HM Revenue & Customs v Benchdollar Ltd & Ors

This judgment follows the simultaneous trial of 14 claims by the Commissioners for Her Majesty’s Revenue and Customs (“the Revenue”) for unpaid National Insurance Contributions from the defendant employers. The claims were originally made in the Newcastle upon Tyne County Court but, because all of them raise common issues of limitation and estoppel, they have been transferred to the High Court and ordered to be tried together. The evidence before me suggests that these 14 cases represent the tip of a substantial iceberg, amounting in total to more than 70 cases, with an aggregate potential value to the Revenue in excess of £3 million, the outcomes of which are likely to turn on the same issues.

With effect from 6^th October 1985 the upper limit for an employer’s liability for secondary Class 1 National Insurance Contributions (“NICs”) was removed, with the consequence that, thereafter, employers faced a substantial increase in their NIC liabilities in relation in particular to higher paid employees. This led to the devising of numerous schemes for the avoidance or minimisation of employers’ NIC liability, the common feature of which was that they sought to take advantage of the exclusion from the definition of relevant earnings of payments in kind (see Regulation 19(1)(d) of the Social Security Contribution Regulations 1979).

Generally, these schemes featured a three-cornered arrangement under which the promoter of the scheme sold valuable assets to the employers, the employers transferred them as purported payments in kind to employees, and the employees turned them into cash by selling them back to the promoters. The types of asset used included platinum sponge, gold coins and book or trade debts. Thousands of employers used these schemes, and the 14 cases before me relate to their use by one or more of the defendant employers during the tax years 1994/95 through to 1997/98. Thereafter, amending legislation brought the use of these schemes to an end.

Issues as to an employer’s liability to pay NICs fall to be resolved (like issues as to liability for tax) by means of the statutory scheme, pursuant to which (at least after 1999) a decision as to liability fell to be made by an officer of the Revenue, subject to appeal to the General or Special Commissioners, and thereafter to the High Court on a point of law. Determinations and appeals occurred in respect of thousands of cases and, mainly as the result of the identification of appropriate test cases, the Revenue were, in the end, uniformly successful in establishing that the schemes were ineffective in eliminating or reducing the relevant employers’ NIC liabilities.

Claims by the Revenue for recovery of employers’ NICs, which are made by litigation in the ordinary way, are subject to a 6 year limitation period, pursuant to section 9 of the Limitation Act 1980 as actions “to recover any sum recoverable by virtue of any enactment”. Time runs from the date upon which each relevant payment of NICs became due. It is common ground that NICs become due and payable on the 15^th day after the end of the income tax month during which the relevant earnings were paid to the employee. Income tax months run from the 6^th of each calendar month to the 5^th day of the following month.

The running of time under section 9 in relation to claims against employers for NICs is not postponed by the existence of appeal proceedings before the General or Special Commissioners concerning liability. Rather, section 117A(5) of the Social Security Administration Act 1992 (as amended in 1999) provides that where an appeal against liability has been brought but not determined then:

“The court shall adjourn the [recovery] proceedings until such time as the final decision is known; and that decision shall be conclusive for the purpose of proceedings.”

It readily became apparent to the Revenue that issues as to liability for NICs in cases where employers were relying upon payment in kind schemes would be unlikely to be finally determined before the General or Special Commissioners, or on appeal to the High Court, before the expiry of the limitation period affecting recovery of all or at least part of the NICs in issue. Entirely understandably, the Revenue did not view with enthusiasm the prospect of having to commence and then have adjourned for substantial periods thousands of recovery claims all round the country, while entirely separate litigation was proceeding to resolve questions of liability. The process of issuing and then adjourning thousands of claims involved potential both for wasted costs and time-consuming administration, both for the Revenue and for the employers concerned. In most cases employers might be expected to pay arrears if their liability was established, so that the issue of recovery proceedings purely to ward off the limitation period would be likely to be a disproportionately expensive and cumbersome procedure.

Accordingly, the Revenue looked for some more cost-effective and proportionate way of preserving their potential recovery claims from becoming statute barred, pending the resolution of the liability issues. To a litigation lawyer, the obvious solution is what is generally known as a “tolling agreement”, namely a contract between the parties to the relevant dispute that the defendant will not raise a limitation defence to a claim started after the expiry of the limitation period, during a specific further period identified in the contract. Tolling agreements are a common feature in the resolution of commercial disputes, all the more so since the general recognition among the litigation community of the desirability of seeking to settle disputes, if at all possible, without recourse to court proceedings, which is a fundamental plank of the reforms to civil procedure introduced by Lord Woolf.

Unfortunately, the Revenue devised their own home-made scheme for the avoidance of the running of time under section 9 of the Limitation Act, based upon the erroneous notion (about which proper legal advice was not taken at the time) that advantage could be taken of section 29(5) of the Act by the obtaining from employers of written acknowledgments of the Revenue’s claim, on terms which were expressed not to constitute an acknowledgement of liability. The Revenue were also persuaded that by accepting payment by an employer of £1 on account of an NIC claim from an employer expressly denying liability, the same result could be achieved. While it is just about possible to discern from the language of section 29(5) of the Act how the Revenue came to form those mistaken views, it is now common ground that they were indeed erroneous, and that neither of the methods which I have described were on their face effective to postpone the running of time, so as to cause the right of recovery to be treated as having accrued on and not before the date of the relevant acknowledgement or payment.

For the purposes of the cases before me, counsel helpfully analysed the way in which the attempt to take advantage of section 29(5) manifested itself in the correspondence between the parties as falling into three broad types. I must describe each of them in a little detail.

In Type (1), the Revenue wrote to the employer (and/or its agent) referring to the fact that the issue as to liability for the NIC claimed might take a considerable time to resolve, and continued:

“Since we are bound in the collection of National Insurance arrears by way of the Limitation Act, we are required to obtain a protective writ, in order to protect the [relevant] tax year debt from becoming time barred. However, if you or your authorised agent were to provide either a signed acknowledgement of the debt or make a payment or part payment of the debt, the Limitation period would run from the date of the acknowledgement or payment/part payment and there would therefore be no requirement at this stage for us to obtain a writ.

