EUROPEAN COMMISSION DG Competition

(Only the English text is authentic)

MERGER PROCEDURE

REGULATION (EC) 139/2004

Article 8(2) Regulation (EC) 139/2004

Date: 27/03/2017

This text is made available for information purposes only. A summary of this decision is published in all EU languages in the Official Journal of the European Union.

Parts of this text have been edited to ensure that confidential information is not disclosed; those parts are enclosed in square brackets.

EUROPEAN COMMISSION

Brussels, 27.3.2017 C(2017) 1946 final

Public version

COMMISSION DECISION

of 27.3.2017

declaring a concentration to be compatible with the internal market and the EEA Agreement (Case M.7932 – Dow/DuPont)

(Only the English text is authentic)

COMMISSION DECISION

of 27.3.2017

declaring a concentration to be compatible with the internal market and the EEA Agreement (Case M.7932 – Dow/DuPont)

(Only the English text is authentic)

THE EUROPEAN COMMISSION,

Having regard to the Treaty on the Functioning of the European Union,

Having regard to the Agreement on the European Economic Area, and in particular Article 57 thereof,

Having regard to Council Regulation (EC) No 139/2004 of 20 January 2004 on the control of 1 concentrations between undertakings, and in particular Article 8(2) thereof,

Having regard to the Commission's decision of 11 August 2016 to initiate proceedings in this case,

Having given the undertakings concerned the opportunity to make known their views on the objections raised by the Commission,

2 Having regard to the opinion of the Advisory Committee on Concentrations,

3 Having regard to the final report of the Hearing Officer in this case,

Whereas:

SECTION I: I NTRODUCTION

(1) On 22 June 2016, the Commission received a notification of a proposed 4 concentration pursuant to Article 4 of Council Regulation (EC) No 139/2004 ("the Merger Regulation") by which The Dow Chemical Company ("Dow", United States, ("US")), the ultimate parent company of the undertaking comprising the Dow group, and E.I. du Pont de Nemours and Company ("DuPont", US), the ultimate parent company of the undertaking comprising the DuPont group, enter into a full merger within the meaning of Article 3(1)(a) of the Merger Regulation by way of an Agreement and Plan of Merger of 11 December 2015 ("the Transaction"). Dow and DuPont are collectively referred to in this Decision as "the Parties", whilst the

1 OJ L 24, 29.1.2004, p. 1. With effect from 1 December 2009, the Treaty on the Functioning of the European Union ("TFEU") introduced certain changes, such as the replacement of "Community" by "Union" and "common market" by "internal market". The terminology of the TFEU will be used throughout this Decision. 2 OJ C ...,...200. , p.... 3 OJ C ...,...200. , p.... 4 OJ L 24, 29.1.2004, page 1.

1

undertaking that would result from the Transaction is referred to as "the merged entity".

(2) Dow is a diversified chemicals company with its headquarters in the US. It is active in plastics and chemicals, agricultural sciences, and hydrocarbon and energy products and services. In 2014, Dow generated global sales of approximately EUR 46 billion.

(3) DuPont is also a diversified company with its headquarters in the US. It produces a variety of chemical products, polymers, agro-chemicals, seeds, food ingredients, and other materials. In 2014, DuPont generated global sales of approximately EUR 21 billion.

SECTION II: THE OPERATION AND THE CONCENTRATION

(4) On 11 December 2015, Dow and DuPont announced a “merger of equals”. The merged entity would have a market capitalisation of approximately USD 130 billion. At a later stage, Dow and DuPont intend to create from their combined activities three separate publicly traded companies focusing on agriculture, material science and specialty products respectively.

(5) In agriculture, the merged entity would become the number 1 global integrated crop protection and seeds player, the number 2 global seeds player (number 1 in the European Economic Area (EEA) with a small increment from Dow).

(6) Dow and DuPont announced that the Transaction would create around USD 3 billion 5in cost synergies, with potential growth synergies of USD 1 billion.

Figure 1 - Three planned spin-offs from the Dow/DuPont merger

Source: Public announcement of the Dow/DuPont agreement, 11 December 2015

(7) The Transaction constitutes a concentration within the meaning of Article 3(1)(a) of the Merger Regulation.

(8) The Transaction is also notifiable in more than 20 other jurisdictions.

SECTION III: UNION DIMENSION

(9) The undertakings concerned have a combined aggregate world-wide turnover of 6 more than EUR 5 000 million[Dow: EUR 45 654 million; DuPont: 7 EUR 21 382 million]. Each of them has an aggregate Union-wide turnover in excess

5 Dow DuPont's investor presentation, dated 11 December 2015. 6 Turnover calculated in accordance with Article 5 of the Merger Regulation and Section C of the Commission Consolidated Jurisdictional Notice (OJ C 95, 16.4.2008, page 1). 7 2015 figures: Dow: EUR 45 592 million; DuPont: EUR 22 589 million.

2

8 of EUR 250 million [Dow: EUR 12 018 million; DuPont: EUR 4 074 million], but they do not achieve more than two-thirds of their aggregate Union-wide turnover within one and the same Member State.

(10) The Transaction therefore has a Union dimension pursuant to Article 1(2) of the Merger Regulation.

SECTION IV: THE PROCEDURE

1. COMMISSION' S PROCEDURE

(11) On 31 December 2015, the Parties informed the Commission of the planned notification of the Transaction by filing a case allocation request.

(12) On 5 February 2016, the initial meeting between the Commission and the Parties took place. A number of subsequent meetings between the Commission and the Parties took place during the pre-notification stage, in particular on 17 February, 26 and 27 April, 20 and 30 May, and 8, 9 and 15 June 2016.

