England and Wales Court of Appeal (Civil Division) Decisions

IN THE SUPREME COURT OF JUDICATURE
COURT OF APPEAL (CIVIL DIVISION)
ON APPEAL FROM COMPETITION APPEAL TRIBUNAL

		Royal Courts of Justice Strand, London, WC2A 2LL
		19th February 2004

Vice-Chancellor :

Introduction
1. On 23rd July 2003 iSoft Group plc ("iSoft") offered to acquire the issued share capital in Torex plc ("Torex") in exchange for iSoft shares. Both iSoft and Torex were and are engaged in the supply of software and systems to the healthcare applications market on such a scale that the offer, if accepted, would lead to a relevant merger situation as defined in s.23 Enterprise Act 2002 ("EA") which had come into force on 20th June 2003. The offer was notified to the Office of Fair Trading ("OFT") by iSoft on 1st August 2003. On 15th August 2003 IBA Healthcare Ltd ("IBA"), a company incorporated in the State of Victoria, Australia and also engaged in the same market complained to OFT about the effect of the anticipated merger.
2. S.33(1) EA provides that

"The OFT shall, subject to subsections (2) and (3), make a reference to the Commission if the OFT believes that it is or may be the case that –

(a) arrangements are in progress or in contemplation which, if carried into effect, will result in the creation of a relevant merger situation; and

(b) the creation of that situation may be expected to result in a substantial lessening of competition within any market or markets in the United Kingdom for goods or services."

3. OFT initiated an investigation. It is not and never has been disputed that the terms of s.33(1)(a) are satisfied. On 30th September 2003 OFT sent to iSoft and Torex, but not IBA or any other third party, what is called an 'issues letter' ("the Issues Letter"). It set out what were described as hypotheses which did not necessarily represent the views of OFT all of which tended to demonstrate that s.33(1)(b) was satisfied too. On 3rd October 2003 officials of OFT met representatives of iSoft and Torex. On 6th October 2003 OFT received written submissions from solicitors acting for iSoft and Torex. They relied, amongst other considerations, on the effect of the National Programme for IT ("NPfIT"). NPfIT is a new regime, proposed by the Department of Health in June 2002, to update IT systems as used in the National Health Service in England. It will provide for cross-referencing of patients' records by creating a complete electronic medical record for each patient across all NHS providers in England.
4. In its written decision dated 6th November 2003 OFT concluded that whilst the strong base of installed systems might give the parties a large market presence it was unlikely, in itself, to confer significant market power in view of the changes being brought about by the NPfIT. OFT considered that such a fundamental change had altered the future competitive landscape so that competitive constraints must be viewed under a new scenario. It added

"OFT does not believe that it is or may be the case that, if carried into effect, the creation of this relevant merger situation may be expected to result in a substantial lessening of competition within any market or markets in the United Kingdom for goods and services."

Accordingly OFT did not refer the proposed merger to the Competition Commission.

5. IBA was dissatisfied with this conclusion. On 21st November 2003 it applied to the Competition Appeal Tribunal ("CAT") under s.120 EA for a review of the decision of OFT. S.120(4) EA requires CAT in determining such an application to

"apply the same principles as would be applied by a court on an application for judicial review."

6. The application was heard and determined by CAT with commendable speed. In their judgment dated 3rd December 2003 CAT described in detail the basis on which they concluded that they should quash the decision of OFT and refer the matter back to it for reconsideration. That basis was summarised in paragraph 266 in these terms:

"...we are not satisfied that the OFT applied the right test, or that the OFT reached a conclusion that was reasonably open to them. We are not satisfied that the facts are sufficiently found in the decision or that all material considerations have been taken into account. We are unable to verify whether there was material on which the OFT could reasonably base important findings in the decision."

7. OFT, iSoft and Torex (collectively "the Appellants") now appeal to this court, with the permission of CAT, under s.120(6) EA. Such an appeal lies on a point of law only. The points of law are (1) whether CAT correctly interpreted and applied s.33(1) EA and (2) whether CAT properly applied the principles of judicial review as required by s.120(4) EA. Before dealing with either of those issues it is necessary to describe the relevant facts and the statutory background and framework in some detail.
The Facts
8. iSoft provides software systems to hospitals and other healthcare providers in the United Kingdom. Between 1999 and 2002 it acquired a number of businesses in the healthcare sector. In the year to 30th April 2003 it had a worldwide turnover of £91.5m of which £74m represented sales in the UK and other EU states. Torex is concerned in both the primary, i.e. GPs, and the secondary, i.e. hospitals, healthcare sectors. It provides both software and systems and hardware, including installation and support. In the year to 31st December 2002 its worldwide turnover was £161.8m of which £65.6m was in respect of healthcare technology sales in the UK and Republic of Ireland and £41.8m for retail sales of software and systems in UK, Republic of Ireland and other parts of Europe.
9. The principal software systems supplied by iSoft and Torex to the secondary healthcare sector, that is to hospitals, are Electronic Patient Records ("EPRs") and Laboratory Information Management Systems ("LIMS"). OFT recorded that the combined share of iSoft and Torex in respect of such systems as are installed in UK hospitals, described as "legacy", is 44% of EPRs and 66% of LIMS. It described iSoft and Torex as "key suppliers in each country of the UK, particularly in the supply of LIMS (where in Scotland and Wales their legacy systems will account for 100% of the installed base)" and "clearly the two leading suppliers of IT software to the healthcare sector in the UK".
10. Formerly such systems were bought by hospitals or their strategic health authorities on an individual basis as and when required. Consequently the NHS had many different installed IT systems thereby giving rise to problems of compatibility. In the summer 2002 the Department of Health proposed the new regime now known as NPfIT. This will allow for cross-referencing of patients' records by the creation of a complete medical record for each patient across all NHS providers in England. National projects will create a national spine of archived records and introduce an electronic system for appointments. The proposal envisaged the creation in England of five regions with a single local service provider ("LSP") as project manager to oversee the implementation of NPfIT. LSPs and their preferred application providers, known as PAPs, are to be appointed by the Department of Health. They will be responsible for developing and managing the transition from legacy systems to the new systems. OFT described the consequence as a fundamental change to the procurement process, significantly reducing the number whilst increasing the size of contracts available in England.
11. The first phase of NPfIT was announced by the Department of Health in January 2003. It involved funding of £2.3bn spread over three years. In February 2003 the Department invited applications from those who wished to be considered as an LSP. Initially Torex applied but, on 30th June 2003, withdrew its application. In May 2003 the Output Based Specification ("OBS") for the integrated care record service ("ICRS") was issued. It was revised in August 2003. It contained a number of passages in which the importance of legacy systems and their continued use is emphasised. Thus para 980 is headed Legacy Management. Para 980.2 deals with the continuation of legacy systems. Para 980.2.2 requires LSPs who assume responsibility for a legacy system to continue the same level of service. If legacy systems are to be replaced then detailed migration plans are required to be produced for approval, para 980.4. One of the assumptions is that "existing infrastructure/services will be used wherever possible in order to minimise duplication and enable earlier implementation of the IRCS".
12. At the commencement of its investigation OFT invited comments from third parties and got them from thirty interested parties, including representatives of the purchasing departments of the national health authorities. As I have already indicated the Issues Letter was sent to the merger parties on 30th September 2003. It set out what were described as main background assumptions with regard to EPRs and LIMS.
13. On the basis of those assumptions the letter set out nine competition concerns. It described them as hypotheses which OFT was still in the course of evaluating. They may be summarised as:

(1) The proposed merger would result in the loss of direct bidding competition between iSoft and Torex, which, since 1998, had occurred on 21 out of 39 EPR contracts and 16 of 31 LIMS contracts.

(2) Though Torex had not won a relevant contract in the previous three years its updated and extended product range as well as its strong installed or legacy base would enable it to be an active competitor in the future.

(3) iSoft and Torex would between them hold more than 50% of the installed base of EPRs and LIMS thereby giving rise to a significant structural change and substantial lessening of competition in the market because of the smallness of the next competitor and the significant advantages in market coverage and potential incumbency the merged company would enjoy.

(4) As the EPR and LIMS systems were specific to the UK and conversion of foreign systems would be expensive there were high barriers to the entry on the UK market of foreign competition.

(5) It was unclear whether the existing competitors in the market could provide competition to the merged company given their current lack of success in winning contracts.

(6) It might be difficult for an LSP to exercise buyer power with respect to the larger contracts to be expected under NPfIT in view of the existence of the PAPs.

(7) The broader product range of the merged company might encourage "one stop shopping" by NHS hospitals.

(8) It is inappropriate to judge the effects of the proposed merger only by reference to the NPfIT programme as such programme covers only England, will not exclude purchases by individual hospitals and is uncertain both as to timing and effect.

(9) iSoft and Torex are key suppliers of EPRs and LIMS to the National Health Service but, if the merger proceeds, will not compete with each other in the development and supply of new and improved systems.

The letter concluded by inviting the parties to put forward any evidence they wished on any of those issues and to consider appropriate undertakings to remedy the potential competition concerns.

14. As I have indicated these issues were considered by representatives of iSoft and Torex with officials of OFT at a meeting held on 3rd October 2003. This was followed by detailed submissions in writing sent to OFT by the solicitors for iSoft and Torex on 6th October 2003. OFT also had the benefit of the answers given by IBA on 7th October 2003 to certain questions it was asked by OFT. These materials were considered by OFT at an internal meeting held on 8th October 2003 at which the relevant decision was reached.
15. The decision was put into writing and is dated 6th November 2003. In paragraphs 1 to 13 OFT described the parties, the proposed merger, the changes to be made by NPfIT, the product market and the geographic market. Paragraphs 14 and 15 contain what is described as the "Competition Assessment". They are in the following terms

"14. The main suppliers of secondary healthcare software currently installed in UK hospitals are iSOFT, Torex/IBA, McKesson and Siemens. The parties' share of installed ("legacy") systems is significant, with the parties supplying 44 per cent of EPRs and 56 per cent of LIMS to the UK public sector. They are key suppliers in each country of the UK, particularly in the supply of LIMS (where in Scotland and Wales, their legacy systems will account for 100 per cent of the installed base). The pace of innovation in healthcare IT systems and changes to the procurement process suggest, however, that the installed base is not the best guide as to whether the parties will have market power in the future.

15. Since most public sector contracts are awarded following a competitive tender, a better measure of potential market power may be the parties' success in winning competitive bids in the past few years. While the existence of an installed base may give incumbent bidders reputational or informational advantages in bidding for new contracts, if the system required is substantially different from existing systems these advantages are unlikely to be significant. The presence of other bidders should act as a competitive constraint on the parties as they bid for new contracts, requiring them to put forward innovative solutions at competitive prices."

16. In paragraphs 16 to 18 OFT dealt with the effect of NPfIT. They pointed out that NPfIT had attracted bids from two major US companies. They had been selected as PAPs by several of the short-listed LSPs. Consequently the installed or legacy base of other suppliers would be displaced to that extent. OFT considered it to be uncertain whether individual NHS Trusts would have the funds to enable them to acquire systems outside NPfIT but that if they did the value of such purchases would be small. OFT noted that iSoft's EPR system had been selected by half of the short-listed LSPs but that no LSP had selected any system of Torex. Further, though over the previous three years Torex had been short-listed as the preferred supplier on number of occasions and selected on three of them, none of them proceeded because they were incompatible with NPfIT. OFT also noted that if contracts were awarded to the two competitors from the US for their integrated systems, modular systems such as those supplied by iSoft and Torex would be excluded.
17. In paragraphs 19 and 20 OFT considered barriers to entry. Although they might be expected to be high OFT pointed out that the significant amounts provided by the Treasury for NPfIT had in fact attracted bids from the two US competitors. They thought that the presence of such suppliers in the market would have a knock-on effect on the market outside NPfIT, namely £850m pa in England, £25m in both Northern Ireland and Wales and £125m in Scotland and provide opportunities for smaller suppliers with innovative solutions.
18. In paragraphs 21 to 23 OFT considered buyer power and thought that it would be increased as it would be concentrated in the hands of 5 LSPs rather than 177 NHS Trusts. In paragraph 24 OFT pointed out that there were none of the difficulties usually associated with a vertical merger. In paragraphs 25 to 28 OFT noted the concerns of a number of third parties. Generally the third parties considered that the merger would lead to a substantial lessening of competition because the merger parties had both knowledge of existing systems and a broader portfolio of products, in particular they were concerned at the effect of the "bundling" of products. IBA was concerned at the high market share of the merged company. Several hospitals were worried that they would be encouraged to abandon existing and useful systems in order to adopt the new ones at an increased cost. Except for Northern Ireland, the national health authorities were unconcerned because, in their view, there was sufficient competition.
19. The assessment and conclusion of OFT contained in paragraphs 29 to 34 is in the following terms:

"29. In terms of their legacy contracts to the UK public sector, iSOFT and Torex are clearly the two leading suppliers of IT software to the healthcare sector in the UK. In a bidding market, competition is for the market rather than in the market so that the competitive advantage acquired from the legacy base is unlikely to be strong, especially where a new procurement strategy is being introduced.

30. The NPfIT has created five LSP regions, and bidders for the five regions have pre-selected their preferred sub-contractors. Torex's products have not been selected (although in line with its claim that its strengths lie in this area it has been selected as a service provider providing support and installation services) (see note 3). Absent the merger, this means that Torex is likely to face significantly reduced opportunities to sell its products (or those of IBA) to hospital users in England. Expenditure elsewhere in the UK is significantly lower and may not justify the costs involved in updating Torex's existing portfolio of products.

31. The NPfIT is a high profile strategy, supported by government, which gives effect to a commitment to increase spending on updating IT healthcare systems in England. The increase in funding has attracted international LSP bids from well known and established global companies and has allowed for partnerships between the LSPs and US IT healthcare providers, Cerner and IDX, as well as iSOFT. The presence of these international competitors makes it likely that competition for future contracts will remain active. There is a reasonable prospect that international competitors with a UK base will bid for contracts in the regions with the likely effect of increased competition for contracts in Northern Ireland, Scotland and Wales.

CONCLUSION

32. iSOFT and Torex have been the two leading suppliers of IT software to the healthcare sector in the UK. While a strong legacy base may give the parties a large presence it is unlikely, in itself, to confer significant market power in view of the changes being brought about by the NPfIT. Such a fundamental change has altered the future competitive landscape with the effect that competitive constraints must be viewed under a new scenario.

33. For these reasons, the OFT does not believe that it is or may be the case that, if carried into effect, the creation of this relevant merger situation may be expected to result in a substantial lessening of competition within any market or markets in the United Kingdom for goods and services.

DECISION

34. This merger will therefore not be referred to the Competition Commission under section 33(1) of the Act."

20. As I have already indicated the application of IBA to CAT for a review of that decision was successful. CAT remitted the matter to OFT for reconsideration in the light of the CAT's decision. OFT, iSoft and Torex applied for permission to appeal. This was granted on 18th December 2003 for reasons given by CAT in a further detailed judgment. There was some issue as to whether we could or should have regard to CAT's explanations in that judgment of what they intended by statements in their first judgment. I see no reason why we should not, but, as will be seen I have not found it necessary to do so. For completeness I should record that the merger was completed on 23rd December 2003 and undertakings were then given by iSoft under s.71 EA. We were told that OFT has completed its reconsideration and is now in a position to give its revised decision. It will not do so pending a decision of this court. We were also told that some 31 cases are pending 3 or 4 of which are likely to be affected by our conclusion.
The statutory background and framework
21. Part 3 of EA superseded Part V of the Fair Trading Act 1973 ("FTA"). Under FTA the Secretary of State might refer a merger situation to the Monopolies and Mergers Commission "where it appears to him that it is or may be the fact that" two or more relevant enterprises had ceased to exist with three alternative consequences (s.64). In that event the Commission investigated and reported to the Secretary of State whether a merger situation, as defined, had been created and if so, whether it would or might be expected to operate against the public interest (s.69). Matters to which the Commission was to have regard in considering questions of public interest were enumerated in s.84. The Secretary of State might exercise his powers to remedy any adverse effect disclosed by a report of the Commission (s.73). A power was conferred on the Secretary of State by s.75 in the case of an anticipated merger similar to that conferred by s.64 in the case of an actual merger and with the same consequences. By s.76 it was the function of the Director General of Fair Trading to keep himself informed about actual or prospective mergers and to make recommendations to the Secretary of State as to the action he should take.
22. Thus the Secretary of State had the power to make a reference to the Commission for investigation but was under no duty to do so. The statutory test was whether the merger might be expected to operate against the public interest though, in practice, references were made on competition grounds. The Commission had no power to prohibit the merger or remedy its consequences, only the Secretary of State might do so. The function of the Director General of Fair Trading was, in the relevant respects, advisory only. In each of those respects EA made important changes.
23. First, the decision whether or not the Commission should investigate a completed or anticipated merger is to be made by OFT. Second, the responsibility is cast on OFT by means of a statutory duty, as opposed to a statutory power or discretion, in the case of actual mergers by s.22 and in the case of anticipated mergers by s.33. S.22 is, mutatis mutandis, in the same terms as s.33. I have quoted s.33(1) in paragraph 2 above and need not repeat it. However the duties imposed by ss.22(1) and 33(1) are in each case subject to subsections (2) and (3). Ss.(2) entitles the OFT not to make a reference "if it believes that" the relevant markets are not sufficiently important, in the case of an anticipated merger the arrangements are not sufficiently far advanced and in both cases any relevant customer benefits, as defined in s.30, outweigh the substantial lessening of competition arising from the merger. Ss.(3) precludes any reference if, inter alia, OFT is considering whether to accept undertakings under s.73 instead.
24. Ss.35 and 36 set out the questions to be decided by the Commission in relation to, respectively, completed and anticipated mergers. Again, mutatis mutandis, they are the same so that I need only quote the questions posed by s.36(1). They are

"(a) whether arrangements are in progress or in contemplation which, if carried into effect, will result in the creation of a relevant merger situation; and

(b) if so, whether the creation of that situation may be expected to result in a substantial lessening of competition within any market or markets in the United Kingdom for goods or services."

25. S. 42 confers power on the Secretary of State to intervene in certain public interest cases. Ss.1(a) and (2) provide a contrast between two alternative legislative formulae. The first imposes the condition that "the Secretary of State has reasonable grounds for suspecting that...". The second applies "if [the Secretary of State] believes that it is or may be the case that...". If the Secretary of State has intervened under s.42 then by s.44(2) OFT is required to report to the Secretary of State in relation to the case. By s.44(4) the report is required to include decisions as to

"whether the OFT believes that it is, or may be, the case that –

[(a) a relevant merger situation has been or will be created]

(b) the creation of that situation has resulted, or may be expected to result, in a substantial lessening of competition within any market or markets in the United Kingdom for goods and services;"

By s.45 power is given to the Secretary of State if he has given an intervention notice and received a report from OFT to make a reference to the Commission

"if he believes that it is or may be the case that –

(a) a relevant merger situation has been created;

(b) the creation of that situation has resulted, or may be expected to result, in a substantial lessening of competition within any market or markets in the United Kingdom for goods or services;

[(c) [and d[?]] there is a relevant public interest consideration and on balance the merger would be against the public interest.]

26. By s.73(2), if the OFT considers that it is under a duty to make a reference under ss.22 or 33 it may instead accept undertakings from the parties "for the purpose of remedying, mitigating or preventing the substantial lessening of competition concerned". S.103 requires OFT in deciding whether to make a reference under s.22 or s.33 to have regard with a view to the prevention or removal of uncertainty to the need to make its decision as soon as is reasonably practicable. S.104 obliges OFT, so far as practicable, to consult any person who appears to them likely to be adversely affected by their decision whether or not to make a reference under s.22 or s.33.
27. S.106 obliges OFT to publish general advice and information "about the making of references by it under section 22 or 33". Such advice is to be prepared

"with a view to –

(a) explaining relevant provisions of this Part to persons who are likely to be affected by them; and

(b) indicating how the OFT...expects such provisions to operate."

Such advice may include advice about the factors which the OFT may take into account in considering whether, and if so how, to exercise a function conferred by that part. Clearly such advice cannot control the interpretation of the Act: that is a matter of law for the court. But, given the duty to provide it, the skill and experience of those required to do so and the purpose for which it is to be provided, the court should, in my view, treat with caution any suggestion that the Act should be interpreted in a sense contrary to such advice. S.107(1) requires OFT to publish, inter alia, any decision made by it not to make a reference under s.33. Such obligation includes the obligation also to "publish [OFT's] reasons for the decision...": s.107(4).

28. S.120(1) enables any person aggrieved by a decision of, inter alia, OFT in connection with a possible reference to apply to CAT "for a review of that decision". Subsection (4) provides that

"In determining such an application the Competition Appeal Tribunal shall apply the same principles as would be applied by a court on an application for judicial review."

An appeal lies from a decision of the CAT to the Court of Appeal "on any point of law arising" from that decision.

29. I should also note the power conferred on OFT by s.131 to make a market investigation reference. In this case the power is exercisable "if the OFT has reasonable grounds for suspecting" the relevant matters. S.154 authorises OFT to accept undertakings for the purpose of remedying, mitigating or preventing any adverse effect on competition instead of making a reference under s.131. But in that event s.139 entitles the Secretary of State to intervene by notice to OFT if, inter alia, he "believes that it is or may be the case that one or more than one public interest consideration is relevant to the case". If the Secretary of State gives such a notice then s.150(1) precludes OFT accepting undertakings without the consent of the Secretary of State. S.150(2) provides that the Secretary of State shall not withhold his consent "if he believes that it is or may be the case that the proposed undertaking will, if accepted, operate against the public interest".
30. In May 2003 OFT published guidance, as required by s.106 EA, to companies and their advisers on the criteria applied by OFT when considering whether to refer a merger to the Commission for further investigation. In paragraph 1.4 it is claimed that the guidance represents the policy and practice of the OFT "reflecting current legal and economic thinking". Paragraph 3.2 is in the following terms:

"The test for reference will be met if the OFT has a reasonably held belief that, on the basis of the evidence available to it, there is at least a significant prospect that a merger may be expected to lessen competition substantially. The OFT considers that this threshold is the same as that against which FTA reference advices were prepared. It differs from that used by the CC in its merger enquiries, reflecting the fact that the OFT is a first-phase screen while the CC is determinative: hence, the test for making a merger reference is lower than the CC's test for deciding that a merger may be expected to substantially lessen competition."

Paragraphs 8.3 and 8.4 indicate the view of OFT that undertakings in lieu of a reference authorised by s.73 are only appropriate in cases where both the substantial lessening of competition and the requisite remedy are clear cut.

Did CAT correctly interpret and apply s.33(1) EA?

31. In their careful and comprehensive judgment CAT set out the background and the statutory framework of EA in detail. In paragraph 63 in the section dealing with the statutory framework CAT expressed the view that a key issue in the case is the intended balance between the two stage procedure, that is a reference by OFT and an investigation by the Commission. CAT then described the procedure followed by OFT in this case, OFT's decision and the arguments for the parties. CAT's own analysis starts at paragraph 168 by posing the question whether OFT was presented with a real question as to whether it is or may be the fact that the iSoft/Torex merger may be expected to lead to a substantial lessening of competition. If the answer was in the negative then no reference would be made and the question of the interpretation of s.33 would not arise. CAT considered that the answer was in the affirmative largely for the reasons set out in the issues letter which I have summarised in paragraph 13 above.
32. CAT dealt with the interpretation of s.33(1) in paragraphs 178 to 214. They divided the subsection into three elements, namely "the OFT believes", "that it is or may be the case" and "may be expected to result". They considered (paras 182 and 183) that the last of those elements indicated a more than 50% chance. In respect of the first element they assumed (para 184 to 188) that the belief is not subjective but must be held on reasonable grounds which presupposes a sufficient investigation. CAT contrasted the belief of OFT under s.33(1) with the decision of the Commission under s.36(1). They considered that "the role of the OFT is primarily that of a first screen, to identify where competition issues may arise".
33. They described (para 189) the remaining element "it is or may be the case" as "the double may" and considered that it was central to the appeal. Their reasoning and conclusion on that issue is set out in paragraphs 190 to 198 which, in fairness to CAT, I must quote in full. CAT said

"190. The use of the word "may" in the second line of section 33(1) seems to us to signify that, even if those responsible at the OFT are themselves of the view that a merger may not be expected to result in a substantial lessening of competition, it still "may be the case", within the meaning of section 33(1), that the merger may be expected to lead to a substantial lessening of competition, if there is, in fact, an alternative credible view that cannot be reasonably rejected by the OFT on the basis of a "first screen".

191. In other words, putting the matter less technically, if there is genuinely "room for two views" on the question whether there is at least a significant prospect that the merger may be expected to lead to a substantial lessening of competition, then in our opinion the requirement in section 33 (1) that "it may be the case" that … [the merger] may be expected to lead to a substantial lessening of competition, is satisfied.

192. In our opinion, in such circumstances, the statutory duty of the OFT under section 33(1) is not to decide, definitively, which of those two views, it, the OFT, prefers. Under the scheme of the Act, the definitive decision maker, in a case where there is room for two views, is not the OFT but the Commission. If there is room for two views, the statutory duty of the OFT is to refer the matter to the Commission, whose duty is to decide on the question whether the merger may be expected to lead to a substantial lessening of competition, as section 36(1) expressly provides.

193. When we refer to the possibility of there being "room for two views" in a given case, we do not envisage a case in which the alternative view is merely fanciful, or far fetched. We envisage a case in which the alternative view is credible. It must be a view which cannot be confidently dismissed on the basis of a "first screen" investigation.

194. There is also in our view a certain asymmetry under section 33(1) between the situation which arises when the OFT makes a reference, and the situation which arises when the OFT decides not to do so. Even in a case where a substantial lessening of competition seems a likely outcome, in making a reference the OFT does not decide whether, in fact, a substantial lessening of competition may be expected. The OFT simply "believes" that such "may be the case", without prejudging or pre-empting the Commission's investigation.

195. Where, however, the situation is the other way round, and the OFT decides not to make a reference it is deciding that the merger does not even reach the threshold of "it may be the case". In other words in such circumstances the OFT decides that the merger does not even reach "the grey area" where there may be room for more than one view. In its practical effect, a decision not to make a reference effectively decides the issue of substantial lessening of competition in the negative. It not only prejudges, but also excludes, any further investigation by the Commission.

196. In the vast majority of cases no practical consequences arise from this asymmetry. An initial search by the Tribunal showed 56 published merger cases considered by the OFT under the Act, of which 21 did not qualify and 31 were cleared in short, clear decisions. Similarly, in the decisions made to refer (such as Unum/Swiss Life and P&O/Stena) the OFT shows shortly and clearly why the OFT felt that it was under a duty to refer.

197. What is the correct approach in cases in the "grey area" in between? In a case where real issues as to the substantial lessening of competition potentially arise, it seems to us that the words "it may be the case" imply a two-part test. In our view, the decision maker(s) at the OFT must satisfy themselves (i) that as far as the OFT is concerned there is no significant prospect of a substantial lessening of competition and (ii) there is no significant prospect of an alternative view being taken in the context of a fuller investigation by the Commission. These two elements may resemble two sides of the same coin, but in our view they are analytically distinct.

198. It is, as we have said, implicit that the OFT in any event must have sufficient material to support its view. It also seems to us implicit in the second limb of the test that the OFT must be able reasonably to discount the possibility of the Competition Commission coming to a different view after a more in-depth investigation. It must be borne in mind throughout that the role of the OFT under the Act is "a first screen"."

34. CAT applied the test it formulated in paragraph 197 in paragraphs 228, 232 and 233 in the following terms:

"228. Secondly, on the proper construction of section 33(1), and in particular the words "it may be the case", the OFT had to satisfy itself not only (i) that in its own mind there was no significant prospect of a substantial lessening of competition, but also (ii) there was no significant prospect of the Competition Commission reaching an alternative view on the basis of a fuller investigation.

"232. In this case, the Tribunal is unable to be satisfied, on the material before it, that the OFT asked itself the right question, namely whether the OFT was satisfied not only that there was no significant prospect of a substantial lessening of competition, but also that there was no significant prospect of the Competition Commission reaching an alternative view after a fuller investigation. There is no indication in the decision that the OFT considered the second limb of that test.

233. In the Tribunal's view, the tenor of the decision read as a whole is that the OFT decided that the effect of the NPfIT was to rebut the inference of a substantial lessening of competition resulting from the increase in market share of the parties following the merger. In other words, the OFT's approach was to seek to decide which of two plausible views the OFT preferred, rather than adopting the correct approach, namely to ask whether there were, reasonably, two views which could be taken. By failing to ask itself that latter question, the OFT failed correctly to ask itself whether "it may be the case" that the merger may be expected to result in a substantial lessening of competition within the meaning of section 33(1)."

35. OFT accepts that if the test formulated by CAT is correct then OFT's decision cannot stand because they did not apply the second part of it. OFT contends that CAT's formulation is wrong. They submit that it is inconsistent with the wording of s.33(1) both when read alone and in the context of other parts of the Act, is impracticable and contrary to the evident intention of Parliament. The test for which OFT contends is that set out in paragraph 3.2 of their guidance quoted in paragraph 30 above.
36. iSoft and Torex also contend that CAT was wrong but they do not support the test propounded by OFT in paragraph 3.2 of their guidance. They object, in particular, that too low a degree of likelihood is implicit in the word "may" being reformulated as "a significant prospect that". But they contend that there is no justification for importing the two-part test which CAT favoured in either the wording of s.33 or the context of the Act as a whole. The formulation on which they rely is that of Sir John Donaldson MR in R v Monopolies and Mergers Commission [1986] 1 WLR 763, 769 "knows or suspects".
37. IBA seeks to uphold the test as formulated by CAT substantially for the reasons CAT gave. Thus IBA points to the different roles of OFT and the Commission. It describes that of OFT as secondary. It points out that the Commission's powers of investigation are far greater than those of OFT and that the ultimate decision is for the Commission not OFT. In its written argument IBA submits

"The second use of the word "may" in section 33, together with the context in which section 33 operates, (as descriptive of the test of a first phase assessor) led the Tribunal to express its conclusions, derived from an analysis of the statutory language, in broader and less technical terms. Accordingly, in paragraph 191 the Tribunal adopts the perfectly sensible "room for two views" test. This formulation or pithy encapsulation of the statutory test fits neatly and accurately into the statutory framework. It has explained that this alternative second view must be a credible view and one which the CC could reasonably adopt. When such a case arises the OFT must refer. This, it is submitted, is entirely logical and consistent with the statutory language."

In oral argument counsel for IBA accepted that there was a problem of expression, as he described it, with paragraph 197 of the judgment of CAT. He supported a single test.

38. I have no hesitation in preferring the submissions of the Appellants on this issue. The statutory test, so far as relevant, imposed by s.33(1) is

"whether OFT believes that it is or may be the case that the [merger] may be expected to result in a substantial lessening of competition…"

Thus the relevant belief is that the merger may be expected to result in a substantial lessening of competition, not that the Commission may in due course decide that the merger may be expected to result in a substantial lessening of competition. Further, the body which is to hold that belief is OFT not the Commission.

39. If the test as formulated by CAT is right then, for the reasons advanced by counsel for OFT, other comparable provisions in the Act become unworkable if OFT does not hold the requisite belief but considers that the Commission may. Thus, unless OFT itself holds the relevant belief it cannot conduct the balancing exercise required by s.33(2)(c). And if OFT does not hold the relevant belief it has no power to accept undertakings in lieu as provided for in s.73 and so could not refuse to make a reference as permitted by s.33(3)(b).
40. Similar problems would arise in connection with the interventions by the Secretary of State as permitted by s.42. I have set out the relevant provisions in paragraph 25 above. Not only does s.42 point a clear contrast between a belief and a suspicion but the jurisdiction of the Secretary of State to make a reference to the Commission under s.45 depends on the OFT holding the relevant belief and expressing it in its report under s.44. It would be absurd if the jurisdiction of the Secretary of State to make a reference to the Commission should depend on the belief of OFT as to what the Commission might decide.
41. In paragraph 29 I have summarised provisions relating to market investigation references. These provisions also point to the contrast between a belief and a suspicion. Whether or not the Secretary of State gives his consent under s.150(2) must depend on his own belief, not that of others. It would be contrary to the statutory test if the Secretary of State had to consent notwithstanding that he did not himself believe the relevant fact but could not dismiss as fanciful an alternative view that others might hold.
42. For all these reasons I would reject the two part test formulated by CAT in paragraph 197 of their judgment and applied in paragraphs 228 and 232. Accordingly it is not necessary to consider, assuming it would be permissible to do so, whether the test suggested by CAT would lead to more references, nor whether it would be contrary to certain parliamentary statements. However it is necessary to go further and reach a conclusion on what is the right test in order to deal with the other grounds, summarised in paragraph 266 of the judgment of CAT quoted in paragraph 6 above, on which CAT concluded that the decision of OFT should be quashed.
The test to be applied under s.33(1)
43. The short (and correct) answer to the question is that the test to be applied is that stated in s.33(1). The words are ordinary English words; they should be applied in accordance with their ordinary meaning; the Court should not substitute other words for those used by Parliament nor paraphrase nor gloss them. Nevertheless in view of the evident importance of the test and the range of meaning the word "may" can connote it may help to explain the statutory test by reference to a series of propositions.
44. First, it is apparent from s.33(1) and the contrast between belief and suspicion demonstrated in ss.42 and 131 that it is necessary for OFT to form the relevant belief. Thus some form of mental assent is required as opposed to the less positive frame of mind connoted by a suspicion. As pointed out in the Shorter Oxford English Dictionary 3rd Edition a suspicion is but a "slight belief". In R v Monopolies Commission, ex parte Argyll plc [1986] 1 WLR 763, 769 Sir John Donaldson MR recast in what he described as simpler language the provision in s.75 empowering the Secretary of State to make a merger reference to the Commission

"where it appears to him that it is or may be the fact that arrangements are in progress or in contemplation which, if carried into effect, will result in the creation of a merger situation qualifying for investigation".

The test he adopted was that the Secretary of State might make a merger reference "if he knows or suspects" that a merger qualifying for investigation has been created or is in contemplation. In my view the slightly different wording of s.33(1) and the different context of EA, in particular the imposition of a duty rather than the conferment of a power and the distinction drawn in ss.42 and 131, do not warrant paraphrasing "believes it ...may be the case that" as "or suspects".

45. Second, the belief must be reasonable and objectively justified by relevant facts. In Education Secretary v Tameside BC [1977] AC 1014 the question was whether the Secretary of State "is satisfied". At p.1047 Lord Wilberforce pointed out that

"This form of section is quite well known, and at first sight might seem to exclude judicial review. Sections in this form may, no doubt, exclude judicial review on what is or has become a matter of pure judgment. But I do not think that they go further than that. If a judgment requires, before it can be made, the existence of some facts, then, although the evaluation of those facts is for the Secretary of State alone, the court must inquire whether those facts exist, and have been taken into account, whether the judgment has been made upon a proper self-direction as to those facts, whether the judgment has not been made upon other facts which ought not to have been taken into account."

It was not disputed that the belief must be reasonably held as accepted in paragraph 3.2 of OFT guidance quoted in paragraph 30 above.

46. Third, by themselves, the words "may be expected to result" in paragraph (b) of both s.33(1) and 36(1) involve a degree of likelihood amounting to an expectation. In paragraph 182 of its judgment CAT expressed the view that these words connoted more than a possibility and adopted what they described as a crude way of expressing the idea of an expectation as a more than 50% chance. No doubt this is right when applied to the single question which the Commission is required to answer under s.36(1)(b).
47. Fourth, however, the belief that must be held by OFT under s.33(1)(b) is "that it is or may be the case that". This introduces two alternatives, the certainty posed by the word "is" and the possibility envisaged by the words "may be". These alternatives are to be considered in relation to the circumstances set out in sub-paragraphs (a) and (b) combined and imported by reference to "the case that". If these alternatives are applied to the circumstance set out in sub-paragraph (a) and then compared with the question the Commission has to answer under s.36(1)(a) if a reference is ordered it is apparent that the degree of likelihood required by the word "may" is less than that required by the answer to question (a) in s.36(1). The answer in accordance with s.36 will be that the anticipated merger "will result in the creation of a relevant merger situation" or not as the case may be. The test for OFT is only whether the anticipated merger "may result in a relevant merger situation" or not. This is consistent with the respective functions of OFT and the Commission. The former is a first screen, the latter decides the matter. Accordingly, although the word "may" appears in the opening phrase of s.33(1) and in paragraph (b) of both s.33(1) and 36(1) it is clear that the opening phrase "believes that it ...may be the case" imports a lower degree of likelihood than paragraph (b) in ss.33(1) or 36(1) would by itself involve. That lower degree of likelihood might, for example, exist in circumstances where the work done by the OFT did not justify any positive view, but left some uncertainty, and where OFT therefore believed that a substantial lessening of competition might prove to be likely on further and fuller examination of the position (which could only be undertaken by the Competition Commission).
48. At the other end of the scale it is clear that the words "may be the case" exclude the purely fanciful because OFT acting reasonably is not going to believe that the fanciful may be the case. In between the fanciful and a degree of likelihood less than 50% there is a wide margin in which OFT is required to exercise its judgment. I do not consider that it is possible or appropriate to attempt any more exact mathematical formulation of the degree of likelihood which OFT acting reasonably must require. As Lord Mustill observed in R v Monopolies Commission, ex p S.Yorks Ltd [1993] 1 WLR 23, 29

"The courts have repeatedly warned against the dangers of taking an inherently imprecise word, and by redefining it thrusting on it a spurious degree of precision."

49. In paragraph 3.2 of OFT's published advice the requisite likelihood is described as "a significant prospect that a merger may be expected to lessen competition substantially". This substitutes "significant prospect" for "may" and is open to criticism on that account. Further I consider that the word "significant" tends to put the requisite likelihood too far up the scale of probability. With that qualification, I agree with the first and third sentences of that paragraph. It is not necessary to reach any conclusion as to the validity of the observation in the second sentence save to point out the danger of a too ready assumption that nothing has changed.
Did CAT properly apply principles of judicial review as required by s.120(4) EA?
50. In paragraphs 215 to 220 CAT pointed out that applications for judicial review to a court arise in a great variety of circumstances such that in some the Court is reluctant to trespass into areas primarily the responsibility of the executive or legislature but in others, such as pure questions of law or procedure, the court will intervene more readily. They pointed to judicial statements in Secretary of State for Education v Tameside MBC [1977] AC 1014, 1047, R v Secretary of State for Health, ex p. LB of Hackney (25th April 1994 unreported) and R v Secretary of State for the Home Department ex p. Daly [2001] UKHL 26 to the effect that the principles of judicial review depend on the context in which they fall to be applied.
51. But in paragraph 220 CAT went on to suggest that its constitution by Parliament as a specialist tribunal

"is in contrast to the more normal situation where a non-specialised court is called upon to review the decision of a specialised decision maker. For that reason we are unpersuaded that there is necessarily a direct "readover" to section 120 from cases such as Cellcom, Interbrew, T-Mobile, and the Rail Regulator that have been cited to us."

52. This conclusion was criticised by the Appellants on the grounds that s.120(4) clearly required CAT to apply the same principles "as would be applied by a court on an application for judicial review". It was submitted that those principles were well known and, though developed from time to time, applicable to all cases. The Appellants recognised that as the circumstances of the cases to which the principles had to be applied are so diverse their application is dependent on the facts. They pointed out that each of the judicial statements relied on relates to the nature of the body whose decision is being reviewed, not the specialist expertise of the court carrying out the review.
53. Counsel for IBA accepted that the principles to be applied by CAT are the ordinary principles of judicial review. In my view he was right to do so. I would accept the submissions of counsel for the Appellants that if and in so far as CAT did not apply the ordinary principles of judicial review as would be applied by a court whether on the ground that CAT is a specialist tribunal or otherwise then they failed to observe the mandatory requirements of s.120(4). The question is whether it has been shown that the Tribunal erred in this respect. The two specific points relied on are (1) reversal of the burden of proof and (2) failure to apply the Wednesbury test of reasonableness. I will examine them in turn.
Reversal of the burden of proof
54. It is common ground that the legal onus or burden of proof on an application for judicial review rests on the applicant throughout. In paragraphs 214, 230 and 253 CAT made observations which, according to the Appellants, indicate that they did not appreciate this cardinal requirement. The relevant statements are

"214. If the above analysis is correct, it also seems to us that where there is a real issue as to substantial lessening of competition, the onus is firmly on the OFT to satisfy the Tribunal that it had solid, logical and properly reasoned grounds for not complying with its duty to refer under section 33(1). That involves showing with a sufficient degree of certainty that it was entitled to come to the view that even the lower "may be the case" threshold was not met. In other words, the OFT must show that it had good grounds for believing that the matter was not even "grey", but "white"."

"230. Fourthly, in a case such as the present, where there is a real issue as to substantial lessening of competition, the onus is on the OFT to satisfy the Tribunal that it applied the right test, and that it had solid, sufficiently certain, and properly reasoned grounds for deciding that the relatively low threshold of "may be the case" under section 33(1), was not met."

"253. In those circumstances the Tribunal is simply not in a position to find that the OFT has discharged the burden of satisfying the Tribunal that there was material on the basis of which it could reasonably have come to the conclusion that it did in the decision."

55. I do not accept this criticism of CAT. All three statements were in the context of there being a real issue as to substantial lessening of competition. This had been dealt with by CAT in paragraphs 168 to 177. In that section of their judgment CAT considered whether there was such an issue. If there was not, no question of the belief of OFT being unreasonably held could arise. In paragraphs 169 to 174 CAT said that there obviously was a real issue arising from the combined market shares of iSoft and Torex, the existence of high barriers to entry, the lack of supply side substitutability and the relative smallness of the next competitor. CAT then observed that it would be difficult to put the matter more cogently than the way it was put by OFT in the Issues Letter.
56. CAT continued in paragraphs 176 and 177

"176. Nonetheless the OFT decided, apparently at a decision meeting on 8 October 2003, just over a week after sending the issues letter, that it was under no duty to make a reference to the Commission under section 33 of the Act. That decision was taken following a meeting with the parties on 2 October and a submission made by the parties on 6 October. As appears from the decision, the basis of the OFT's conclusion was, in broad terms, that the potential competition concerns did not after all arise, as a result of the countervailing effects of the NPfIT interpreted by the OFT.

177. We are not required to decide, and should not decide under section 120, whether the OFT's decision was correct on its merits. We do, however, have to decide whether the decision was lawful."

57. The concept of a shifting onus is well known, see, for example, Halsbury's Laws of England 4th Ed Vol 17(1) Reissue para 420. The onus will fluctuate from time to time in the course of a hearing such that if a prima facie case is made out by a claimant, the onus may shift to the defendant. But the onus which shifts to the defendant is only that unless he can displace the prima facie case by evidence or argument the claimant is likely to discharge the legal onus which rests on him. A good example in the field of judicial review is provided by Padfield v Minister of Agriculture, Fisheries and Food [1968] AC 997 to which I refer in paragraph 66 below. In my view it is clear beyond doubt that, read in context, this is all CAT meant. Their comments are readily understandable in the light of the OFT's apparent change of view in the course of a few days. If the hypotheses set out in the Issues Letter were well founded then OFT was bound to refer. CAT was entitled and bound to examine with care why such hypotheses were rejected in so short a time and whether their rejection was justified, particularly in view of the statutory duty to give reasons imposed by s.107. I would reject this criticism.
Failure to apply the Wednesbury test
58. As is well known "Wednesbury unreasonableness" is shorthand for an act or decision which is so unreasonable as to be an act or decision which no person or tribunal properly instructed and taking account of all but only relevant considerations could do or make. In paragraphs 223 to 225 CAT said

"223. As far as the specific context of a decision by the OFT not to make a reference under section 33(1) is concerned, it seems to us that, unlike Wednesbury itself and many leading cases on judicial review, the issue before us does not involve controlling the exercise of a discretion. Hence a test geared to controlling a discretionary power does not seem to be appropriate. The issue before us is whether the OFT has complied with a duty, and in particular whether the OFT acted unlawfully in taking the view that the underlying circumstances giving rise to the duty were not present.

224. Moreover, in the present context, the Tribunal's task is not to take a decision itself, but primarily to decide which of two other specialised decision makers, the OFT or the Commission, should take the decision. As already emerges from the earlier part of this judgment, we see this case primarily in terms of statutory construction and the process to be followed, and not in terms of deciding factual disputes.

225. As a matter of general approach, the broad question we ask ourselves is whether we are satisfied that the OFT's decision was not erroneous in law, and was one which it was reasonably open to the OFT to take, giving the word "reasonably" its ordinary and natural meaning."

59. Counsel for the Appellants contend that these statements betray a series of errors. They contend that the Wednesbury test applies to the formation of its belief by OFT which is a condition of the duty under s.33(1) arising, that the task of OFT was not to decide which Tribunal should decide the issues, and that the test of unreasonableness as being that connoted by its ordinary and natural meaning is wrong in law.
60. I will consider the second objection first. The suggestion that the task of CAT was to decide which of two specialist bodies should decide whether if the merger went through it might be expected to result in a substantial lessening of competition appears to hark back to the two part test espoused by CAT. To that extent, for the reasons I have already given, I agree that paragraph 224 betrays an error of law. But I do not think that at that stage in CAT's judgment the error is of any consequence. CAT was not then dealing with the test to be applied by s.33(1). But, equally, it can be read as a recognition of the consequence of the performance of OFT's duty under s.33(1). To that extent it is correct.
61. Of more significance are the first and third objections. Plainly unreasonableness in the ordinary and natural meaning of the word is different from Wednesbury unreasonableness. If CAT was seeking to apply the former meaning as the test of Wednesbury unreasonableness they were wrong to do so. But, once again, I do not think that a fair reading of their judgment leads to that conclusion.
62. The question of reasonableness was carried forward, through paragraph 226, to paragraph 233 where CAT said

"233. ...In other words, the OFT's approach was to seek to decide which of two plausible views the OFT preferred, rather than adopting the correct approach, namely to ask whether there were, reasonably, two views which could be taken. By failing to ask itself that latter question, the OFT failed correctly to ask itself whether "it may be the case" that the merger may be expected to result in a substantial lessening of competition within the meaning of section 33(1)."

63. CAT then continued with a section headed 'Could the OFT have reasonably excluded the alternative view?'. The first paragraph of that section, paragraph 234, states

"234. Secondly, if, which is not established before the Tribunal, the OFT believed not only that there was no significant prospect of the merger resulting in a substantial lessening of competition, but also that there was no significant prospect of the Commission coming to an alternative view after a fuller investigation, the Tribunal is not satisfied that was a view that the OFT could reasonably have reached."

This idea is carried forward into subsequent passages in which CAT considers not whether they can interfere with OFT's decision as being unreasonable but whether OFT asked itself the right questions and gave adequate reasons for its answer.

64. For these reasons I reject the first and third objections too. I do not consider that CAT adopted the wrong standard of unreasonableness when seeking to apply the Wednesbury test. They were considering a different question, whether if there are two or more tenable views as to the likelihood of a substantial lessening of competition OFT was reasonable to reject that which produced an affirmative answer. This question arose in the context that it was and is common ground that the belief of OFT referred to in s.33(1) had to be reasonably held, a requirement which is emphasised in OFT's own guidance quoted in paragraph 30 above. It is in this sense that I read the passage in paragraph 266 quoted in paragraph 6 above that

"we are not satisfied that the OFT applied the right test, or that the OFT reached a conclusion that was reasonably open to them."

Did the reasons given by OFT justify their conclusion?
65. This, in summary, is the last of the reasons given by CAT in its conclusions set out in paragraph 266 for quashing the decision of OFT. It does not appear to me to be dependent on the two-stage test formulated and applied by CAT in earlier sections of their judgment. Nevertheless it is necessary to consider it on the basis of the correct test as to likelihood to which I have referred in paragraphs 43 to 49 above.
66. I take the test to be that formulated by Lord Wilberforce in Education Secretary v Tameside BC [1977] AC 1014, 1047 which I have quoted in paragraph 44 above. A good example is provided by the decision of the House of Lords in Padfield v Minister of Agriculture, Fisheries and Food [1968] AC 997. In that case the Minister had the power to direct a committee of investigation to consider and report to him in respect of any complaint as to the operation of any marketing scheme for agricultural produce. Milk producers in the South Eastern Region complained about the price paid by the milk marketing board for their milk when compared with prices paid to producers in the Far-Western region. They asked the Minister to appoint a committee of investigation to investigate their complaint. The Minister refused.
67. At p.1058 Lord Upjohn noticed that the Minister had a discretion so that the real question was how far it was subject to judicial control. Having summarised the four conventional heads under which the exercise of such a discretion may be attacked he said

"In practice they merge into one another and ultimately it becomes a question whether for one reason or another the Minister has acted unlawfully in the sense of misdirecting himself in law, that is, not merely in respect of some point of law but by failing to observe the other headings I have mentioned.

In the circumstances of this case, which I have sufficiently detailed for this purpose, it seems to me quite clear that prima facie there seems a case for investigation by the committee of investigation. As I have said already, it seems just the type of situation for which the machinery of section 19 was set up, but that is a matter for the Minister.

He may have good reasons for refusing an investigation, he may have, indeed, good policy reasons for refusing it....So I must examine the reasons given by the Minister, including any policy on which they may be based, to see whether he has acted unlawfully and thereby overstepped the true limits of his discretion, or as it is frequently said in the prerogative writ cases, exceeded his jurisdiction. Unless he has done so, the court has no jurisdiction to interfere."

Lord Upjohn then proceeded to consider the Minister's various reasons individually and in detail.

68. A similar course was taken by CAT in this case. In paragraphs 240 to 247 they considered whether the facts are sufficiently found in the decision. They considered them to be defective in a number of respects, including the lack of any clear exposition of the working of the market concerned (para 240) and the lack of any satisfactory description of what NPfIT would involve and when (para 241). CAT concluded (para 243) that the description of the market was so scanty that it was extremely difficult to be satisfied that all material considerations had been taken into account and all material facts ascertained. In paragraph 246 they said

"The present matter is a case of a direct merger between two companies who compete horizontally and who are identified in the decision as numbers 1 and 2 in the market, with combined market shares in the 45%/55% range. A decision by the OFT to the effect that on no reasonable view could such a merger be expected to lead to a substantial reduction of competition in our view needs a proper factual basis and exceptional clarity of analysis. We do not find such a basis in the decision."

69. Finally in paragraphs 248 to 265 CAT considered the question whether there was material on which OFT could base their decision. In paragraph 248 CAT listed 10 matters in issue arising from OFT's decision. In paragraphs 254 to 262 CAT dealt in detail with the issue concerning the significance of the legacy base. CAT concluded in paragraph 263

"Self evidently, issues like these cannot be resolved, or even gone into, on an application for judicial review. But the resulting picture the Tribunal has is one of considerable confusion, in which the exact nature of the competitive process in relation to new, extended or upgraded contracts, and the significance of incumbency, is far from clear. Particularly in a sector of national importance, where large amounts of public money are at stake, a decision such as the present should in our view clearly set out the OFT's reasoning on issues such as these, together with sufficient material to show that the conclusion can be supported and that the matter has been properly investigated. The Tribunal has been unable to satisfy itself that such is the case here."

70. In my view the issue for this court is whether that conclusion can stand, notwithstanding the mistaken adoption of the two-stage test under s.33(1). In my view it is not dependent on that test and must be considered on its own merits. I bear in mind the warning given by Lord Brightman in R v Hillingdon BC, ex p. Pulhofer [1986] AC 484, 518B-F that

"Where the existence or non-existence of a fact is left to the judgment and discretion of a public body and that fact involves a broad spectrum ranging from the obvious to the debatable to the just conceivable, it is the duty of the court to leave the decision of that fact to the public body to whom Parliament has entrusted the decision-making power save in a case where it is obvious that the public body, consciously or unconsciously, is acting perversely."

Equally it is appropriate to bear in mind that by perversity is meant that class of case which Lord Radcliffe described in Edwards v Bairstow [1956] AC 14, 36 as those where

"...the facts found are such that no person acting judicially and properly instructed as to the relevant law could have come to the determination under appeal. In those circumstances, too, the court must intervene. It has no option but to assume that there has been some misconception of the law and that this has been responsible for the determination."

Later he described such a case as one where "the only reasonable conclusion contradicts the determination". By those standards I consider that CAT's conclusion should be upheld. I can explain my reasons relatively shortly.

71. The Issues Letter set out 10 matters in respect of which a significant lessening of competition might be expected to result from the proposed merger. The letter explained that they were hypotheses which OFT were still evaluating in the light of evidence put before them by iSoft and Torex as well as by third parties. Seemingly some eight days later all these hypotheses had been rejected or discounted to such an extent that OFT felt able reasonably to believe that there was no likelihood that the proposed merger would lead to a significant lessening of competition.
72. Not only did IBA not see this letter but they had no opportunity to comment on iSoft and Torex's responses to it. Moreover OFT did not apparently see the OBS to which I referred in paragraph 11 above. Further in each of paragraphs 29 to 32 in which OFT expresses its conclusions it does so in terms of likelihood. Thus,

"the competitive advantage acquired from the legacy base is unlikely to be strong, especially where a new procurement strategy is being introduced" (para 29)

"Absent the merger, Torex is likely to face significantly reduced opportunities to sell its products to hospital users in England" (para 30)

"The presence of these international competitors makes it likely that competition for future contracts will remain active." (para 31)

"There is a reasonable prospect that international competitors with a UK base will bid for contracts in the regions with the likely effect of increased competition for contracts in Northern Ireland, Scotland, and Wales." (para 31)

"While a strong legacy base may give the parties a large presence it is unlikely, in itself, to confer significant market power in view of the changes being brought about by NPfIT." (para 32)

73. None of these conclusions excludes a likelihood to the opposite effect so as reasonably to justify the belief that the anticipated merger "may" result in a significant lessening of competition. No doubt OFT is right to conclude (para 32) that

"Such a fundamental change (NPfIT) has altered the future competitive landscape with the effect that competitive constraints must be viewed under a new scenario."

But that does not, in my view, justify the further conclusion that such a scenario overcomes the anti-competitive features which do exist to such an extent as to remove the requisite likelihood of a significant lessening of competition.

74. Such features include the facts that the merger is horizontal, the merger parties enjoy a market share well above the usual 25% level, the next competitor enjoys a market share well below that enjoyed by the merger parties, supply substitutability is limited, the merger parties enjoy the benefit of a network effect and information asymmetry. All of those features are referred to by OFT in its guidance as important in assessing the effect of a proposed merger on competition.
75. In these circumstances, it appears to me that either OFT applied too high a test of likelihood when forming their belief or they failed adequately to justify the belief they formed in accordance with the proper test. In either event, notwithstanding the fact that CAT adopted a wrong test as to likelihood, their ultimate conclusion is right and should be upheld.
Conclusions
76. For all these reasons I conclude that

a) the two-stage test formulated by CAT is not the test of likelihood required by s.33(1);

b) CAT did not wrongly reverse the burden of proof;

c) CAT did not wrongly apply the Wednesbury test of unreasonableness;

d) CAT were right to conclude that either OFT had applied the wrong test as to likelihood or they had failed adequately to explain or justify their conclusion in accordance with the right test.

I would dismiss this appeal.

Lord Justice Mance:

77. I have had the benefit of studying in draft the judgment given by the Vice-Chancellor and the supplementary observations made by Carnwath LJ, and I agree with both. I specifically agree with the Vice-Chancellor's examination in paragraphs 43 to 49 of the test to be applied under s.33(1).

Lord Justice Carnwath:

78. I agree that the appeal should be dismissed for the reasons given by the Vice-Chancellor. Given the novelty and complexity of the statutory framework, I shall add some comments of my own, under three heads: the statutory test, the principles of judicial review, and the duty to give reasons.
The statutory test
79. The annotations to the Current Law edition of the Enterprise Act 2002 refer to the changes made by Part 3 of the Enterprise Act 2002 as "dramatic", in two ways: -

"First the Secretary of State is removed from the process. Under the previous position the OFT merely referred cases to the Secretary of State who made the final decision whether to refer. Second, the previous 'public interest' test is replaced with a 'competition-based' substantial lessening of competition test."

(The latter test is referred to in the annotations as the "SLC test". For convenience I shall follow the same approach and use the term "SLC" to mean "a substantial lessening of competition". This is perhaps one of those rare cases where such shorthand is permissible as an aid to finding a way through the density of the statutory language.)

80. These major changes make it unwise, in my view, to seek assistance in the caselaw and practice under the previous statutes, even though some of the wording of the 2002 Act echoes phrases used in the Fair Trading Act 1973. The 2002 Act must be construed in accordance with its own terms and apparent purpose.
81. I see the key to the present issue in the contrast between the respective roles of the OFT under section 33 and the Commission under section 36 (confining attention for these purposes to anticipated mergers). To highlight the critical point, it is helpful to simplify the other elements. The ultimate questions for both authorities are defined in effectively the same terms, by paragraphs (a) and (b) of sections 33(1) and 36(1) respectively. In this case, (a) does not arise; we can assume a "merger situation". We are concerned only with (b): whether the creation of the merger situation "may be expected to result in" SLC. The Tribunal (without dissent) has interpreted "may be expected" as implying a "more than 50% chance" (para 182). Thus, in effect, the factual judgment has to be made on the balance of probabilities. With that qualification, and reduced to its essentials, the question for each body under (b) is simply: will the merger situation result in SLC?
82. The difference between ss 33 and 36 lies in the nature of the conclusion to be arrived at. The question for the OFT is whether it "believes that (SLC) is or may be the case"; the Commission is required to "decide" whether there will be SLC. Thus for the OFT, unlike the Commission, belief in the possibility of SLC is enough to trigger the next stage.
83. At first sight, it might be thought that the OFT is required to do no more than form a preliminary view of the matter, so that any case raising an arguable issue of SLC should be referred to the Commission. This is one possible view of a "first-screen" role (to use the Tribunal's term: heading to para 73ff). If that were the correct view, it would be difficult to see any answer to the case for a reference on the facts of the present case. As the OFT's evidence makes clear, this case passed the initial screening stage for cases which could be cleared as "not raising any complex or material competition concerns". It was selected for treatment under the procedure for cases "identified as potentially raising complex or material competition issues…" leading to the preparation of the "issues letter". (Mr Gaddes' second witness statement para 4(e)-(f)). At that point the OFT clearly considered that there was at least a possibility of SLC. If the matter had to stop there, a reference would have been inevitable.
84. However, other indications in the statute show that the formation of a "belief" by the OFT requires something more than that. The Vice-Chancellor has referred to the provisions relating to public interest cases, in which (in relation to the Secretary of State's role) the word "belief" is used in contrast to having "reasonable grounds for suspecting" (s 42 (1)(2)). Furthermore, the "beliefs" of the OFT on the question of SLC are treated as "decisions" which are binding on the Secretary of State (see ss 44(4), (46 (2)), as to whether SLC is expected to result, or not to result (see s 45(2)(b),(3)(b)). They are also treated as "relevant decisions" on which the OFT is under a duty to carry out prior consultation with relevant parties upon what is proposed (s 104 (2), (6)).
85. These pointers, combined with the undoubted expertise of the OFT, indicate something more than a purely "first-screen" role. The precise extent or timescale of what is required by way of investigation is not defined by the statute. It is no doubt implicit that the OFT must take "reasonable steps" to inform itself within the limits of its statutory role (see Secretary of State for Education v Tameside BC [1977] AC 1014, 1077). However, it is not suggested that the OFT failed in that respect. The Tribunal was entitled to assume that, at least by the time of the hearing, the OFT had done whatever it thought necessary to reach the stage of decision required by its statutory role.
86. Having reached that stage, there were three possible views for the OFT to take under the wording of section 33(1): (1) that it believed that there would be SLC; (2) that it believed there might be SLC; (3) that the risk of SLC was sufficiently low for the OFT to believe there neither was nor might be a SLC. (1) and (2) follow from the words "is or may". (3) is their implicit obverse. I agree with the Vice-Chancellor (para 49) that the word "significant", as used in the OFT advice, is capable of being misread as setting the test too high. No doubt the possibility must be more than fanciful, but subject to that I would prefer not to qualify the statutory wording.
87. This analysis helps to highlight what to me is the basic distinction between the roles of the OFT and the Commission. The Commission has to reach a "decision" on the SLC question, and for that purpose is given extensive powers for gathering information, calling witnesses and making a detailed investigation (see s 109). The OFT is required only to form a belief as to the possibility of SLC, and its powers are accordingly more limited. This relationship seems to me entirely consistent with what was said in the White Paper ("Productivity and Enterprise"), which preceded the statute:

"5.20 The new regime retains a two-stage approach to merger investigations. The OFT will conduct the first-stage investigation to decide whether a reference is necessary. The Competition Commission will continue to carry out such references via a second-stage, in-depth investigation."

The material placed before the Tribunal represented the results of the first-stage investigation. The issue for the Tribunal was whether on that material the OFT could reasonably take the view that the issues (so clearly identified by it in the "issues letter") had been sufficiently resolved for it to be satisfied that there would not be SLC. If not, it was its duty to refer the matter for "in-depth" investigation by the Commission.

Principles of Judicial Review
88. The Tribunal was required to apply the principles which would be applied "by a court on an application for judicial review" (s 120(4)). On its face, this seems a clear indication that, notwithstanding the Tribunal's specialised composition, the review was not to take the form of an appeal on the merits, but was limited by the ordinary principles applied in the Administrative Court.
89. The Tribunal expressed their difficulty in interpreting this duty (para 217). In these circumstances, they might have found help in the leading textbooks on the subject. The case-law on the principles of judicial review, even at the appellate levels, is now so voluminous, and the subject-matter so diverse, that the assistance of an authoritative guide (such as de Smith or Wade, to name only two of many) should in my view be regarded as indispensable when considering the application of those principles to a new statutory regime. (References below are to De Smith, Woolf and Jowell Judicial Review of Administrative Action 5^th Ed; and Wade and Forsyth Administrative Law 8th Ed.. Another invaluable source-book for comparison of relevant case-law on particular issues is Michael Fordham's Judicial Review Handbook 3rd Ed.)
90. For example, the CAT was right to observe that its approach should reflect the "specific context" in which it had been created as a specialised tribunal (paras 224); but it was wrong to suggest that this permitted it to discard established case-law relating to "reasonableness" in administrative law, in favour of the "ordinary and natural meaning" of that word (para 225). Its instinctive wish for a more flexible approach than Wednesbury would have found more solid support in the textbook discussions of the subject, which emphasise the flexibility of the legal concept of "reasonableness" dependent on the statutory context (see De Smith para 13-055ff "The intensity of review"; cf Wade p 364ff "The standard of reasonableness", and the comments of Lord Lowry in R v Secretary of State ex p Brind [1991] 1AC 696, 765ff).
91. Thus, at one end of the spectrum, a "low intensity" of review is applied to cases involving issues "depending essentially on political judgment" (de Smith para 13-056-7). Examples are R v Secretary of State, ex p Nottinghamshire CC [1986] AC 240, and R –v- Secretary of State ex p Hammersmith and Fulham LBC [1991] 1AC 521, where the decisions related to a matter of national economic policy, and the court would not intervene outside of "the extremes of bad faith, improper motive or manifest absurdity" ([1991] 1AC at 596-597 per Lord Bridge). At the other end of the spectrum are decisions infringing fundamental rights where unreasonableness is not equated with "absurdity" or "perversity", and a "lower" threshold of unreasonableness is used:

"Review is stricter and the courts ask the question posed by the majority in Brind, namely, "whether a reasonable Secretary of State, on the material before him, could conclude that the interference with freedom of expression was justifiable." (De Smith para 13-060, citing Brind –v- Secretary of State [1991] AC 696)."

92. A further factor relevant to the intensity of review is whether the issue before the Tribunal is one properly within the province of the court. As has often been said, judges are not "equipped by training or experience or furnished with the requisite knowledge or advice" to decide issues depending on administrative or political judgment (see Brind [1991] 1AC at 767, per Lord Lowry). On the other hand where the question is the fairness of a procedure adopted by a decision-maker, the court has been more willing to intervene:

"Such questions are to be answered not by reference to Wednesbury unreasonableness, but 'in accordance with the principles of fair procedure which have been developed over the years and of which the courts are the author and sole judge'" (R –v- Takeover Panel ex parte Guinness plc [1990] 1QB 146, 184, per Lloyd LJ).