If this is the course of action you would prefer to take, please sign and return the enclosed declaration as soon as possible.

It should be emphasised that this would be an acknowledgement of the debt, and not an admission of liability.

We look forward to hearing from you.”

The enclosed declaration was in the following form:

“FOR PURPOSE OF THE LIMITATIONS ACT 1980

Re: [name of employer]

Section 29(5) of the Limitations Act 1980 (fresh accrual of action on acknowledgement or part payment), states,

“Where any right of action has accrued to recover,

(a)

any debt or liquidated pecuniary claim

and the person liable or accountable for the claim acknowledges the claim or makes payment in respect of it, the right shall be treated as having accrued on and not before the date of the acknowledgement or payment.”

We, on behalf of/as the defendants, hereby acknowledge the claim of the Inland Revenue for National Insurance Contributions for the years [identified], for the purpose of S29[5] of the Limitations Act 1980.

This is not an admission of liability regarding the claim for National Insurance Contributions.

Signed (etc)”

Declarations in the form of the Acknowledgment which I have set out above were signed by or (subject to certain issues as to authority which I will have to resolve) on behalf of the employers in Cases 1, 3, 6, and 8 to 14, although Case 12 was preceded by a letter from the Revenue in a slightly different form from that which I have described.

Under Type (2), the Revenue wrote to the relevant employer (and/or its agent) in the following form:

“When arrears of National Insurance Contributions (NIC’s) are owed for a period approaching six years ago, it will be for the Inland Revenue to seek or commence proceedings in the Civil Court to safeguard the debt from the effect of the Limitations Act 1980.

Section 29(5) of the Limitations Act permits an “acknowledgement” of the NIC’s debt to be given without the need to take legal action to protect the arrears through the courts system. You will note the drafted acknowledgement is not an admission of liability.

If you are in agreement, to prevent the necessity for Civil Court action, an appropriate person should sign and return the enclosed forms of acknowledgement within the next 21 days.

If no response is received we intend to place the matter before the Civil Court for recovery of the debt. However, as there is a matter to be resolved outside this Court’s jurisdiction by your appeal, it is our intention to notify the Court of your appeal and to request immediate adjournment until your appeal has been cleared.”

Under Type (2), the accompanying form of acknowledgement took the following much shorter form:

“Re [relevant company- relevant asset]

For the purposes of Section 29(5) of the Limitation Act 1980 I acknowledge your claim for Class 1 Contributions amounting to [amount] for the [relevant period]. This is not an admission of liability.

Signed etc”

Acknowledgements of the Type (2) form were signed by or on behalf of the relevant employers in Cases 2, 4 and 5.

Type (3), of which Case 7 is the only example before me, was initiated by a letter from the employer’s agents DPC (a firm of Chartered Accountants and Registered Auditors) containing the following material terms:

“We enclose a copy of an extract from the Inland Revenue’s Enforcement Office Manual which states that the six year time limit starts afresh from the date the person liable (or their agent) acknowledges the debt in writing or makes part payment.

Accordingly, although we dispute the debt, we are conscious of your position and, without prejudice, we enclose our cheque in the sum of £1 in part payment.

We would therefore ask you to consider refraining from taking action at this time since, according to the enclosed, you will now have 6 years from this date in which to commence proceedings by which time we hope the matter will be finally resolved.”

The letter included an extract from the relevant Manual together with DPC’s cheque for £1.

The Revenue replied to DPC, acknowledging receipt, and continuing in the following terms:

“2.

The cheque for £1 has been accepted as acknowledgement by the company of the Class 1 debt and the interest accrued to date. Mr Redrup has been notified and will be copied this letter as confirmation.

3.

The company now have 6 years from the date the cheque was received in the Inland Revenue, 8 August 2000, to clear the Class 1 debt. I must, however, re-iterate that interest on the debt will continue to accrue at the daily rate previously advised until payment in full is received.”

Some of the Type (1) cases gave rise to additional correspondence beyond the common features which I have described. For example, in Case 1, the employer’s agents (DPC again) stated in the fax transmission form by which the signed acknowledgement was sent:

“Client anxious to avoid court proceedings. … Please confirm claim will not be issued.”

They repeated this request by letter and the Revenue replied:

“I can confirm that the County Court proceedings previously needed to protect this amount will now not need to be issued.

The Commission of Inland Revenue reserve the right to issue new proceedings to enforce collection of the debt should your client fail to provide payment for any amount held to be due as a result of this appeal.”

By way of further example, in each of Cases 3, 9, 10 and 12, the letters enclosing the signed acknowledgements all specifically requested that the Revenue ensure that no further action was taken against the relevant employer.

The only other relevant variation consists of the differently worded Revenue letter in Case 12, which was, in its material part, in the following terms:

“Until the appeal has been decided we are required to issue protective proceedings to the County Court appropriate to your business/company address. The requirement to issue a protective claim can be avoided by either: −

−completing the statement which acknowledges the NICs debt outstanding or,

−making a part payment towards the debt

In doing so the limitation period for the collection of the debt is effectively extended for another 6 years. The requirement to issue the Protective Claim will therefore be negated.

It must be stressed that failure to respond within 14 days will leave the department with no alternative but to proceed with the County Court Claim.”

The reason why none of the purported acknowledgements nor the £1 payment engaged section 29(5) of the Limitation Act is, as the Revenue now accept, that acknowledgement or part payment postpones the running of time only if it is an acknowledgement of liability or, in the case of a part payment, evidence of an admission of liability for the debt claimed: see generally Surrendra Overseas Ltd v. Government of Sri Lanka [1977] 1 WLR 565. In all the cases before me, the purported acknowledgements and part payments expressly denied liability for the debt claimed.

Nonetheless, the Revenue maintained that the written exchanges which I have described were nonetheless all sufficient to prevent the relevant employers now relying upon section 9 of the Limitation Act 1980. In its final form, as presented by Mr Philip Jones QC for the Revenue in opening, the case is advanced upon the basis of three alternative submissions. I shall take them in their logical order.

The first is that in each case the Revenue and the relevant employer contracted by correspondence on terms which included, first, a promise by the Revenue not to start recovery proceedings until the final determination of the employer’s appeal as to liability or, if earlier, just before the expiry of a further 6 years from the date of the relevant acknowledgement or part payment, and secondly, a promise by the relevant employer not to rely upon the Limitation Act in relation to any claim brought by the Revenue within that extended 6 year period. Mr Jones submitted that the promises which I have described are to be implied into the parties’ agreement, so as to give it business efficacy.

Mr Jones’ second (and main) submission was that the correspondence gave rise in each case to a unilateral contract in which the Revenue made the promise which I have described, in consideration of receiving the employer’s acknowledgement or part payment, but which included an express mutual averment by each of the parties to the other that the acknowledgement or (as the case may be) part payment was effective to trigger the commencement of a new 6 year limitation period pursuant to section 29(5) of the Act. The contract therefore gave rise in each case to an estoppel by convention founded and sufficiently supported by the contract itself, of a type originally known as estoppel by deed, but which has since escaped from that straitjacket.

Thirdly, and against the possibility that the Court might not identify any contract between the parties in the correspondence relied upon, Mr Jones submitted that the exchanges which I have described manifested a mutual understanding, expressed by each of the parties to the other, that signed acknowledgements or part payments did postpone the running of time pursuant to section 29(5), and that, coupled with the Revenue’s decision not thereafter to bring proceedings within the primary limitation period, and with the obtaining by the relevant employers of the benefit of not being subjected to civil proceedings while their statutory appeals remained to be determined, this was sufficient to give rise to an estoppel by convention of the non-contractual type.

Before addressing the issues raised by those submissions, it is necessary for me to say a little more about the common facts, most of which are ascertainable from the Revenue’s disclosed documents, together with a review of them in a witness statement by a Revenue Policy Officer, Mr Robin Wythes, who was cross-examined.

The decision to seek to preserve the Revenue’s NIC claims from becoming statute barred while appeals were pending by recourse to section 29(5) appears to have originated from a Mr Max Smith, supported by a Ms Vicky Passant, who according to Mr Wythes were at the time the two Revenue officers with principal responsibility for the achievement of this objective. Mr Smith was a member of the Revenue’s Business and Management Services Division, and Ms Passant was a Policy Adviser. It is evident that they were both oppressed by the cost, administrative effort and time which the Revenue would need to devote to protecting claims by means of issuing what they called protective writs, and they appear also to have been motivated by a concern that such a procedure would also expose relevant employers to similarly disproportionate cost and management time. After brief discussions with a trainee solicitor employed by the Revenue, one Michelle Turkie, who appears to have been persuaded to swallow her initial reservations, but not to give an unqualified assurance that it would work, the scheme was introduced in mid-2000. The earliest example of a relevant acknowledgement in the cases before me occurred in September 2000, and the latest in July 2001.

It did not take long before reservations about the effectiveness of the purported acknowledgements came to the attention of the Revenue. As early as 5^th October 2000 the Surrendra case was referred to the Revenue by Eversheds, solicitors, on behalf of an employer client, as the reason for their client’s refusal to sign an acknowledgement. That led, not to a review of the efficacy of the acknowledgement scheme, but to the issue of a protective claim against that employer. The same point was taken by Stephen Penny and Partners, a firm of accountants, on behalf of their clients by letter dated 16^th February 2001, which was promptly drawn to Mr Smith’s attention. By 22^nd March 2001 Mr Stephen Norman of the Revenue’s solicitor’s office had advised informally, on the basis of Surrendra, that there was “a strong possibility” that employers which had already signed acknowledgement letters would not be held responsible to pay the debts, if proceedings for recovery were started more than 6 years after the primary period.

Nonetheless the Revenue continued at Mr Smith’s direction to use the acknowledgement scheme, despite growing concerns as to its effectiveness, until on 9^th August 2001 Mr Norman provided detailed and unqualified written advice that the acknowledgement scheme was ineffective, fully endorsed by his superior Mr Thirkell.

By that date the primary limitation period in respect of claims arising from the 1994/95 tax year had expired, but numerous claims arising under the subsequent two tax years could still have been preserved from becoming statute barred by the prompt issue of protective claims in relation to them. This is what Mr Norman and Mr Thirkell advised should be done. But, at a meeting of the Technical Support Managers held in Gloucester on 10^th and 11^th September 2001, Mr Smith decided, with support of a Technical Adviser Mr Fernadez, that in relation to claims which had by then been made the subject of purported acknowledgements or part payments, no further action should be taken, whether or not they were by then time-barred, and that ‘protective writs’ should be issued only in respect of new cases. This was in the event the policy adopted, with the result that none of the claims (including several of those before me) in respect of which the primary limitation period had yet to expire, were made the subject of protective claim forms. Mr Wythes acknowledged in cross-examination that this appeared from the documents to have been the result of a positive decision, rather than of inadvertence, but his research into the history (in which he had not himself been personally involved) did not enable him to offer any conclusion as to the reasons for that decision. Neither the documents nor Mr Wythes’ research suggested that Mr Smith, or anyone else within the Revenue considered that the exchanges with the relevant employers placed the Revenue under any contractual restraint from issuing protective claims if they wished to do so. In any event, the result was that all the present claims were instituted well outside the primary limitation periods relating to them, so that they are all statute barred unless the defendant employers are disabled from relying upon section 9, either by contract or by estoppel.

The first question is whether the exchanges between the parties which I have described were contractual in nature. An essential feature of a contract is that one or both parties must have offered, and had accepted, a promise to do something or, as in the present case, to abstain from doing something. The common feature of both Mr Jones’ contractual submissions was that the Revenue promised, upon receipt of an acknowledgement or part payment, not to bring recovery proceedings against the relevant employer.

The express terms of the written exchanges between the parties do not disclose any such promise by the Revenue. Although differently worded, the initiating letters from the Revenue in each of Types (1) and (2) referred, first, to the Revenue’s duty to protect its claims for NICs from becoming time-barred and secondly to there being “no requirement” or no “necessity” for the Revenue to bring such claims in cases where the running of time was postponed by an acknowledgement or part payment. In Type (3), the relevant employer’s agent simply proffered a part payment and asked the Revenue “to consider refraining from taking action at this time since, … you will now have 6 years from this date in which to commence proceedings …”. In all three types, the common assumption was not that it was necessary for the Revenue to promise not to bring proceedings but that, upon receipt of an acknowledgement or part payment, the discharge of the Revenue’s duty to protect its NIC claims would no longer make the early bringing of proceedings necessary.

Mr Jones sought to meet this difficulty by submitting that if, after any of the three types of exchange, the Revenue had then immediately commenced recovery proceedings, the relevant employers would have been justifiably outraged. It was necessary he submitted either to imply a contract so as to give teeth to that sense of outrage, or more simply to imply a promise by the Revenue not to bring proceedings in order to give the assumed contract business efficacy. Similarly, to meet the difficulty that nothing in the exchange of correspondence specified the period during which the Revenue promised not to bring proceedings, it was necessary to spell that out by imaginative construction or implication so as to avoid an obviously intended contract from being found to be void for uncertainty. His solution was, as I have described, that the promise was not to bring proceedings until the final determination of the statutory appeal process, unless necessitated by the earlier imminent expiry of the extended 6 year period, running from the date of the receipt of the acknowledgement or part payment.

It is necessary to address that attractive submission by reference to first principles. Leaving aside the special cases of terms implied by law, and cases covered by the officious bystander principle, the governing criterion for the implication of terms is necessity. Similarly, in cases where parties have not contracted expressly, so that it is necessary for the claimant to assert an implied contract, the same necessity test applies: see Baird Textile Holdings Ltd v. Marks & Spencer plc [2001] EWCA Civ 274. As Mance LJ put it at paragraph 62:

“It could not be right to adopt a test of necessity when implying terms into a contract and a more relaxed test when implying a contract – which must itself have terms.”

Addressing the matter in terms of intention to create legal relations, Mance LJ said this at paragraph 61:

“An intention to create legal relations is normally presumed in the case of an express or apparent agreement satisfying the first requirement (an agreement on essentials with sufficient certainty to be enforceable): see Chitty on Contracts (28^th Ed.) Vol. 1 para.2-146. It is otherwise, when the case is that an implied contract falls to be inferred from parties’ conduct: Chitty, para.2-147. It is then for the party asserting such a contract to show the necessity for implying it. As Morison J said in his paragraph 12(1), if the parties would or might have acted as they did without any such contract, there is no necessity to imply any contract. It is merely putting the same point another way to say that no intention to make any such contract will then be inferred.”

Baird Textile Holdings also sheds valuable light on the differing ways in which the court addresses apparent uncertainty of terms, as between cases where the parties manifestly intended to contract, and cases where they did not. At paragraph 30 Morritt V-C said this:

“This is not a case in which, the parties having evidently sought to make a contract, the court seeks to uphold its validity by construing the terms to produce certainty. Rather it is a case in which the lack of certainty confirms the absence of any clear evidence of an intention to create legal relations…. It cannot be said, let alone with confidence, that the conduct of the parties is more consistent with the existence of the contract sought to be implied than with its absence. The implication of the alleged contract is not necessary to give business reality to the commercial relationship between M & S and Baird.”

In the present case, the starting point is in my judgment that in all three types, the Revenue and the relevant employer both assumed that, if the employer provided an acknowledgement in the requested form or a part payment, it would simply be unnecessary for the Revenue to have to bring recovery proceedings while, at the same time, pursuing its case in the statutory appeals process. It was not necessary for the relevant employer to extract a promise from the Revenue to that effect, still less to identify the intended duration of any such promise. The employer was content to assume that, armed with the necessary acknowledgement or part payment sufficient to postpone the running of time, the Revenue would choose, as a matter of common sense and in its own interests, not to commence proceedings.

There is therefore no necessity to imply a contract and, there being no clear evidence of an intention to contract, no necessity to imply the suggested promise by the Revenue, nor to struggle to find certainty as to the duration of the period during which it is suggested that the Revenue promised not to sue.

That analysis sufficiently disposes of the Revenue’s contractual submissions. For completeness I should add that it is even clearer that there is no basis for implying the suggested promise by the relevant employer (in the bilateral version of Mr Jones’ contractual analysis) not to rely upon the Limitation Act in relation to any claim brought within 6 years of the date of the acknowledgement or part payment. Since both parties assumed that the relevant employer would be unable to run a limitation defence in relation to such a claim, because they believed the acknowledgements and part payments to be effective under section 29(5), there is plainly no necessity for the implication of any such promise by the employers.

There remains therefore Mr Jones’ third submission, the question being whether any of the three types of exchange between the parties gave rise to an estoppel by convention, the convention being neither contained in nor supported by contract.

A concise summary of the relevant principles is to be found in the following passage from the speech of Lord Steyn in Republic of India v. India Steamship Co Ltd (No 2) [1998] AC 878 at 913E-G:

“It is settled that an estoppel by convention may arise where parties to a transaction act on an assumed state of facts or law, the assumption being either shared by them both or made by one and acquiesced in by the other. The effect of an estoppel by convention is to preclude a party from denying the assumed facts or law if it would be unjust to allow him to go back on the assumption: K Lokumal & Sons (London) Ltd v. Lotte Shipping Co Pte Ltd [1985] 2 Lloyd’s Rep 28; Norwegian American Cruises A/S v. Paul Mundy Ltd [1988] 2 Lloyd’s Rep 343; Treitel, The Law of Contract, 9^th ed. (1995), pp. 112-113. It is not enough that each of the two parties acts on an assumption not communicated to the other. But it was rightly accepted by counsel for both parties that a concluded agreement is not a requirement for an estoppel by convention.”

For present purposes, that general statement needs to be supplemented by the following passage from the judgment of Peter Gibson J in Hamel-Smith v. Pycroft & Jetsave Ltd (unrep) Feb 5^th 1987, cited with approval by Bingham LJ in the Norwegian American Cruises case (“The Vistafjord”) [1988] 2 Lloyd’s Rep 343, at 352, and further approved by the House of Lords in Hiscox v. Outhwaite [1992] 1 AC 562 at 575 per Lord Donaldson:

“Thus the court is not so rigid and inflexible as to insist on the parties being held to an assumed and incorrect state of fact or law when there is no injustice in allowing a party to resile therefrom (see, for example, Multon v. Cordell (1988) 277 Estates Gazette 198). Further, if the estoppel applies it will do so only “for the period of time and to the extent required by the equity which the estoppel has raised” (per Ralph Gibson LJ in Troop v. Gibson at p.1144). Thus, once a common assumption is revealed to be erroneous the estoppel would not apply to future dealings between the parties (per Purchas LJ in the same case at p.1144).”

Lord Donaldson summarised the point in Hiscox v. Outhwaite as follows:

“Once a common assumption is revealed to be erroneous, the estoppel will not apply to future dealings.”

The precise circumstances in which it will be regarded as unjust or unconscionable for a party to the common assumption subsequently to resile from it are, as counsel agreed, infinitely various. Nonetheless I was pressed with a number of authorities relevant to this question, and must deal with some of them. The earliest attempt at a comprehensive analysis of the question of injustice is to be found in a celebrated passage in the judgment of Dixon J in the High Court of Australia in Grundt v. The Great Boulder Proprietary Goldmines Ltd (1937) 59 CLR 641 at 675-6:

“Before anyone can be estopped, he must have played such a part in the adoption of the assumption that it would be unfair or unjust if he were left free to ignore it. But the law does not leave such a question of fairness or justice at large. It defines with more or less completeness the kinds of participation in the making or acceptance of the assumption that will suffice to preclude the party if the other requirements for an estoppel are satisfied. A brief statement of the recognised grounds of preclusion is contained in the reasons I gave in Thompson v. Palmer (1933) 49 CLR at page 547, and it is convenient to repeat it: - “whether a departure by a party from the assumption should be considered unjust and inadmissible depends on the part taken by him in occasioning its adoption by the other party. He may be required to abide by the assumption because it formed the conventional basis upon which the parties entered into contractual and other mutual relations, such as bailment; or because he has exercised against the other party rights which would exist only if the assumption were correct, …; or because knowing the mistake the other laboured under, he refrained from correcting him when it was his duty to do so; or because his imprudence, where care was required of him, was a proximate cause of the other party’s adopting and acting upon the faith of the assumption; or because he directly made representations upon which the other party founded the assumption.””

Counsel before me both invited me to treat that statement as a valuable guide, rather than an exhaustive code. They were right to do so.

Dixon J’s analysis suggests that injustice may be caused not merely by detrimental reliance upon the assumption, but also where the party sought to be estopped has gained some benefit from it. An example of a case in which benefit to the defendant rendered it unjust for him to depart from the shared assumption is Colchester Council v. Smith [1991] Ch 448: see per Ferris J at 496C to D, albeit that the case was (with the same outcome) subjected to an analysis not requiring recourse to estoppel in the Court of Appeal.

Much time was taken up with analysis of a passage from the judgment of Oliver LJ in Keen v. Holland [1984] 1 WLR 251. The primary basis of the decision in that case was that it was not possible to avoid the protection afforded by the Agricultural Holdings Act 1948 either by contract, or by an alleged convention estoppel contained in a contract. Thus, Oliver LJ said, at page 261D:

“The terms of section 2(1) are mandatory once the factual situation therein described exists, as it does here, and it cannot, as we think, be overridden by an estoppel even assuming that otherwise the conditions for an estoppel exist … having regard to the purpose of the Act of 1948, it cannot be said to be unconscionable for the tenant who is protected by it to rely upon the protection which the statute specifically confers upon him.”

It was common ground that no such protection is afforded by the Limitation Act 1980, the parties being free in effect to contract out of it, by substituting a longer or shorter limitation period for that provided by the Act.

Nonetheless Mr Bartley Jones QC for the defendant employers relied upon the following passage, at page 261F to 262B:

“That is sufficient to dispose of the argument but there seems to us to be other insuperable obstacles to a successful plea of estoppel. This is not strictly a case of the parties having established, by their construction of their agreement or their apprehension of its legal effect, a conventional basis upon which they have regulated their subsequent dealings as in the Amalgamated Investment case [1982] QB 84. The dealing alleged to give rise to the estoppel is the entry into the agreement itself in the belief that it would produce a particular legal result. In fact, for reasons which had nothing to do with the defendant, the plaintiffs got it wrong: and what Miss Williamson appears to us to be contending for is a much wider conventional estoppel than has yet been established by any authority, namely that where parties are shown to have had a common view about the legal effect of a contract into which they have entered and it is established that one of them would not to the other’s knowledge have entered into it if he had appreciated its true legal effect, they are, without more, estopped from asserting that the effect is otherwise than they originally supposed.

So broad a proposition cannot be deduced from the actual decision in the Amalgamated Investment case and although it may be supported on the basis of the very wide proposition of Lord Denning M.R. in the Amalgamated Investment case referred to above, it cannot, in our judgment, be right. If, for instance, the parties had been negligently advised by a solicitor that a yearly tenancy was not protected by the Act of 1948 and had entered into one accordingly, it would, we should have thought, be an impossible contention (quite apart from Johnson v. Moreton [1980] A.C.37) that the tenant was estopped from invoking the protection which the Act confers on such a tenancy.”

The precise meaning of that passage in Keen v. Holland has been addressed in a number of subsequent first instance decisions, and in academic writings. In the light of my conclusion that the parties’ exchanges in the cases before me were not contractual in nature, the Keen v. Holland dictum is not strictly applicable, since that was a case about an estoppel alleged to have been created within a contract. Nonetheless, since there was force in Mr Bartley Jones’ submission that if an estoppel could not be created in such circumstances in a contract, it would be strange if it could more easily be created by an exchange falling short of contract, I must briefly address those subsequent cases, in order to ascertain whether, for present purposes, Keen v. Holland affords the defendant employers any assistance.

In Hamed El Chiaty & Co v. The Thomas Cook Group Ltd (“the Nile Rhapsody”) [1992] 2 Lloyd’s Rep 399, Hirst J applied the Keen v. Holland principle without further explanation, in the context of an alleged jurisdiction clause in a commercial contract. In CP Holdings Ltd v. Dugdale (unrep) 22^nd May 1998 Park J said that the effect of the Keen v. Holland principle was that a shared incorrect assumption as to the true construction of an agreement could not be used, without more, to found an estoppel by convention in relation to it, the only remedy being rectification. In Wilson v. Truelove [2003] EWHC 750 (Ch) Mr Simon Berry QC treated the Keen v. Holland principle as applicable to a common assumption in relation to the meaning of an agreement reached individually by each of the parties to it, rather than mutually expressed to each other, and therefore shared.

In PW & Co v. Milton Gate |Investments Ltd [2004] Ch 142, at 185, Neuberger J expressed his understanding of the Keen v. Holland principle as follows, at paragraph 165:

“Mr Hodge’s suggestion that the words “without more” indicate that a convention could be established in the circumstances described by Oliver LJ provided that the essential additional element of unconscionability could be established, does not appear to me to accord naturally with the meaning of the words in that context. In my opinion, Oliver LJ was indicating that some course of dealing after the contract in question had been entered into was necessary.”

A similar conclusion is to be found in Ferris J’s analysis of the effect of Keen v. Holland in Colchester v. Smith (supra), at page 496.

Finally, the editors of Spencer Bower on The Law Relating to Estoppel by Representation (4^th Ed) express a similar view to that of Mr Berry QC in Wilson v Truelove (supra), in the following passage in paragraph VIII.5.1 at page 188:

“However, the element of an estoppel by convention (given the authorities which establish that such an estoppel may be as to law), that is lacking in their example, is communication by the party to be estopped to the estoppel raiser that he shares the relevant view, such as to render him accountable to the estoppel raiser for its correctness.”

This is developed in the following passage from paragraph VIII.5.7:

“It is this issue of the assumption of responsibility for the correctness of the relevant proposition that … the requirement that the party estopped actually (or as reasonably understood by the estoppel raiser) intended the estoppel raiser to act in reliance on the representation, is designed to address. In the context of estoppel by convention, the question here is whether the party estopped actually (or as reasonably understood by the estoppel raiser) intended the estoppel raiser to rely on the subscription of the party estopped to their common view (as opposed to each, keeping his own counsel, being responsible for his own view).”

In my judgment, the principles applicable to the assertion of an estoppel by convention arising out of non-contractual dealings, to be derived from Keen v. Holland, and the cases which comment upon it, are as follows:

Applying all those principles to the present facts, I consider that, first, there was a sufficient sharing of the common assumption (namely that acknowledgements of the claim or part payments without an admission of liability would start a new 6 year limitation period running under section 29(5)) in all three types of case before me. In Types (1) and (2) the Revenue unequivocally expressed its assumption in the letters which enclosed the form of acknowledgement requested. Each of the forms of acknowledgement (in due course signed by the relevant employers) specifically spelt out an assumption that the acknowledgement was intended to be, and was, sufficient for the purposes of section 29(5). As Mr Bartley Jones acknowledged, the case for a mutually shared and exchanged assumption was if anything stronger in Type (3), where the initiative came from the employer rather than from the Revenue.

Secondly, I consider that in each of the three types of case, the employers assumed a sufficient degree of responsibility for the shared assumption, in the sense that (in Types (1) and (2)) by returning acknowledgements specifically referring to section 29(5) as a positive response to the Revenue’s expressed wish that they should do what was necessary to start time running again, the employers sufficiently expressed a desire and intention that the Revenue should rely upon that assumption. In Type (3) the employer initiated the assumption.

Thirdly, I consider that, albeit to a subordinate extent, the Revenue did rely upon the shared assumption, in deciding thereafter not to issue protective recovery proceedings against those employers who had returned signed acknowledgements or made part payments. For this purpose I see no reason why reliance should mean anything different from that which it means in connection with the law of misrepresentation. It is not necessary that the Revenue should have relied solely or exclusively upon the employer’s expression of the shared assumption, but only that it should have been a significant part of the Revenue’s thinking, in the sense of having been actively present in the Revenue’s mind at the time when it decided not to bring immediate protective recovery proceedings. On the facts, I consider that the Revenue relied primarily upon the combination of the receipt of the signed acknowledgments and part payments, coupled with its own (ill-informed) view of the law. Nonetheless the fact that the employers who signed acknowledgements or made part payment did so in terms in which they expressed their own concurrent (but equally ill-informed) understanding as to their effect was in my judgment a material consideration for the Revenue in deciding not to bring recovery proceedings against them.

It became common ground between counsel that a decision by the Revenue not to protect their claim from becoming statute barred by bringing recovery proceedings against employers who provided acknowledgements or part payments was properly to be regarded as an aspect of their mutual dealings. In my judgment that is correct. A decision not to take an available step in dealings between parties may be as much an aspect of their common dealings as a decision to take that step. For example, in Amalgamated Investments & Property Co Ltd v. Texas Commerce International Bank Ltd [1982] QB 84, an important aspect of the detrimentally reliant conduct of the bank was its decision not to call in the loans or enforce other securities for them at a time when it believed that the plaintiffs were liable as guarantors.

All the other elements for an estoppel by convention being therefore established, I turn to the question of injustice or unconscionability, and to the question as to the effect upon any estoppel of the Revenue’s discovery in August 2001 that the acknowledgements and part payments were ineffective for the purposes of section 29(5).

In relation to claims for NICs arising from the 1994/95 tax year, the Revenue had in reliance upon the shared understanding that the acknowledgements and part payments were effective to postpone the running of time in fact allowed the primary limitation period for those claims to expire by, at the latest, the end of May 2001. By then, the Revenue had been given reason to think that the efficacy of those acknowledgements and part payments was in doubt, but did not know, prior to obtaining advice to that effect on 9^th August, that they were in fact ineffective. The Revenue therefore suffered a detriment in reliance upon the shared assumption, in relation to claims arising from the 94/95 tax year, from which it could not extricate itself in or after August 2001. At the same time, the employers subject to disputed claims for that tax year had enjoyed the full anticipated benefit of the Revenue’s reliance upon the assumed effectiveness of the acknowledgements and part payments, in the sense that they had not been visited with protective recovery proceedings in the meantime and, subject to certain isolated possible exceptions, continued to enjoy that benefit until the resolution of the statutory appeals.

In my judgment the combination of a detrimental reliance by the Revenue coupled with the obtaining by the relevant employers of the anticipated benefit of the Revenue’s reliance upon the shared assumption is sufficient to render it unfair and unjust for those employers now to advance a limitation defence in relation to NICs arising out of the 94/95 tax year. In short, in relation to NICs for that tax year, the employers are estopped by convention from asserting the ineffectiveness of the acknowledgements or part payments.

It may be, but the papers before the court do not enable this to be ascertained with any precision, that there are NICs arising from early receipts of benefits by employees in the 1995/96 tax year which would also qualify for protection by estoppel by having become statute barred, if time for the pursuit of those claims had already passed, by August 2001. There may also be claims which became statute barred between 9^th August 2001 (when the Revenue received advice that the acknowledgements were ineffective) and 10^th/11^th September 2001, when the Revenue decided not to take steps to protect claims from becoming statute barred. Again, the effect of a party becoming aware of the untruth of the shared assumption is not necessarily to kill the estoppel stone dead there and then. The reliant party is commonly afforded a limited time in which to protect itself from the consequences of a discovery of the true legal or factual position: see for example Arden LJ’s analysis in London Borough of Hillingdon v. ARC Ltd [2000] EWCA Civ 191, at paragraph 64, where she said:

“If a common assumption existed up to this moment in time, ARC were entitled to a reasonable time to re-act to the disappearance of their assumption. In the circumstances of this claim that period might be measured in weeks rather than days, but not in months.”

On the facts of the present cases, a reasonable time for the Revenue to react did include the period between 9^th August and 10^th/11^th September 2001. There is no need artificially to extend the period beyond that date, since by then the Revenue had made a conscious decision not to take steps to protect itself in relation to any claims for which the primary limitation period was still running at that later date.

In relation to claims which only became statute barred after 10^th/11^th September 2001, the analysis of the question of injustice or unconscionability assumes an altogether different aspect. The Revenue could have taken steps to protect those claims from becoming statute barred either by issuing protective claim forms or by seeking to make effective tolling agreements with the relevant employers. Those NIC claims did not therefore become statute barred by virtue of the Revenue’s reliance upon the shared assumption.

Against that, as Mr Jones submitted, is the fact that the Revenue’s decision in September 2001 to take no further steps meant that employers who had returned acknowledgments or made part payments continued to enjoy in full the benefit which they anticipated that they would receive, namely not being made the subject of protective recovery proceedings while liability to pay the NICs remained in issue in the statutory appeals process. Mr Jones submitted that, in cases where the receipt of the anticipated benefit made it unjust for the recipient to resile from the shared assumption, the fact that a case in detrimental reliance could not also be established was neither here nor there.

In my judgment, questions of injustice or unconscionability must be addressed in the round, and due weight given to all relevant factors. In that context the following observations of Neuberger J in the PW & Co v. Milton Gate Investments case, (supra) at paragraph 222 are of real force:

“In almost all cases, such unconscionability must be based on the prejudice which would be caused to the claimant if the strict legal position applied. As I see it, the claimant must also establish that that prejudice arises from its reliance upon the convention. In other words, the court generally must be satisfied that (a) the claimant will suffer real prejudice, and (b) the prejudice arises from its reliance (upon) the convention. It should be emphasised that, even if the claimant satisfies these criteria, there may be no estoppel, because there may be other, more powerful, factors pointing the other way.”

In the present cases, in relation to claims which became statute barred only after September 2001, the prejudice occasioned to the Revenue by the loss of the ability to pursue those claims was in no sense reliant upon the convention. The Revenue had been advised in terms that the acknowledgements and part payments did not have the effect assumed by the convention, and knew that, nonetheless, those claims could still be saved by the taking of prompt protective steps.

While it is tempting to speculate that the Revenue might have held its hand in relation to those claims out of some perception that the exchanges of correspondence which I have described would have made it unconscionable for it to do otherwise, I consider it quite wrong to draw any inference to that effect. Mr Smith was, so I was informed, still employed by the Revenue at the time of the trial. He was not called to give evidence, and it did not appear that Mr Wythes even asked Mr Smith what had been the reason for the decision in September 2001 not to take any protective steps. Where a party declines without explanation to call an apparently available witness who could have given favourable evidence on a relevant matter, it is not for the court to exercise a sympathetic imagination in filling the consequential void.

While I recognise that the continued enjoyment by employers of the benefit (i.e. not be sued) which they sought to obtain from providing the acknowledgements and part payments on the shared assumption as to their effectiveness is a relevant element to be considered on the question whether it would now be unjust, unfair or unconscionable for them to rely upon the true legal position under section 29(5), it is in my judgment a relatively modest factor, in particular when compared with the Revenue’s decision not to protect claims which it knew could have been protected by other means. In relation to those claims, the Revenue was, quite simply, the author of its own misfortune.

It follows in my judgment that it is not on balance unfair, unjust or unconscionable for employers facing claims which were not statute barred as at 11^th September 2001 to assert a limitation defence to those claims, arising out of the Revenue’s decision not to protect them once aware of the true legal position.

In the light of the conclusion to which I have come on the common issues, only one particular issue affecting some, but not all defendants, needs to be considered separately. That is the issue, raised by the defendants in Cases 6, 7, 8, 11 and 14, in which each of them deny that the acknowledgement or (as in Case 7) part payment was made with that defendant’s authority. In each of those cases, the acknowledgements and part payment were signed or made by DPC, purportedly on behalf of the relevant employer. In none of those cases do the Revenue hold letters of authority in terms sufficient to confer ostensible authority on DPC for that purpose. In each case, by its defence and evidence in support, the relevant defendant denies having conferred actual authority on DPC for that purpose.

By the time of closing speeches, the issue had come down to a simple question of fact in each case. Did DPC obtain their relevant client’s authority for the making of the acknowledgements or part payment in each case?

I shall take Cases 6, 7 and 8 together. In Cases 6 and 7 the employer was Fletcher Bickerton (Holdings) Ltd. In Case 8 the employer was Fletcher Bickerton Manufacturing Co Ltd. They were all companies beneficially owned by the Fletcher family, and Mr Malcolm Paul Fletcher was a major shareholder in Fletcher Bickerton (Holdings) Ltd which (I assume) wholly owned Fletcher Bickerton Manufacturing Co Ltd. The relevant acknowledgements in Cases 6 and 8 were provided by DPC on 6^th and 28^th June 2001. The part payment in Case 7 was made on 3^rd August 2000, by a cheque drawn on a DPC account.

Mr Fletcher told me that DPC had been retained by the Fletcher Bickerton group of companies for many years in connection with accountancy and auditing matters and regularly dealt with Companies House on the group’s behalf. He said that the extent of its authority was not to be found in any document, but depended on custom and practice. He said that he was not a director of any of the companies at the time, but only a major shareholder, his wife, son and daughter being the directors. Nonetheless he said that he was a signatory on company cheques.

The gist of his evidence was that he had no recollection of having authorised either of the acknowledgements or the part payment, or of having been told by the director members of his family that they had done so, and he regarded it as unlikely that this would have occurred without him being informed. He said that it was possible that he had authorised or been informed about the transactions, but thought it unlikely. He said that he had not complained to DPC about their having exceeded their authority, and his explanation was that this issue was a relatively new one for him.

When it was pointed out that the plea of lack of authority had been set out in Fletcher Bickerton (Holdings) Ltd’s Defence in Case 6, signed with a statement of truth by the company’s solicitors and dated 18^th April 2008, his evidence was that he could not recall this Defence being lodged or authorised at the time and that, if he had given authority, he could not recall having done so.

As Mr Jones rightly submitted, the natural probability in the case of a step taken by a reputable firm of accountants for companies in a group by which they had been retained for many years is that the necessary authority for the making of the acknowledgements and part payment was in fact obtained. I do not consider that Mr Fletcher’s evidence, which amounted in substance to no more than having no recollection of giving or being told about the giving of the relevant authority, came anywhere near to disturbing that natural probability. His ability to forget relevant matters was as ample in relation to what occurred a little more than a year ago as it was in relation to matters which occurred in 2000 and 2001. It was, I am afraid, a classic example of the fallacy that because a witness cannot recall something, it did not happen.

In my judgment, the giving of an authority to bring about a postponement in the running of time sufficient to avoid the Revenue from having to bring an otherwise wholly unnecessary protective claim is not the sort of event likely to stick in a businessman’s mind for 8 years. It was, I find, simply something which Mr Fletcher or (more probably) the one or more of the director members of his family thought it sensible to approve, no doubt on advice from DPC that the companies had something to gain and nothing to lose from going along with the Revenue’s suggestion (in relation to the acknowledgements) and in proposing the part payment, designed to achieve the same effect.

Accordingly, I find that in fact DPC had the requisite authority to make the acknowledgements and part payment in Cases 6, 7 and 8.

I have reached the same conclusion in relation to Case 11, concerning McEvoy Foods International Ltd, where the acknowledgement was given on 6^th June 2001 by DPC, and in relation to which Mr Howard John McEvoy gave evidence that, in effect, he had no recollection of providing the requisite authority, either orally or in writing.

He said that DPC had acted for his company from about 1995 or 1996, preparing audited accounts, and advising on all matters financial and fiscal in relation to the company. He said that they continued to act for the company at the time of the trial. His evidence (which must have been derived from someone at DPC) was that the person who had in fact signed the relevant acknowledgement (and who appears from the documents to have been a Ms Vicky Hulse) had subsequently left the firm. He said that she had been supervised by the senior partner Mr Kane, or a colleague of his, and that she had never had any direct dealings with Mr McEvoy or his company. He readily acknowledged that he gained his information about what had happened from Mr Kane, but said in cross-examination that Mr Kane had not at any time said to him that there had been a serious mistake made by DPC. He said that the matter had been the subject of a discussion between him and Mr Kane, but not a formal complaint.

In my judgment, that evidence fails to displace the natural probability that DPC duly obtained the authority to make the acknowledgements in question. In particular I consider that the absence of any admission by Mr Kane that his firm had acted without authority (which would have been a serious mistake) makes it most unlikely that authority had not been obtained. Again, as with Mr Fletcher, the obvious explanation is that a minor event which occurred many years ago had simply been forgotten by Mr McEvoy by the time that the matter became the subject of contention in hostile litigation.

In relation to Case 14, concerning Waverley Care Homes Ltd, the acknowledgment was provided by DPC on 28^th June 2001, and the evidence as to lack of authority came from a Mr Richard Britten, who was a director of the company at the material time. He said that his company had retained DPC about 15 years ago, that they dealt with annual tax returns, Companies House filing requirements, corporation tax calculations and payments and other accounting matters of a similar nature. He had no recollection of his company appealing the NIC claim, and thought that, because of this, DPC might even have filed the statutory appeal without his authority. He said that, neither in relation to the statutory appeal nor the acknowledgement had he made any complaint to DPC, but just left the matter in abeyance until the outcome of these proceedings.

He suggested, rather unrealistically in my judgment, that if the Revenue’s suggested choice between providing the acknowledgement or having protective proceedings brought immediately against the company had been referred to him he would have taken legal advice. Apart from that, his evidence simply was that he had no recollection that the requisite authority had been provided at the time.

I did not find Mr Britten’s evidence on this point any more convincing than that of Mr Fletcher or Mr McEvoy. By that I do not mean that any of them were seeking actively to mislead the court. I have little doubt that they could not recollect, in 2009, supplying authority for this minor matter in 2001. In my judgment, on the balance of probabilities, the requisite authority was given.

It became apparent early in the trial that the parties were seeking, at this stage, declaratory relief or the determination of preliminary issues, rather than a final determination of all matters of liability and quantum in each of the 14 cases. This was in part because documents relevant to issues affecting some of the parties were still unavailable. I shall therefore express my conclusions in the following general terms, and invite counsel to agree, or make submissions, as to an appropriate form of order.

My conclusions are therefore as follows: