Cérélia Group Holdings SAS & Anor v Competition and Markets Authority

This is the single judgment of the Court. The appeal concerns the judgment of the Competition Appeal Tribunal (“CAT”) dated 1 September 2023 (“the Judgment”) upholding the decision of the Competition and Markets Authority (“CMA”) which found that the completed acquisition by Cérélia Group Holding SAS (“Cérélia”) of certain assets relating to the United Kingdom (UK) and Ireland dough business of General Mills, Inc. (“GMI”), operated under the “Jus-Rol” brand (“Jus-Rol”), resulted in a substantial lessening of competition (“SLC”) in the wholesale supply of dough-to-bake (“DTB”) products to grocery retailers in the UK, harming the interests of these retailers and, potentially, end-consumers of such products. The CMA, in the exercise of statutory powers under the Enterprise Act 2002 (“EA 2002”), concluded that the only remedy sufficient to remedy the SLC was divestiture.

On 31 January 2022, Cérélia completed the acquisition of the Jus-Rol business. On 1 February 2022, the Mergers Intelligence Committee of the CMA informed Cérélia of its intention to investigate the acquisition. The CMA commenced a Phase I investigation on 31 March 2022 and issued a Phase I decision on 30 May 2022 to the effect that, on the information currently available to it, it was or may have been the case that the merger may be expected to result in a SLC within a market or markets in the UK.

In consequence, on 15 June 2022, the CMA referred the merger for a Phase II investigation and report by a group of CMA panel members (under section 22(1) EA 2002). On 5 October 2022, the CMA issued a statutory notice under section 39(3) EA 2002 extending the reference period by eight weeks to expire on 24 January 2023, instead of 29 November 2022.

The CMA issued its “Provisional Findings” (“PF”) and notice of possible remedies on 4 November 2022, and continued to collect evidence. It issued a Remedies Working Paper on 16 December 2022 and a Consultation Paper on 19 December 2022 with a summary of the additional evidence gathered. It invited the parties to make any representations on the additional evidence by 3 January 2023.

The final decision is contained in a CMA document dated 20^th January 2023. This is referred to as the Final Report (“FR”). It runs to 423 pages in length. It found that the merger had (i) given rise to a “relevant merger situation”; and (ii), resulted in a SLC in the UK market for the wholesale supply of DTB products to grocery retailers. The CMA determined that Cérélia should be required to divest the entire Jus-Rol UK business to a suitable purchaser albeit, and subject to approval, that it would have the right to carve out brand and IP rights in relation to the Ireland business. In a nutshell, the CMA found that Cérélia and GMI were the two largest suppliers of DTB products at the wholesale level to retailers and were in competition with each other, even though Cérélia concentrated on supplying retailer own-brand product and GMI concentrated on branded product (i.e. Jus-Rol). The merger would remove a significant amount of competition from the marketplace, which could have a detrimental effect upon competition and in particular price. In relation to potential sources of competition elsewhere in the market which might serve to counteract the SLC, the CMA concluded that retailers did not have sufficient buyer power to negate any lessening of competition and that third party entry into the market was improbable. The only potential source of competitive constraint would therefore have to come from third party rivals to the merged entity. The third and fourth largest suppliers were Henglein and Bells respectively. They were significantly smaller in scale than the merged entity and for a variety of reasons, individually or collectively, would not create a sufficient competitive constraint to prevent the merged entity being able to exploit market power to the detriment of customers. It followed that the merger would give rise to a SLC for which the only remedy, the transaction having been completed, was divestiture.

There are five grounds of appeal.

Grounds I and II concern the findings of the CAT that the CMA correctly concluded that neither Bells nor Henglein would serve to keep the merged entity in competitive check. The appellants argue that the analysis of the CMA was deficient and should not have been endorsed by the CAT.

Grounds III – V raise a significant procedural issue relating to the decision of the CMA, during the administrative procedure leading up to the FR, to extend the time for the completion of the examination by two months. According to section 39(3) EA 2002 the CMA is empowered to extend time only when there are “special” reasons. Before the CAT the CMA set out in a witness statement its reasons for extending time. The appellant did not challenge either the admissibility or accuracy of the reasoning. Instead it argued that, as a matter of statutory construction, these reasons did not reveal anything “special” and therefore the extension was unlawful and it followed that the CAT erred in law in agreeing that the reasons were “special” (Ground III). This being so, the CMA lacked the power or jurisdiction to extend time and there was accordingly no lawful decision of the CMA in existence as at the end of the statutory time limit for completion of the investigation, which meant that there was nothing capable of prohibiting the merger (Ground IV). In any event the reasoning of the CMA was irrational (Ground V). Shortly prior to the appeal hearing Cérélia applied for permission to add a new ground of appeal and modify one of its existing grounds. These new arguments concerned (i) an alleged failure by the CMA properly to analyse market shares and (ii) a failure properly to consult. The Court indicated that it would hear full argument on the new arguments and would deal with the issue of whether permission to appeal should be granted in this judgment.

The procedure governing the regulation of mergers is set out in the EA 2002. Section 35, so far as relevant, provides as follows:

“35 Questions to be decided in relation to completed mergers

(1)

Subject to subsections (6) and (7) and section 127(3), the CMA shall, on a reference under section 22, decide the following questions—

(a)

whether a relevant merger situation has been created; and

(b)

if so, whether 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.

(2)

For the purposes of this Part there is an anti-competitive outcome if—

(a)

a relevant merger situation has been created and 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; or

(b)

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

The CMA shall, if it has decided on a reference under section 22 that there is an anti-competitive outcome (within the meaning given by subsection (2)(a)), decide the following additional questions—

(a)

whether action should be taken by it under section 41(2) for the purpose of remedying, mitigating or preventing the substantial lessening of competition concerned or any adverse effect which has resulted from, or may be expected to result from, the substantial lessening of competition;

(b)

whether it should recommend the taking of action by others for the purpose of remedying, mitigating or preventing the substantial lessening of competition concerned or any adverse effect which has resulted from, or may be expected to result from, the substantial lessening of competition; and

(c)

in either case, if action should be taken, what action should be taken and what is to be remedied, mitigated or prevented.

(4)

In deciding the questions mentioned in subsection (3) the CMA shall, in particular, have regard to the need to achieve as comprehensive a solution as is reasonable and practicable to the substantial lessening of competition and any adverse effects resulting from it.”

Section 38 imposes a duty upon the CMA to publish a report on the reference:

“38.

Investigations and reports on references under section 22 …

(1)

The CMA shall prepare and publish a report on a reference under section 22 […] within the period permitted by section 39.

(2)

The report shall, in particular, contain—

(a)

the decisions of the CMA on the questions which it is required to answer by virtue of section 35 […];

(b)

its reasons for its decisions; and

(c)

such information as the CMA considers appropriate for facilitating a proper understanding of those questions and of its reasons for its decisions.

(3)

The CMA shall carry out such investigations as it considers appropriate for the purposes of preparing a report under this section.”

Section 103(1) requires the CMA to have regard to the need to make decisions on references as soon as reasonably practicable.

Where the CMA’s proposed decision on the questions mentioned in sections 35(1) or (3) is likely to be adverse to the interests of a relevant party, section 104 imposes duties upon the CMA to consult and to give reasons for its proposed decision to that party:

“104.

Certain duties of [the CMA] to consult

(1)

Subsection (2) applies where the relevant authority is proposing to make a relevant decision in a way which the relevant authority considers is likely to be adverse to the interests of a relevant party.

(2)

The relevant authority shall, so far as practicable, consult that party about what is proposed before making that decision.

(3)

In consulting the party concerned, the relevant authority shall, so far as practicable, give the reasons of the relevant authority for the proposed decision.

(4)

In considering what is practicable for the purposes of this section the relevant authority shall, in particular, have regard to—

(a)

any restrictions imposed by any timetable for making the decision; and

(b)

any need to keep what is proposed, or the reasons for it, confidential.

(5)

The duty under this section shall not apply in relation to the making of any decision so far as particular provision is made elsewhere by virtue of this Part for consultation before the making of that decision.

[…].”

Section 120, so far as relevant, governs challenges to CMA decisions:

“120.

Review of decisions under Part 3

(1)

Any person aggrieved by a decision of the CMA … under this Part in connection with a reference or possible reference in relation to a relevant merger situation or a special merger situation may apply to the Competition Appeal Tribunal for a review of that decision.

…

(4)

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.

(5)

The Competition Appeal Tribunal may—

(a)

dismiss the application or quash the whole or part of the decision to which it relates; and

(b)

where it quashes the whole or part of that decision, refer the matter back to the original decision maker with a direction to reconsider and make a new decision in accordance with the ruling of the Competition Appeal Tribunal.”

We summarise below those parts of the FR relevant to the grounds advanced. A great deal of the material in the FR is commercially sensitive and treated as confidential. For this reason the analysis in this judgement is necessarily expressed in broad terms without recourse to detail which would otherwise have been included to aid understanding. We have, nonetheless, endeavoured to provide as detailed an explanation as we are able within the confines of the agreed confidentiality regime.

The merger concerns “DTB” products. These include ingredient pastry dough (i.e., shortcrust, puff and filo pastry dough), pizza dough and other ready-to-bake (RTB) dough products (including RTB croissant dough, pain au chocolat dough, cinnamon swirl dough, gingerbread dough, cookie dough). DTB products are manufactured by combining flour with a liquid (e.g. water) and/or fat (butter, olive oil etc.) and sometimes with flavoured toppings. They are primarily sold in grocery retailers in the chilled shelves as part of the butter, spreads and margarine (BSM) category of products. A smaller proportion are also available as frozen products, or at ambient temperatures.

Most of the largest retailers in the UK stock both private label (“PL”) and branded DTB products. PL products are also known as “own brand” or “own label”’ products and are products sold exclusively by a given retailer with their own packaging and branding. This is referred to as the “PL channel”. Branded products are sold under the brand name of the suppliers that sell them to retailers. This is referred to as “branded channel”. In the FR the CMA concluded that PL and branded products formed part of the same product market, and were in competition with each other, even though they were supplied through different channels.

The Jus-Rol business is by far the largest supplier of DTB products to grocery retailers in the UK and the only full range brand with a national presence. Prior to the merger, the Jus-Rol business was owned by GMI, a US-based global manufacturer and supplier of consumer and pet food. The Jus-Rol business supplies branded DTB products to grocery retailers and foodservice customers primarily in the UK, and to a lesser extent in Ireland. The UK product range is available either chilled or frozen in sheets, block and RTB forms. Specific products within the range include ingredient pastry dough, pizza dough, sharing bread dough and certain breakfast DTB products supplied in cans such as croissant dough, pain au chocolat dough and cinnamon swirl dough.

Cérélia is the second largest supplier of DTB products in the UK. It is the largest supplier of PL products to grocery retailers. It is a joint stock company headquartered in Paris, France. It is controlled by funds affiliated with the private equity firm Ardian France SA (Ardian). It produces pies, pizzas, pastry dough, crepes, pancakes, waffles, cookie dough and ready to eat cookies for its own brands and for PL brands of its customers from nine manufacturing sites in Europe. In the UK, Cérélia operates under the name ‘BakeAway’. It has a manufacturing plant in Corby, Northamptonshire. Its predominant activity in the UK is the supply of DTB products to grocery retailers who sell these products to end-consumers as own-brands. Cérélia currently also manufactures a large proportion of the Jus-Rol branded products sold in the UK. The Corby plant manufactures ingredient pastry dough, pizza dough, cookie dough, brownie dough and gingerbread dough. Cérélia is the second largest supplier of DTB products to grocery retailers in the UK (after Jus-Rol) by value with a share of supply that is more than double the size of the next largest supplier.

Table 9.1 of the FR sets out estimates of wholesale supply shares for DTB products by value to grocery retailers in the UK in 2021-2023. We set out below a non-confidential version attributing share of supply by range:

The actual share of supply figures lie on the range. The Court has seen the actual data. The merged entity will have a combined market share in the range of 60-80%. Henglein will have a market share in the range 10-20%; and Bells, a market share in the range 5-10%. The combined market share of Henglein and Bells is therefore in the range of 15-30%. In FR paragraph [9.9] the CMA stated:

“9.9

Based on the above methodology, the Parties currently have a very high combined share in the wholesale supply of DTB products to grocery retailers in the UK. Further, that high share has been sustained over time and sustained at least up to forecast for 2023, and the only other suppliers with notable (although significantly lower) shares are Henglein and Bells.”

During the administrative procedure the appellants argued that the merger was to be evaluated upon the basis that it had vertical, not horizontal, effects. On this basis the focus should be upon the existence, or otherwise, of vertical foreclosure effects, as to which on the evidence there were none. The view of the CMA differed. The merger was to be treated as horizontal, i.e. as between actual competitors. That conclusion was unsuccessfully challenged before the CAT. There is no challenge, on this appeal, by Cérélia to the conclusion of the CAT on this, and therefore to the analysis in the FR as to the removal of competition brought about by the merger itself.

Horizontal, adverse, effects upon competition are more likely where the merging firms are close competitors or where their products are close substitutes. This is because the merged entity is able to recapture a larger set share of the sales lost following a price increase, or other worsening in the offering, making the price rise more profitable (FR paragraph [9.15]). According to the CMA’s Merger Assessment Guidelines (“MAG”), where the CMA finds evidence that competition primarily occurs between a small number of firms, any two would normally be sufficiently close competitors such that the removal of competition between them would raise competition concerns “subject to evidence to the contrary”. In the FR, at paragraph [9.16], the CMA stated: “… the smaller the number of significant players, the stronger the prima facie expectation that any of the two firms are close competitors. In such a scenario, the CMA will require persuasive evidence that the merging firms are not close competitors in order to allay any competition concerns”.

The CMA set out at length, in Chapter 9 of the FR, its analysis of horizontal effects. The analysis is qualitative, not quantitative (Footnote: ¹), as was confirmed by Mr Palmer KC for the CMA during the hearing. This observation is not intended as a criticism (Footnote: ²). It means that the exercise of the CMA when evaluating whether competition from alternative sources, such as competitors, was sufficient to counteract the loss of competition brought about by the merger, involved a complex evaluation balancing any offsetting competitive constraints against its conclusion as to the extent of the anticompetitive harm brought about by the merger.

It is unnecessary to set out in detail the reasoning in the FR on this issue because the grounds of appeal focus upon the adequacy, or otherwise, of third party constraints, in particular from Bells and Henglein upon the merged entity, not upon the nature of competition between the merged entities. At paragraphs [37]-[44] FR the CMA set out, in broad sweeps, its conclusions concerning the extent of competitive interactions between the merging parties. The gist of the analysis is that the presence of GMI and Cérélia represent competitive “tension” or “constraint” in the marketplace which was important in the eyes of retailers in ensuring that prices offered to retailers were competitive. Removal of that tension or constraint (by the merger) was therefore harmful to competition:

“37.

Large grocery retailers, which account for the large majority of DTB products sold in the UK, told us that their ability to trade off the Parties in their negotiations is an important constraint which enables them to get a good deal when purchasing DTB products.

38.

These grocery retailers told us that they may not typically explicitly pit their PL supplier against their branded supplier but that the availability of both is a source of competitive tension that would be lost by the Merger, thereby reducing their ability to protect against potential price rises (or other kind of worsening in the Parties’ DTB offerings). The Parties’ internal documents also show some evidence of this kind of constraint operating in practice.

39.

We consider that the constraint between the Parties is important for both channels, noting that PL in particular (for which Cérélia is the leading supplier) operates as a pricing discipline on Jus-Rol. Post-merger, the strong market positions held by each of Cérélia and Jus-Rol will be consolidated within the Merged Entity, resulting in the loss of the constraint between the Parties which will, in turn, affect retailers’ ability to resist a price rise (or other worsening in the Parties’ offerings).

40.

There is significant overlap in the product ranges that the Parties supply to grocery retailers. While some retailers might only buy some of the products within the Parties’ ranges at present, we found that the DTB product category should, for the purposes of assessing competition, be considered as a whole (because retailers consider all DTB products together and suppliers are able to alter and expand the types of DTB product that they offer to grocery retailers).

41.

While, as noted above, there are important differences in the offerings of the Parties, the relative importance of the competitive constraint offered by the Parties upon each other also depends on the available alternatives. As discussed further below, we found that there were few credible alternatives for grocery retailers purchasing DTB products, which makes the loss of the competition between the Parties particularly important.

42.

The Parties told us that because Cérélia already manufactures most of the Jus-Rol products sold in the UK, there could be no existing competition between the Parties which would be lost by the merger. We note that this submission is not fully supported by the data that the Parties have provided. We also note, more broadly, that Cérélia’s role in manufacturing Jus-Rol products is based on a contractual relationship, which is materially different in nature to a merger. A contractual relationship does not result in a lasting change in market structure, has limited duration and may be renegotiated or terminated even before its initial term.

43.

In this regard, the Merger would result in material changes in competitive dynamics and market structure:

(a)

Post-Merger, Cérélia would have control over all aspects of the wholesale offering to retailers across both channels, which it does not have at present. In particular, Cérélia would have control over pricing of both the PL products bought by retailers from Cérélia and Jus-Rol products and could determine pricing to maximise joint profits (which is not the case at present).

(b)

The Merger would also ‘cement’ Cérélia’s role as the manufacturer of Jus-Rol products. As a result of the transaction, GMI would lose its ability to independently decide its commercial strategy, including whether to terminate the agreement with Cérélia and appoint an alternative supplier, take the production back in-house, or take any other course of action relating to its Jus-Rol products.

44.

We therefore found that the Parties’ submissions, that it would not be profitable to raise prices or degrade the quality of both Jus-Rol products and Cérélia’s manufacturing services to retailers for the PL channel, were not supported by the evidence available to us.”

Cérélia argues that no SLC arises because competition from third parties amounts to a sufficient constraint or check upon the merged entity to ensure the continuation of effective competitive pressure and obviates the risk of harms identified by the CMA eventuating. In other words, it is sufficient to offset any loss of competition which arises because the merging parties are now one entity, not two.

There is also no challenge, in this appeal, to the conclusion of the CMA that retailers had insufficient competitive power to exert discipline upon their suppliers. If a merged entity confronts powerful purchasers, then it might be relevant to the analysis of constraints upon the merged entity upstream that it confronts buyers with sufficient muscle and heft to ensure that the terms and conditions, including as to price, of the product supplied to them remains competitive. Here the CMA examined whether any retailer had the ability to impose a competitive discipline upon its suppliers. The CMA examined for instance the extent to which retailers sought to make suppliers participate in competitive tendering as a structural device to force suppliers to offer competitive terms and conditions. The CMA found that the normal method of obtaining supplies was through negotiation, not tendering processes; and that in negotiations the existence of GMI and Cérélia as rivals was important. There was no suggestion that in a post-merger world the use of a competitive tendering process would be sufficient to mitigate the lessening of competition brought about by the merger itself. Table 5.1 of the FR sets out each principal retailers share of supply. A non-confidential version, based on share of sales ranges, is set out below.

At paragraphs [10.124ff] FR the CMA examined the likelihood that the merger would encourage new entry to the market, for instance to take advantage of the high prices that the merged entity might seek to impose. The conclusion was that new entry was improbable.

Before turning to the grounds of appeal we address the nature of the role played by the CAT in this sort of case. In paragraphs [48] – [49] of the Judgment, the CAT described an approach which implied that the CAT would apply an intensive level of review. In paragraphs [47] and [50]-[60], it described an approach which implied a light touch, hands off, approach. And in paragraph [343] it observed that a detailed review of evidence was “unusual” in a judicial review under section 120 EA 2002. There is thus a degree of uncertainty in the CAT’s analysis as to the role it plays in relation to decisions of the CMA.

In paragraph [48] the CAT cited with approval the judgment of the Court of Appeal in Office of Fair Trading and others v IBA Healthcare Limited [2004] EWCA 142 (“IBA”) at paragraphs [91]-[93]. This case concerned the merger control provisions under the EA 2002 and repays some consideration. It related to the approval by the OFT (the relevant regulator), without there being any reference to the Competition Commission for a full phase II investigation, of a merger between two software companies. IBA Healthcare, a competitor to the merging entity, complained to the OFT that the merger would create a SLC. The OFT initiated an investigation at the culmination of which it concluded that the merger would not be expected to result in a SLC. The case file was closed. The OFT decision relied upon the effect of the National Programme for IT ("NPfIT"), 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 which was intended to provide for cross-referencing of patients' records by creating a complete electronic medical record for each patient across all NHS providers in England. The OFT considered that the NPfIT would give rise to substantial buyer power sufficient to counteract any anticompetitive effects of the merger and for this reason it decided not to refer the merger to the Competition Commission for full investigation. IBA Healthcare applied for judicial review under section 120 EA 2002. It argued that the OFT had failed to review the facts giving rise to the NPfIT and had, exaggerated, materially, its importance as countervailing buyer power and had therefore erred in its analysis of SLC. The OFT, and the intervening merging parties, contended that the assessment of countervailing buyer power and SLC was a factual matter falling substantially within the evaluative discretion of the decision maker. There was no basis upon which the CAT should interfere in a case involving such a broad, multifaceted, economic evaluation.

The CAT ([2003] CAT 27) disagreed and adopted an approach which was subsequently endorsed by the Court of Appeal. It emphasised that the concept of reasonableness or rationality varied with circumstances (paragraph [218]). Judicial review arose in an “extraordinarily diverse range of circumstances”, from cases involving broad issues of policy or discretion where a court was rightly conscious of the risk of trespass into the realm of legitimate decision making, through to issues involving pure questions of law or procedure which were “pre-eminently” within the judicial function. In other cases, the court might be scrutinising facts to determine whether the decision maker had acted lawfully in a particular context and this was also a core judicial function. The CAT then examined the facts and concluded that the OFT had failed to conduct a proper investigation.

The OFT decision was quashed and remitted for reconsideration. The OFT and the merging parties appealed. The appellants contended that the CAT had exceeded the proper bounds of judicial review under section 120 EA 2002.

The appeals were dismissed and the findings of the CAT upheld. The Court, echoing the CAT, differentiated between low and high intensity reviews. At one end of the scale (the low intensity review) were cases involving issues depending essentially upon political judgment, for instance a case concerning a matter of broad national economic policy. At the other end of the spectrum (high intensity reviews) were, for instance, decisions concerning fundamental rights. In determining where, on the spectrum, a particular case fitted, important factors were (i) the nature of the decision being taken and (ii) whether the court was “equipped by training or experience or furnished with the requisite knowledge or advice”. In paragraph [93] the Court observed that the IBA case was not concerned with questions of policy or discretion which were the normal subject matter of a low intensity challenge. Under the legislative regime the issue for the regulator was one of factual judgment (high intensity). The Court observed: “… there is no doubt that the court is entitled to inquire whether there was adequate material to support” the conclusion of the decision maker.

The CAT in the present case also cited, with approval, the observations of Carnwath LJ in IBA (at paragraph [96]) endorsing observations of the House of Lords in Edwards v Bairstow [1956] AC 14 (“Edwards”) at pages [38]-[39], to the effect that:

“...there is no reason to make a mystery about the subjects that commissioners deal with, or to invite the court 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.”

We agree with the CAT that the test laid down in IBA is the appropriate test to apply. None of the parties to the appeal disagreed with the correctness of the test described above.

However, the CAT also held in the present case that it was not concerned with the correctness of facts found (paragraph [47]). In paragraph [51] it said that there was a “high hurdle” to meet to persuade the CAT to disturb a fact finding of the CMA. In paragraphs [52] – [60] on rationality, in particular in relation to remedies, it stated that the test was whether the CMA acted “manifestly without reasonable foundation” and that this was the same as the rationality test. All of this could suggest that the CAT eschewed an approach based upon a detailed scrutiny of the evidence and applied a uniform (undifferentiated) approach as to the degree of deference it accorded the CMA.

As we explain below, we have ultimately endorsed the conclusion of the CAT. But we consider it is helpful to summarise some basic propositions about the role of the CAT when conducting a judicial review of a decision of the CMA.

First, Parliament created the CAT as a tribunal comprising specialist lawyers, economists and others with specific relevant expertise, to oversee the decisions of regulators. In addition to a review of a regulatory decision on questions of vires and law, Parliament entrusted the CAT with the responsibility for reviewing findings of fact and the evaluation of those facts by regulators. This point was made expressly by the Court of Appeal in IBA at paragraph [93] in the context of a judicial review.

In a given case therefore it may be the task of the CAT to determine whether there is “adequate material” before the CMA to support its conclusion, an exercise the CAT is singularly well equipped to perform. It can be expected to examine closely the complaints made about a decision and its evidential underpinning. Such a deep dive into the evidence equips the CAT with the information necessary, then, to make an informed judgement as to whether the decision under challenge was properly justified by the evidence. The extent to which the forensic sleeves must be rolled up the judicial arm is not to be confused with the margin of appreciation to be accorded to the decision maker. It is at the point that the CAT is seized of a detailed understanding of the evidence that it can then decide whether the CMA was acting within legitimate bounds in its determination and evaluation of the facts.

Secondly, the degree of deference to be accorded by the CAT to the CMA is fact and context specific, as IBA makes clear. If, for example, the dispute concerns the interpretation of a contract or letter then the view of the CAT on a question of interpretation might be as equally valid as that of the CMA. If the issue concerns the inferences to be drawn from statistical data, then the conclusions drawn by the CAT might again be as valid as those drawn by the CMA. The CAT is also well placed to measure the adequacy of the decision maker’s fact finding. In IBA the CAT disagreed with the OFT over its analysis of a complex set of documents and facts and as to the extent to which inferences that could be drawn from this material entitled the OFT to conclude that the new system would create sufficient countervailing buyer power. The analysis of the CAT on the evidence and the facts as recorded in the judgment far exceeded that set out in the OFT decision. That detailed review enabled the CAT to conclude that the OFT had made errors in its analysis of the facts and it remitted the case to the OFT. At the other end of the scale, if the CMA has evaluated a wide variety of complex evidence, not all of which is consistent, a broader margin will be accorded to the CMA in relation to its findings of fact and the inferences to be drawn therefrom.

It also follows that in a given case the breadth of the deference to be accorded to the decision maker may vary as between different grounds of challenge. It is, however, important to recognise that, because of its expertise, it is quite possible that the CAT will be critical of relatively complex evaluations by the decision maker, even where a non-specialist court might not be. That is a necessary corollary of the CAT having been instituted as a specialist body tasked to conduct precisely that sort of exercise.

It is, though, important not to let semantics obscure the nature of the exercise. If, following a detailed review, the CAT concludes that the decision maker erred because, for example, it misconstrued the evidence or data, or failed properly to inquire into the evidence, then it is a matter of words only to say that the decision is in error because it was not supported by the evidence, or alternatively, that the decision was “irrational”. Finally, none of this involves the CAT substituting its own view for that of the decision maker. It is simply holding the CMA to a proper standard.

With this in mind we turn to the first ground of appeal. This concerns a complaint that the CAT erred in endorsing the analysis by the CMA in the FR that Bells did not amount to a countervailing, competitive, force sufficient to negate the SLC brought about by the merger. Cérélia argues that the CMA’s conclusions were irrational and unsupported by evidence; and that the evidence on the CMA file did not support the conclusions set out in the FR.

These criticisms were to be seen in the context of findings in the FR which it was said were supportive of the grounds of appeal: (i). “switching costs are not a major barrier to [brand owners] switching PL supplier” (FR, paragraph [9.167]); (ii) Bells already held a significant contract with a major retailer for the manufacture of that retailer’s own label products (FR paragraph [9.177]); (iii) even after taking on that retailer’s contract, Bells had capacity, sufficient to enable it to take on the manufacturing of the entire own label requirements of another large retailer (FR paragraph [9.197]); and (iv), Cérélia perceived Bells as a competitive constraint (FR, paragraph [9.192]). Mr Hoskins KC argued that whilst individual pieces of evidence on the CMA file might not, standing alone, be dispositive; nonetheless, considered in combination, and in the light of these findings in the FR, the evidence factually undermined the conclusions of the CMA.

Mr Hoskins KC referred the Court to a series of pieces of evidence on the CMA file which, he argued, demonstrated that Bells amounted to a significant potential constraint upon the merged entity. He grouped the evidence under four headings: (i) the internal view of Bells by Cérélia; (ii) the views of retailers about the ability of Bells to constrain the merged entity; (iii) market share analysis; and (iv), Bells’ ability to expand its sales, including with new capacity to respond to increased demand.

In relation to each heading the Court was taken to the actual evidence placed before the CMA emanating from, inter alia: transcripts and notes of meetings between the CMA and market participants; answers to questionnaires sent by the CMA to the parties and others; evidence from internal Cérélia documents assessing Bells as a competitor; evidence given during the hearing held by the CMA with the parties; emails emanating from within major retailers in which they assessed the pros and cons of Bells as an actual or potential source of supply; and, evidence from within Bells relating to its ability and willingness to expand and/or invest in new manufacturing capacity.

That evidence provided information as to: the present scope of operations of Bells (mainly within Scotland); its commercial incentives and ability to expand south of the border; the effect of the fact that Bells had, recently, secured a contract with a major retailer in England and Wales and as to the scope of that contract, including as to the views of the retailer as to its strengths and weaknesses and as to its importance over time; the views of retailers generally as to how, if at all, Bells fitted into their perceptions of competitive constraints upon the merged entity both in the past and prospectively; and the ability of Bells, in the longer term, to secure further contracts with major retailers and the implications of its ability to expand upon capital expenditure in new capacity. It is fair to describe the evidence as painting a mixed and nuanced picture. It does not all point in a single direction.

In the FR the CMA concluded that Bells exerted only a “limited” competitive constraint, summarising its reasoning as follows:

“…even if Bells can take on a contract equivalent to the size of [a substantial supermarket chain], we note that the market position of Bells would remain modest and that it would be significantly smaller than the Merged Entity. We note, in addition, that the majority of grocery retailers do not consider Bells to be a credible alternative supplier to the Parties. On this basis, we consider that Bells provides a limited constraint on the Merged Entity.”

(FR paragraph [9.195]).

The CMA’s analysis can be summarised as follows. Bells is a Scottish manufacturer of DTB products. Its main products are branded ready-rolled puff pastry and puff pastry blocks. It also supplies such products to food manufacturers. Until September 2022 it supplied a modest quantity and value of PL products to two retailers. These contracts came to an end due to commercial considerations: FR paragraph [9.173]. In January 2023 Bells commenced supplying PL puff pastry to a supermarket chain: FR paragraph [9.177]. In relative terms the combined market share of the merged entity is substantially higher than the share of supply of Bells at 5%-10%. This indicates that the merged entity had a strong position in the wholesale supply of DTB products to retailers in the UK: FR paragraph [9.14].

Bells had a maximum DTB production capacity of [x] tonnes per annum. The production capacity of the merged entity was substantially in excess of that. The ability of Bells to exploit that capacity was affected by operational matters, switching costs and its commercial strategy. Bells provided evidence to the CMA about its ability to utilise spare capacity. The CMA, upon the basis of the evidence before it, concluded that there were limitations upon the ability of Bells to utilise its spare capacity. Bells might be able to meet the PL needs of one additional large retailer: FR paragraph [9.193]. The CMA considered the picture that would be presented were Bells (hypothetically) to win a further significant contract. In such circumstances Bells’ market share, whilst enhanced, would remain less than a quarter of the size of the market share of the merged entity. A majority of grocery retailers did not consider Bells to be a credible alternative supplier to the merged entity. Bells also outlined to the CMA its intentions as regard manufacturing capacity. The CMA concluded that market shares were an effective indicator of market power. It rejected the argument of Cérélia that the market was characterised by competitive bidding exercises (whereby purchasers could impose a competitive discipline upon suppliers). Tender processes were not used for the supply of branded DTB products: FR paragraphs [9.35] and [7.28]. The supply of PL DTB products typically occurred through bilateral negotiation not formal tender: FR paragraphs [7.20] and [7.23].

The CAT’s conclusions are found between paragraphs [168]-[171]. There is no reference to individual pieces of evidence or as to how they related to the CMA’s analysis. This has made it difficult for this Court to understand, save at the broad level set out in the Judgment, what the CAT’s views of the evidence were:

“168.

Given the diverse mix of factors taken into account by the CMA in considering Bells’ effectiveness as a constraint, it is entirely possible that the CMA might have reached a different outcome on this assessment. It is possible that, had it done so, this might have affected its overall SLC assessment regarding a post-Merger increase in PL prices. It was open for the CMA to have done more to explore whether any given loss of Cérélia’s PL volume to a rival PL supplier would (if credible) have been sufficient to provide an effective competitive constraint on Cérélia’s pricing of PL products, thereby placing less emphasis on simple market share considerations. However, it is evident from the long list of factors at play here that the CMA is entitled to considerable discretion when making this assessment, and we cannot say that it was unreasonable to draw the conclusion that it did and not make further specific enquiries. Cérélia’s argument about the existence of Bells’ spare capacity is contentious in its own right, and also addresses only part of the picture that was relevant to the CMA’s deliberations on this point. There was sufficient information for the CMA to conclude that Bells’ spare capacity was only a limited competitive constraint for the reasons set out in the Decision. Bells had left its options open and naturally would want to see how its new contract would develop before taking a decision on whether to explore taking on any new large contract.”

The CAT noted the CMA’s entitlement to take into account the perceived commercial motivations of competing suppliers when considering SLC concerns:

“169.

It also bears mention that the CMA is entitled to take into account the commercial motivation of competing suppliers when addressing their submissions on SLC concerns. In contrast to customers, who generally have a direct interest in avoiding an SLC, rival suppliers would stand to gain from any relaxation in competition in the market in which they are also sellers. This is not to suggest that Bells took a strategic approach to its responses to the CMA, but it does justify the CMA in taking a more rounded view of the evidence it received from competitors to the post-Merger firm.”

The CAT further noted that it agreed with Cérélia’s arguments that it was necessary to consider the sum total of constraints offered by competitors, but it did not consider that this made the CMA’s assessment irrational:

“171.

Cérélia also stressed that the actual constraint on Cérélia’s PL business from alternative competitors would comprise the sum total of the constraints offered by Bells and Henglein combined. We agree. However, since the CMA’s assessment of neither constraint was obviously irrational or wrong, this does not change our conclusion.”

The CAT concluded:

“175.

Having reviewed all the arguments in relation to Ground 1A, the Tribunal dismisses this Ground which was a root and branch attack on significant aspects of the Decision. The CMA gathered a great deal of information and evidence for its assessment and carefully weighed the evidence that it had gathered. Whilst on some aspects it may have been open to it or a reasonable decision maker to take a different view on certain aspects, it did not make any error of fact or law. In dealing with Ground 1A the Tribunal has reviewed both the various sections of the Decision and the underlying evidence that the parties have referenced for its consideration.”

We can express our conclusion briefly. Prior to the appeal we directed that the parties include in a special bundle all the evidence relevant to the challenge to the CAT’s judgment on Bells and Henglein. We have read that material in its totality and have paid special attention to the documents drawn to our attention in oral argument by the parties. We cannot refer to this material in this judgment because it is overwhelmingly marked as confidential. We have however systematically considered the CMA’s findings in the light of that evidence. In coming to a conclusion, and in accordance with the indications given in IBA, we take account of the following when considering the nature of the margin of appreciation: (i) that the exercise engaged in by the CMA involved the evaluation of a disparate, and not entirely consistent, body of evidence; (ii) that the CMA interpreted this evidence in the context of both its own (horizontal) and Cérélia’s (vertical) competing theories of harm; (iii) that the evaluation involved the CMA in making a qualitative, prospective, assessment of the extent to which the acknowledged ability of Bells to impose some degree of constraint (ultimately described as “limited”) could serve now and into the future to counter or off-set to a sufficient degree the SLC created by the merger. We conclude that the approach adopted was logical and rational. We have not identified any document that we consider has been misconstrued by the CMA. We are clear that, bearing in mind the nature of the issue, there was sufficient material before the CMA, as contained on the file and as recorded and summarised in the FR, for the CMA to find as it did in relation to Bells.

We therefore uphold the conclusion of the CAT that the CMA did not err in relation to Bells. Insofar as it is necessary to label the analysis of the CMA, it conducted an adequate fact finding exercise, had sufficient evidential support for its factual findings, and reached a conclusion on the degree of competitive constraint that was rational.

We would briefly comment upon certain criticisms made by Cérélia, in particular in written submissions, of the CAT’s reasoning in the Judgment. Three main points are made.

First, it is said that the CAT highlighted inadequacies in the analysis of the CMA and that Cérélia’s own complaints were strengthened by this. The appellants point to the concerns expressed by the CAT in paragraph [168]: It was “open for the CMA to have done more” to explore the potential competitive constraint posed by Bells and that it was “entirely possible that the CMA might have reached a different outcome on this assessment”. Had it done so it was “possible that... this might have affected its overall SLC assessment”. It is unclear to us whether in this paragraph the CAT was levelling a criticism at the CMA upon the facts of the case; or, simply pointing out that the exercise performed by the CMA was a rounded one in which it had to exercise administrative judgement as to which strands of evidence were examined to a greater or lesser degree and that it followed that the CMA had a certain discretion as to how it conducted the investigation. On balance we share the analysis of Mr Palmer KC, for the CMA, that the observations of the CAT amounted to the latter and were not a specific criticism of the CMA about its approach to the actual evidence. He pointed out that under the EA 2002 the CMA had a discretion as to the manner in which it conducted an investigation in any given case, subject, of course, to the requirements imposed by public law. We agree.

Secondly, it is said that the CAT erred in paragraph [169] in taking into account the commercial motivation of competing suppliers when addressing their submissions on SLC because rival suppliers would stand to gain from any relaxation in competition in the market in which they were also sellers. Mr Hoskins KC argued that this was improper speculation on the part of the CAT and in any event counter-intuitive because, rationally, competitors would wish to exaggerate the competitive threat posed by the merged entity, rather than underplay it. Mr Palmer KC argued to the contrary that the statement was in fact logical because the CAT was probably alluding to the possibility that Bells might seek to capitalise upon concerns on the part of retailers about the market power held by the merged entity and encourage switching to Bells (with its extra capacity) in order to create competitive tension. Mr Palmer suggested, therefore, that the observations of the CAT were explicable.

We have difficulty in construing paragraph [169]. Whatever it means it does not alter our conclusion that the CMA had an adequate evidence base before it upon which to found its conclusions. We would, though, observe that insofar as it is contended that the CAT is not entitled to use its own experience and forensic skills to evaluate the probative value of evidence by reference to its perceptions of the incentives upon parties who give evidence, any such suggestion is misplaced. The CAT is entitled to consider commercial incentives when weighing evidence, as of course is the CMA.

Thirdly, Cérélia complained that the CAT substituted its own view of the evidence for that of the CMA in order to uphold the decision. We disagree. The task of the CAT in a case such as this is to determine whether the approach taken in the decision is rational and that the decision maker has adequate evidence before it upon which to found its factual conclusions. The CAT thus needs to compare the underlying evidence base with the decision. This is what the CAT did in IBA. This will frequently involve the CAT considering evidence which might not be recorded in the decision being challenged. In arriving at a judgment, the CAT is entitled to ignore infelicities in language or expression in the CMA decision. It may take into account that in the decision the evidence may also be summarised in a relatively propositional and concise manner. It may further take into account that the decision maker is not bound to refer to each and every piece of evidence. There may, therefore, be many circumstances where in evaluating the adequacy of a decision the Tribunal necessarily adopts an analysis of its own. This remains the proper province of judicial review. Such an approach does not however imply that the CAT is substituting its own judgment for that of the CMA.

Cérélia sought permission to appeal a new point. It was invited by the Court to address the matter at the oral hearing upon the basis that the Court would rule upon the application in its judgment.

The proposed new ground was framed in the following way. At paragraphs 9.193-9.194 FR the CMA recognised that Bells might be able to meet the PL needs of an additional large retailer. As such Bells’ share of the market would be between 10-20%. The FR stated that, if this occurred, its market share would remain less than one quarter of that of the merged entity (50-60%). It is now argued that the CMA erred because it failed to consider Bells’ market share in relation to that of Cérélia alone. Bells and Cérélia both manufacture PL products for retailers, whilst Jus-Rol supplies branded products to retailers. If Bells won a PL contract with an additional large retailer, both Bells and Cérélia would have market shares of 10-20%, reflecting their manufacture of PL products. The close market share of Bells and Cérélia in this respect would amount to an important competitive constraint and the CMA’s failure to take this into account was a further indicator of irrationality.

Mr Hoskins KC accepted candidly that this was an argument that Cérélia could have raised before the CMA and the CAT, but did not. He contended however that the argument does not require further evidence to be evaluated since it arises solely out of the content of the FR.

Mr Palmer KC submitted that the point had no merit but that, in any event, it was inappropriate to grant permission for it to be raised for the first time on appeal. As to the merits he pointed out that the CMA defined the product market as including both PL and own-brand product and that the principal vehicle for competition was cross-channel competition between PL and branded product. These findings were not challenged upon the appeal and, in and of themselves, amounted to an answer on the substance of the new complaint because they struck at the heart of the proposed ground of challenge which assumed that PL was, in some way, to be analysed as economically different. He also, in a variety of respects, identified facts or evidence which the CMA might have wished to investigate had this argument been raised during the administrative procedure.

We deal with this point briefly. It is trite that the Court of Appeal may, in principle, consider new points not raised in proceedings below. These might, characteristically, concern pure points of law or where the issue concerns the inference to be drawn from undisputed facts. In our judgement the market share argument now raised is a variant of economic arguments advanced before the CMA where Cérélia mounted a “robust” defence of its position, as of course it was entitled to do. There is therefore no suggestion that it lacked the opportunity to raise this matter. Had it been raised it might not have generated entirely new lines of inquiry or investigation; nonetheless, it is a factual and evidential matter that the CMA might well have needed to address in a discrete way, and it is therefore unfair to allow it to be raised during this appeal, so long after the event. We can see no basis why we should grant permission to appeal to such a point, and we decline to do so.

We would add only that, as explained by Mr Palmer KC for the CMA, the short answer appears to be that the product market identified by the CMA in the FR, which is a key determinant of the way in which competition operates in the market, includes both PL and branded product. As such the suggestion that a focus should lie only with PL product seems misplaced.

For the above reasons we refuse permission to appeal upon this proposed ground.

We turn now to the second ground of appeal which concerns the CAT’s endorsement of the CMA’s finding that Henglein was not a sufficient constraint. Prior to the hearing Cérélia also sought permission to amend its second ground of appeal to add a further new argument not raised below. Again, the Court indicated that it would hear full argument upon the point and then decide whether to grant permission to appeal on the amended ground. The ground as amended is in the following form with (iii) below being the proposed new (amended) ground of appeal: the CAT committed errors of law in: (i) concluding that the CMA’s assessment of competition from rival contract manufacturer Henglein was not irrational; (ii) failing to address Cérélia’s formulated challenge; and (iii) not finding that the CMA’s assessment in relation to Henglein did not comply with the statutory duty to consult.

The context is as follows. It is argued that in the PF the CMA concluded that Henglein amounted to a “material constraint” upon the merged entity and set out the evidence justifying that conclusion. However, in the FR, the CMA, without explanation, performed a volte face and concluded that Henglein amounted to no more than a “limited” constraint which was insufficient to offset the SLC. This is set out in FR paragraph [9.246] where the CMA stated: “... the CMA does not consider that this evidence alters its assessment that Henglein offers only a limited competitive constraint on the parties for a range of reasons”.

This change of position was contended to be unjustified, unexplained and irrational. The CAT records the argument in the final sentence of paragraph [165] of the Judgment but when addressing the position of Henglein in paragraph [170] completely ignores it. The gloss raised by the amendment is that if there was such a “volte face” Cérélia was entitled to be consulted on this change in position. The CMA is under a statutory duty to consult. The relevant principles are summarised at paragraph [63] of the judgment. The relevant question is whether an “adequate gist” was provided on an issue on which the CMA proposed to take a decision adverse to Cérélia. As to this, it is contended that the CMA never put the gist of the volte face to Cérélia. It is said that this argument can be advanced solely by reference to a comparison between the PF and the FR in the context of material on the CMA file. As such it is a point that is capable of being fully assessed by the Court of Appeal and that the CMA, as respondent, is not prejudiced by this new point being raised.

There is a substantial overlap between the proposed new ground and the existing grounds. If this Court concludes that, upon a fair reading of the PF and FR, there was no change of position and that the position adopted throughout by the CMA was supported by adequate evidence, then the rationality challenge falls away. It would then also follow that there was no cause for the CMA to consult Cérélia and provide it with an appropriate gist, since there was no volte face.

In this context we conclude that it is appropriate, first, to consider the substance of the dispute which is whether there is, as a matter of fact, any material or relevant change in position as between the PF and the FR. To frame the analysis, it is relevant to record that consistently, between the PF and the FR, it is conceded that Henglein constitutes “a”, and generates “some”, constraint. That much is set out in both documents. The only question, therefore, is whether it has been the consistent view of the CMA that this constraint is “limited”, that being the final conclusion of the CMA.

In paragraph [63] the CAT summarised the law on consultation as it applied to competition proceedings. It is unnecessary to address that analysis in this appeal. It suffices to observe that the most important principle is that the entity concerned must be given adequate information to enable it to understand the position of the decision maker and to respond fairly and fully. Whilst there is a discretion accorded by the CMA, this is not unlimited. In balancing the need to preserve confidentiality (of third party material) against the right of the merging entities to respond, it cannot be assumed that confidentiality trumps procedural fairness. The obligation upon the decision maker to provide adequate gist is an important procedural safeguard for parties whose freedom of commercial action is threatened with regulatory intervention, and is not lightly to be alleviated. Where material that is relevant to the CMA’s position is confidential, the CMA must still find a way of conveying the gist.

The issue arising for determination on the facts of the instant case is narrow: was there a difference between the PF and FR and, if there was, did the CMA give adequate gist to Cérélia to enable it to respond? Since it is common ground that, outside the contents of the PF, no gist was proffered the second part of the question is otiose. The only question therefore is whether, in the PF, the CMA gave adequate notice of the conclusion it arrived at in the FR.

The starting point is the PF. We set out below a summary of the position of the CMA. This is based both upon the contents of the PF and the material before the CMA which we have carefully reviewed. Again, we endeavour to describe the position without divulging confidential information.

Henglein is a large supplier of DTB products based in Germany. It manufacturers, in Germany, PL shortcrust, puff pastry, and filo-pastry products for three UK retailers. It has a relationship with a distributor in the United Kingdom, Golden Acre Foods, which distributes products to UK customers and manages its relationships with retailers. There have been fluctuations in the company’s share of supply for the United Kingdom from 2013 onwards. The PF records its maximum production capacity which is geared towards covering supplies in Europe and the UK and compares this with the production capacity of Cérélia dedicated to supply in the UK. The total capacity of Henglein to cover Europe and the United Kingdom is not substantially in excess of the maximum capacity of Cérélia to cover only the United Kingdom. There is a brief reference to the ability of Henglein to reconfigure capacity (i.e. away from Europe and to the UK) but no conclusions on this are set out.

The PF recorded submissions made by Cérélia to the effect that Henglein was an effective actual and potential competitor for DTB products in the UK.

The PF and internal material on the CMA file describe evidence provided to the CMA by Henglein as to the range of products it supplies, its limitations, and its use of alcohol in its products (which was a problem for at least one major UK retailer).

As to evidence from major UK retailers, some indicated that they viewed Henglein as an alternative PL supplier they would turn to if they were switching; others took the opposite view. The PF records answers from retailers to a questionnaire prepared by the CMA. Question 18 asked: “if you were to switch to an alternative supplier of private label products, who would you switch to? Please indicate their suitability and provide reasons for your choices”. Addressees were invited to score their view on suitability on a range from 1 (barely able to meet needs satisfactorily) to 5 (fully meets requirements). None of the retailers who responded considered Henglein to be a DTB supplier able to compete with Jus-Rol in the UK.

The CMA also asked: “Have you considered switching away from/to Cérélia for the supply of private label products in the last three years? Why/why not? Please also include in your response whether you have considered sponsoring entry and if so, how this would work in practice”. Answers indicated that at least a number of retailers viewed the fact that Henglein was based in Europe as a competitive disadvantage due to such matters as: increased transportation costs, border frictions when sourcing from Europe (Brexit), a desire to source from UK-based manufacturers; and concerns over sustainability.

The PF also examines internal documents of Cérélia to gauge whether it viewed Henglein as a competitor. The PF records the total number of documents reviewed which mentioned the company. The evidence indicated that Cérélia monitored Henglein but neither the internal documents nor the analysis in the PF indicate that Cérélia treated Henglein as a really significant threat.

In the PF, paragraphs [9.197]-[9.199], the CMA pulled the strands together. In paragraph [9.199] PF it concluded:

“Overall, we consider Henglein to represent a material constraint on the parties, particularly in relation to competition for smaller supply contracts (of the sort it has now). However, we believe the constraint exerted by Henglein is likely to be weaker when it comes to competition to supply large retailers, for the reasons described immediately above. We provisionally conclude that the constraint is likely to be confined to PL supply, with minimal impact on Jus-Rol”.

In paragraph [9.220], in relation to “the degree of constraint from PL suppliers”, the CMA stated:

“We provisionally conclude that PL suppliers, individually and in aggregate, exert only a limited competitive constraint on Cérélia, and even less of a constraint on Jus-Rol given the branded nature of its product.”

In relation to PL suppliers generally based in the EEA the CMA stated that they offered “some” competitive constraint; and in respect of Henglein specifically, the CMA stated that the competition it could offer to its current retail customers was material but it was less likely to constrain the parties when it came to contracts to supply large retailers i.e. the supermarkets that comprise the preponderant part of supply to consumers.

In our judgment there is no significant difference between the analysis in the PF and the FR. Both describe the competitive threat posed by Henglein in terms which may properly be categorised as “limited”. When the CMA used the expression “material” in the PF it did so in a balanced and nuanced way which focused upon the ability of the company to supply smaller retailers; not larger retailers. Further, the underlying evidence generated during the administrative procedure provided a sufficient factual basis supporting that conclusion in both the PF and FR.

It follows that not only are the findings in the FR properly founded in the evidence, but that when it was provided with the PF during the administrative proceedings Cérélia was sufficiently informed of the provisional view of the CMA that competition from Henglein was “limited” that it was able to address those concerns fairly and fully. There was therefore no need for any additional gist statement from the CMA to explain the basis upon which it came to the conclusion in the FR that competition was “limited”.

The CAT does not in the Judgment set out any of the above facts or matters but instead provides summary conclusions on rationality which are to the same effect:

“170.

As regards Henglein, the factors behind this assessment are different but again, we do not find that the CMA’s conclusion was unreasonable. The Decision showed that Henglein had become a less active supplier to the UK since Brexit. Although Cérélia cited its very large total DTB capacity relative to the size of the UK market, that alone does not address legitimate questions as to how much of Henglein’s capacity it would choose to devote to supplying UK customers, or indeed the extent to which those customers would perceive the logistical complications of depending on a supplier whose production was based outside the UK. This is not to say that all the indicators around the threat from Henglein as a PL supplier pointed unequivocally in favour of the CMA’s assessment, but the criticisms raised by Cérélia come nowhere close to proving its case to the requisite standard for establishing irrationality.”

We dismiss the challenge to the CAT Judgment based upon rationality. In relation to the consultation point this was a variant or refinement of the rationality challenge in respect of which we had already granted permission to appeal. The issue was important and we are fully able to determine it upon the material before us. The respondent did not have any difficulty in addressing it. For this reason, we grant permission to appeal but we reject the ground of challenge.

We turn now to Grounds III – V. These raise novel points of law and procedure of potentially broader significance to the operation of the merger control regime. We were presented with detailed analysis and argument from the parties. The first ground concerns the meaning of “special” in section 39(3) EA 2002.

Cérélia argues that the CAT erred in holding that the CMA had “special reasons” to extend the time for completing its investigation under section 39(3) EA 2002. The CAT implicitly acknowledged that the CMA did not give special reasons in its notice of extension (“the Extension Notice”). Accordingly, it was bound to find that there were no such reasons. In law neither the reasons set out in the Extension Notice nor those set out by the CMA in subsequent correspondence and in witness statement evidence are capable of amounting to proper reasons under the legislation. The CMA therefore acted unlawfully when extending time for the investigation.

The Extension Notice was communicated to the parties on 5 October 2022. It stated:

“The Inquiry Group has decided that the reference period should be extended by eight weeks under section 39(3) of the Act as it considers that there are special reasons (set out below) why the final report cannot be prepared and published within the reference period. The revised reference period will expire on 24 January 2023.

In taking this decision, the Inquiry Group had regard to the complexity of the inquiry, the need to consider the issues raised by the Parties and by third parties, including the broad scope of the submissions made by the Parties in response to the Annotated Issues Statement and Working Papers and the need to reach a fully reasoned final decision in the statutory timeframe. The Inquiry Group considers that completion of its investigation and the publication of its final report will not be possible within the original reference period.”

On 6 October 2022, Cérélia wrote to the CMA objecting upon the basis that there were no special reasons. It argued that, contrary to what was stated in the Extension Notice: (i) the issues were not complex; (ii) the responses to requests for information were not exceptionally broad and Cérélia’s submissions in relation to the Annotated Issues Statement (“AIS”) and Working Papers (“WP”) contained no material new evidence and were not exceptional; and (iii), the need to reach a fully reasoned final decision did not constitute a special reason.

The CMA responded on 13 October 2022. It continued to believe there were special reasons to justify an extension:

The CMA’s Defence filed on 20 March 2023 before the CAT was accompanied by a signed witness statement dated 17 March 2023 from Margot Daly, Chair of the CMA Inquiry Group (“the Statement”). Ms Daly had earlier signed the CMA’s Extension Notice. Cérélia did not challenge either the admissibility or the accuracy of the contents of the Statement. In oral argument Mr Hoskins KC accepted that the Statement did not advance new purported “special” reasons, but merely elaborated upon reasons already given.

The Statement insofar as relevant provides:

“19.

By 14 September 2022 the Group had sight of the combined submissions from both GMI and Cérélia, received by the CMA on 13 September following a short extension. These submissions comprised:

(a)

Eleven separate written submissions:

(i)

Two written submissions in response to the AIS (one from each of Cérélia and GMI);

(ii)

Eight written submissions in response to the disclosed working papers;

(iii)

One additional submission on internal documents.

(b)

506 detailed individual comments from Cérélia on the text of the disclosed working papers, provided by way of a markup of each individual working papers;

(c)

Two new economic analyses from Cérélia along with two underlying economic models (on each of Foreclosure Incentives and Entry and Expansion respectively);

(d)

Market Structure Charts and an explanatory note;

(e)

A brand consultancy agency presentation from Cérélia; and

(f)

A multi-country report from Cérélia.

20.

The additional material was voluminous, and the economic papers included complex analysis on the basis of vertical as opposed to a horizontal lens. At the time, there was a serious appetite amongst the Group to be able to have the time and space needed to properly consider the additional materials…

21.

At a Group Meeting on 15 September 20229, the Project Director provided an update in which he confirmed that the responses to the AIS and working papers had been received and the case team’s review of the responses was ongoing. In this meeting, the Project Director highlighted to the Group that at this early stage it may be necessary to extend the statutory timetable to allow the case team time to fully consider the submissions made by the Parties in response to the AIS and disclosed working papers.

…

23.

[On 21 September 2022] [t]he Project Director indicated that there was a considerable volume of material to be reviewed, necessitating an extension. It was clear from this discussion that additional time was required, and that it would be impossible to give fair consideration to the evidence submitted by the Parties without giving sufficient time to the task...

24.

… During the Group Meeting [on 27 September], the Project Director requested that the Group grant an 8-week extension to the Inquiry timetable and approve the draft revised administrative timetable… the overriding desire to ensure that the additional material was properly reviewed pointed towards the full 8-week extension.

…

26.

… In my view, no other course would have allowed the Group to give fair consideration to the materials submitted by the Parties, and to take it fully into account.”

Given that there was no challenge to the admissibility of the Statement the question arising upon this appeal is whether the reasons given therein constituted “special reasons” for the purposes of section 39 EA 2002, the relevant wording of which provides as follows:

“(1)

The CMA shall prepare and publish its report under section 38 within the period of 24 weeks beginning with the date of the reference concerned.

….

(3)

The CMA may extend, by no more than 8 weeks, the period within which a report under section 38 is to be prepared and published if it considers that there are special reasons why the report cannot be prepared and published within that period.”

Under section 40(4) EA 2002, no more than one extension is possible. Under sections 40(8) and (9), the Secretary of State may, by statutory instrument, amend the section 39 time periods, but only to reduce, not increase, them.

Section 40(1) EA 2002 provides that the CMA shall take action to impose a remedy (under section 41(2)): “where a report of the CMA has been prepared and published under section 38 within the period permitted by section 39 and contains the decision that there is an anti-competitive outcome”.

The usual timetable involved in a merger assessment is broadly as follows. A period of pre-notification discussions will usually take place for some weeks or months before a Merger Notice is issued. This triggers Phase 1. Where the CMA makes a reference under section 22(1) EA 2002, it must complete the Phase I stage of the review process within 40 working days of commencement (section 34ZA). There is scope for extension in limited circumstances. Where the CMA believes the merger results in a realistic prospect of a SLC, it will make a reference for an in-depth Phase II assessment (section 33) which assessment must be completed within 24 weeks. This can however be extended to 32 weeks (i.e. by 8 weeks) if the CMA considers there are special reasons (section 39). Absent an extension, the process will therefore usually comprise 32 weeks (for Phases I and II) together with such period of discussion between the parties and the CMA as occurred pre-notification.

The appellants draw attention to the Explanatory Notes before Parliament at the Bill stage relating to section 39:

“The section permits the CC to extend the 24-week period for the report for one further period of no more than 8 weeks where it is satisfied that there are special reasons for a delay. The section does not further define ‘special reasons’, but it is anticipated that they would include matters such as the illness or incapacity of members of a reporting group that has seriously impeded its work, and an unexpected event such as a merger of competitors.”

Mr Hoskins KC also rejected the notion that the following amounted to anything “special”: (i) any action or conduct which could be expected in the ordinary course of an investigation involving a “robust” defence by the parties concerned; or (ii), any of the material provided to the CMA by Cérélia in the present case which, he accepted, entailed a full blown refutation of more or less every point taken by the CMA. These were simply incidents of the ordinary, normal, workings of the system. Mr Hoskins KC argued that if the CMA experienced difficulty in producing a report within the 24-week period, it was obliged to acquire more resource but otherwise and by whatever means it was still required to produce its report within the 24 weeks. The scheme of the Act established that Parliament intended, by imposing a strict timetable and requiring special reasons for an extension, to balance the public interest in the CMA intervening in anticompetitive mergers with the public interest in mergers being dealt with expeditiously. The Act had to be construed strictly for strong policy reasons.

The CAT held that the CMA did have special reasons. It set out its analysis at paragraphs [317]-[318] of the Judgment:

The CAT concluded at [319]-[320] that:

“319.

Ideally these features would have been more explicitly particularised and dealt with by the CMA, whether in the extension notice or in correspondence with Cérélia and that might, perhaps, have addressed Cérélia’s objections at the time. Nevertheless they support the proposition that this was an unusual case in terms of how little common ground there was at that stage of the inquiry. It was both necessary and appropriate for the CMA to carry out the further investigation and analysis that it in fact did. No doubt Cérélia would have criticised it if the CMA had not.

320.

The fact that, in the event, the CMA conducted the Further Consultation into matters raised and particularly the issue of alternative competitive constraint involved the further testing of the CMA’s thinking and Cérélia’s responses in a way that partially (albeit ex post) justified the extension.”

The CAT held (Judgment paragraph [292]) that the meaning of “special reasons” in section 39 EA 2002 was for it to determine as a matter of law. Although the CMA had a “margin of appreciation” to determine what amounted on the facts to a special reason, if the CMA made an error of law the CAT could intervene. Whether a matter was sufficiently special to trigger the power to extend was “…a matter of fact for the CMA”. The CAT could intervene only “…on the usual judicial review grounds”.

As to the meaning of “special reasons”, the CAT considered that the provision required special reasons “in the form of good, case-specific reasons which justify an extension of the normal time limit for the management and conclusion of Phase 2 cases”. It was neither necessary nor helpful to put a gloss upon the meaning of “special”. The expression balanced Parliament’s desire for the CMA to perform its functions with appropriate expedition with its desire that the CMA perform its duties in a thorough manner. The need for expedition did not therefore require “an unduly narrow approach to the concept of special reasons” (Judgment paragraph [303]). However, to avoid a circular approach in which “special” applied whenever the CMA needed an extra eight weeks, it was nevertheless important that “special reasons” had a real meaning and imported a degree of qualification that separated such cases from the norm.

In our judgment the CAT did not err. Like the CAT, we have come to the conclusion that it is undesirable to lay down any canonical definition of “special”. What amounts to special will be fact and context specific and Parliament deliberately accorded an undefined discretion to the CMA. Nonetheless, we are able to express conclusions about certain arguments raised before us.

First, we consider why Parliament created a power for the CMA to extend time. The statute creates a one-size-fits-all regulatory framework which does not distinguish between different types of case or different levels of complexity. It does not cater for circumstances where, for instance, the CMA receives late submissions by the parties to be considered as a part of their defence and cannot do justice to them before the deadline expires. This explains why some sort of a safety valve is needed. Without the facility for an extension a variety of perverse incentives might arise. Parties might “game the system” by overloading the CMA with materials shortly before the deadline, leaving the CMA unable to complete an appropriately thorough report by the end of 24 weeks. Alternatively, if the CMA was up against an impossible deadline in a borderline case it might feel pressurised or constrained to take a cautious (negative) decision upon incomplete evidence to obviate the risk that the time limit expired and there was then nothing to prevent the continuance of a transaction which the CMA had cause to believe was anticompetitive. This might be so even if a solution or compromise might have been achieved during an extension, for instance as to remedy.

Next, the choice of language is important. There are two phrases in section 39(3) to consider: “special” and “if it considers”. We take each in turn.

As to “special”, Parliament’s use of the phrase is unusual. The more obvious phrases to use to qualify “reasons” would be “exceptional” or “extraordinary”, or even “good”. Parliament however chose not to use “exceptional” or “extraordinary” phrases which might, consistent with the appellant’s argument, have imported a requirement that facts and circumstances said to be “special” had to be out of the ordinary. Neither of the examples given in the Explanatory Notes are extraordinary or exceptional events. Illness, for instance, is a relatively commonplace occurrence. Equally, given that an investigation running into Phase II can last a reasonably lengthy amount of time (see paragraph [94] above) it will not be unusual for some development in a marketplace to occur which might cause the CMA to need extra time to factor that development into account, including disclosing its analysis to the parties for their considered response. We therefore accept that a reason which is extraordinary or exceptional might qualify as a “special” reason; but we do not accept the converse, namely that the only reasons that are “special” are those which are exceptional or extraordinary. The latter may be a subset of “special” reasons but do not describe its legal boundaries.

The CAT described special reasons as being related to the “good, case-specific reasons”. There is an obvious common sense attractiveness to this formulation. However, care is required. The CAT did not define what it meant by “case-specific” and whether this was to be construed broadly or narrowly. Many reasons, properly defined as “special”, might relate to the case; but that is a qualification not found in the legislation and it is hard to say that, for example, illness is specific or special to the “case”. Equally, we have no doubt that, for a reason to be special, it must be “good” in the sense that we can infer that Parliament did not intend a bad reason to be “special”. However, Parliament did not use the expression “good”; instead preferring the expression “special”.

There are a number of particular arguments advanced by the appellant that we would address:

Next there is the phrase “if it considers”. Parliament has, by using these words, conferred a discretion upon the CMA. Further, it has not set out a list of facts and matters that the CMA is required to take into account. It has, to the contrary, left the discretion open-ended. Parliament did not wish to pre-empt or predict the sorts of circumstances which might justify an extension and was content to leave this (subject to CAT supervision) to the judgment of the regulator. We see the logic behind this approach. The combination of facts and matters which might give rise to the need for an extension are unpredictable. Importantly, an extension is unlikely materially to disadvantage the parties. If the effect is to enable the CMA to conduct a more comprehensive, and thereby fairer, evaluation then it might increase the possibility that a transaction is cleared because, for example, lingering doubts are overcome or a remedy is identified. Conversely, if the extension serves only to confirm the CMA in a pre-existing provisional, negative, view of a transaction, all the parties have lost is two months during which it has not been possible to implement remedial measures. Either conclusion serves the public interest better than a hurried CMA decision that might be incorrect.

All of the above supports the approach adopted by the CAT. Whether reasons given are “special” is ultimately a matter of legal classification, and hence one for the CAT. But the CAT will take into account that Parliament has not sought to define what amounts to “special” which, when considered in conjunction with “if it considers”, accords a relatively broad discretion upon the CMA. A decision of this sort will involve a balancing of available resources against the nature and complexity of the work outstanding at any one particular time in the context of an appreciation by the CMA of the need to be as comprehensive, thorough and fair as it possibly can be within the tight timeframe imposed upon it by Parliament.

We turn to the present facts. The starting point is the reasons set out in the Statement. As a matter of public law a decision maker is not ordinarily allowed to advance new reasons to justify a decision when a matter is challenged before a court. This is to prevent the decision maker unfairly seeking to pull itself up by its bootstraps after the event. In the present case there is no challenge to either the admissibility or the accuracy of the contents of the Statement. It is accepted that the Statement is an elaboration, but not an extension, of the reasons in the Extension Notice. We therefore take the Statement as identifying the motivating reasons. We do, however, share the concerns of the CAT as to the paucity of detail set out in the Extension Notice. The notice bore all the hallmarks of an identikit document and communicated very little of real substance. In the future we would expect the extension notice to be the repository of all relevant reasons justifying an extension, concisely expressed. As the CAT noted, that approach might have allayed Cérélia’s objections at the time.

We agree with the CAT as to its assessment of the reasons as disclosed by the Statement. In order for this Court to understand the background to the proffered reasons we invited the parties to prepare an agreed, blow by blow, chronology of the steps occurring during the Phase I and II investigations. This was produced. The critical moment for the CMA appears to have emerged on receipt of the submissions from the parties to the AIS and WP on 13-14 September 2022. Over the following week, it became evident to the CMA that the combination of (i) an unusual lack of narrowing of the issues; and (ii), the very substantial volume of material provided in the response which related to the full gamut of issues, including whether the merger was to be examined as horizontal or vertical, created acute problems of time. To deal with these issues necessitated an extension. The CMA concluded that it would be “impossible” to produce the Final Report by the original deadline of 29 November 2022. The decision to extend was ultimately taken on 27 September 2022 (week 15 of the 24-week period).

In our judgment these reasons fall within the broad umbrella of “special” and the CAT was correct to so conclude.

We agree with Mr Hoskins KC that parties are entitled to defend their position robustly. No criticism attaches to Cérélia for the rigour of its defence. But if this involves a decision not to make concessions or narrow the issues then a consequence might be that the CMA comes under intense pressure of time and considers that it is unable to complete its review fairly and thoroughly absent an extension, so that this might be relevant to a decision to extend time.

Evidence was before the Court that c. 50% of merger cases before the CMA now result in an extension. We note that in a Ministerial Statement made at the time of the Bill it was stated that it was expected that the “vast majority” of cases would be determined without an extension. It is possible that the number of extensions has increased as the growth of digitisation enables parties to submit ever more detailed submissions. But we have no evidence on this and are loathe to draw conclusions or inferences based upon speculation. All we can say is that if there is evidence that the system of extension is not working (and it does not necessarily follow that the increase indicates that it is failing) then there could be cause for a review by Parliament.

For all the above reasons we reject Ground III.

We turn next to Ground IV. If we are correct in our analysis of Ground III the issue arising under Ground IV is academic. Nonetheless, it was addressed by the CAT, argued fully before us, raises an important point of principle, and we are clear as to the conclusion. The starting premise is that the CMA acted unlawfully in extending time. The question which then arises is as to the consequences. Cérélia argues that Parliament, in the EA 2002, laid down a tight and restrictive timetable for the completion of merger inquiries. Read purposively, any decision taken in breach of the stipulated time limits was intended by Parliament to be unlawful and a nullity. Given that – on this premise - the negative decision adopted (out of time) was a nullity it followed that no lawful decision had been taken by the expiry of the initial Phase II investigation and nothing in law therefore existed to prohibit the transaction. The CMA disagrees and submits that whether a decision taken in breach of a procedural requirement is a nullity is for the discretion of the CAT which has a statutory power over the grant of relief. It argues in this case that the (ex hypothesi) prior procedural breach did not result in the decision losing legal effect. Provided a decision was taken in good faith (which it was) then it should not be quashed (by the CAT) and remains valid even if, technically, unlawful.

Both parties accept that an analysis of Parliamentary intent is relevant. If the legislation indicates how a discretion should be exercised then the court must respect that direction. The CMA’s obligations under the EA 2002 are expressed in mandatory terms:

Nothing however expressly addresses the legal consequences of a final decision taken beyond the permitted timeframe, in the absence of “special reasons”. However, the legislation does provide, under section 120 EA 2002, for decisions of the CMA to be subject to judicial review on normal principles and it follows that it is to ordinary principles of administrative law that we must look for an answer. Under section 120(5) if the CAT finds that the CMA acted unlawfully it has a discretion (c.f. “may”) as to the relief to be granted:

“(5)

The Competition Appeal Tribunal may—

(a)

dismiss the application or quash the whole or part of the decision to which it relates; and

(b)

where it quashes the whole or part of that decision, refer the matter back to the original decision maker with a direction to reconsider and make a new decision in accordance with the ruling of the Competition Appeal Tribunal.”

In considering how the discretion must be exercised the point of departure is first principle. Decisions by public bodies are presumed valid until set aside or otherwise held to be invalid by a court of competent jurisdiction: See e.g. De Smith's Judicial Review (9th ed., 2023) paragraph [4-066]. Public decisions have legal effect at the time they are made. The public must be able to rely upon their validity so a decision that is not challenged within the relevant (limited) time period for judicial review retains legal effect, regardless of whether it would have been declared unlawful had it been challenged: e.g. McLaughlin v His Excellency the Governor of the Cayman Islands [2007] UKPC 50; [2007] 1 W.L.R. 2839, per Lord Bingham at paragraph [14]). In R. v Panel on Takeovers and Mergers Ex p. Datafin Plc [1987) Q.B. 815 at 840 it was said that it was "a very special feature of public law decisions” that:

"…however wrong they may be, however lacking in jurisdiction they may be, they subsist and remain fully effective unless and until they are set aside by a court of competent jurisdiction."

A court has a discretion as to remedy which includes in relation to an unlawful decision: IRC v National Federation of Self Employed and Small Businesses [1982] AC 617 at page [656]). Section 120 EA 2002 replicates that common law jurisdiction. Where there is a discretion it: “…must be exercised judicially and in most cases in which a decision has been found to be flawed, it would not be a proper exercise of the discretion to refuse to quash it”: R (Edwards) v Environment Agency [2008] UKHL 22 (“Edwards”) at paragraph [63]. In some cases there will therefore be a strong pressure in favour of quashing. The overarching legislative regime might lead to the conclusion that a discretion not to quash is non-existent or very limited. That might be the case because statute wholly or substantially circumscribes the discretion. To foreshadow a point addressed below it is evident from case law that the good faith of the decision maker will be an irrelevant consideration if the impugned decision lacks a proper legal basis.

In R v Soneji [2006] 1 AC 340 (“Soneji”) Lord Steyn indicated in relation to the modern approach to be taken by courts, that the “rigid” historical approach which distinguished between “mandatory and directory” requirements, and which had “many artificial refinements”, had outlived its usefulness (at paragraph [23]). Instead: “… the emphasis ought to be on the consequences of non-compliance, and posing the question whether Parliament can fairly be taken to have intended total invalidity”. Subsequent case law has applied this formulation by posing two questions: (i) did Parliament intend total invalidity to result from failure to comply with the statutory requirement; and (ii), if not, and despite invalidity not being the inevitable consequence of a failure to comply, does it nonetheless have that consequence in the circumstances of the given case and, if so, on what basis? If the answer to the first question is in the affirmative that is the end of the analysis. If the answer is in the negative, and invalidity is not the inevitable consequence of the failure, then the court considers broader public interest considerations including the concept of substantial compliance: See e.g. North Somerset DC v Honda Motor Europe Ltd [2010] EWHC 1505 (QB) at paragraph [43]; and, Home Department v SM (Rwanda) [2018] EWCA Civ 2770 at paragraphs [51] and [52]. Where a discretion exists the “nature of the flaw in the decision” is a relevant matter: Edwards (ibid).

Even where a decision is to be quashed the Court may fashion a temporal remedy. It can declare nullity from inception or prospectively only: e.g. R (British Academy of songwriters, composers and authors musicians limited small v Secretary of State for Business, Innovation and Skills [2015] EWHC 2041 (Admin) paragraphs [12] – [21].

The CAT found for the CMA. The CAT set out a number of reasons. Two are central. First, in relation to the argument of Cérélia that Parliament intended a tight and rigid timetable to be applied to the regulation of mergers and thereby condemned any regulatory decision which was non-compliant with statutory strictures, the CAT disagreed. Parliament’s intention that there be expedition was balanced against its intention that the CMA investigate and remedy anticompetitive harm arising from a merger in the wider interest of protecting markets and consumers. The conferral upon the CAT, by section 120(5) EA 2002, of a power over remedy was a clear indication by Parliament that a balancing exercise was required when fashioning a remedy (Judgment paragraphs [339] and [340]). This was not a case where Parliament had predetermined the outcome of the exercise to be carried out. Here, the balance would lie in preserving a decision reflecting the importance of remedying, mitigating or preventing the SLC that had been identified. Secondly, absent the extension it was inevitable the CMA would still have found an SLC requiring divestiture: R (Plan B Earth) v Secretary of State for Transport [2020] EWCA Civ 214 at paragraph [267]. This was evident from the PF which was, in substantial measure, reflected in the FR.

In our judgment both are compelling reasons. In exercising the judicial discretion, a balance must be struck between the private and public interest. There is undoubtedly an interest in the merger control regime being operated expeditiously. This is a private interest in that it meets the needs of commercial parties for speed and legal certainty. But it is also a public interest in that the state has no proper interest in restricting commercial activity for longer than is necessary to meet a regulatory objective. Equally, there is a strong public interest in the preservation of a decision intended to protect the integrity of competition in a market and consumers. In balancing these interests, it is relevant that the prejudice to the merging parties of an unlawful extension is limited (see paragraph [112] above). If an extension permits the CMA to approve a merger, or find a solution (such as a remedy short of prohibition or divestiture) which permits it to proceed, the extension has worked in favour of the parties. Their interests are protected. It is only if the extension leads to a negative decision that there is prejudice to the private interest. But in such a case it is to be inferred (certainly on the facts of this case) that had the CMA been forced into completing its work within the timetable it would have come to the same (negative) conclusion, even if, as the Tribunal noted, it might not have been as well reasoned and worded having been “rather rushed” (Judgment paragraph [342]). So the outcome is the same. All that the merging parties have lost is 2 months within which they could have implemented the divestiture remedy.

We are less convinced by the other reasons given by the CAT. First it is said that Cérélia chose to complete the transaction prior to the final decision and accepted the risk of an adverse outcome. The statutory regime does not prohibit completion of a transaction pending a regulatory decision and the acceptance of the risk of rejection is unaffected by the existence or otherwise of an extension. It appears to us that this factor is neutral in the judicial balance. Secondly, reference is made to the substantial volume of material submitted by Cérélia. No criticism attaches to the company for having robustly defended its position. The volume of submitted material might be relevant to whether there be “special” reasons for extending time. But if the volume of evidence submitted is found not to constitute “special” reasons, and not therefore to justify an extension of time, which is the working hypothesis for Ground IV, it is hard to see why it should then be resurrected as a factor adverse to the parties when it comes to the determination of a remedy. We would add, finally, that we do not accept the argument of the CMA that any decision promulgated in good faith thereby becomes valid, notwithstanding the nature and severity of any prior unlawfulness. Whilst the absence of good faith might be a factor relevant to the exercise of judicial discretion (in favour of quashing), the mere fact that a decision is taken in good faith (which we assume is the more or less inevitable default position), albeit unlawfully and erroneously, does not mean that, in an appropriate case, it should not be quashed. This is clear from case law: see e.g. the quotation from Edwards (ibid) set out at paragraph [124] above.

For the above reasons we dismiss the appeal on ground IV.

We deal with Ground V briefly. We have concluded that the CMA identified reasons falling within the broad expression “special”. We have also concluded that the CMA has a discretion as to what will amount to special reasons justifying an extension. We have also concluded that the reasons upon which the CMA relied were capable of amounting to special reasons within the scope of its discretion. It follows that the CAT was correct to conclude that there was nothing irrational about the decision to extend time. We reject Ground V.

For all the above reasons we dismiss the appeals.