue

In my judgment this is, at least potentially, the most serious and significant breach of the Order. Indeed, Mr Steinfeld accepted, entirely realistically, that if I am persuaded that the deficiencies of the OCR scanning were as alleged by the respondents then by definition there will have been a failure by the Liquidators to carry out a reasonable search at a level that is both serious and significant and that the application for relief could not succeed by reference to the second and third stages identified in Denton.

  1. I have described already the nature of the OCR Issue. I need not repeat or expand upon that description. Mr Steinfeld submitted that on a proper analysis of the disclosure process I could not conclude that there had been a breach of the obligation to make a reasonable search by reference to the OCR Issue. What follows is a summary of what Mr Steinfeld says the disclosure process consisted of. He argues that these submissions are supported by the witness statement of Mr Kilby, who describes himself in his witness statement in these proceedings as “one of the owners and Chief technology and Operations officer of Unified OS Ltd”. He maintains that once this is understood then it becomes clear that the OCR Issue is simply another attempt by the respondents to take advantage of a difficult situation in order to avoid this claim.

  2. Mr Steinfeld submits that the disclosure process undertaken by the Liquidators’ solicitors involved six stages. Stages 1 and 2 involved eliminating obviously irrelevant material from the material that had been disclosed originally as Schedules C and D to the 2012 List. This largely consisted of primary accounting material either consisting of or including the material that it had been agreed would not be disclosed because the scripts were to be disclosed. Stage 3 involved the delivery of what was left (180,302 documents) to Unified for scanning onto the master database. It was this material that was at this stage subjected to OCR scanning in order to enable key word searching then to take place.

  3. Stage 4 consisted of exclusionary word searching. Exclusionary word searching was designed to filter out irrelevant or privileged material by using words such as the names of the Liquidators’ solicitors. What was then left was subjected to the stage 5 process, which involved using inclusionary key words to identify potentially disclosable documents. The effect of these two processes says Mr Steinfeld was to reduce the 180,302 documents first to 94,247 documents (being those left after completion of stage 4) and then to 28,686 documents (being the documents left after stage 5). These documents were then examined in their non-OCR scanned format by lawyers for disclosability and 5,535 documents were identified for disclosure by this process. It follows from this that 53% of the documents that were scanned by Unified responded either to exclusionary or inclusionary key word searches but the balance of the documents (86,055 documents) did not respond to either of the key word searches.

  4. The only evidence that has been filed by the Liquidators going to this issue (which it should be remembered was first raised by the respondents well before the hearing of the appeal from Birss J’s order and was the subject of the Ladd v. Marshall application to the Court of Appeal that was stood over to be heard with the appeal but in the end was never determined because it was unnecessary for the reasons already identified) is the evidence of Mr Kilby.

  5. Mr Steinfeld concedes in his supplemental written submissions that Mr Kilby’s evidence does not “… provide an evidential basis … for reaching conclusions about the number of disclosable documents that might have been picked up by a manual (i.e. non key word led) disclosure review of the remaining …” 86,055 documents. He acknowledges that this is so because it is unknown whether the remaining documents were unresponsive to the key word searches because their scanned images were so compromised that keyword searches were ineffective or because the remaining documents did not contain any of the keywords. He maintains however that on proper analysis of the evidence, I can safely conclude that these remaining documents did not respond to key word searches because they did not contain key words not because the OCR conversion process failed to identify key words that were present.

  6. It is necessary that I now summarise what Mr Kilby says in his statement in order to test whether it supports Mr Steinfeld’s submissions. At paragraph 17 of his statement Mr Kilby confirms that the total number of “hardcopy” documents that were supplied to Unified was 180,302 and it was these documents that were first scanned and then had OCR technology applied to them. He distinguishes the hardcopy documents from some electronic documents being the contents of the hard drive of a lap top computer used by Mr English.

  7. Mr Kilby then says that of the 180,302 hardcopy documents “… 28,686 were responsive to the search terms.”. He does not describe the exclusionary and then inclusionary searches that Mr Steinfeld refers to in his submissions. Mr Kilby then says at paragraph 22 of his witness statement:

“22. The responsive hardcopy documents were then reviewed for relevance, however given the understanding that there were limitations to OCR technology …. Isadore Goldman undertook an individual review of a considerable number of additional hardcopy documents to assure themselves of the validity of the search responsiveness and the reasonableness of the search. In total 94,247 hardcopy documents were reviewed. That is to say over 3 times more hardcopy documents were reviewed than those which were returned as responsiveness to search terms. This represents over half of all the hard copy documents in the database.

23. Of the hardcopy documents that were reviewed, 5,535 were identified as relevant and falling to be disclosed. This represents 3% relevance rate of the hardcopy documents …”

  1. I am bound to say that I read this material as saying that 28,686 documents were responsive to the keyword search, that a further very substantial number of documents were reviewed by the Liquidators’ solicitors (whether the 94,247 included or excluded the 28,686 is not clear, though I suspect it is the former rather than the latter) and that in the aggregate 5,535 documents were identified as disclosable. What the evidence does not say is how many of the documents that were physically examined but which had not been identified by keyword search in fact contained key words, or how many of the 5,535 documents were revealed by the OCR search and how many were not but were found as a result of the individual review of the documents other than the 28,686 that responded to the keyword search. What the evidence also does not elaborate on is whether the additional documents selected for physical review that had not been identified by keyword search were selected at random or whether there was some method to the selection process. Assuming the selection process was a random one then it would seem a reasonable working hypothesis that up to 3% (or about 2,500) of the 86,055 documents that have not been examined may be disclosable.

  2. The omissions from Mr Kilby’s evidence are significant since if either the documents that had been physically inspected had been selected other than randomly and/or none of those physically inspected had either contained keywords or were disclosable, it can safely be assumed that Mr Kilby would have said so since that evidence would have provided powerful support for the Liquidators’ case concerning the OCR Issue. These omissions are all the more significant because (a) of the lapse of time that has occurred since this point first became in issue between the parties – that is significantly before the appeal was heard on 30 July 2014, (b) that no further evidence has been served going to this issue notwithstanding that the Liquidators have been represented throughout by experienced solicitors and leading and junior counsel, or, in the case of counsel, had the funding to enable them to obtain advice from leading or junior counsel, and (c) no application was made on behalf of the Liquidators either to adduce additional evidence going to these issues or for an adjournment to allow such evidence to be adduced. This last point is important because it is difficult to see how a short adjournment of a few days necessary to adduce evidence going to a point as critical as the one I am now considering could sensibly have been refused even allowing for the delay that has occurred in this case over the years and is all the more so in the light of the concession made by Mr Steinfeld that I was bound to refuse relief if I concluded that the OCR Issue demonstrated that there had been a serious and significant breach of the obligation to carry out a reasonable search.

  3. Notwithstanding these points, Mr Steinfeld submits that the onus rests on the respondents to establish the failure for which they contend. He submits that they have plainly failed to discharge the evidential burden that he says rests upon them. In summary his submissions as to why I should conclude that the failure for which the respondents contend has not been made out are as follows.

  4. First, he submits that the number of documents identified using keyword search of the OCR converted hardcopy documents was a 16% response rate whereas the electronic documents when searched using keywords produced a 15% response. This he submits shows that it is likely that the keywords search of the scanned hard copy documents was effective at identifying all the relevant documents within the 180,302 hardcopy documents.

  5. Aside from its speculative nature, there are a number of obvious difficulties about that submission.

i) First, it ignores the fact that the evidence is entirely silent as to how many of the 5,535 hardcopy documents that were ultimately disclosed were documents that were not part of the 28,686 documents that were revealed by the keyword search or how many of the physically examined documents that had not been identified by keyword search contained inclusionary keywords;

ii) Secondly, it ignores the fact that, if correct, it would mean that all the relevant documents had been identified in the documents physically inspected even though (a) 48% of the hardcopy documents have not been examined, (b) the documents that were physically examined appear to have been selected at random since the contrary is not asserted in Mr Kilby’s evidence and (c) inferentially at least some of the documents ultimately disclosed were identified by physical inspection but not by keyword search since the contrary is not asserted in Mr Kilby’s evidence either;

iii) Thirdly, I accept Mr Crow’s submission that this is a false point because it is not comparing like with like. The electronic documents all come from a lap top formerly belonging to Mr English the former finance director of the companies that postdates the insolvency of the companies and the retirement of the respondents. By definition therefore it is to be expected that the electronic database would provide fewer documents that respond to the relevant keywords not more and thus if any party is to obtain support from this it is the respondents not the Liquidators; and

iv) Finally this submission ignores the fact that if either no or no significant number of documents that contained keywords, or which were disclosable, had been disclosed by the physical examination of documents not identified by the keyword search process, then Mr Kilby could and would have said so, thus rendering unnecessary the speculation that this submission requires.

  1. Mr Steinfeld’s next submission relies on Mr Kilby’s evidence at paragraphs 26-28 of his witness statement concerning what Mr Kilby calls “Fuzzy Searching methodology“. Mr Kilby explains what this is at paragraph 9 of his statement in these terms:

“This technology will instruct the search engine to return documents that have an agreed level of discrepancy, at the character level, from the search terms. This can assist in returning documents that have characters incorrectly recognised by the OCR engine. This approach does not add substantially to cost or timescales as it is incumbent in most industry standard search engines. It is however likely to return many “false positives” – that is documents that do not respond to the search terms but that had character differences to the search terms not caused by the misinterpretation of characters by OCR. Fuzzy searching is extremely common in most litigation support scenarios.”

  1. Mr Kilby says at paragraphs 26-28 of his statement that a fuzzy search has been carried out. He describes that examination in these terms:

“On conducting a fuzzy word search test of +1 on the entire document population (hardcopy and electronic combined), it is apparent that the results are not only similar in terms of percentage hits (hardcopy 49% and electronic 57%) but that the electronic documents return a greater number of “hits” than the hardcopy. Again this indicates to me that the results of searching are not compromised by the quality of the OCR.”

He does not make clear when this test was carried out but inferentially it was not as part of the disclosure exercise but some time later, probably sometime after the issue I am now considering first surfaced. Mr Kilby says that misrecognition cannot be a factor with electronic documents, and thus that the closeness of return suggests that misrecognition is not an issue.

  1. This seems not to be addressing, or to be avoiding, the real issue because it fails to distinguish between those documents that were not identified by keyword search which were examined physically, and those that were not identified by keyword search and to date have not been examined physically. If (as by implication appears to be the case from Mr Kilby’s evidence) there were documents not identified by key word search that the physical examination of documents had revealed (a) to contain keywords and (b) to be disclosable, then I do not see how comparing the outcome of a fuzzy words search of all the hardcopy scanned documents on the one hand and all of the electronic documents that come from Mr English’s lap top on the other demonstrates that there is not an issue.

  2. It strikes me that the one thing that might have helped – a fuzzy word search of the documents that had not been physically examined at the time of the disclosure search compared with a similar search of the documents that were physically examined but which had not been identified as a result of the keyword search – has not taken place. Had such an examination taken place it might have been possible to arrive at some statistical conclusions as to the probability of disclosable documents being amongst the documents that have not been physically examined and which have not been identified by keyword searches, particularly if Mr Kilby had identified how many disclosable documents had been identified amongst those that had been physically examined but had not been identified by the keywords search.

  3. Thus in the result the court is being invited to accept that no documents or no significant numbers of documents or no significant number of documents containing keywords or which are disclosable would be revealed from amongst the documents that have not so far been examined physically on the basis of a statistical analysis that does not support such a conclusion.

  4. The final point that Mr Steinfeld makes is to argue that “the fact that the Respondents have not been able to point to any specific documents which they would have expected to see disclosed but which have not been …” strongly supports the inference that in truth there are no potentially disclosable documents within the 86,055 documents that have not been physically examined. In my judgment this point lacks reality.

  5. First I do not accept that the respondents can sensibly be expected to recall the existence of documents that have not been disclosed that relate to events that took place in 2003 to 2005. Aside from the effect of the passage of the years, the sheer volume of documentation that has been scanned makes this proposition inherently unlikely. The businesses operated by the companies were substantial as is apparent not merely from the volume of documentation generated but from the sums being claimed. The respondents are not alleged to have had any direct responsibility for the day to day activity that is alleged to give rise to this claim thereby making it even more unrealistic to suppose that they could identify particular documents that should have been but have not been disclosed.

  6. Secondly, the solicitors acting for the respondents have attempted to identify the categories of documents that they say are missing from the disclosure exercise. They did so in paragraph 14 of their letter to the Liquidators’ solicitors dated 5 July 2013, which was the formal response to the lists served on behalf of the Liquidators in purported compliance with the Default Order. Whilst not all the categories there referred to are material to the point I am now making two stand out. The first concerns operating manuals. These are likely to be of obvious potential significance to a case where a claimant seeks to infer knowledge of wrong doing on the part of a director and company secretary, neither of whom were involved in the activity that is said to have been wrongful. The sole response to this issue from the Liquidators’ solicitors was that the manuals “… did not appear in the pleaded case …” and so there was no obligation to disclose. It is unclear to me whether the Liquidators are asserting that the manuals were scanned and did respond to the keyword search or not.

  7. The other category concerns management accounts. These are known to exist because they are referred to in the evidence filed in support of the application for freezing relief at the start of these proceedings. However they did not appear anywhere in the list served in purported compliance with the Default Order. The response from the Liquidators’ solicitors was that all relevant management accounts had been disclosed. This response does not answer the point I am now considering. Unanswered, this problem suggests either that the management accounts were not scanned or the keyword search failed to pick up the management accounts.

  8. Thirdly, the reason why disclosure is important in a case like this is because the respondents need to have access to documentation that by definition they have not seen before – that is the material generated by those with direct responsibility for the wrongful acts that it is said gives rise to this claim. Since the respondents deny knowledge of what is relied on by the Liquidators, demonstrating that material generated by those with direct responsibility for what happened was not addressed, copied or disclosed to them at the time these events were occurring is of potential importance.

  9. Finally this point is an artificial one in this sense. By a letter dated 26 May 2013, the Liquidators’ solicitors informed the respondents solicitors that by 5 July 2013, access would be provided to the respondents’ solicitors to (a) the database containing the documents that appeared in the list to be served in compliance with the Default Order and (b) “…all non-listed documents, except those subject to privilege”. Not merely did the Liquidators’ solicitors not give access to the disclosed documents database on 5 July 2013 or at any time until a further order had been obtained from Norris J at the end of November 2013, but they have not given access to the non listed documents database at all. If and to the extent that there was any doubt about what the respondents’ solicitors expected, that was eliminated by their letter of 20 November 2013. All this notwithstanding, access has never been given to the non-listed documents. It is thus in my judgment unreal for the Liquidators to submit as they do by paragraph 6 of Mr Steinfeld’s supplemental Note that if the respondents “… were able to show actual problems with the disclosure (i.e. by identifying specific disclosable documents that have not been disclosed) they would have done so by now”.

  10. There is one point that Mr Steinfeld made that has given me pause for thought. It was that if in truth there are potentially disclosable documents within the 86,055 documents that have not been examined then their existence would have become apparent from the documents that have been disclosed because they would have been referred to in those documents. To an extent that is a fair point but in my judgment is not one that ultimately assists in resolving the issue I am now considering. Whilst non disclosed disclosable documents might be referred to in documents that have been disclosed, whether that is so is a matter of chance and does not sensibly lead to the conclusion that because there are no such references (assuming that to be the position) there are no disclosable documents within the 86,055 documents that have not been examined.

  11. I return now to the issue that at this stage I am considering – namely whether the defects referred to in this section of the judgment constitute a failure to carry out a reasonable search. The relevant factors are those identified in CPR r.31.7(2) that is

” (a) the number of documents involved;

(b) the nature and complexity of the proceedings;

(c) the ease and expense of retrieval of any particular document; and

(d) the significance of any document which is likely to be located during the search.”

  1. In considering this issue when this case was before the Court of Appeal, and in relation to the question whether there had been a reasonable search in relation to the scripts, Lewison LJ said in relation to (b):

“The claim is one for £50 million or thereabouts, based on allegations of dishonesty and fraud arising out of events that are a decade old. The liquidators had already been criticised for the previous failings and disclosure. There had been a number of slippages in the timetable and they were under the shadow of an Unless Order, so it was obviously important to make sure that the disclosure list was accurate. “

In relation to the other factors identified in CPR r.31.7(2):

i) As to (a), the volume of documents that have not been examined number 86,055. The evidence suggests by implication that a number of documents were not identified using the keywords search method and that a number of documents containing keywords which were disclosable were identified as a result of the physical examination of a number of documents that had not been identified using the keyword process and are not part of the 86,055 documents that have not been physically examined. The number of documents that should have but have not been disclosed is unknown. It might have been possible to extrapolate if the relevant information had been provided by Mr Kilby but it has not been;

ii) As to (c), it would not have been any more expensive or difficult to carry out a physical examination of the 86,055 documents than it was to carry out an examination of the 94,247 documents that were physically reviewed. It would have taken a significant amount of time, or more people, to carry out the exercise but once it was appreciated that there was a problem, the Liquidators’ solicitors could have informed the respondents’ solicitors of that fact and sought consent (subject to the approval of the court) to an extension or applied to the court for an extension. If the application had been made before the time for giving disclosure under the Default Order had passed it is likely that an extension necessary in order to overcome the effects of an unanticipated technical problem would have been granted – see CPR rr. 3.1(7) , Tibbles v. SIG Plc [2012] EWCA Civ 518 [2012] 1 WLR 2591 and Mitchell v. NGN [2013] EWCA Civ 1537[2014] 1 WLR 795, unaffected on this point by Denton. It is not suggested that the Liquidators did not have the financial resources to carry out this exercise. As I have said the Liquidators were and are the beneficiaries of an indemnity from HMRC. It is not suggested that the necessary human resources were not available or could not have been recruited. In any event, the task is not uniquely demanding. Whilst not common place it is certainly not unusual for disclosure exercises of similar magnitude to be carried out in commercial litigation. The scale of the task that had to be completed was obvious from well before the date when the Default Order was made; and

iii) As to (d), the significance of any document that is likely to be located during the search is necessarily unknown for the reasons already explained but given the nature of the allegations made by the Liquidators against the respondents – which invites the court to infer fraud on the part of each in circumstances where neither had a direct involvement with the wrongdoing that is the foundation of the claim – it is difficult to conclude that there is no realistic prospect of material helpful to the respondents’ defence are not contained within the 86,055 documents that have not been physically examined.

All this leads me to conclude that a reasonable search has not been carried out. It further leads me to conclude that this breach was both serious and significant. Mr Steinfeld accepted (correctly and realistically) that the only question that could arise in relation to this issue was whether in the circumstances there has been a failure to carry out a reasonable search for disclosable documentation. If the mechanism that had been devised has not worked correctly then there could not have been a reasonable search and that would be both serious and significant because even today the disclosure exercise will not have been completed correctly and what remained to be done would take significant further time.


The judge then went on to consider the second and third stage of the Denton test.

The Second and Third Stages

  1. I can take each of these stages in relation to each of the ways in which the Default order has been breached rather more quickly than my consideration of the first stage.

The OCR Issue – Stage 2

  1. Mr Steinfeld accepted (with respect both correctly and realistically) that if I concluded that there had been a breakdown of the mechanism for searching for relevant documentation that was serious and significant, then this could only be because inadequate care had been taken to ensure that the documents that were subjected to OCR scanning produced electronic versions of the documents that were of sufficient quality to enable a key word search to be undertaken. I agree but would go further. Once it became apparent to the Liquidators’ solicitors that the OCR scan was defective (as must have been the case at the latest once physical examination of the documents that had not been identified by keyword search started to reveal documents that contained keywords which were disclosable) then that was the time to increase the resources being made available to complete a physical examination of the documents that have so far not been physically examined and/or approach the respondents’ solicitors and/or the court for a variation to the terms of the Default Order relying on the principles that I refer to earlier in this judgment. That was not done. There is thus no satisfactory explanation for the failure that has occurred.

The OCR Issue – Stage 3

  1. Mr Steinfeld accepted that he would have great difficulty in these circumstances in persuading a court for stage three purposes that relief ought nonetheless to be granted. I agree for the following reasons, which I set out in summary terms in the light of the concession that has been made.

  2. The court is required to consider all the circumstances of the case. In my judgment those factors include the gravity of the allegations made against the respondents, the sums claimed, the length of time that has elapsed since the commencement of these proceedings, the fact that the Liquidators sought and obtained freezing orders at the commencement of these proceedings (that were4 later replaced by undertakings in equivalent terms) which of itself imposes an obligation to prosecute the claims efficiently, and the serial failures in the disclosure process that led to the Default Order being made. The reasons why Henderson J considered that it was necessary to impose the sanction that he imposed are highly material to an assessment of whether relief ought to be granted in relation to the failure to carry out a reasonable search that I am now considering. Those are set out in the quotation in paragraph 22 above but in particular include what Henderson J characterised as the lamentable history of the Liquidators’ disclosure exercise down to the date when he made the Default Order.

  3. There is no excuse for any of this, particularly having regard to the fact that throughout the Liquidators have been represented by experienced and well resourced firms of solicitors with substantial experience of major litigation that have been funded in effect at the tax payers’ expense throughout. Although in the course of his submissions in relation to the other failures relied on by the respondents, Mr Steinfeld suggested that there ought to be weighed in the balance in favour of granting relief the facts that the sums claimed were large and that the proceedings were brought primarily to make good losses suffered by the Crown, quite rightly Mr Steinfeld did not seek to rely on these points in relation to the issue that I am now considering. The reality is that proceedings brought on behalf of the Crown and funded by the tax payer ought to be conducted throughout to best practice standards. These proceedings have not been. The other ways in which the Liquidators have breached the default Order demonstrate a failure to approach this litigation appropriately. The failure to include the scripts in the list and the failure to ensure that categories of documents were appropriately described shows that even when operating in the shadow of the Default Order insufficient care was taken to ensure that appropriate resources were applied to ensuring compliance.

  4. In any event, the errors that have occurred to date show that as a result of the failures of the Liquidators, these proceedings have not been conducted efficiently or at proportionate cost. On the contrary they have been conducted in a manner that has exposed the respondents to unnecessary or substantially increased costs that could have been avoided. A perusal of the index to the hearing bundle shows that there have been no less than 10 Orders made in these proceedings leaving out of account consent orders. This is obviously disproportionate when it is remembered that disclosure has not yet been completed satisfactorily. A particular example of the way in which these proceedings have not been conducted efficiently or at proportionate cost is the failure of the Liquidators’ solicitors to give access to the disclosure database until ordered to do so by Norris J in November 2013. This caused delay, expense and also incidentally meant that further court time was avoidably taken up by these proceedings. Indeed, a similar point can be made in relation to the hearings leading to both the November Order and the Default Order and to the proceedings before Birss J and the Court of Appeal.

  5. In those circumstances I consider it is inappropriate to grant relief from the sanction imposed by the Default Order. That is sufficient to dispose of this application. In those circumstances, I set out my conclusions in relation to the remaining issues shortly below.

The Scripts Issue

  1. Had the only ways in which the Default Order been breached been the failure to include the scripts within the list served in purported compliance with the Default Order then I make clear that I would have granted relief from sanctions. As I have explained the failure was serious in the sense that a substantial number of primary documents had deliberately been omitted from disclosure on the basis that the scripts would be included within the disclosure database. The documents were not included in the list served in purported compliance with the Default Order as they should have been and there was no explanation for this failure other than error on the part of the Liquidators’ solicitors. However, the impact of this error was limited for the reasons that I have explained and easily rectifiable and as far as I can see was rectified by the service of a supplemental list.

The CPR r.31.10 Issue

  1. The failure to comply with CPR r.31.10 was serious and significant but for reasons already set out above was much less significant than it would have been the case had a conventional disclosure exercise been undertaken. So far as stage 2 is concerned, there is no exculpatory explanation offered for this breach. It is therefore necessary to turn to stage 3 and consider all the circumstances.

  2. With some hesitation I would have granted relief in respect of this failure as well but would have done so only on condition that the Liquidators used their best endeavours to supply the information missing in relation to the generically described documents by service of an amended list to be served by a short fixed future date with a sanction in similar terms to that imposed by the Default Order applying in the event of default. All the circumstances includes the ease with which the default can be remedied. The reality is that assuming the Liquidators complied with an Order in the terms I have mentioned (assuming that they have not complied with Birss J’s directions to similar effect) the problem could have been eliminated speedily and at minimal expense, and without any significant additional inconvenience or expense for the respondents. My hesitation in reaching this conclusion is caused by the factors summarised in paragraphs 94-96 above, by the serial failures that have occurred and the particular requirements to enforce efficient conduct of litigation at proportionate cost and compliance with rules and orders. Weighing those factors against the fact that this default is at a relatively low level of seriousness and significance, that it can be corrected easily and at minimum expense and that correction would have minimal adverse effects on either the respondents or the litigating public generally leads me to conclude, as I say with some hesitation, that on balance relief should be granted in relation to this failure.


  1. The application for relief fails and is dismissed.”


Earlier posts on this case: