England and Wales High Court (Queen's Bench Division) Decisions

IN THE HIGH COURT OF JUSTICE
QUEEN'S BENCH DIVISION

		Royal Courts of Justice 7 Rolls Building, Fetter Lane London, EC4A 1NL
		04/07/2014

The Hon. Mr Justice Popplewell :

Introduction

1. This claim arises out of a share purchase agreement dated 1 July 2010 ("the SPA"), under which the Claimant ("Ageas") acquired from the First Defendant, Kwik-Fit (GB) Limited ("KFGB"), the entire issued share capital in Kwik-Fit Insurance Services Limited ("KFIS") for a consideration of £214.75m. Under the SPA, KFGB warranted the truth, fairness, accuracy and compliance with relevant accounting standards of the KFIS accounts which were relied upon by Ageas for the purpose of valuing KFIS and its subsidiaries, subject to a £5m cap on liability. As is now common ground, KFGB breached those warranties. At the same time as entering into the SPA, Ageas took out a Warranty and Indemnity insurance policy with the Second Defendant ("AIG"), then called Chartis Insurance UK Limited, to protect it against losses resulting from breaches of warranty in excess of the £5m cap under the SPA. The measure of loss recoverable under the policy is that which would be recoverable from KFGB under the SPA, subject to the £5m excess.
2. The claim against KFGB has been settled. AIG now admits liability under the policy. Ageas claims that the proper quantum of its warranty claim is £17,635,000, giving rise to a claim under the policy of £12.635m; AIG contends that the proper quantum of the warranty claim is £8,792,000 giving rise to a claim under the policy of £3.792m. The quantum issues narrowed considerably in the course of the hearing, enabling me to set out relatively briefly the background necessary to address the issues which remain.

The background

3. Ageas (formerly named Fortis (UK) Limited) is the holding company for the group's insurance operations in the United Kingdom. At all material times Ageas has predominately been engaged in providing personal lines insurance through a network of intermediary brokers, focusing mainly on the underwriting of motor insurance, but also providing household insurance, travel insurance and a small amount of commercial insurance.
4. KFGB is a well-known company providing automotive repair, testing and maintenance services, together with associated services. In addition to these services, KFGB developed an insurance broking business under the Kwik-Fit brand, through which it offered motor and household insurance (with various forms of add-on cover, such as breakdown cover and legal expenses insurance). This insurance services division of the Kwik-Fit group comprised KFIS and its subsidiaries, Express Insurance Services Limited ("EIS") and The Green Insurance Company Limited ("TGIC"). Prior to the acquisition by Ageas, KFIS was a wholly owned subsidiary of KFGB and EIS and TGIC were wholly owned subsidiaries of KFIS. I shall refer to the three insurance companies collectively as KFFS, as the parties did at the time and during the hearing.
5. KFIS was established in 1995 to sell motor insurance products adopting a strategy to take advantage of the customer service calls offered to Kwik-Fit customers following a visit to a branch. Thereafter KFIS's business model evolved to address changes in the insurance market place. There were significant changes in the few years prior to the acquisition, including in particular a rapid rise in the use of comparison websites (or in the jargon of the business "internet aggregators as a distribution channel"). Over the three years prior to the acquisition, the amount of new motor business in the UK sourced through aggregator distribution channels rose from about 30% to about 65%. This was Kwik-Fit's reason for selling the business. In the five years prior to the acquisition the proportion of new business which came from exploiting the automotive customer base dropped from about 80% to about 10%. At the time of the acquisition, the market, and KFFS's position in it, was one of recent and rapid change, which made valuing the business based on projections of future performance all the more difficult.
6. An important aspect of the KFFS business model was its reliance on an invoice discounting agreement with Barclays Bank to provide it with cash flow. KFIS and TGIC sold insurance policies to customers on behalf of a number of different underwriters. Premiums could either be paid in full at the start of their policy or by monthly direct debit instalments. KFFS was usually required to pay the full premium to the underwriter within 30 days of the commencement of the policy. To cover the funding shortfall on instalment policies, KFFS entered into Debt Purchase Facility Agreements ("DPFAs") with Barclays Bank plc. The precise terms and operation of the DPFAs are no longer in issue in these proceedings; in summary they were invoice discounting arrangements secured on the stream of future income which enabled KFFS to obtain cash from the Bank in respect of policies which were to be paid for by direct debit instalments. As with all invoice discounting schemes this had cash flow advantages for KFFS. It also gave KFFS the ability to offer financing terms to its customers and to generate additional income from this activity. This finance income was a significant element in KFFS's revenue, increasing from about 24% of the total revenue of the business in 2007 to about 44% in 2010.
7. Ageas was alerted to the proposed sale of KFFS when it received an Information Memorandum from the seller's investment banking advisors, Credit Suisse, on or around 12 March 2010. Ageas regarded it as a potentially attractive investment opportunity, and put together a team of senior personnel from its own staff. Within this team there was considerable past experience of acquisitions and a wealth of knowledge of the operation of the UK insurance market, including both underwriting and intermediary activities. Ageas also retained Deutsche Bank and Deloittes to advise them in relation to the purchase, completing an impressive M&A team.
8. After conducting due diligence and a valuation exercise (see below), on 6 June 2010 the Ageas M&A team submitted a paper on the proposed acquisition of KFFS to the Management Committee for approval. On 14 June 2010, the Ageas M&A team submitted an updated paper to the Holdings Board. On 15 June 2010, the Holdings Board approved an offer for KFFS in the range of £210m to £220m. On or around 22 June 2010, Ageas made its "final" offer of £215m, which was revised down to £214.75m in order to take into account an additional liability incurred by KFIS. This was accepted and on 1 July 2010 the parties entered into the SPA, whereby KFGB agreed to sell to Ageas the entire issued share capital in KFIS for £214.75m. Completion took place on 2 August 2010 and the purchase price was paid in full by Ageas to KFGB.
9. For the purposes of assessing an appropriate acquisition price, the Ageas team adopted and relied on a discounted cash flow ("DCF") analysis, as is common in valuing companies. As is well known, DCF modelling involves using projections of the future free cash flow of a company (the amount of cash not required for operations or reinvestment) discounted to arrive at a present day value of a future income stream. The DCF model created by Deutsche Bank and used by Ageas was relatively sophisticated, and involved starting with the projections put forward by KFFS, identifying their key performance indicators and the assumptions which drove them, adjusting those KPIs and assumptions for Ageas' own views of the future for the business and the market, and applying the revised assumptions to the model taking the historical figures from the warranted accounts. Cross checks were also made by reference to other common methods of company valuation, in particular by considering price/earnings ratios and other market transactions. The final version of the DCF model used the figures in the 2010 management accounts and produced an ultimate valuation of £214.5m.
10. By clause 7.1 and Schedule 3 of the SPA, KFGB gave warranties that the audited annual accounts of KFIS and its subsidiaries for the calendar years 2007 to 2009 were prepared in accordance with GAAP, and gave a true and fair view of the assets, liabilities and financial position of the relevant companies; and that the management accounts for the first five months of 2010 had been prepared on a consistent basis with the annual accounts and did not materially misstate assets or liability or items of income and expenditure or profits or losses for that period.
11. These warranties were breached by the treatment in the accounts of two aspects of bad debt. One was "the Bad Debt Percentage", a figure used monthly for the purposes of the DPFAs, the result of which was that the revenue and assets were understated in the accounts. The other was "Time on Cover Bad Debt" ("TOCBD"), the result of which was that the revenue and assets were overstated. Overall, the TOCBD errors were greater than the Bad Debt Percentage errors, with the result that the net effect of the errors was an overstatement of revenue and assets. The net overstatement of revenue for 2009 was £1,404,639 and that for the 5 months January to May 2010 was £905,157.
12. It is not necessary to explain the Bad Debt Percentage error in order to resolve the remaining issues. At the heart of the dispute is TOCBD. This is bad debt which arises out of customers remaining on cover for a period for which they have not paid the premium. This can only arise for customers who have taken out instalment contracts: there is no risk of such bad debt on policies for which the annual premium is paid in full in advance. TOCBD was calculated monthly for the purposes of the DPFAs. The error was in thinking that it was a bad debt risk assumed by Barclays, whereas it was in fact to be borne by KFFS.
13. Although the errors in relation to TOCBD resulted in the overstatement of the profits and assets in the warranted accounts, there was no discrete figure for TOCBD which was identified or separated out as a distinct line or entry in those accounts. TOCBD was recorded by KFFS in management accounting ledgers from which the management and annual accounts were compiled. The TOCBD figures were combined with other individual ledger entries to give a single balance for a particular line item in the warranted accounts. The balance of the numerous daily entries in the ledger (including TOCBD) formed the basis of the "Commission Write Off" line in the 2010 management accounts and the "Turnover" line in the annual accounts. In this way the errors were in figures embedded within the accounts. If the errors had been corrected in the warranted accounts, the figures for profit and net assets would have been reduced but there would still have been no identifiable item in those accounts for TOCBD.

The issues

14. The measure of loss for breach of warranty in a share sale agreement is the difference between the value of the shares as warranted and the true value of the shares: Lion Nathan Ltd. v C-C Bottlers Ltd [1996] 1 WLR 1438 1441F-H; Eastgate Group Ltd v Lindsey Morden Group Inc [2002] 1 WLR 1446. Each involves valuing the company. It was common ground that the value of the shares as warranted was the price which Ageas paid, namely £214.75m, which was very close to the figure produced by the DCF model using Ageas' assumptions and key performance indicators. The dispute was what value was to be attributed to the error in relation to TOCBD.
15. Ageas' case, based on the evidence of its expert accountant Mr Mesher, was that adjustment for the understatement of TOCBD reduced the value of the company by £21.9m, which involved a reduction in value, net of the Bad Debt Percentage adjustment, of £17.635m. The figure was calculated as follows. Mr Mesher identified the actual TOCBD incurred for the five months covered by the 2010 management accounts, which were the latest trading figures available for the period prior to the sale. He grossed it up to give an annual figure for 2010 of £2.866m. He then applied that figure to all the years in the DCF model, by adjusting the commission write off cell, which flowed through the model to give the adjusted value. He was therefore valuing the company by using the DCF model used by Ageas but projecting that TOCBD would remain constant to 2014 at the absolute level at which it had been for the first 5 months of 2010. The model projected that after 2014 all revenue assumptions would remain constant, so that this involved assuming TOCBD at 2010 levels until 2020.
16. AIG's case, in its final iteration at the conclusion of the trial, was that this overcompensated Ageas, because the actual TOCBD figures between June 2010 and April 2014 can be seen to be much lower than those assumed by Mr Mesher. It can now be seen that the rates of TOCBD in mid 2010 were at a level which represented an historical peak and fell back thereafter. AIG's expert, Mr Mainz, produced a calculation in the course of Mr Mesher's cross examination which identified the relevant adjustment based on actual TOCBD figures to be £8.792m (after adjustment for Bad Debt Percentage). He too made his recalculation up to 2014 and assumed, as had the model, constant levels thereafter. For the period 2010 to 2014 he did not simply apply the actual TOCBD figures, because he recognised that the actual level of TOCBD had been influenced by the fact that there had been a reduction in the number of policies written compared with the number predicted at the time of acquisition and used in the model. He therefore sought to remove this variation as a factor, by calculating from the actual TOCBD post acquisition data a rate of TOCBD per policy, and then applying that rate to the forecast volume of policies used in the model at the date of the acquisition. Although these figures were put to Mr Mesher in cross examination before he had had a chance to consider or check them, he was able to follow the exercise which they purported to carry out, and he and Ageas' advisers have had ample subsequent opportunity to check their accuracy, which has not been challenged. The issue is whether they reflect the correct methodology for making the relevant adjustment.
17. AIG had advanced three other ways of putting its case which were abandoned at the conclusion of the evidence. The first was that if damages were to be assessed on forward looking assumptions, with no account being taken of the actual experience of TOCBD after the acquisition, the value of the company would have required a downward adjustment of £3.973m. This was based on the approach of Mr Mainz in his first report, before sight of the actual TOCBD experience post acquisition, which predicated that TOCBD would fall from its 2010 level to 2008 levels after a year, and thereafter remain at 2008 levels, as more reasonably reflecting what would have been projected on the information available to the parties at the time. The second was that the actual TOCBD figures post acquisition could be used as a cross check to support the reasonableness of Mr Mainz's figures based on 2008 levels as what would reasonably have been predicted by the parties on the information available to the parties at the time (see Buckingham v Francis [1986] 2 All ER 738 per Staughton J at 740a-b, 742b). The third was that damages should not be assessed at the date of breach but at some later date.
18. AIG's abandonment of the first of these points was a realistic recognition of the effect of the evidence. The evidence of Mr Mesher was more persuasive than that of Mr Mainz. Significantly, this part of AIG's case had relied heavily on the predictions by KFFS in the pre acquisition materials that rates of cancellation were forecast to fall. On the basis that cancellation projections could be treated as a proxy for TOCBD projections, this was said to justify a forecast of lower TOCBD following the acquisition. This was flawed for two reasons. The better assumptions to rely on were those of Ageas, not KFFS, in the light of the wealth of experience of Ageas' M & A team and the fact that Ageas' assumptions were what drove the DCF model which in fact produced a figure almost identical to the agreed price. The assumption that the Ageas M & A team put into the model was that rates of cancellation would remain constant at 2010 levels. Secondly using KFFS's assumptions rather than Ageas' assumptions did not assist this aspect of AIG's case. Both Mr Mesher and Mr Mainz used absolute figures in their adjustment, whereas the KFFS assumptions about falling cancellations were of rates of cancellations as a percentage of the volume of policies written, not numbers of policies cancelled. Since the model projected increased numbers of policies written, a projected fall in the cancellation rate did not imply a reduction in the number of policies cancelled. A calculation in the course of the hearing applying the KFFS projected fall in cancellation rates to the projected increased volumes revealed that even on KFFS's assumptions the number of cancelled policies was projected to rise.
19. AIG therefore conceded that if the assessment of the value of the company is to be judged by reference to information available at the date of the acquisition, Mr Mesher's adjustment is reasonable and appropriate. This reflects what a willing buyer and willing seller would have agreed upon as the price had they appreciated that TOCBD was a cost to the business, on the basis of historic information as to TOCBD then available. This of course required AIG to abandon the second point, that actual experience provided a useful cross check to support reasonable projections at the date of acquisition.
20. The main issue which remains for determination, therefore, is whether in valuing KFIS at the date of acquisition, account is to be taken of the TOCBD experience of the business since the date of acquisition. Can one use the benefit of hindsight?
21. On behalf of AIG Ms Robertson QC argued that the valuation of the company at the date of the acquisition involves an assumption having to be made about the future incidence of TOCBD. This is an assessment which if viewed prospectively at the date of acquisition involves future uncertain contingencies. But those contingencies are not now uncertain or in the future. She submitted that where an assessment of damages depends upon a contingency existing at the date by reference to which damages fall to be assessed, the Court can and should take into account what it knows about the outcome of the contingency by the time the assessment falls to be made, using its knowledge of events occurring subsequent to the breach. To do otherwise would be to value the company on the assumption that TOCBD would continue at 2010 levels, when it is now known that such assumption is a false one, which would result in Ageas receiving a windfall and would offend the compensatory principle.
22. On behalf of Ageas, Mr Matovu QC and Mr Scannell argued that there is no good reason to depart from the breach date rule, assessing damages by reference to the value of the company at that date without reference to subsequent events; and that following this conventional approach does not involve any injustice to AIG or windfall to Ageas.

Discussion

23. My attention was drawn to only two authorities which have directly addressed the permissibility of using hindsight or subsequent events in order to value a company for the purpose of a warranty claim. Neither provides a clear and authoritative answer to the legal issue before me.
24. In Senate Electrical Wholesalers Ltd v Alcatel Submarine Networks Ltd [1999] 2 Lloyd's Rep 423 the company had been sold for a price of £20m for the goodwill and £70m for the assets. The warranted 1990 management accounts had overstated the earnings for that year by almost £1.7m by erroneously including "rebate reserves", which were delayed trade discounts allowed to the business by manufacturers in reduction of their published prices. The reserves were meant to represent rebates due on purchases which had already been made. The position was complex because there was no uniform system for earning or recording such rebates: [18]-[20]. At first instance the plaintiffs sought to calculate their claim by reference to a price earnings calculation applied to the £1.7m overstatement of earnings in the 1990 management accounts. This approach was rejected at first instance and a cross appeal on the point was dismissed. The judge had awarded damages on the basis that if the accounts had disclosed the correct figure the negotiations would have been conducted at a lower figure and the price would have been £5m less. Amongst many points taken in support of a successful appeal, the sellers argued that there was in fact no difference between the warranted figure for profit and the actual profits for 1990 (see [11(c)], [38(B)]). One part of this argument was that a £750,000 overestimate of the 1989 rebate reserve was available to boost the 1990 profits: [38(B)(a)]. In this connection Stuart Smith LJ observed at [56]:

"Although the 1990 management accounts did not show a true and fair view because rebate reserves were overstated by £1.7m., in order to see if the plaintiff has suffered any loss and, if so, how it should be quantified, it is necessary to establish the actual profit for that year. Thus, if some credit or profit has been omitted which can properly be taken into account in the 1990 profit, the apparent loss is pro tanto extinguished or diminished. For this purpose, in our judgment, it is permissible to take into account hindsight to arrive at the actual figures."

25. I do not find it easy to understand from the report quite what hindsight was being referred to. In particular it is not apparent that what was meant by hindsight was the taking into account of matters which had not yet occurred at the time of the sale, rather than retrospective accounting treatment. Whilst paragraph [60] addresses and dismisses a hindsight argument by reference to what was known or expected at the date of sale about actual payment of the 1990 rebates, paragraph [57] appears to address the £750,000 overestimate of 1989 rebate reserve as giving rise to an understatement of profit in the 1990 accounts as a matter of accounting treatment. I have not therefore found this dictum of great assistance in resolving the current dispute.
26. In Infiniteland Ltd v Artisan Contracting Ltd [2004] EWHC 955 (Ch), Infiniteland brought a breach of warranty claim arising from its purchase of a number of companies from Artisan including in particular a company called Bickerton, whose trade collapsed within a matter of months of the purchase. Park J found against Infiniteland on liability, but went on to address the measure of loss which would have applied had liability been established. In this obiter passage, he observed at [132] to [134]:

"[132] Before I move on from this part of my judgment, there is a more general point which I wish to make. So far I have looked at the question of the quantum of any loss in the way in which the experts looked at it, making hypothetical estimates both of the but for valuation and of the actual valuation. Both experts were at pains to leave hindsight out of account in both of their valuations. That is plainly correct as regards the but for valuation, which is a hypothetical exercise, but I question whether it is also correct as regards the actual valuation. The actual valuation is not a hypothetical valuation like, for example, a valuation made for capital gains tax or inheritance tax purposes. It is designed to establish what the true loss to Infiniteland has been. Why should hindsight be excluded in that context?

[133] In Mr Downes' written closing submissions he wrote:

"Mr Palmer's evidence is that the actual value was nil. Given the absence of a track record and recent historic losses, it is submitted that Mr Palmer's opinion has an inherent probability about it."

I agree with that, but I am in any event inclined to go further. Infiniteland bought the companies in July 2001, and within a matter of months Bickerton's trade collapsed. Although I do not believe that Artisan realised how bad things were at the time, and although I accept that Artisan would have preferred Infiniteland to make a success of Bickerton, I think that it is obvious now - admittedly with hindsight - that Bickerton was doomed from the time that Infiniteland acquired it. I would in any event agree with Infiniteland (and its expert witness, Mr Palmer) that Bickerton was worthless at the time of acquisition, but even if I was uncertain about that I do not see why I should close my eyes to the undoubted facts that the companies for which Infiniteland agreed to pay over £1.4m turned out in the event to be completely worthless. There were a few hearsay suggestions in the evidence that Mr Aviss milked Bickerton during his brief period in control, so that the failure of Bickerton may have been Mr Aviss's fault, but the point was not pressed, and I repeat my view that Bickerton was doomed from the start.

[134] I do not need hindsight to come to my conclusion, but as a generalisation and subject to principles of remoteness of damage in contract or in tort (which do not in my view arise in this case), losses recoverable in damages are quantified at what they turn out in the event to have been, not at what they might have been forecast to be at the time of the transaction which subsequently goes wrong. Given my earlier conclusions on liability this does not help Infiniteland, but if I had taken a different view on liability, then either without or with hindsight I would agree with Mr Downes and Mr Palmer that the actual valuation of the companies which Infiniteland acquired — the valuation which falls to be compared with the but for valuation of some £1.4m — was nil."

27. These obiter comments seem to reflect a view which went beyond what was argued by Mr Downes for the claimants and may therefore not have been the subject of adversarial argument. The point which was argued was that the subsequent events were a useful cross check on the prospective forecast espoused by the first claimant's expert, a point which had considerable force on the striking facts of the case. This point was not addressed on the appeal which upheld the decision on liability for different reasons: [2005] EWCA Civ 758. The case is therefore slender authority in support of the permissibility of AIG's approach in this case.
28. Nor have I derived any great assistance from cases dealing with the valuation of companies in other contexts. In some cases subsequent events have been stated to be irrelevant to the exercise of valuing a company: see for example Re Holt [1953] 1 WLR 1488 (Danckwerts J), and Ng v Crabtree [2011] EWHC 1834 (Ch) at [17], [32] (Arnold J). In Buckingham v Francis [1986] 2 All ER 738 Staughton J adopted something of a half way house, holding that the valuation had to be undertaken prospectively at the valuation date but that "regard may be had to later events for the purpose only of deciding what forecasts for the future could reasonably have been made" on the valuation date (740a-b), for which exercise he "resolutely exclude[d] after-events, save for the purpose of checking what was a proper estimate at that date" (p742b). A point of this kind was advanced in Joiner v George [2002] EWCA Civ 160. The claimants were entitled to damages for the defendants' failure to transfer 51% of a company pursuant to an option to purchase which had been validly exercised. They failed in a claim for specific performance, and appealed against the assessment of damages, which involved valuing the company at the date on which the option should have been honoured. The claimants' argument, advanced at first instance and on appeal by one of them as a litigant in person, was not that later trading results should be adopted as such, but that such trading results supported the reliability of a forecast which they had made some time prior to the option date, upon which their expert had based his forecasts at the option date for valuation purposes, in preference to the forecasts used by the defendant's expert who had used the trading results for the period immediately prior to the option date: see [41]-[42]. Giving the leading judgment Sir Christopher Slade did not criticise the decision or approach in Buckingham, but treated it as casting no doubt on the correctness of the decision of the first instance judge in that case of preferring one expert's forecast figures over another's forecast figures: see [72]-[73]. At [74] he rejected the claimants' criticism of the first instance judge's short statement that "hindsight ought to be excluded" saying that it was a "statement of the general principles of share valuation so well established as to require no amplification".
29. Ms Robertson rests AIG's case on a broader legal submission. It is that whenever an assessment of damages depends upon a contingency existing at the date by reference to which damages fall to be assessed, the Court can and should take into account what it knows about the outcome of the contingency by the time the assessment falls to be made, using its knowledge of events occurring subsequent to the breach, where to do so gives effect to the compensatory principle and avoids the recovery of a windfall.
30. In considering this submission, two uncontroversial propositions can be taken as the point of departure:

(1) Damages for breach of contract are intended to put the innocent party in the same financial position as if the contract had been performed. This is the compensatory principle which "has been enunciated and applied times without number and is not in doubt" per Lord Bingham in Golden Strait Corpn v Nippon Kubisha Kaisha (The Golden Victory) [2007] 2 AC 353 at [9] (see also per Lord Scott at [29]). It may also be expressed as the claimant being entitled to the value in money of the contractual rights he has lost (per Lord Scott at [30]). The lodestar is that the damages should represent the value of the contractual benefits of which the claimant is deprived by the breach of contract, no less but also not more (per Lord Scott at [36]).

(2) Damages for breach of contract will normally be assessed by reference to the position at the date of breach. But a later date may be used for the assessment if in all the circumstances of the case to do so would more accurately reflect the overriding compensatory principle: County Personnel (Employment Agency) Ltd v Alan R Pulver & Co [1987] 1 WLR 916, at 925-6; The Golden Victory at [11] and the authorities there cited.

31. Ms Robertson puts the decision of the House of Lords in The Golden Victory in the forefront of her argument. In that case charterers had repudiated a seven year time charter when it had almost four years left to run. The charterparty provided at clause 33 that both owners and charterers should have a right to cancel the charter if war or hostilities were to break out between any two or more of a number of countries, including the United States of America, the United Kingdom and Iraq. This clause would have been triggered by the outbreak of the Second Gulf War in March 2003 some 33 months before what would have been the earliest date for contractual redelivery of the vessel. The arbitrator held that charterers would have terminated the charter on that date pursuant to the clause if it had still been alive. Viewed at the date of breach, the chance that any relevant clause 33 hostilities would break out was no more than a possibility and was certainly not a probability. The House of Lords held by a majority that the owners' damages were to be calculated by reference to the period ending on the outbreak of the war in March 2003, albeit at a rate of hire fixed at the date of the breach. Lord Scott said:

"[36] The same would, in my opinion, be true of any anticipatory breach the acceptance of which had terminated an executory contract. The contractual benefit for the loss of which the victim of the breach can seek compensation cannot escape the uncertainties of the future. If, at the time the assessment of damages takes place, there were nothing to suggest that the expected benefit of the executory contract would not, if the contract had remained on foot, have duly accrued, then the quantum of damages would be unaffected by uncertainties that would be no more than conceptual. If there were a real possibility that an event would happen terminating the contract, or in some way reducing the contractual benefit to which the damages claimant would, if the contract had remained on foot, have become entitled, then the quantum of damages might need, in order to reflect the extent of the chance that that possibility might materialise, to be reduced proportionately. The lodestar is that the damages should represent the value of the contractual benefits of which the claimant had been deprived by the breach of contract, no less but also no more. But if a terminating event had happened, speculation would not be needed, an estimate of the extent of the chance of such a happening would no longer be necessary and, in relation to the period during which the contract would have remained executory had it not been for the terminating event, it would be apparent that the earlier anticipatory breach of contract had deprived the victim of the breach of nothing.

………

[38] The arguments of the owners offend the compensatory principle. They are seeking compensation exceeding the value of the contractual benefits of which they were deprived. Their case requires the assessor to speculate about what might happen over the period 17 December 2001 to 6 December 2005 regarding the occurrence of a clause 33 event and to shut his eyes to the actual happening of a clause 33 event in March 2003. The argued justification for thus offending the compensatory principle is that priority should be given to the so-called principle of certainty. My Lords, there is, in my opinion, no such principle. Certainty is a desideratum and a very important one, particularly in commercial contracts. But it is not a principle and must give way to principle. Otherwise incoherence of principle is the likely result. The achievement of certainty in relation to commercial contracts depends, I would suggest, on firm and settled principles of the law of contract rather than on the tailoring of principle in order to frustrate tactics of delay to which many litigants in many areas of litigation are wont to resort. Be that as it may, the compensatory principle that must underlie awards of contractual damages is, in my opinion, clear and requires the appeal in the case to be dismissed."

Lord Brown said:

"[78] …My more fundamental conclusion, as I shall shortly explain, is that the breach date rule does not require contingencies - such as the likely effect of a suspensive condition – to be judged prior to the date when damages finally come to be assessed. ……..Could it really be thought to be in the interests of certainty to address these questions prior to that on which damages are in fact being assessed? To my mind not. Must the judge really shut his eyes to the known facts and speculate how matters might have looked at some earlier date? Again, not without compelling reason and none appears to me. Lord Bingham, at para 12, and Lord Carswell, at para 65, have already explained the "Bwllfa principle": Bwllfa and Merthyr Dare Steam Collieries (1891) Ltd v Pontypridd Waterworks Co [1903] AC 426. There is no need to repeat it. Suffice it to say that I see no good reason to depart from it here."

32. The speeches of each member of the majority in that case (Lord Scott at [36], Lord Carswell at [65] and Lord Brown at [78]) placed reliance on the approach of the House of Lords in Bwllfa and Merthyr Dare Steam Collieries (1891) Limited v The Pontypridd Waterworks Company [1903] AC 426. In that case a coal owner claimed statutory compensation against a water undertaking which had, pursuant to statutory authority, given notice preventing him mining a particular seam of coal over a period during which the price of coal had subsequently risen. The question was whether the coal should be valued prospectively at the date of the notice, based on an assessment of its value on what was then known, or as its value during the currency of the period, taking into account a known rise in coal prices. The Court of Appeal determined that it was the former. The House of Lords held that it was the latter. Lord Macnaghten said at p. 431:

"If the question goes to arbitration, the arbitrator's duty is to determine the amount of compensation payable. In order to enable him to come to a just and true conclusion it is his duty, I think, to avail himself of all the information at hand at the time of making his award which may be laid before him. Why should he listen to conjecture on a matter which has become an accomplished fact? Why should he guess when he can calculate? With the light before him, why should he shut his eyes and grope in the dark? The mine owner prevented from working his minerals is to be fully compensated – the Act says so. That means that so far as money can compensate him he is to be placed in the position he would have been if he had been free to go on working."

Lord Robertson said at p. 432 that "estimate and conjecture are superseded by facts as the proper media concludendi" and at p. 433 that "as in this instance facts are available, they are not to be shut out."

33. The Bwllfa principle, to the effect that the court should not speculate when it knows, and should prefer available facts to prophecies, has been expressed in a number of subsequent cases: see In Re Bradberry [1943] Ch 35, 45 (Uthwatt J), Curwen v James [1963] 1 WLR 748, 753 (Harman LJ); Kennedy v Van Emden [1996] PNLR 409 at 413C-D, 414C (Nourse LJ), 419F-420D (Schiemann LJ); Phillips v Brewer Dolphin Bell Lawrie Ltd [2001] 1 WLR 143 at [26] (Lord Scott); Whitehead v Searle [2009] 1 WLR 549 at [25] (Laws LJ), [78]-[80] (Rix LJ). In Murfin v Campbell [2011] EWHC 1475, HHJ Pelling QC, sitting as a Judge of the High Court, applied these principles to a claim against a solicitor for negligent advice which had given rise to a client's liability to a warranty claim.
34. Whilst one way of giving effect to the compensatory principle is to take a date for assessment of damages which is later than the breach, there is no conceptual difficulty in using subsequent events to inform an assessment of value at an earlier date in an appropriate case. Such was the approach adopted by the House of Lords in Phillips v Brewin Dolphin when determining whether a share sale agreement was a sale at an undervalue for the purposes of section 238(4) of the Insolvency Act 1986. It was recognised as a permissible approach by Romer LJ in Bwllfa. In Bwllfa the argument which had succeeded in the Court of Appeal (see [1902] 2 KB 135 at p138) was that the giving of the notice by the waterworks company was the equivalent of completing a contract by which the waterworks company purchased the coal, or the right to work the coal, other than as to price; in those circumstances the price would fall to be ascertained by agreement or arbitration based on estimates of future coal prices which would have been made at that time. This analogy was treated as a false one by the House of Lords (the Earl of Halsbury LC at p. 429, Lord Macnaghten at p. 431, and Lord Robertson at p. 433) because the assessment required was one of statutory compensation which involved a different test. The Earl of Halsbury LC said at p. 428:

"My Lords, I think in this case that Phillimore J. stated the question for debate with perfect accuracy when he said that "the true inquiry here is not what is the value of the coalfield or of the coal, but what would the colliery company, if they had not been prohibited, have made out of the coal during the time it would have taken them to get it."

In carefully distinguishing the sale analogy as a false one, the House of Lords might be thought to have endorsed the conclusion of the Court of Appeal on the question had the analogy been a good one, which was that subsequent fluctuations in coal prices would have been irrelevant to a valuation if there had been a sale. But Romer LJ, at least, thought that there was no objection in principle or conceptual difficulty in using subsequent events as an aid to valuation at the earlier date, although in common with the other members of the Court he did not regard subsequent fluctuations in coal prices as assisting in that exercise. He said at p. 141:

"….it is common ground that the compensation to be awarded by the umpire was the true value of the coal on October 15, the day on which notice to treat was given. I agree that in ascertaining that value the umpire may look at facts occurring subsequent to the notice to treat so far as they are material as shewing what was the value of the coal on that date, but not for any other purpose."

35. The Bwllfa approach, as applied in The Golden Victory, supports the proposition that when assessing damages for breach of contract by reference to the value of a company or other property at the date of breach, whose value depends upon a future contingency, account can be taken of what is subsequently known about the outcome of the contingency as a result of events subsequent to the valuation date where that is necessary in order to give effect to the compensatory principle. In an appropriate case, the valuation can be made with the benefit of hindsight, taking account of what is known of the outcome of the contingency at the time that the assessment falls to be made by the court. This is so not merely as a cross check against the reasonableness of prospective forecasting, as Staughton J regarded as permissible in Buckingham. It is so whatever view might prospectively be taken at the breach date of the outcome of the contingency.
36. This seems to me consistent with principle and justice. In the course of argument I posited an example of the sale of a racehorse, which the seller warranted to be free from disease; its value at the date of sale was to be measured by reference to an assessment of the races it might win and its consequent stud value; at the date of sale it had a latent disease which increased the risk of it suffering a career ending lameness at some stage; if the parties had known the true position at the date of sale the horse would have been valued at half the price because of this increased risk of lameness; by the time damages came to be assessed, however, the horse's racing days were over and it was known that there had been no incidence of career ending lameness despite the increased risk. Would the buyer still be able to claim half the price of the horse on the basis that its value without the benefit of hindsight was half what he paid? I am inclined to think not. By the time damages come to be assessed, it is known that the buyer received a horse which was every bit as valuable at the date of sale as the horse as warranted; with the benefit of hindsight it is known that the horse was as capable of winning the same number of races over its racing career as a horse without the latent disease. To award the buyer half the price of the horse would offend the compensatory principle and provide the buyer with a windfall.
37. I would, however, sound a note of caution. There are, in my view, two qualifications to the adoption of such an approach. The first is that it can only be justified where it is necessary to give effect to the overriding compensatory principle. The prima facie rule, from which departure must be justified, is that damages are to be assessed at the date of breach and that only events which have occurred at that date can be taken into account.
38. Secondly, it is important to keep firmly in mind any contractual allocation of risk made by the parties. Party autonomy dictates that an award of damages should not confound the allocation of risk inherent in the parties' bargain. It is not therefore sufficient merely that there is a future contingency which plays a part in the assessment. It is necessary to examine whether the eventuation of that contingency represents a risk which has been allocated by the parties as one which should fall on one or other of them. If the benefit or detriment of the contingency eventuating is a risk which has been allocated to the buyer, it is not appropriate to deprive him of any benefit which in fact ensues: it is inherent in the bargain that the buyer should receive such benefit. In The Golden Victory the contractual allocation of risk, in relation to the period of the charter, formed no impediment to the majority's approach to assessment of damages. The contingency was the outbreak of war, which the parties had provided for by the right to cancel conferred by clause 33. The risk of that happening had been foreseen and had been agreed to be one which if it eventuated would entitle the other party to terminate the charter. The owners had contractually undertaken the risk of such an event allowing the charterers to terminate the charterparty. Adopting an approach to the assessment of damages which took into account the eventuation of the contingency to reduce the owners' damages did not cut across the contractual allocation of risk.
39. Before turning to the application of these principles, I should mention that Ms Robertson also sought to rely upon clause 9.2 of the policy, which provides:

"[Ageas] shall, and to the extent reasonably possible shall cause [KFFS] to, act at all times as if uninsured and take all action reasonably necessary or advisable to mitigate any [insured loss] or potential [insured loss]."

40. As I understood the argument it was not that there had been a failure to mitigate or a breach of clause 9, but that a duty to mitigate could only arise if damages were capable of being assessed by reference to a date subsequent to the sale, or at least taking into account conduct subsequent to the sale. This does not advance the argument. Both are permissible approaches in an appropriate case, but neither approach is dictated by the existence of the clause. The policy provides cover for the costs of defending third party claims against KFFS which arise from a breach of warranty by KFGB. Clause 9 can be given content in such circumstances without requiring a departure from the breach date for assessment of loss in the current context or requiring subsequent TOCBD experience to be taken into account.

Application

41. With these principles in mind, there are two potentially overlapping matters which AIG must establish in order for the post acquisition incidence of TOCBD to be used in valuing KFIS at the date of acquisition. The first is that it is necessary to do so in order to give effect to the overriding compensatory principle. Or as Ms Robertson put it, that failure to do so would confer on Ageas a windfall. The second is that any benefit Ageas receives by the reduced post acquisition incidence of TOCBD is not one which the parties have conferred on Ageas by the allocation of risk in their contractual bargain. In my judgment AIG has failed to surmount either of these hurdles.

Windfall

42. Mr Mesher's calculation is based on an assumption about future incidence of TOCBD which is one the parties would reasonably have made at the time of acquisition if they had known that the TOCBD was for KFIS' account and had erroneously been excluded from the warranted accounts. Mr Mainz's calculation, using TOCBD figures for the period since acquisition, does not show that a value based on Mr Mesher's reasonable prospective assumption would result in a windfall or offend the compensatory principle.
43. Mr Mainz's calculation recognises that it would not be appropriate simply to take actual TOCBD figures. AIG recognises that that would be to compare apples and pears, because the 2010 TOCBD experience was on a volume of business which subsequently declined, whereas the valuation was based on a forecast of increased volumes. Mr Mainz therefore corrected for projected volumes of business rather than actual experience of volumes written post acquisition by calculating the post acquisition incidence of TOCBD per policy and applying it to forecast volumes.
44. However volume levels are not the only variable that will affect the incidence of TOCBD. In the course of the trial, four other variables were identified which may affect the level of TOCBD. One is the number of policies which are entered into on an instalment basis. This is a function not merely of the overall volume of business written but of the mix of business and in particular the proportion of business written on an instalment basis. Mr Mainz makes no adjustment for the mix differing from that projected at the time of sale. No analysis is undertaken of the number of instalment policies written in the relevant period after acquisition nor how it compares to 2010 levels or forecast levels. If, hypothetically, the post acquisition fall in TOCBD matched the fall in the number of instalment policies from 2010 levels, subsequent results would vindicate rather than cast doubt on the previous forecast of constant TOCBD at 2010 levels if what was forecast was the maintenance of the same number of instalment polices written. The drop in TOCBD would not in any sense be a windfall, but merely the product of expectations being confounded as to the number of instalment policies written.
45. A second variable affecting the incidence of TOCBD is the number of instalment policies which are terminated early by cancellation (either formal cancellation or simple cessation of payment). Mr Mainz again does not address or adjust for the experience of cancellation of instalment policies differing from that in 2010 or that projected at the time of sale. His calculation does not purport to show whether or to what extent the rate of cancellation per instalment policy has fallen during the period.
46. The third variable identified is the number of cancelling instalment policyholders who fail to pay the debt for time on cover. Again, Mr Mainz's calculation does not address or adjust for actual experience of debtor default by comparison to historic or forecast experience in this respect.
47. The fourth variable is the efficient management of TOCBD so as to reduce, as far as possible, the period for which the policyholder remains on cover in the event of cancellation, and to maximise the recovery of the debt from potentially defaulting debtors. Again, Mr Mainz's calculation does not adjust actual experience to historic or forecast experience in this respect.
48. Moreover, Mr Mainz's analysis ignores the effects of factors associated with reduction of TOCBD on other aspects of the business. For example, whilst a reduction in TOCBD is desirable as such, if it results from a reduction in the number of instalment policies written, there will also be a reduction in the financing revenue which comes from instalment contracts, which formed an important part of the revenue of the business. The net effect of a reduction in TOCBD and reduction in finance revenue may mean an overall loss rather than a gain to the business, despite the fall in TOCBD. Valuing the company in 2010 on the basis of a continuation of 2010 levels of bad debt does not involve a windfall if a subsequent reduction in bad debt is part of a trading pattern which as a whole results in a greater reduction in revenue than that forecast.
49. In short Mr Mainz has still compared apples with pears, despite his correction for total volume of business written. AIG has simply not shown that the conventional prospective approach of assessment at the breach date offends the compensatory principle or results in a windfall to Ageas.

Allocation of risk

50. The SPA was for a fixed price based on what Ageas was prepared to pay, and KFGB to accept, for a business which was thereafter Ageas' to do what it wanted with. There was no provision, as there sometimes is in such agreements, for any post acquisition adjustment of the price based on subsequent trading performance. Each party would have to determine an acceptable price based on forecasts reached prior to completion in what was a fast moving and competitive market facing new challenges in the grip of a major recession whose effect on customers remained uncertain. Upon completion, the contract was fully executed. The outcome of all the contingencies inherent in the forecasts were risks conferred on Ageas. If the business did better than the parties projected when calculating a price, that was for Ageas' benefit. If it did worse, that was its loss. The bargain embodied in the SPA was the allocation of risk to Ageas of any benefit or loss arising either as a result of the way Ageas chose to run the business or as a result of external influences on the success of the business.
51. I have identified above four of the variables which may influence the incidence of TOCBD in addition to the total volume of business written. All five of these (total volume, mix, cancellation rates, bad debtor profile and TOCBD management) are in turn influenced by both internal and external factors:

(1) The total volume of business which Ageas aims to write will depend upon its pricing strategy, upon its target customer base, and upon its administrative and cost structures. There is, for example, a spectrum between high volume low premium/margin business and lower volume business with higher premiums and margins. It is for the business under Ageas' ownership to decide where to seek to position itself. The volume of business written following acquisition will in part be the result of Ageas' own strategy in running the business. Further, whatever Ageas' strategy in these respects, the insurance market and wider macro economic circumstances are likely also to have an effect on the number of policies written. These may influence Ageas' market share, the level of premiums and margins, and possibly the size of the market as a whole in a recession.

(2) The mix of business, and in particular the number of instalment policies written, again may be the result of the way Ageas chooses to run the business in a number of respects. A higher number of instalment contracts might increase the incidence of TOCBD; but it brings with it an increase in the opportunity for finance income, and for add on products. Conversely, as already observed, a reduction in the number of instalment contracts might bring a welcome reduction in TOCBD but an unwelcome reduction in finance income. Whatever Ageas' strategy in these respects, again the insurance market and wider macro economic circumstances may have an effect on the number of instalment policies written.

(3) Ageas may take steps to reduce the incidence of cancellation of policies, which again may have ramifications for other parts of the business model. The evidence suggested that cancellation rates tend to be lower for lower premium business and the evidence of Mr Gilmurray, KFIS' Finance Director since May 2011, was that one of the reasons TOCBD amounts fell after acquisition was a fall in premium rates. Lower premium business could be part of a business strategy or the effect of market influences or both. Lower premiums might lead to a reduction in TOCBD through a lower cancellation rate, but any such benefit would be offset or outweighed by the consequences in lower revenues or profit margins.

(4) The number of cancelling policyholders who default on the debt may be capable of being influenced by the company's administrative arrangements and credit risk inquiries, but may also be a product of macro economic conditions. One might expect more defaulters in a recession.

(5) The efficient management of TOCBD may be down to the way Ageas runs the business; it may also be a function of the regulatory environment and terms and conditions available from insurers.

52. What happened to TOCBD after the acquisition was therefore part and parcel of the way Ageas chose to run the business following acquisition and the interaction between those business decisions and the effect of the market and macro economic conditions on the business. Those contingencies are all matters which the parties agreed are for Ageas' risk. The incidence of TOCBD was just one element inextricably bound up with the way the business was run and the external influences on its success, and was subject to the same allocation of risk.
53. That is sufficient to dispose of AIG's case. It is not necessary to undertake an analysis of why the incidence of TOCBD in fact turned out to be lower than the historical 2010 levels in the following years. It is sufficient that whatever the reasons, any resultant benefit was to be for Ageas to enjoy, just as Ageas would have to shoulder any resultant burden. As it happens, I did not have the evidence to enable me to determine with any degree of confidence what were the reasons for the extent of the fall. The only evidence was to be found in paragraph 29 of a statement of Mr Gilmurray which said: "TOCBD has been on a downward trend since August 2011. This is wholly due to external factors, including fewer sales and a fall in motor premiums, resulting in fewer cancellations and less bad debt." AIG did not require Mr Gilmurray to be called for cross examination. The statement does not purport to be exhaustive as to the reasons for the decline. It was a statement prepared in response to a disclosure application, not specifically for use at trial, and at a time when AIG's expert Mr Mainz had not yet identified that he was going to rely on the actual incidence of TOCBD post acquisition. It did not therefore purport to be a considered or careful analysis of the reasons for the fall. The passage is fairly to be characterised as a throw away line expressing an unconsidered opinion of some of the reasons for a fall in TOCBD since the acquisition. But even taken at face value, it does not assist AIG. If a reduction in sales and a reduction in premiums as a result of external market factors is a reason for the reduction in TOCBD from 2010, those are benefits accorded to Ageas by the contractual allocation of risk, just as would have been a loss resulting from an increased TOCBD caused by an increase in sales or premiums or any other internal or external factors.

Conclusion

54. Ageas' claim against AIG is entitled to succeed in the principal sum of £12,635,000. The amount of the judgment to be entered is affected by issues of the tax treatment of the principal sum, and interest, which were left over for further argument.