93. The present case, as the Tribunal observed (para 223), is not concerned with questions of policy or discretion, which are the normal subject-matter of the Wednesbury test. Under the present regime (unlike the 1973 Act) the issue for the OFT is one of factual judgment. Although the question is expressed as depending on the subjective belief of the OFT, there is no doubt that the court is entitled to enquire whether there was adequate material to support that conclusion (see Tameside case, [1977] AC at 1047 per Lord Wilberforce).
94. Again, even in relation to factual issues, the cases show considerable variation in the "intensity" of review. The well-known case of Edwards –v- Bairstow [1956] AC 14 is a good example of the flexible approach of the court to such issues, outside the sphere of political judgment.
95. The case is sometimes misrepresented as a mere application of Wednesbury unreasonableness. This may owe something to the fact that it was cited by Lord Diplock in the CCSU case [1985] AC 374, 410) in the context of his famous re-statement of Wednesbury unreasonableness in terms of "irrationality": that is –

"… a decision which is so outrageous in its defiance of logic or of accepted moral standards that no sensible person who had applied his mind to the question to be decided could have arrived at it."

However, his reference to Edwards –v Bairstow was simply in relation to his view that "irrationality" could now "stand upon its own feet", without resort to the inference of a mistake of law. The actual decision in Edwards –v- Bairstow could certainly not be described as "outrageous" in any sense (at least without gross unfairness to the Tax Commissioners who made it). The issue was whether a particular transaction was "an adventure in the nature of trade". Although the House of Lords accepted that this was "an inference of fact", they held that on the primary facts as found by the Commissioners "the true and only reasonable conclusion" contradicted that decision ([1956] AC at p36 per Lord Radcliffe).

96. The concluding remarks of Lord Radcliffe's speech are often overlooked. He criticised the tendency of the courts to treat such questions as "pure questions of fact", so as to exclude review:

"As I see it, the reason why the courts do not interfere with the Commissioners' findings or determinations when they really do involve nothing but questions of fact is not any supposed advantage in the Commissioners of greater experience in the matters of business or any other matters. The reason is simply that by the system that has been set up the Commissioners are the first tribunal to try an appeal, and in the interest of the efficient administration of justice their decisions can only be upset on appeal if they have been positively wrong in law. The Court is not a second opinion where there is a reasonable ground for the first. But there is no reason to make a mystery about the subjects that Commissioners deal with or to invite the courts to impose any exceptional restraint on themselves because they are dealing with cases that arise out of facts found by the Commissioners. Their duty is no more than to examine those facts with a decent respect for the tribunal appealed from and if they think that the only reasonable conclusion on the facts found is inconsistent with the determination come to, to say so without more ado." (pp 38-9, emphasis added).

97. At the other extreme is R v Hillingdon ex p Puhlhofer [1986] AC 485, another case which is often cited without sufficient regard to its special context. The subject-matter was the duty of a local housing authority towards the homeless, the imposition of which depended on the authority's conclusion on various factual questions defined by the statute. Lord Brightman, giving the leading speech, was concerned by the "mass of litigation" affecting authorities who were "endeavouring to cope with intractable housing problems and to balance competing claims to limited housing resources" (p 511B). It was against that background that he referred to the limited role of the judicial review, which he defined by reference to the narrowest version of the Wednesbury test ("unreasonableness verging on absurdity"), as applied in cases such as the Nottinghamshire CC case (see above). It is clear that this was a response to the particular context, and there is no reason to see it as intended as a statement of more general application.

(In fact, his comments on the Wednesbury test, though of course of great authority, were not essential to the decision. This turned on a pure issue of construction: whether the term "accommodation" was to be read as qualified by some word such as "appropriate" or "reasonable". The House of Lords answered that question in the negative, holding that Parliament had "plainly and wisely" added no such qualifying adjective, and none was to be implied (p 517E). The difficulty of divining the Parliamentary intention in such cases is underlined by the fact that the decision was reversed later the same year by a statutory amendment adding the word "suitable": Housing and Planning Act 1986, s 14(2)).)

98. Another illustration of the different approaches to review - closer to the present context - is the decision of Lightman J in R v Director General ex p Cellcom [1999] ECC 314 (one of the cases referred to but discarded by the Tribunal: para 220). The duty of the Director was to exercise his functions in the manner which

"he considers best calculated to secure… such telecommunications services as satisfy all reasonable demands for them…" (Telecommunications Act 1984, s 3(1)).

Lightman J held that the section drew a distinction between "means" and "ends":

"… whilst the Director is expressly made the arbiter of the means to the ends, he is not made the arbiter of the ends. Section 3 recognises that there is a public interest in reasonable demands for telecommunications services being met and the court is intended to be the guardian of that public interest." (p 330.)

His subsequent exposition of Wednesbury principles (p 331) was directed only to the policy issues (relating to "means") on which the decision-making power was conferred on the Director, rather than the court.

99. Finally, reference may be made to a recent case in the House of Lords (Moyna v Secretary of State [2003] UKHL 44), where Lord Hoffmann discussed the difficulty of finding a clear or logical division, in the case-law, between issues of fact and issues of law. Edwards-v-Bairstow was seen as exemplifying the proposition that -

"the question whether the facts as found fall on one side or the other of a conceptual line by the law is a question of fact"

and that, on an appeal limited to questions of law, the court was able to interfere if the decision –

"falls outside the bounds of reasonable judgment" (para 25).

In his view the lack of a clear division caused no difficulty –

"… as long as it is understood that the degree to which an appellate court will be willing to substitute its own judgment for that of the tribunal will vary with the nature of the question." (para 27, citing his own judgment in In re Grayan Building Services [1995] Ch 241, 254-5).

100. I have referred to these cases in some detail, because they show that the Tribunal did not need to rely on some special dispensation from the ordinary principles of judicial review. Those principles, whether applied by a court or a specialised tribunal, are flexible enough to be adapted to the particular statutory context. No doubt the existence of such a special jurisdiction will help to ensure consistency from case to case; and the expertise of the Tribunal will better fit it to deal with such cases expeditiously and with a full understanding of the technical background. However, the essential question was no different from that which would have faced a court dealing with the same subject-matter. That question was whether the material relied on by the OFT could reasonably be regarded as dispelling the uncertainties highlighted by the issues letter. That question was wholly suitable for evaluation by a court. It involved no policy or political judgment, such as would be regarded as inappropriate for review by the Administrative Court.
101. On that basis, I agree with the Vice-Chancellor, for the reasons given by him, that the Tribunal was fully justified in reaching the conclusion that the uncertainties had not, on any reasonable view, been adequately resolved.
Adequacy of reasons
102. Finally, although inadequacy of reasons is not a ground of challenge as such, it may be helpful to comment briefly on the Tribunal's observations on this aspect.
103. The Tribunal expressed concern at having to consider material outside the decision letter. It noted that the OFT was under a statutory duty to give its reasons, and referred to what it called the "well-known principle" that -

"the Court should at the very least be circumspect about allowing material gaps to be filled by affidavit evidence or otherwise" (para 257, citing R–v- Westminster City Council ex parte Ermakov [1996] 2AllER 302, 312E per Hutchinson LJ).

It commented:-

"If a material element is not set out in the decision, it is very difficult for the reviewing court or tribunal to be satisfied that the matter was properly investigated or that the supplementary reasons did in fact form part of the decision making process." (para 258).

In other parts of the judgment, the Tribunal criticised the failure of the OFT to set out all the underlying material (see para 211, 252).

104. With respect, I think this concern, and the associated criticisms, were misplaced. The statutory duty to give reasons is an important one, but it is not the same as a duty to give a "judgment" (such as that of a court) or a duty to make a "report" (such as that of an inquiry inspector). Again reference to the textbooks might have assisted. The numerous cases on the subject lay down no general test, other than the requirement that reasons must be "intelligible and must adequately meet the substance of the arguments advanced" (see Re Poyser and Mills Arbitration [1964] 2QB 467, 477-478; cited in de Smith para 9-049 as "the most frequently cited judicial articulation of the test"; see also Wade pp 916-9).
105. In a case such as the present, where the subject-matter is complex and the supporting material voluminous, there is no statutory requirement for all the evidence to be set out in the decision letter. However when a challenge is made, there is, as the Tribunal noted, an obligation on a respondent public authority to put before the Court the material necessary to deal with the relevant issues; "all the cards" should be "face upwards on the table" (see R –v- Lancashire County Council ex parte Huddleston [1986] 2AllER 941 ).
106. There is certainly nothing unusual, particularly in a case which has to be dealt with in a relatively short timescale, for the stated reasons to be amplified by evidence before the Court. While in some areas of the law, the Court may need to be "circumspect" to ensure that this is not used as means of concealing or altering the true grounds of the decision, that does not arise in this case. As I understand it, no objection had been taken to any of the evidence being put before the Tribunal (or to the additional evidence adduced in the Court of Appeal). The question for the Tribunal was not whether the reasoning was adequately expressed in the decision, but whether the material ultimately before it, taken as a whole, disclosed grounds on which the Tribunal could reasonably have reached the decision it did.