Denham C.J.
O’Donnell J.
Clarke J.
MacMenamin J.
Dunne J.
Charleton J.
O’Malley J.

Between/

Judgment of Mr. Justice Clarke delivered the 27th February, 2017.

1. Introduction
1.1 At the heart of this appeal lie questions concerning the proper approach which a court should adopt to the valuation of a company and in particular the way in which both trial and appellate courts should consider and deal with expert evidence tendered by the parties in support of their respective cases on such valuation.

1.2 For reasons which are no longer relevant to the issues which arise on this appeal, it has been determined that the respondents/appellants (collectively “Danbywiske”) should purchase the shareholding of the petitioner/respondent (“Donegal”) in the company named in the title of these proceedings (“Elst”). On the 11th April, 2014, the High Court (Kelly J.) directed the trial of an issue as to the price at which Danbywiske might purchase the shares of Donegal in Elst. There were, at the same time, also issues as to the precise level of shareholding which Donegal held in Elst. That question depended on issues arising out of an option agreement. There was an order of the High Court (McGovern J.) in different proceedings (2014 47 COM) which resolved that question (Danbywiske & anor v Donegal Investment Group plc [2015] IEHC 126) and also a separate order made by McGovern J. in these proceedings which found that the appropriate relief to which Danbywiske was entitled under this petition was an order compelling the purchase of the shares in Elst (Re Elst, Donegal Investment Group plc v. Danbywiske & ors [2015] IEHC 439.

1.3 In passing, and so as to make sense of the numbers, it should be noted that the order of Kelly J. directing that the valuation issue, which is the subject of this appeal, should be separately tried required that the valuation exercise be carried out on the assumption of a 35% shareholding in Elst. As a result of the decision of McGovern J. on the option question, already referred to, it transpired that the proper valuation was required to be carried out on the basis of a 30% shareholding. The ultimate order required the purchase of the shares concerned for €26,228,571. The numbers which appear in the calculations of the High Court and in this judgment were, however, based on the larger assumed shareholding of 35%. It should be made clear that the difference in numbers simply reflects the change in the scale of the shareholding. The issues which arise in the valuation exercise are not in any way affected by that difference.

1.4 However, for the purposes of this appeal, the key decision of the High Court was the judgment of McGovern J. on the 5th September, 2014 fixing the price at which Danbywiske might purchase Donegal’s shares in Elst at €30.6m on the assumption that the relevant shareholding was 35%. I mention some of the other judgments given in these and related proceedings simply for the purposes of recording that there has been a lot of activity in this litigation but that the issue before this Court is confined to the valuation question.

1.5 On the 8th June, 2016 the Court of Appeal (Finlay Geoghegan J., Hogan and Creegan JJ.) overturned the judgment of McGovern J. in respect of price (see Donegal Investment Group plc v. Danbywiske & ors [2016] IECA 193). The Court of Appeal ordered that the matter be remitted back to the High Court and gave certain directions as to the proper approach to the calculation of amount to be paid for the relevant shares.

1.6 It is against that finding of the Court of Appeal that Danbywiske sought leave to appeal to this Court. In that context the Court, in a determination (Donegal Investment Group plc v. Danbywiske & ors [2016] IESCDET 124), set out the basis on which leave to appeal was granted in the following terms:-

“(a) Whether the principles set out in Hay v O’Grady as to the limits of an appellate court’s review of fact apply both generally and to expert testimony and, as such, constitute a complete code which cannot be departed from?

(b) Whether these principles were departed from in the rulings of the Court of Appeal on the findings of fact in the High Court relating to share valuation?

(c) Does the costs order of the Court of Appeal require to be reviewed?”

1.7 The appeal is, therefore, concerned only with those issues.

2. The Position of the Parties
2.1 In order to specify with greater particularity the precise issues which arise on this appeal in the light of the written and oral procedure it may be useful to identify, at least in very broad terms, the position adopted by the parties.

2.2 On the first question of whether the principles identified in Hay v. O’Grady apply in respect of expert evidence, the parties adopted positions which did not differ to any great extent. At the level of principle it was accepted that Hay v. O’Grady did apply to expert testimony. It also seemed to be accepted by both parties that the proper approach which should be adopted by an appellate court in scrutinising a finding of fact by a trial judge, where that finding involved a consideration of expert evidence, might be somewhat different to that applicable in a case where an appeal related to a straightforward finding of fact. On one view it might be said that there was at least a difference of emphasis between the parties as to the proper approach in such circumstances and it will, therefore, be necessary in due course to touch briefly on the proper approach of an appellate court to findings based on expert evidence.

2.3 I now turn to the second issue. While it will be necessary to consider the judgment of Finlay Geoghegan J., speaking for the Court of Appeal, in due course, it is fair to say that a key finding of that Court is to be found at paras. 79 and 80 which are in the following terms:-

“The Court has concluded that, although there was before the trial judge significant complex evidence in relation to valuation, much of which he took into account and made decisions in respect of certain elements from which there is no appeal, there were, nevertheless, key elements of the expert evidence in relation to the proper approach to determining the multiple to be used with which he simply did not engage and in respect of which he gave no reasons for departing from what appears to this Court to have been a consensus amongst the experts, namely, the necessity to form a judgment or considered expert view as to the appropriate multiple following upon an assessment of multiples derived from relevant transaction and trading comparables. In those circumstances, consistent with the approach of the Supreme Court in Doyle v. Banville [2012] IESC 25 and Hay v. O’Grady [1992] 1 I.R. 210, it is a finding which cannot be upheld.

In reaching this conclusion, the Court is not intending to indicate that the trial judge was bound to follow the conclusion reached by expert testimony on the multiple offered by one or other party. As is clear from the judgment of Blayney J. in Irish Press plc. v. Ingersoll Irish Publications Ltd. [1995] 2 I.R. 175, a trial judge is entitled to reach his own conclusion, albeit that that is different to the conclusions reached by the expert evidence of either party. Nevertheless, in a matter such as this, where there has been a consensus of approach to market valuation as between the experts - albeit differing in its application to the particular facts - then if a trial judge is going to depart from and use a different approach, this is a matter must be explained and explained by reference to the key elements of the evidence before him.”

2.4 The Court of Appeal then went on to analyse the way in which the trial judge had approached the question of determining the correct multiplier. For the reasons set out, most particularly at para. 76 of the judgment, Finlay Geoghegan J. came to the view that there was no evidence to support what she found to be the method adopted by the trial judge. It will be necessary to consider that aspect of the judgment of the Court of Appeal in due course. Clearly if that finding by the Court of Appeal is correct then this appeal would have to be dismissed and the order of the Court of Appeal affirmed.

2.5 Essentially the case made by Danbywiske in relation to this issue suggested that the Court of Appeal was wrong to regard the trial judge as having come to conclusions which were not sustainable under the jurisprudence relating to appellate review of facts which is to be found in Hay v. O’Grady [1992] 1 I.R. 210 and Doyle v. Banville [2012] IESC 25.

2.6 The third issue, relating to costs, was, as is normal practise, left over until the substantive issues have been determined. It follows that the key issue for resolution on this appeal is as to whether, as Danbywiske argues, the Court of Appeal went beyond the permissible remit of an appellate court when scrutinising findings of fact made by a trial judge. That question, in turn, requires an assessment of the evidence and the findings of the trial judge which are specifically at issue in this case.

2.7 It will again be necessary in due course to address the expert evidence given by both sides on the valuation question. Interestingly, Danbywiske called two experts in relation to valuation while Donegal called one. The actual approach adopted by the trial judge to the valuation exercise did not, however, follow the precise methodology suggested by any of the experts. It should be noted that the experts themselves differed materially on their overall approach as well as on specific elements of the valuation. However, the trial judge approached the matter in a way which was different to each of them.

2.8 As the argument developed it became clear that there were, in reality, three separate questions which were identified at the oral hearing by counsel for Donegal. They were:-

(a) Whether the actual methodology adopted by the trial judge in approaching the question of valuation was clear from his judgment;

(b) Whether there was evidence to justify the overall approach or methodology of the trial judge and his specific findings not least because the trial judge had approached the matter in a somewhat different way to any of the experts; and

(c) Whether the reasons given by the trial judge for approaching the issue in the way in which he did were sufficiently clear from the judgment.

2.9 It seems to me that it is appropriate to approach this appeal by considering those three questions. Counsel for Danbywiske asserted that, while some elements of the reasoning of the trial judge as to the methodology adopted in valuing the company were not expressly set out in the judgment, the methodology was said to be sufficiently clear so as to permit this Court to infer the approach adopted. Counsel did accept that the methodology of the trial judge was different from that of any of the experts but suggested that there was nothing wrong in principle with the trial judge adopting a different approach to the experts provided that the trial judge’s approach and the specific findings made in determining the value on the basis of that approach were rooted in the evidence. Counsel for Danbywiske asserted that each of the findings either actually made by the trial judge or which it could reasonably be inferred had been made by the trial judge, were rooted in the evidence. Finally, counsel argued that the reasoning of the trial judge both in adopting the approach which he did and in applying it in the manner in which he did, was again either clear from the judgment or could reasonably be inferred.

2.10 On the other hand counsel for Donegal questioned whether there was sufficient clarity in the judgment of the trial judge to determine even the basis on which the trial judge approached the valuation task. Even if it were possible to determine the approach of the trial judge with a sufficient degree of clarity, it was argued that the trial judge had gone outside the range of permissible approaches having regard to the evidence. In that regard counsel argued that the Court of Appeal was correct to find that there was no evidence to support the method adopted by the trial judge for his valuation.

2.11 Finally it was argued that, even if it might have been permissible on the evidence for the trial judge to come to a conclusion of the type which he did, the judgment, it was said, lacked any reasons from which it was possible to discern why the trial judge adopted the approach which he did. On that basis alone it was said that the Court of Appeal was correct to overturn the judgment of the High Court and remit the matter back.

2.12 In order to analyse in more detail the arguments under those three headings it is necessary to address first the approach of the trial judge and in particular his analysis of the expert evidence and then address the judgment of the Court of Appeal.

3. The High Court Judgment
3.1 In addressing the judgment of the High Court it is important to recall that there were a wide range of issues concerning the valuation of Elst still alive at that stage. As will become clear, the central issue which remained in contention by the time of this appeal was the question of the appropriate multiplier to be adopted. That is a matter to which it will be necessary to return.

3.2 However, as already noted, the respective experts differed as to the overall approach which ought to be adopted. The evidence before the High Court was that there were two broad approaches to the valuation of companies, being a market approach and a discounted cash flow approach. A Mr. Tom Lindsay of Spayne Lindsay gave evidence for Donegal and based his valuation on the market approach. The two experts called on behalf of Danbywiske, being Mr. David Tynan of PriceWaterhouseCoopers and Mr. David O’Flanagan of Deloitte suggested a so-called hybrid approach being partly based on a market approach and partly on a discounted cash flow (“DCF”).

3.3 The trial judge noted the distinction between the two methods. In a market based approach the issue of valuation is determined by considering the valuation which the market place has put on comparable companies. Some such companies may have been the subject of takeovers or acquisitions and thus have been valued at arms length by the purchaser. Other companies may have their shares dealt with on various markets with the value of the company being determined by the price which the market places on the shareholding. In general terms it may be possible to identify, by looking at such comparators, a multiplier which the market places on EBIDTA (Earnings Before Interest Depreciation Tax and Amortisation) to arrive at a view as to the value of the enterprise which the company operates. EBIDTA is, of course, a measure of the inherent earnings of a company. Such earnings are what a purchaser of a company might expect to obtain on acquisition. An analysis of the EBIDTA of comparable companies in comparison to the value placed by the market on those companies (together with other factors such as debt) may allow an assessment to be made of the multiple of such earnings which the market is prepared to pay for acquisition.

3.4 There does not appear to have been any real dispute between the parties that this market based approach is at least one of the matters to which regard should be had in valuing a company. In addition, it was agreed that the value of the company itself required adjustments to be made to the enterprise value to reflect items such as debt for the shareholding in a company carries with it, as it were, any debt which is owed by the company. There were, indeed, disputes, which it is unnecessary to specify here, about some of the deductions which were suggested as being appropriate in this case. However, at the very broadest level, it was accepted that the market based approach involved ascertaining EBIDTA, multiplying that figure by a multiplier (which was derived by exercising a professional judgement by reference to comparable companies) and then making appropriate deductions from the figure thus arrived at.

3.5 On the other hand, a DCF approach involves estimating the cash flow of the company into the future and discounting the total value of that cash flow by an appropriate discount rate to reflect the fact that, by definition, much of that cash flow will be delayed. While the two experts called on behalf of Danbywiske differed as to the precise weighting to be given to the DCF valuation, they both arrived at their ultimate valuation of Elst by giving a significant weighting to DCF. On the figures in this case, the giving of a significant weighting to DCF had the effect of quite significantly reducing the overall valuation. For reasons which it is unnecessary to analyse for the purposes of this appeal (for it is no longer an issue), the trial judge rejected the partial DCF approach and came to the view that a market based approach was appropriate in all the circumstances of Elst. Given that finding, the question of the appropriate multiplier loomed particularly large in the overall approach. It also has to be said that the trial judge largely found for Donegal and against Danbywiske on most of the valuation issues which arose at the trial. Thus, by the time the case came to be heard on appeal, and in particular on appeal to this Court, counsel for Danbywiske was left with the task of having to base his case on the overall approach which had been advocated by the expert called by the other side but on supporting the one significant variation from the approach advocated by Donegal which the trial judge had adopted being the trial judge’s view as to the appropriate multiplier to be applied.

3.6 When the trial judge came to consider the appropriate multiplier he first identified that there had been a dispute as to whether a company called Walkro ought be included as a comparator. The evidence of Mr. Lindsay had suggested that Walkro should not be included. Mr. Lindsay had also suggested that the appropriate multiplier to use, having regard to a wide range of considerations, was 7.25.

3.7 In addressing that issue the trial judge said the following at para. 15 of his judgment:-

“The respondents argue that Mr. Lindsay’s multiple of 7.25 is called into question because he ignored Walkro as a comparable transaction. Applying a multiple of 4.75 to Walkro’s EBITDA in 2014 of €20.9m gives a value of approximately €99m to Walkro and that part of the Monaghan business. The respondents argue that applying Mr. Lindsay’s multiple of 7.25 to Walkro’s EBITDA of €20.9m would give rise to a value of approximately €152m equating to an 86% uplift in value in two years. Furthermore, the respondents argue that Walkro should not have been ignored because it produces compost, as does Monaghan and the majority of Monaghan’s profits are derived from compost production. The purchase was a recent transaction which took place after an extensive marketing campaign, and had Mr. Lindsay included Walkro in his comparable fruit and vegetable transactions, it would have reduced his average multiple to 5.9x and the median multiple to 6.1x.”

3.8 The trial judge went on to state that he had come to the view that Walkro should be included and that its inclusion would have “reduced Mr. Lindsay’s average multiple to 5.9x and the median multiple to 6.1x”. The trial judge further went on to indicate that a Japanese mushroom producer, Hokuto, and a further company called Adelaide Mushrooms “should have been considered”.

3.9 Having dealt with various other issues the trial judge reached his conclusion on valuation in para. 48 in the following terms:-

“Valuation of a shareholder’s interest in a case such as this is not an exact science. There is a discrete issue before the Court, namely, a determination of the price at which the respondents might purchase the petitioner’s shares. Having regard to what I have said in paras. 45 and 46 above, my determination on the valuation is as follows:-

Normalised 2014 EBITDA €44.9m

Applying a Median Multiplier of 6.1 (see para. 15) €273.9m

Net Cash (Debt Adjustments) €114.3m

CAPEX Debt-Like Adjustment €42.4m

Buyout of Minority Shareholders:

GIMV 45% in Walkro €26.2m

David Johnson 40% in Tunnel Tech €3.5m

Equity Value of Monaghan Mushrooms €87.5m

Equity Value of Petitioner Shareholding 35% €30.6m”

3.10 As can be seen the trial judge applied a median multiplier of 6.1 and made express reference back to para. 15 (which I have already cited) as justification for that figure.

3.11 One further matter commented on by the trial judge is worth recording at this stage. The valuation placed on a 35% shareholding in Elst by Mr. Lindsay was €64.5m. That reflected Mr. Lindsay’s market based approach and the application of a multiplier of 7.25. However, the trial judge noted the evidence which suggested that, in December 2014, immediately prior to the commencement of the proceedings, Donegal had demanded that Danbywiske purchase their stake for only €34m. This was argued by Danbywiske to undermine the valuations of almost twice that sum which was suggested in Donegal’s evidence. The trial judge did note that the valuation suggested by the experts called on behalf of Danbywiske (who respectively valued the relevant shareholding at €26.4m and €26m) were much closer to the figure which Donegal had originally put on their shareholding than the figure which their expert asserted at the trial.

3.12 Be that as it may it is next necessary to turn to the findings of the Court of Appeal.

4. The Court of Appeal Judgment
4.1 It must again be emphasised that there were a range of other issues, beyond the single substantive issue which is before this Court, still alive at the time of the appeal to the Court of Appeal.

4.2 However, on the multiplier question the Court of Appeal noted the finding of the trial judge, which I have cited earlier, which sets out the median multiplier of 6.1. The Court of Appeal stated that a particular basket of comparables (to which it will be necessary to refer later and which gives a 6.3 multiplier which in turn was reduced to 6.1 by the inclusion of Walkro) was but one element of the comparables considered by Mr. Lindsay. The Court of Appeal then noted the argument of Donegal that there was no expert evidence “which supported or even permitted the use of the arithmetically derived median of 6.1 of one subset of transaction comparables”.

4.3 In that context Finlay Geoghegan J. said the following at para. 76 of her judgment:-

“The Court has concluded that it must uphold that submission. In its view, the trial judge’s finding of an EBITDA multiplier of 6.1 cannot be sustained on the evidence. There was no evidence to support the use of an arithmetically derived median of multipliers from a subset of transaction comparables as decided. Firstly, all the experts were agreed that arriving at the relevant EBITDA multiplier involves the exercise of professional judgment based on transaction and trading comparables and deriving a range of EBITDA multipliers based on these comparables. It is a qualitative not mathematical exercise. Therefore, the error of approach by the learned High Court judge, for which there was no evidential support, was to derive an EBITDA multiplier by reference to a purely mathematical approach, when all the experts agreed that this was not an appropriate manner in which to generate an EBITDA multiplier. Furthermore, all experts were also agreed that in arriving at the appropriate multiplier to be used in the valuation of shares in the Company using the market approach, it was necessary to consider and assess both transaction and trading comparables. The conclusion of the trial judge is, however, based only upon an assessment of one data set of the transaction comparables and does not make any assessment by reference to the trading comparables at all.”

4.4 The Court noted, indicating that it was applying the principles set out in Hay v. O’Grady [1992] 1 I.R. 210, as further explained in Doyle v. Banville [2012] IESC 25, that there was no relevant evidence before the High Court to sustain that finding.

4.5 The Court went on to make the comments which I have already cited at para. 2.3 of this judgment. However, another key finding appears to have been that, in the view of the Court of Appeal, there was no evidence which would have permitted the trial judge to adopt the approach to valuation which he did.

4.6 Having analysed the reasoning of the Court of Appeal but before going on to address the three central questions which arise on this appeal, it is appropriate to say something about the proper approach of an appellate court when assessing a factual decision made by a trial court on the basis of expert evidence.

5. Findings based on Expert Evidence – The Role of an Appellate Court
5.1 A starting point has to be to identify the proper role of a trial judge in assessing expert evidence. Charleton J. explained that role in James Elliott Construction Limited v. Irish Asphalt Limited [2011] IEHC 269, (para. 12 of the judgment) in the following terms:-

“Every expert witness has to be evaluated on the basis of sound reasoning. An expert witness is, however, no different to any other witness simply because he or she is entitled to express technical opinions; all of us are subject to human frailty: exaggerated respect based solely on a witness having apparent mastery of arcane knowledge is not an appropriate approach by any court to the assessment of expert testimony. Every judge has to attempt to apply common sense and logic to the views of an expert as well as attempting a shrewd assessment as to reliability.”

5.2 In setting out the reasons why he preferred certain expert testimony over others in that case Charleton J. went on to say that:-

“Of these criteria, the most important reasons whereby I have chosen one expert over another have been the manner in which an opinion has been reasoned through and the extent to which opposing views have been genuinely and objectively considered on the basis of their merit. A judge must bear in mind that, notwithstanding that an expert may firmly declare a duty to the court, it is a natural aspect of human nature that even a professional person retained on behalf of a plaintiff or defendant may feel themselves to be part of that side’s team. Of particular importance in this case, therefore, has been the extent to which an expert has been able to step back and to consider and to think through an opposing point of view. As with demeanour, this is not readily demonstrated on a transcript of evidence. Rather, to a trial judge, it can be possible to see the degree to which a witness is thinking through the potential for an opposing theory before giving a reasoned answer. Experience in other cases demonstrates that there is a danger that experts may erect a barrier of apparent learning in order to disguise what would be an answer awkward to their side were it to be expressed plainly. Apart from the attractions of logic and reasoning, therefore, assessing an answer based on what is seen and heard in the courtroom remains important.”

5.3 It follows that the assessment of expert testimony does require a trial judge to assess the way in which that testimony is given. As Charleton J. pointed out, the way in which an expert responds to questioning or to the views of an expert witness tendered by the other side, can play an important role in the assessment by the trial judge of the extent to which the expert’s views may truly be said to be uninfluenced by the case which his or her side is seeking to put forward. Furthermore, experience has shown that it is much easier to engage with the detail of evidence which is explored and explained (and, indeed, challenged) at an oral hearing by being present at that hearing rather than reading a transcript of what transpired.

5.4 For those reasons it seems to me that counsel on both sides were correct to accept that the principles in Hay v. O’Grady do apply to the role of an appellate court in scrutinising findings made by a trial judge with the assistance of expert testimony.

5.5 However, as Charleton J. also pointed out in Elliott, an important part in the assessment of any evidence is the application by the trial judge of logic and common sense to the testimony heard. That approach is particularly relevant in the context of expert evidence. Where experts differ the position adopted by the other side will be put to each of the experts in cross-examination. Their reasons for maintaining their view can be examined in some detail. The trial judge can, therefore, assess whether the reasons given by one expert or the other stand up better to scrutiny.

5.6 While it is true, therefore, that the assessment of all evidence, whether expert or factual, requires both the application of logic and common sense, on the one hand, and an assessment of the reliability or credibility of the witness gleaned from having been in the courtroom, on the other, it may be fair to say that it is likely that a decision based on expert evidence will be significantly more amenable to analysis on the basis of the logic of the positions adopted by the competing witnesses and the assessment of the trial judge of their evidence on that basis.

5.7 Precisely because a decision to prefer the evidence of one expert over another is likely to be influenced, to a much greater extent than might be the case in respect of factual evidence, by the rationale put forward by the competing witnesses, there may be somewhat greater scope for an appellate court to assess whether the reasons given by a trial judge for preferring one expert over another can stand up to scrutiny. That being said it must remain the case that an appellate court should show significant deference to the views of a trial judge on the question of findings based on expert evidence because the trial judge will have had the opportunity to see the competing views challenged and scrutinised at the hearing. Having made those observations it seems to me appropriate to go on to consider the first of the three questions posed by counsel for Donegal being whether it is possible to determine, from the judgment of the High Court, the approach actually adopted by the trial judge in the valuation exercise.

6. Is it possible to infer the approach of the Trial Judge?
6.1 As can be seen from the table which appeared in the judgment of the High Court, the basic methodology adopted by the trial judge is clear. McGovern J. found that the EBITDA of the company as of 2014 was €44.9m. He then applied a multiplier of 6.1 referring back, in terms, to paragraph 15. This gave an enterprise value of €273.9m from which various deductions were taken so as to give an equity value of the company of €87.5m and a resulting value of Donegal’s shares at €30.6m (on the assumption of a 35% shareholding).

6.2 As noted earlier there were disputes between the parties at the trial about a number of other aspects of the valuation but, with the exception of the determination of the relevant multiplier, those disputes are not before this Court on this appeal.

6.3 As also noted the evidence before the High Court differed on whether it was appropriate to base the valuation of the company solely on a market approach or whether there ought be a significant weighting attached to the figure derived from a DCF. The method adopted by the trial judge is clearly based solely on a market approach.

6.4 While it might, superficially, appear that differences in a multiplier between figures around six or seven might not make too great a difference to the overall valuation of a company (perhaps even lying within the margin of error, as it were) a closer look at the table from which the trial judge derived his ultimate valuation shows why that is not so in the circumstances of this case. Ultimately the value of the company comes down to the value of the enterprise less appropriate deductions for matters such as debt. The deductions in this case, as found by the trial judge, came to €186.4m. Thus the value of the company was based on an enterprise value less that sum. If the enterprise value had come out at approximately €186m, a calculation of the type conducted by the trial judge (and by Mr. Lindsay) would have given a zero value to the company, although it may well be that the professional judgement of experts would not quite have valued the company in that way in those circumstances. That being said, it is clear that it is only when the enterprise value begins to move materially above €186m that the calculation begins to give a significant value to the company. Given the judge’s finding for EBITDA of €44.9m then the calculation would only move into positive territory when a multiplier of the order of 4.2 or above is used.

6.5 Thus it is the extent to which the multiplier exceeds a figure of that order which directly impacts on the bottom line. Seen in that context a multiplier of 6.1 (which exceeds that threshold by just under 2) is very different from a multiplier of 7.25 (Mr. Lindsay’s evidence) which exceeds that threshold by over 3. Seen in that way a multiplier of 7.25 gives a value to the company which is more than 50% greater than that which would be obtained by using a multiplier of 6.1. That analysis explains why the question of the appropriate multiplier looms so large in the final analysis. The ultimate conclusion of the trial judge was that the company was worth €87.5m. A change of even 0.1 in the multiplier would alter that figure (whether upwards or downwards) by €4.4m.

6.6 In any event, as noted earlier, the market approach requires identifying the sort of multiplier that the market is likely to apply to EBITDA in the context of a company of the type under consideration. The market itself can be divided into two sectors. First, there are companies which are publicly quoted. Such companies are referred to as trading comparables. The value of such a company is, of course, fixed by market forces in the light of the views which investors in shares take. In the case of such companies it is possible to calculate the multiplier which, from time to time, the market considers appropriate.

6.7 Second, there are specific transactions where companies or businesses are taken over in a single transaction and where there may be information available to experts as to the price actually paid and other relevant details of the companies’ financial position. Such companies are referred to as transaction comparables. Again, in such circumstances it is possible to calculate the multiplier which operated.

6.8 One of the important issues in the valuation of any company will be the selection of appropriate comparators. There was undoubtedly conflicting expert evidence on the comparators which should be used. But the method actually adopted by the trial judge seems to be relatively clear by reason of his reference back, in the table setting out his calculations, to para. 15 of his judgment. Mr. Lindsay had given a multiplier of 7.25. However, there had been some debate about the correctness of that figure. It will be necessary to return to that key issue in due course. However, the way in which the trial judge came to the multiplier which he used (being 6.1) seems relatively clear from paragraph 15. The trial judge indicated that had Mr. Lindsay “included Walkro in his comparable fruit and vegetable transactions, it would have reduced his average multiplier to 5.9 and the median multiplier to 6.1”. There had been a debate about whether it was, in fact, appropriate to include the company Walkro as an appropriate comparator.

6.9 Part of the evidence concerning comparables made reference to multiplier figures published in respect of what is described as the “UK and European fruit and vegetable” basket of companies. Mr. Lindsay had identified that group or basket (“the basket”) as being “the most directly comparable transactions” for valuation. The relevant multiplier derived from the basket gave a median figure of 6.3. Mr. Lindsay had given evidence, under cross-examination, that, as a mathematical exercise, the inclusion of Walkro would reduce the multiplier of 6.3 to 6.1. The calculation that gives the multiplier of 6.1 seems, therefore, to take the multiplier figure for the basket and add in Walkro. Subject to one matter to which I will shortly turn, it seems to me that it is possible to infer the approach adopted by the trial judge. He adopted the basket multiplier but added in Walkro.

6.10 The one matter which casts some doubt on that conclusion is that the basket involved transaction comparables only. It might, therefore, be inferred that the trial judge took the view that dealing with the multiplier on the basis of a set of companies which did not include any quoted companies was an appropriate approach. Certainly the method which the trial judge appears to have adopted as the basis for his calculation (subject to the Walkro adjustment) was of that type. On the other hand, the trial judge, at para. 17, does state that he considered that a Japanese mushroom producer called Hokuto “should have been considered” as a comparable. First, however, Hokuto is a quoted company and the fact that trial judge seems to have considered that it should have been considered for inclusion in the calculation must cast at least some doubt on whether the trial judge was satisfied to use only transaction comparables. Second, the calculation of a median multiplier of 6.1 derives from simply adding Walkro to the basket. If it was considered appropriate to also include Hokuto, it is by no means clear how that found its way into the calculation for the figure actually used in the trial judge’s final assessment of the value was an unadjusted 6.1.

6.11 In conclusion, under this heading, it can be said that it is possible to infer the approach of the trial judge with a reasonable level of confidence but there are some minor questions, which I have sought to analyse, which do cast at least some doubt on whether that was the exact approach actually adopted.

6.12 It is then necessary to turn to the question of whether there was evidence to support that method and the conclusions reached as a result of it.

7. Was there evidence to support the findings?
7.1 There was some emphasis placed at the hearing on passages from the judgment of the Court of Appeal (particularly para. 76 as already cited) which found that there was no evidence for certain aspects of the trial judge’s findings, especially the method used to assess the multiplier. However, I have already cited what Finlay Geoghegan J. said at para. 80 of her judgment. In that paragraph it was made clear that the Court of Appeal followed the views of Blayney J. in Irish Press plc v. Ingersoll Irish Publications Limited [1995] 2 I.R. 175 where, in a judgment of this Court, it was clearly held that a trial judge is entitled to reach his or her own conclusion as to the proper approach even where it is different to the approach suggested in the expert evidence of either party. It seems to me to be clear that Finlay Geoghegan J. accepted that, at the level of principle, a trial judge is entitled, for rational reasons, to adopt an approach to an exercise which requires the assessment of expert evidence which differs from the approach of each of the experts who have testified.

7.2 In my view Finlay Geoghegan J. was correct in her expression of the broad principle. A trial judge is not bound necessarily to adopt the views of either expert. This is most obviously so in cases where experts differ as to degree and a trial judge has to choose between two differing views but can frequently, and without any need for significant rationalisation, take a position somewhere between the expert views tendered. But even where, as here, the dispute was as to the proper way to go about calculating a multiplier (rather than about a particular component of such an approach) a trial judge is not bound to choose between the expert views given in evidence. There is long standing support for that view in the Scottish judgment in Davie v. Magistrates of Edinburgh [1953] SC 34 where it was held that a court is not bound to accept the conclusions of an expert witness even when not contradicted. It logically follows that a court can adopt a different approach to each of the experts provided that there is an appropriate basis, rooted in the evidence or in logic, for the approach actually adopted.

7.3 However, I also agree with the view expressed in the latter part of the same paragraph by Finlay Geoghegan J. when she said that if “the trial judge is going to depart from and use a different approach, this is a matter which must be explained and explained by reference to the key elements of the evidence before him”.

7.4 As indicated earlier it may not require any great deal of explanation for a trial judge to put forward very brief reasons for adopting a particular view on a straightforward issue governed by expert evidence. But where the trial judge adopts an approach which is significantly different from the positions espoused by the experts in their testimony, there is a greater obligation on the trial judge to explain why a particular approach is favoured. Where a trial judge favours one approach espoused by an expert over another espoused by a different expert then that choice may not require a great deal of explanation in a judgment other than to indicate in brief terms the reason why the views of one expert was preferred. Obviously each of the experts will have given their reasons for preferring their own approach.

7.5 But where the trial judge finds “a third way” then a much greater degree of explanation is required for otherwise the parties (and appellate courts) will not know why that approach found favour for the argument in favour of the approach in question will not have been made by either expert.

7.6 It is against that backdrop that the finding of the Court of Appeal cited earlier needs to be assessed. As noted that finding involved a number of elements. In particular the overall conclusion of the Court of Appeal was that there was no evidence to support the use “of an arithmetically derived median of multipliers from a subset of transaction comparables”.

7.7 It is certainly correct that none of the experts suggested that the valuation exercise required to be carried out in determining the multiplier involved a mathematical calculation. Rather it was agreed that it was necessary to exercise an informed judgment on the appropriate multiplier having regard to the evidence concerning all potential multipliers from comparable companies weighted in accordance with the extent to which such companies might be regarded as close comparators. That expert evidence is hardly surprising. It will very rarely be the case, in any valuation exercise, that there will be an exact comparator. Perhaps in valuing assets (such as residential houses) where there is a regular market for very similar properties, it may be possible to find very close comparators in sufficient numbers to reach a fairly clear view. But significant companies always will have differences between them. Differences in their geographical reach, their product ranges, perception as to their prospects for the future, management strengths and a whole range of other factors may, in the particular circumstances of an individual case, appear to be of especial importance.

7.8 The exercise of expert judgment requires identifying the comparables that are considered most suitable and attaching, perhaps, at least greater weight to them than other comparables which might be considered to have at least some influence on the valuation process but not such as would weigh as heavily.

7.9 A second reason for the Court of Appeal’s determination was its criticism of the trial judge for adopting what it found to be a conclusion “based only upon an assessment of one data set of the transaction comparables…” and not making “any assessment by reference to the trading comparables at all”. Again, it was said that there was no evidence for such an approach on the basis that all of the experts had agreed that both market and transaction comparables should be considered although, of course, it might well be that, in the particular circumstances of an individual case, one or other set might loom much larger in the valuation exercise because the companies in that category might be considered to be much more comparable or, indeed, for other reasons. In the course of argument counsel on both sides did note that there were factors which differentiated market and transaction comparables but that the issue cut both ways. For example, market comparables, because of the significant information which a quoted company must place in the public domain, can be capable of greater analysis. On the other hand their value can be affected by general market conditions which may not have anything much to do with the particular company concerned.

7.10 On the other side a transaction comparable is obviously specific to a particular deal entered into at arms length but undoubtedly on the basis of full due diligence where all of the information which might have emerged in that process may not be publicly available. Such considerations are yet another reason why the valuation process cannot be regarded as an exact science.

7.11 It follows that the Court of Appeal was correct, so far as it goes, in saying that none of the experts suggested either that a mathematical approach or an approach based solely on one particular category of transaction comparables was appropriate. It was on that basis that the Court of Appeal held that the findings of fact in respect of the multiplier could not be sustained.

7.12 The comment in relation to the use by the trial judge of a mathematically derived approach stems from the fact that the multiplier used by the trial judge seems to have been derived from taking the basket multiplier and adding in Walkro. Likewise, the suggestion that the approach of the trial judge was to take a subset of transaction comparables is based on the same analysis. The basket relates to companies which were transaction comparables but there were other transaction comparables referred to in the evidence. On the assumption that it is correct to infer that the trial judge obtained his multiplier of 6.1 by taking the basket multiplier and adding in Walkro, then the analysis of the Court of Appeal would clearly be correct. This is a matter to which it will be necessary to return.

7.13 However, in the light of the analysis by the Court of Appeal (derived from Ingersoll), which noted the entitlement of a trial judge to depart from the approach or methodology suggested by experts, I am not persuaded that it is appropriate simply to say that there was no evidence to support the methodology adopted by the trial judge and to immediately go on to treat the findings resulting from that methodology as unsustainable. For the reasons which I have already sought to identify, I am satisfied that a trial judge can adopt an approach which differs from that of any of the experts who give testimony provided that the trial judge puts forward a sufficient reason rooted either in the evidence or in logic for adopting the approach concerned. It seems to me, therefore, that the real issue which requires to be addressed on this appeal derives from that obligation. Given that the trial judge did depart from the methodology suggested by each of the experts, does the judgment make clear why that was so and provide a sufficient explanation for adopting the methodology concerned? That, in essence, is the third question referred to earlier.

8. Was there an explanation for the methodology?
8.1 I have already indicated that, in my view, a trial judge who proposes to adopt a methodology which differs materially from that suggested in evidence by all of the experts may do so but, in so doing, is required to set out a rational basis, rooted in the evidence or in logic, for adopting the methodology concerned. The parties are entitled to know, at least in broad terms, why the trial judge both rejected the approach suggested by each of the experts and adopted the methodology which was actually utilised. Likewise, an appellate court must be given similar information to enable it to assess whether there was, in reality, an appropriate basis for choosing the methodology used.

8.2 I have already dealt in some detail with an analysis of what appears to have been the methodology adopted by the trial judge. The very fact that a significant part of that analysis involved inferences or assumptions demonstrates that, at a minimum, the judgment is light on explanation. A few examples will suffice but more could be given.

8.3 First, the trial judge draws attention, correct so far as it goes, to the fact that criticism was made of Mr. Lindsay for excluding Walkro. However, it must be recalled that Mr. Lindsay’s proposed multiplier was 7.25 having regard to his judgement based on a range of factors. The basket multiplier of 6.3 was one element, albeit a significant element, of that assessment. But the addition of Walkro to the equation would only have had the effect of slightly reducing Mr. Lindsay’s 7.25 estimate. The figure of 6.1 actually arrived at in para. 15 involves, as the evidence demonstrates, adding Walkro to the basket. As noted earlier, Mr. Lindsay was asked, in cross-examination, to confirm that the figure of 6.1 would, as a mathematical exercise, be arrived at if you added Walkro in that way.

8.4 Given that the trial judge refers back to the analysis in para. 15 when using the multiplier of 6.1 in his final calculation, it is hard to avoid the conclusion that he took the basket multiplier as adjusted for Walkro as his multiplier. If the trial judge had some other basis for coming to the multiplier figure of 6.1 then it is not apparent from the judgment and there is no reasonable basis on which any such other reason for coming to that figure could be inferred.

8.5 Why then did the trial judge appear to base his assessment on the basket alone without having regard to any other comparables? It might, for example, be that the trial judge considered that the basket was of such a high level of comparability that it significantly outweighed all other considerations. More plausibly the trial judge might have considered, agreeing with the evidence of Mr. Lindsay, that the basket provided the best comparator and that the other possible comparators, to which some weight would have to be attached, broadly pointed relatively evenly in opposite directions so that the basket might be the appropriate starting point. Perhaps the trial judge simply felt that 6.1 was appropriate in the light of all of the evidence and it is just a pure coincidence that it happens to be the figure derived by taking the basket and adjusting for Walkro. But if that latter position was the case it is very hard to see how the reference to para. 15 of the judgment would allow for it.

8.6 In reality the judgment does not give any real assistance as to why the trial judge took the approach which he did. It is possible that the approach might have been justified on the evidence (or be rooted in the evidence as counsel for Danbywiske put it), or in logic and could be justified by a rational approach to that evidence even though it departed from the differing approaches advocated by the experts. But it is impossible to assess whether that is so because of the absence of any reasons either express or which can readily be inferred.

8.7 Next, it is necessary to return to the question of the inclusion or exclusion of Hokuto as a comparator. I have already indicated that the finding of the trial judge that some weight should have been attached to Hokuto raises at least some doubt about the approach actually adopted. But that issue raises a number of questions about why the trial judge adopted the approach he did. His finding that Walkro should have been included as a comparator led to him apparently adjusting the basket figure of 6.3 to 6.1. No similar exercise was carried out in adjusting for Hokuto. There may have been a reason for that approach but it is neither apparent nor can it readily be inferred from the judgment. Again, returning to the question of whether the trial judge paid any regard to market comparators, Hokuto creates a problem. Did the trial judge consider that the market comparators were of such little weight that he was prepared to look solely at the basket? But if so why consider Hokuto? And if considering Hokuto why none of the other market comparables? Again this is something for which there might be a rational reason but it is neither set out in or is to be readily inferred from the judgment.

8.8 It is, in my view, important to emphasise that the exercise which an appellate court has to carry out when scrutinising the judgment of a trial judge is not one to be conducted in a mechanical way so as to encourage parties to attempt to find some element of the findings of the trial judge which is said to be insufficiently explained. It must be recalled that a judgment is arrived at the end of a very open and transparent trial process. The case will have been fully pleaded, the evidence fully heard and submissions made on both sides. In many cases, and in particular in the Commercial Court, there will be further procedures including the exchange of witness statements and expert reports. Against that backdrop it will often be possible readily to infer why a particular finding was made even if there is no express statement in the judgment. The parties will know how the case ran. An appellate court can read the record of the case. The judgment needs to be read in the light of the case as made and defended before the trial judge.

8.9 But there can be cases where it is just not possible to ascertain, with any reasonable degree of confidence, the reasons why a trial judge adopted a particular approach in relation to an important part of the facts. Where a finding of fact is of significant materiality to the overall conclusion of the case and where the reasons of the trial judge are neither set out in the judgment or can safely be inferred from the run of the case and the structure of the judgment itself, then an appellate court is unable properly to carry out its task of scrutinising the judgment to see whether the findings of fact are sustainable in the light of the principles set out in cases such as Hay v. O’Grady and Doyle v. Banville. In such circumstances an appellate court will have no option but to allow an appeal to the extent appropriate and take whatever further steps may be required in all the circumstances of the case in question.

8.10 Clearly the Court of Appeal was not persuaded that any sufficient reasons could be gleaned either directly or indirectly from the judgment or the run of the case for the choice made by the trial judge of the methodology to be applied in the valuation of the company in this case. I am persuaded that the Court of Appeal was correct in that regard and, therefore, correct to allow the appeal before it and take the course of action which it did. On that basis I would dismiss the appeal to this Court.

9. Conclusions
9.1 For the reasons set out in this judgment I am satisfied that it is open to a trial judge to adopt a methodology or approach which differs from each of the approaches advocated in the expert testimony tendered by the parties. However, where a trial judge is persuaded to adopt a different approach, it is necessary for the judge to structure the judgment in such a way that either expressly explains why the approach adopted is considered to be appropriate notwithstanding the expert evidence tendered or that, at a minimum, the reasoning of the trial judge in that regard can be inferred with some reasonable level of confidence.

9.2 There is even some doubt as to the precise approach actually adopted by the trial judge in this case. But even if the approach actually adopted can be inferred to a sufficient level of confidence, I am satisfied that the Court of Appeal was correct to hold that the reasons why the trial judge utilised the approach which he did are neither clear from the judgment nor can safely be inferred.

9.3 For those reasons I am satisfied that the Court of Appeal was correct to allow the appeal and make the order which it did referring back the issue of valuation to the High Court. It follows that, in my view, this further appeal should be dismissed.