(13) Between March and early June 2016 the Parties submitted the first draft of the Form 9CO to the Commission in several parts.

(14) On 15 June 2016, the Parties submitted the second draft of the Form CO to the Commission.

(15) During the Phase I investigation the Commission contacted a large number of market participants (mainly customers of the Parties, competitors, stakeholders), requesting information through electronic questionnaires, telephone calls and written requests for information pursuant to Article 11 of the Merger Regulation.

(16) In addition, the Commission also sent several written requests for information to the Parties and reviewed internal documents of the Parties submitted during the Phase I investigation.

(17) On 20 July 2016, the Parties submitted five sets of commitments which consisted of the Herbicide Remedy, Insecticide Remedy, Genome Editing Remedy, Acid Co- Polymer Remedy, and Ionomers Remedy. Those commitments were not market tested.

(18) On 11 August 2016, the Commission found that the Transaction raised serious doubts as to its compatibility with the internal market and the EEA Agreement and adopted a decision to initiate proceedings pursuant to Article 6(1)(c) of the Merger Regulation ("the Article 6(1)(c) Decision").

(19) The Article 6(1)(c) Decision raised serious doubts in relation to (i) Agriculture: selective herbicides targeting broadleaf weeds, most prominently in relation to

8 2015 figures: Dow: EUR 10 575 million; DuPont: EUR 4 030 million. 9 The sections of the first draft of the Form CO were provided in particular on 18 and 23 March (Sections 1-5 and 9-11; Seeds; Fungicides); 23 and 25 March 2016 (EVA; Material Science); 15 April 2016 (Insecticides); 25 and 28 April 2016 (Nutrition, Safety); 5, 9 and 12 May 2016 (R&D in Crop Protection, Herbicides – Introduction, Cereals and Rice); 13, 16 and 17 May 2016 (Elastomers, Vertical Relations in Agricultural Products, Electronics and Communications and Oilseed Rape Herbicides); 22 and 23 May 2016 (Maize Herbicides and Beets Herbicides); 27 and 30 May 2016 (Cereal Herbicides (country by country analysis), Other Crops Herbicides and Miscellaneous); 4 June 2016 (Divested Businesses and Other Vertical Relationships).

cereals but also in relation to other crops; insecticides targeting chewing insects notably for specialty crops (grapes/vines, fruits and nuts, vegetables, and some further sub-segments like citrus, pome and stone fruit, strawberries, tomatoes), as well as corn/maize; innovation in crop protection; gene editing technologies; bundling; and (ii) Material Science: acid co-polymers; ionomers.

(20) The Article 6(1)(c) Decision also highlighted areas for further investigation, such as (i) Agriculture: insecticides for other crops, such as oilseed rape, as well as broad- spectrum insecticides and targeting sucking pests; nematicides; fungicides; licensing and supply of active ingredients; technical tying; (ii) Material Science: vertical link involving glacial methacrylic acid (“GMA") and ethylene/n-butyl acrylate/glycidyl methacrylate ("E/nBA/GMA") ter-polymer; vertical link involving Polyolefin elastomers ("POE") and maleic anhydride ("MAH")/POE; and (iii) Specialty products: vertical link involving DuPont's activities in 248 photoresist polymers and Dow's activities in 248nm photoresists to assess the risk of input foreclosure.

(21) The Article 6(1)(c) Decision also found that the commitments proposed by the Parties on 20 July 2016 were not sufficient to eliminate the Commission's serious doubts as to the compatibility of the Transaction with the internal market, for the reasons described in that decision.

(22) On 12 August 2016, the Commission provided non-confidential versions of certain key submissions of third parties collected during the Phase I investigation to the Parties.

(23) On 26 August 2016, the Parties submitted their written comments on the Article 6(1)(c) Decision ("the response to the Article 6(1)(c) Decision").

(24) On 31 August 2016, following the Parties' response to the Article 6(1)(c) Decision, a State of Play meeting took place between the Commission and the Parties.

(25) During the Phase II investigation, the Commission sent more than 40 requests for information to the Parties. The Commission obtained and analysed substantial amount of information from the Parties, including internal documents, third party industry reports and submissions.

(26) The Commission also sent a number of requests for information to competitors, customers of the Parties, and other stakeholders. The Commission conducted further calls and meetings with some market participants, including customers, competitors of the Parties and other stakeholders such as technical institutes.

(27) On 2 September 2016, the merger review time period was extended by 10 working days following the Parties' request pursuant to Article 10(3) of the Merger Regulation.

(28) On 8 September 2016 and 3 November 2016, the Commission adopted two decisions pursuant to Article 11(3) of the Merger Regulation suspending the merger review time limit due to the failure of the Parties to provide certain requested documents. The first suspension lasted from 1 September 2016 until 26 September 2016 and the second from 13 October 2016 until 7 November 2016, when the requested documents were provided.

(29) On 1 December 2016, the Commission informed the Parties of the preliminary results of the Phase II investigation during a formal State of Play meeting.

4

(30) A Statement of Objections addressed to the Parties was adopted on 7 December 2016. Access to file was given to the Parties via DVD-ROMs on 8 December 2016, followed by access to confidential quantitative and qualitative information in a data room on 14, 15 and 16 December 2016. On 22 December 2016 and 6 January 2017, the Parties were given additional access to the confidential version of their data room reports at the Commission's premises, where they prepared a new confidential version of their data room reports as well as confidential annexes that supplemented their written response to the Statement of Objections. Access to file was subsequently given to the Parties via encrypted emails on 20 December 2016, 6 January 2017, 23 January 2017, 2, 15 and 16 February 2017 and 2, 10 and 13 March 2017.

(31) The Parties replied to the Statement of Objections on 21 December 2016 (“the response to the Statement of Objections”). On 6 January 2017, following additional access to the data room, each Party submitted a supplemental response to the Statement of Objections based on the confidential information contained in the data room. These two supplemental responses to the Statement of Objections were put on the Commission's file.

(32) An oral hearing took place on 9 January 2017.

(33) On 20 January 2017, the Commission, with the agreement of the Parties, extended the merger review time period by additional 10 working days pursuant to Article 10(3) of the Merger Regulation.

(34) On 20 January 2017, the Commission sent the first Letter of Facts to the Parties, who replied on 30 January 2017 (the “response to the first Letter of Facts”).

(35) On 1 February 2017, the Commission sent the second Letter of Facts to the Parties, who replied on 6 February 2017 (the “response to the second Letter of Facts”).

(36) In order to address the competition concerns identified in the Statement of Objections, the Parties submitted commitments on 7 February 2017. Consequently, the period for the adoption of a final Decision was extended by 15 working days pursuant to Article 10(3) of the Merger Regulation.

(37) The Commission launched a market test of those commitments on 8 February 2017. 10The Parties submitted final commitments on 17 February 2017.

(38) By decision of 20 February 2017, the Commission rejected a claim by Dow regarding legal professional privilege ("LPP") in relation to part of an internal e-mail which had been submitted by Dow to the Commission during the proceedings. The said document, and particularly the part for which Dow had claimed LPP, contains relevant evidence which supports several of the findings of the Commission in the present Decision. While the Commission considers that, further to its decision rejecting Dow's LPP claim, it would be entitled to use in full the document at stake, it has decided, without prejudice, not to rely, at this stage, on the content of the contested part of the document, until Dow has had the opportunity, if it so decides, to appeal the decision rejecting its LPP claim before the General Court.

(39) The meeting of the Advisory Committee took place on 14 March 2017.

(40) The Hearing Officer issued his final report on 16 March 2017.

10 The Parties submitted a Corrigendum to the Crop Protection Final Commitments on 27 February 2017.

5

2. PROCEDURAL ISSUES

(41) During its review of the Transaction, including the pre-notification stage, the Commission encountered a number of procedural issues which made it more difficult to conduct an efficient and effective investigation.

(42) The main of those issues are described in this Section IV.2. Those issues are important for understanding the way in which the Commission interpreted specific evidence in its substantive assessment and the reasons for certain procedural measures that the Commission had to take in order to ensure the integrity of its investigation of the Transaction.

2.1. Requests for information and exchanges with the Parties in order to obtain information necessary for the Commission’s investigation and assessment

(43) In its assessment of mergers the Commission routinely relies on internal documents that undertakings prepare in the ordinary course of business. Those documents allow the Commission to gain a quicker and better insight into the relevant markets as viewed by the market participants themselves.

(44) Internal documents can prove to be an especially useful source of evidence in relation to specific areas of investigation. For example, they often allow the Commission to verify factual claims made by the parties and verify data they submit. Internal documents are frequently crucial to understand the factors which affect the incentives of the parties before and after the proposed merger.

(45) Undertakings are well aware of the importance of internal documents as a source of evidence which is likely to be requested by competition authorities during a merger review.

(46) In attributing the probative value to specific internal documents, the Commission takes into account the timing and context in which they were prepared. In particular, internal documents prepared in the ordinary course of business, for example before the proposed merger was agreed upon or without the knowledge of the preliminary competition concerns, may have higher probative value than internal documents prepared for the Commission or influenced by the Commission's review of the proposed merger. Indeed, the former type of internal documents usually reflects the information objectively and independently of the merger control procedure.

(47) It would therefore be particularly harmful for the integrity of a Commission's merger investigation if the parties strategically develop or modify pre-existing evidence composed of internal documents with a view to influencing the investigation of the Commission.

(48) Where the Commission considers that such behaviour may exist, the Commission must carefully assess all internal documents submitted by the parties and otherwise obtained by the Commission.

(49) In this Decision, the Parties' internal documents have also played a key role in the Commission's assessment. They were useful for assessing the impact of the Transaction in practically all of the relevant markets mentioned in this Decision, as well as in the relevant markets in which no concerns have been raised precisely due to the evidence found in internal documents.

(50) Sections IV.2.1.1 and IV.2.1.2 describe certain instances which [details on internal documents request].

6

2.1.1. The incident of January 2016 concerning an existing [internal document] showing direct competition between Dow’s product Spinetoram and DuPont’s product Rynaxypyr

(51) In the ordinary course of business Dow uses the so-called [internal document], in particular for the preparation of pricing of new crop protection products. [Information on internal documents for the analysis of competitive interaction]. Dow introduced [internal document] in early 2015.

(52) The Commission considers [internal document] to be important documents for the analysis of the competitive interaction between the different products of different players in crop protection markets. [Information on internal documents for the analysis of competitive interaction] indicate planned pricing, market sizes, market shares and benchmarking against other competitors. Such information is particularly important, for example, for the assessment of closeness of competition between different crop protection products (for the use of [internal document] by the Commission as evidence, see Sections V.6.3 to V.6.6 on crop protection).

(53) In their response to the Statement of Objections, the Parties submit that [internal document] are only one tool in Dow's marketing toolbox and are not determinative for price setting. Even if true, this statement does not undermine the fact that [internal document] provide valuable information on pricing and closeness of competition. Indeed, the Commission notes that since the beginning of the Transaction, Dow considered [internal document] to be important for the purposes of competitive assessment. Already in early January 2016 Dow was looking internally for [internal document] to assess whether they supported specific arguments to be made in the context of the merger notification. The relevant e-mail exchanges are set out in the following recitals.

11 (54) On 5 January 2016, a Dow employee, Mr. […],asked his colleagues for examples of [internal document] in order to prepare a dossier for the upcoming merger 12review.

13 (55) On 7 January 2016, another Dow employee, […],replied to Mr. […] request by 14sending him a "[title of internal document]".

(56) In his reply on 11 January 2016, Mr. […] noted: "the example shows the direct competition between Spinetoram and Rynaxypyr, and shows kind of a exchangeability, [information on internal documents for the analysis of competitive interaction]."

(57) To this, Mr. […] responded on 11 January 2016: "[information on internal documents for the analysis of competitive interaction]."

11 Mr. […] is Dow's Regional Commercial Unit Leader Crop Protection Europe. 12 Dow's internal e-mail from […] to […] et al of 5 January 2016 at 08:16 AM, file name "[…]" (ID6696- 3172). 13 […] is Dow's Territory Marketing Leader South Zone. 14 Dow's internal e-mail from […] to […] of 7 January 2016 at 11:19 AM, file name "[…]" (ID6696- 3172). 15 Dow's internal e-mail from […] to […] of 11 January 2016 at 8:31 AM, file name "[…]" (ID6696- 3172). 16 Dow's internal e-mail from […] to […] of 11 January 2016 at 10:09, file name "[…]" (ID6696-3172).

7

(58) The Commission may examine in a separate procedure whether the e-mail exchange mentioned in recitals (54) to (57) would suggest that the employee of Dow, Mr. […], was ready to create a document which would show the lack of direct competition between Spinetoram and Rynaxypyr, in contrast to a pre-existing [internal document] prepared in the ordinary course of business.

(59) The Commission notes that, irrespective of the reasons [information on internal document for the analysis of competitive interaction], the Commission's view is that Dow’s Spinetoram competes with DuPont’s Rynaxypyr (see Section V.6.4).

(60) Finally, the Commission considers that the e-mail exchange mentioned in recitals (54) to (57) is not protected by LPP. While in their response to the Statement of Objections the Parties claimed that this e-mail exchange and the related documents are covered by LPP, no substantiation was provided. The Parties also did not react to the Commission's e-mail of 6 January 2017 which dismissed their LPP claim due (i) to the lack of substantiation and, in any event, (ii) the waiver of LPP by submission of the same documents as part of the Parties' response of 7 November 2016 to the Commission's Article 11(3) decision. The Commission also reiterated its position regarding the lack of LPP coverage in relation to this e-mail exchange at a meeting with the Parties on 27 January 2017.

2.1.2. The Parties [details on internal documents requests]

(61) [Details on internal documents requests].

17,18 (62) […][…].

(63) [Details on internal documents requests].

2.2. Delays in providing information and integrity of investigation

(64) Another area of procedural issues in this Decision concerned the delays in the provision of information and documents by the Parties which led the Commission to take measures to protect the integrity of its investigation, as described in this section.

2.2.1. Delayed submission of internal documents and third party reports during the pre- notification phase

(65) During the pre-notification phase, the Parties repeatedly provided the requested internal documents and third party reports with significant delays.

(66) On 2 and 5 February 2016, the Parties submitted to the Commission internal documents which they had previously provided to the US Department of Justice ("DoJ") under items 4(c) and 4(d) of the Hart-Scott-Rodino Form ("4(c) and 4(d) documents"). The number of those submitted documents was 122.

(67) On 5 February 2016, the Commission held an initial meeting with the Parties, where they presented a general overview of the Transaction and areas concerned.

(68) During that meeting, the Commission asked the Parties to provide their internal documents and third party competitive intelligence documents as early in the process

19 as possible.The Commission explained that this was important in order to understand the relevant markets quickly and in a way as viewed by the Parties in the 20 ordinary course of business and by other industry experts.As a starting point, the Commission suggested that the Parties provide the industry reports and supporting internal documents on which the Briefing Paper of 31 January was based since those 21would be readily available to the Parties.

(69) On 16 February 2016, the Commission held a second pre-notification meeting with the Parties, who provided an introduction to the areas of Crop Protection (including R&D) and Material Science. At that meeting, the Commission reiterated that the Parties' internal documents and third party reports would be a key source of evidence that would have to be produced quickly in order to facilitate the Commission's review of the Transaction.

22(70) In follow-up to that meeting, on 19 February 2016, the Commission sent an e-mail to the Parties asking them for their 4(c) and 4(d) documents submitted to the DoJ (to the extent they had not been already provided to the Commission) and analyst reports 23 concerning the Transaction. On 22 February 2016 the Parties repliedconfirming that the 4(c) and 4(d) documents submitted to the Commission on 2 and 5 February 2016 were the complete set submitted to the DoJ. No analyst reports were provided.

(71) On 26 February 2016, the Commission sent to the Parties the request for information RFI 1, in which it requested among other things the Parties' internal documents and third party reports regarding the Transaction and the Parties' activities in the relevant 24areas.

25 (72) On 11 March 2016, the Commission sent an e-mailto the Parties expressing a concern regarding the lack of further information and documents from the Parties in light of their previously stated intention to proceed as quickly as possible. In particular, the Commission asked to receive "as a matter of urgency" the requested industry reports upon which the data in the Briefing Paper was based, and the requested internal documents, […].

(73) Dow submitted the first analyst and industry reports on 14 and 17 March 2016, that is more than a month since they were initially requested by the Commission at the kick-off meeting of 5 February 2016. The production of those reports [details on procedure]. For example, the submitted Phillips McDougall industry reports are the

19 In line with the Commission's Best Practices on the conduct of EU merger proceedings, one of the goals of the pre-notification contacts is "to discuss issues such as the scope of the information to be submitted and to prepare for the upcoming investigation" (paragraph 6). 20 According the Commission's Best Practices on the conduct of EU merger proceedings, "DG Competition recommends that notifying parties should, as early as possible in pre-notification, submit internal documents such as board presentations, surveys, analyses, reports and studies discussing the proposed concentration, the economic rationale for the concentration and competitive significance or the market context in which it takes place. Such documents provide DG Competition with an early and informed view of the transaction and its potential competitive impact and can thus allow for a productive discussion and finalisation of the Form CO" (paragraph 17). 21 Parties' submission entitled "Briefing paper for the European Commission" submitted on 31 January 2017, ("Briefing Paper"). 22 E-mail from the Commission to the Parties, 19 February 2016 at 17:38 (ID28). 23 Two e-mails from the Parties to the Commission, 22 February 2016 at 15:31 and 19:42 (ID34). 24 Commission's request for information RFI 1, questions 12-19. 25 E-mail from the Commission to the Parties, 11 March 2016 at 14:49 (ID82).

9

customary source of reference used by the Parties and other players in the agrochemical industry.

(74)With respect to internal documents responsive to the Commission's request for information RFI 1 of 26 February 2016, the first ones relating to crop protection (one of the key areas of investigation) were submitted by DuPont on 25 April 2016, that is to say two months after the issuance of the Commission's request. On that date, DuPont provided only six internal documents.

(75)It was not until early June that the bulk of other internal documents responsive to the Commission's request for information RFI 1 relating to crop protection was 26 submitted by DuPont. Similarly, Dow submitted most of its internal documents 27 relating to crop protection in May and June,that is to say several months after the Commission's request and a few weeks before notification.

(76)Some of the documents which were responsive to the Commission's request for information RFI 1 of 26 February 2016 were not provided by the Parties until much later in the proceedings. For example,

(a)Several industry reports by a consultancy Kline & Company ("Kline") were provided only on 18 June 2016, a few days before the formal notification, in response to a specific question in a separate request for information (Commission's request for information RFI 8). Kline reports subsequently proved to be particularly important for establishing that the Parties' nematicide products compete (see Section V.6.5 on nematicides), contrary to what was 28argued in the Form CO.

(b)Two Phillips McDougall reports on seeds were submitted by the Parties only on 13 and 18 June 2016 in response to a specific question in a separate request for information (Commission's request for information RFI 8). Those reports were relevant for the Commission's analysis of the seed markets and the subsequent dismissal of competition concerns in relation to them.

(c)In response to the Commission's request for information RFI 13 of 28 June 2016, the Parties provided several relevant industry reports for the affected markets. Many of those industry reports were pulled from the online databases (IHS Chemicals, CRM), [details on procedure]. Those reports have been important in particular for the Commission's assessment of the impact of the Transaction in the area of Material Science.

(77)The Commission consistently reminded the Parties of the importance of providing internal documents and industry reports as early as possible. In addition to the Commission's e-mail of 11 March 2016 (see recital (72)), the Commission expressed its concerns regarding the slow production of documents, for example, during pre-notification calls with the Parties' lawyers on 19 and 31 May 2016 and at the meeting

26 See in particular DuPont's submissions of internal documents responsive to question 18 of Commission's request for information RFI 1 on 23 May (nine documents concerning Seeds and five documents concerning Crop Protection); 24 May (19 documents); 28 May (96 documents); 1 June (80 documents); 9 June (10 documents). 27 See in particular Dow's submissions of internal documents responsive to question 18 of Commission's request for information RFI 1 on 4 May (181 documents); 19 May (29 documents from Dow's agricultural business, also responsive to other Commission's requests for information); and 11 June (also responsive to other Commission's requests for information). 28 Form CO, part B.II, paragraph 60 and Annex B.II.07.

10

with the Parties on 30 May 2016. Also, on 16 June 2016, the Commission sent an e- 29 mail to the Parties reiterating its concerns regarding "the state of the file".The Commission drew the Parties' attention to the delays in the document production also later on, during the Phase I investigation, in particular at the State of Play meeting on 13 July 2016.

(78)In their response to the Statement of Objections, the Parties argue that (i) they acted under extreme time pressure to notify, as they “were aware of other planned concentrations in the agrochemical industry that were likely to be filed quickly in the 30 31 EU” and (ii) they had to “comply with very burdensome requests”,in particular related to contact details and the refusal of the Commission to grant “waivers for 32 affected markets with combined market shares of less than 35%.”The Parties also confirm that “[o]n June 22, 2016, the Parties formally submitted a Form CO of over 1,650 pages along with over 4,100 annexes, after having fully complied with the 33 12 prenotifications RFIs” and conclude that “there is no basis whatsoever to allege 34that the Parties delayed the pre-notification phase”.

(79)The Commission notes that it is often unavoidable to have an intensive pre- notification stage with extensive information requests in a case of a comparable scope and complexity. While this may impose heavy burden on the merging parties, [details on procedure].

(80)In particular, the Commission notes that it first requested the relevant internal and third party documents from the Parties in early February 2016. [Details on internal documents requests]. For example, in two months since the Commission's request for internal documents in RFI 1 DuPont produced […] such documents (see recital (74)). The Commission notes that while the Parties submitted to the Commission the 4(c) and 4(d) documents in early February, those documents were much more limited in 35 36 scopeand numberthan those requested by the Commission.

(81)The internal documents and third party reports eventually obtained by the Commission have been some of the key sources of evidence in this Decision. Given the size and complexity of the case, the delayed submission of those documents and reports in the pre-notification phase made the Commission's review of the Transaction significantly more time-consuming and difficult.

(82)Also, contrary to the Parties’ claim, elements requested in pre-notification requests for information were not provided at the time of notification (see for example Section V.2.2.2 on the market share methodology, Section V.2.2.3 on patent data and Section 2.1 of Annex 2 on data necessary to assess the Parties’ submissions on generics).

29 E-mail from the Commission to the Parties, 16 June 2016 at 18:04 (ID1210). 30 Parties' response to the Statement of Objections, paragraph 1212. 31 Parties' response to the Statement of Objections, paragraph 1211. 32 Parties' response to the Statement of Objections, paragraphs 1217-1218. 33 Parties' response to the Statement of Objections, paragraphs 1219. 34 Parties' response to the Statement of Objections, paragraphs 1220. 35 Besides being US-focussed, another limitation of the 4(c) and 4(d) documents was that they were prepared specifically in relation to the Transaction, whereas the Commission's request was broader covering the industry, relevant markets, and the Parties' activities. 36 The US 4(c) and 4(d) documents provided to the Commission on 2 and 5 February totalled 122.

11

(83)Based on Section 2.2.1, the Commission considers that the production of documents by the Parties during the pre-notification phase was delayed in several instances which complicated the Commission's review of the Transaction.

2.2.2. Delayed submission of data and methodology used for the computation of market shares provided in the Form CO

(84)Market shares provide a useful first indication of the market structure and the competitive importance of market players. Before relying on market shares, it is essential for the Commission to verify the reliability of the sources and methodology used for their computation.

(85)As from the pre-notification stage and during part of the Phase I investigation, the Commission and the Parties had numerous and extensive discussions regarding the data and the methodology used by the Parties for the computation of market shares for crop protection markets, as well as regarding the various results obtained through this process.

(86)In that context, the Commission issued several requests for information partly or entirely devoted to issues related to market shares (see Section 2.1 of Annex 3). Also other exchanges took place, in particular by emails and in conference calls.

(87)The methodology note attached to the Form CO of 22 June 2016 provided limited information on the Parties' data and methodology used for the computation of 37 markets shares provided in the Form CO (see Annex 3),in particular in light of the importance of market shares in this case with, potentially, hundreds of affected relevant (downstream) markets for crop protection formulated products.

(88)In response to the requests for information sent by the Commission, the Parties provided, on 11 July 2016, details on the data sources used for each market as well as an updated methodology note reflecting the different calculations and assumptions 38made when data was available and when data was not available to them.

(89)On 7 July 2016 and on 14 July 2016, the Parties provided details on the methodology used for the allocation of products across the various segments, for, respectively, 39 40herbicides and insecticides, and fungicides.

(90)On 15 July 2016, the Parties completed their response to question 1 of the 41 Commission's request for information RFI 16,dated 30 June 2016 and due by 1 July 2016, by submitting a spreadsheet that, according to the Parties' additional methodology explanations submitted on 14 July 2016, was necessary to compute market shares for fungicides formulated products.

(91)Other elements have also been provided in relation to the discrepancies noted by the Commission between several versions of the market shares provided by the Parties.

37 Annex 3, Section 2. 38 Parties' response to the Commission's request for information dated 6 July 2016), dated 11 July 2016 at 1:01 and at 17:47 (ID5025 and ID5026). 39 Parties' response to the Commission's request for information dated 6 July 2016, question 5, for herbicides (ID4897) and insecticides (ID4896). 40 Parties' response to the Commission's request for information dated 6 July 2016, question 5, for fungicides (ID5114). 41 Parties' response to the Commission's request for information dated 6 July 2016, question 1 (ID5157).

12

(92)In responding to the Commission's request to "provide an updated methodology note […] explaining in detail the preparation of all global data and indicating sources for 42 their respective data point, year, unit",the Parties indicated that they "do not consider it necessary to provide a long description of this incredibly simple and easy-to-understand methodology".In their response to the Statement of Objections, the Parties indicate that “[t]he methodology used for the calculation is the standard methodology. The Parties and the RBB team provided sufficient information in this regard during pre-notification and in the Form CO [… and] that the Commission ultimately disregarded the global shares that the Parties submitted and requested, in 44RFI21, to provide market shares based on the Agrowin database.”

(93)The Commission considers that [assessment why the Commission had several rounds of questions on market shares and methodology].

(94)The Commission also notes that, as explained in Section 2.6 of Annex 3, the decision to use Agrowin data was taken by the Commission in light of the limited information provided at the time of notification on the data and methodology used for the purpose of calculating markets shares in the Form CO, [details on communication regarding market share methodology].

(95)Overall, the Commission considers that [details on communication regarding market share methodology].

2.2.3. Provision of access to patent data and databases during the pre-notification and formal investigation

(96)Given that one of the central areas of the Commission's investigation was the impact of the Transaction on innovation, access to patent data was of key importance for the Commission to evaluate the innovation strength of the relevant players. Nevertheless, the Commission encountered significant difficulties in collecting patent data from the Parties, despite them using such data in the ordinary course of business.

(97)The Commission notes that the Parties did not mention their internal databases on crop protection patents and did not provide their competitive intelligence reports on competitors' crop protection patents in their responses to several initial Commission's requests for information. Annex 1 on patent analysis describes in detail the Commission's requests for information where information on patents was asked from the Parties.

(98)On 31 August 2016, Dow finally submitted its patent database in response to the 45 Commission's request for information RFI 42 dated 25 August 2016.Also DuPont provided a database on its patents as a response to the Commission's request for 46information RFI 42.

(99)The Commission notes that the Parties had several earlier opportunities to submit their patent databases, as well as the relevant competitive intelligence reports, on

crop protection patents (Commission's requests for information RFI 1, RFI 5, RFI 9, RFI 32). In addition, the Commission's requests for information RFI 13, RFI 17, and RFI 26, while focusing on data used in the report assessing the potential Monsanto/Syngenta transaction, were also opportunities for the Parties to mention their internal databases on crop protection patents and the relevant competitive intelligence reports.

(100)In their response to the Statement of Objections, the Parties argue that they provided patent data when specifically requested by the Commission during the pre- notification phase, and mention for example the submission by DuPont of the patent board notices, requested in question 4 of the Commission's request for information 47 RFI 10. However, the Commission notes that these patent board notices are not raw 48 data, which were available to DuPont and were not mentioned in response to the Commission's requests for information during the pre-notification phase. As regards Dow, the Parties did not provide any comment in the response to the Statement of Objections on the delay to provide access to its patent data.

(101)The Commission also notes that the Parties did not provide at least one competitive intelligence report on crop protection patents. Indeed, this specific report was 49provided to the Commission by the DoJ.

(102)Access to patent databases and patent data has been fundamental for the Commission's assessment of the innovation strengths of crop protection players, as described in detail in Section V.8.7.2.1 and Annex 1 on patent analysis.

2.2.4. Delayed provision of economic data underlying economic analysis during the formal investigation

(103)The Parties submitted several economic analyses aimed at assessing the competitive constraint brought by generic products on their own products as well as the switching of customers between Dow and DuPont.

(104)Most of the time the Parties did not provide the underlying data and methodology together with their economic analyses, as required by DG Competition Best Practices 50 for the submission of economic evidence,but submitted those only as the result of separate Commission's request for information.

(105)By way of example, the Parties submitted an economic annex purporting to show that they are not close competitors in their response to the Article 6(1)(c) Decision on 26 August 2016. The underlying data and code were provided by the Parties only in response to the Commission's request for information RFI 44 on 15 September 2016, 14 working days after the submission of the economic analysis.

47 Parties' response to the Statement of Objections, paragraphs 1230-1231. 48 See Annex 1 on patent analysis. 49 Dow's internal document "Evaluation of potential Japanese partner companies in the crop science space", file name "DOW-ORION-33061629.pdf" (ID7973-17). 50 DG Competition Best Practices for the submission of economic evidence, accessible at http://ec.europa.eu/competition/antitrust/legislation/best practices submission en.pdf. See in particular paragraph 46: "[w]here economic submissions rely on quantitative data the parties should provide the data and codes timely, in an appropriate format and in accordance with the criteria laid down in section 3 of this document. In particular, the absence of all the necessary elements needed for replication and assessment of an economic submission can constitute grounds for not taking it further into consideration".

14

(106)Other instances of the delayed submissions of economic data and methodology, which the Commission had to pro-actively request, are described in Section 2.1 of Annex 2 on generics and in Section 2 of Annex 6 on switching.

(107)The economic data and methodology mentioned in recitals (84) to (102) were essential for the Commission to verify the conclusions reached by the Parties in their economic analyses.

2.2.5. Need to adopt two Article 11(3) Decisions in Phase II to preserve the integrity of the Commission investigation

(108)In light of several instances of delayed submission of the requested information by 51the Parties in pre-notification and Phase I, at the start of the Phase II investigation, the Commission reminded the Parties that during Phase II it was particularly important to provide responses to the Commission's requests for information in a timely and complete manner.

(109)Nevertheless, also in Phase II there were instances where the Parties failed to provide certain documents in full within the (extended) deadline. To preserve the integrity of the investigation, the Commission had to adopt two decisions pursuant to Article 11(3) of the Merger Regulation suspending the merger review time limit (the "Article 11(3) Decisions"), as described in the next two sections.

2.2.5.1. First Article 11(3) Decision of 8 September 2016

(110)The first Article 11(3) Decision was adopted by the Commission due to the Parties' failure to provide, in full or partially, several thousands of internal documents in response to three Commission's requests for information within the deadline.

(111)Those internal documents were requested by the Commission during the Phase II investigation through the Commission's requests for information RFI 35, RFI 36 and RFI 37 dated 18 August 2016. Those requests for information asked inter alia for the Parties' internal documents – on the basis of specified search terms, custodians and time range – regarding gene editing, innovation and post-integration R&D plans, which were all important areas of the Commission's investigation.

(112)Following the Parties' request, the Commission extended the deadline for response to the Commission's request for information RFI 35 from the initial six working days to 11 working days, to the Commission's request for information RFI 36 from the initial seven working days to 11 working days, and to the Commission's request for information RFI 37 from the initial six working days to nine working days. No further extensions of the deadline were asked by the Parties.

(113)Following additional requests by the Parties, the Commission also agreed to several amendments to the relevant requests for information, notably by narrowing the timeframe of the e-mails search (for the Commission's request for information RFI 35), reducing the number of custodians (for the Commission's requests for information RFI 35 and RFI 36) and accepting an approach avoiding the need for manual screening of the documents to determine whether they fall within the scope of the request (for the Commission's request for information RFI 37).

51 For example, at the state-of-play meeting of 31 August 2016 after the Parties' response to the Commission's Article 6(1)(c) Decision.

15

(114)The Commission considers that the extended deadline granted to the Parties of nine working days (for Commission's request for information RFI 37) and 11 working days (Commission's requests for information RFI 35 and RFI 36) was sufficient to produce the requested internal documents. This is based on the estimate of three to four working days required to extract the responsive documents from the custodians. This period could have been even shorter if the documents of some custodians were covered under the Second Request of the DoJ and their documents had already been extracted. The remaining six to seven days would be used for screening the extracted documents for LPP, which is usually first done automatically based on key words or authors and then manually with the assistance of contracted outside lawyers or paralegals. These estimates are based on the Commission's experience in other merger review proceedings which involved extensive production of internal documents.

(115)On 31 August 2016, the Parties provided a partial response to the Commission's request for information RFI 37 but also indicated that they had to review whether other potentially responsive documents are covered by LPP under US law. On 2 September 2016 (the extended deadline for a complete response to the Commission's requests for information RFI 35 and RFI 36), the Parties indicated that they were not in a position to provide a response to the Commission's requests for information RFI 35 and RFI 36, for the same reasons.

(116)The Parties informed the Commission that they might be in a position to submit their responses on 9 September 2016 or at the latest on 12 September 2016. Subsequently the Parties specified to the Commission that they should be able to submit a complete set of all responsive documents to all requests for information on 16 September 2016 at the earliest.

(117)In order not to compromise its investigation, on 8 September 2016 the Commission adopted an Article 11(3) Decision suspending the merger review time limit from 1 September 2016 (the extended deadline for the response to the Commission's request for information RFI 37) until the receipt of the complete and correct response.

(118)On 15 and 16 September 2016, the Parties submitted their internal documents in response to the Commission's requests for information RFI 35, RFI 36, and RFI 37, as well as to the Commission's request for information RFI 43. Together with their submissions, the Parties provided privilege logs listing the documents for which they claimed LPP. For Dow, the number of such LPP documents was over 14 000 and for DuPont – around 3 800. Almost all of those documents were fully withheld, with only a dozen provided in a partially redacted form.

(119)Upon the examination of the privilege logs, it became apparent to the Commission that in producing internal documents the Parties seemed to have applied a deficient methodology which had resulted in an overly broad definition of documents covered by LPP. This was despite the fact that the definition of categories of documents covered by LPP was explicitly set out in the relevant requests for information. The Commission communicated to the Parties its concerns regarding incomplete 52 provision of responsive documents by e-mail on 20 September 2016and in a telephone call on 21 September 2016.

52 E-mail from the Commission to the Parties, 20 September 2016 at 20:26 (ID6881).

16

(120)On 23 and 26 September 2016 the Parties provided the updated submissions of internal documents in response to the Commission's requests for information RFI 35, RFI 36, RFI 37 and RFI 43. The number of documents over which the LPP protection was claimed was reduced to around 6 300 for Dow (around half of them withheld and half partially redacted) and 2 800 for DuPont (all of them withheld, except for one document partially redacted).

(121)After its review of the Parties' submissions, the Commission reverted to the Parties 53 on 28 September 2016requesting them to double-check the LPP coverage for around 50 documents of Dow and 90 documents of DuPont.

53 E-mail from the Commission to the Parties, 28 September 2016 at 17:58 (ID7231).

(122)On 29 September 2016, Dow produced additional nine documents in their entirety and three documents in a partially redacted form. On the same date, DuPont produced 13 additional documents in their entirety and 11 in a partially redacted form.

(123)It would appear therefore that the Parties' initial LPP claims in the context of the provision of internal documents in response to the Commission's requests for information RFI 35, RFI 36, RFI 37 and RFI 43 were overly broad. It required the Commission to intervene and spend considerable time and efforts in order to ensure that thousands of responsive internal documents were not unduly withheld under the pretext of LPP.

(124)The Commission decided to end the suspension of the time limit for merger review imposed by the first Article 11(3) Decision from 26 September 2016 when the majority of internal documents were provided by the Parties, despite the fact that certain additional responsive document were submitted three days later, on 29 September 2016.

2.2.5.2. Second Article 11(3) Decision of 3 November 2016

(125)The second Article 11(3) Decision was adopted by the Commission in light of the failure of the Parties' to submit [type of internal document].

(126)As described in Section V.2.1.1, […] is a tool used within Dow for preparation of pricing of new crop protection products. It constitutes important evidence for assessing competitive interaction between different crop protection products (see Sections V.6.3 to V.6.6 on crop protection).

(127)[Information on internal documents for the analysis of competitive interaction].

(128)[Information on internal documents for the analysis of competitive interaction].

(129)The Commission first became aware that [information on internal documents for the analysis of competitive interaction] only during the Phase II investigation. In particular, the Commission identified [information on internal documents for the 54analysis of competitive interaction].

(130)After discovery of the first [type of internal document], the Commission requested from the Parties other [type of internal documents] by way of simple requests for 55 information.Following the Parties' replies, the Commission still had concerns that the Parties had not provided all [type of internal documents].

53 E-mail from the Commission to the Parties, 28 September 2016 at 17:58 (ID7231). 54 For further details, see the Article 11(3) Decsion of 3 November 2016. 55 Commission's requests for information RFI 49 and RFI 58.

17

(131)To preserve the integrity of its investigation, the Commission considered it necessary to require the Parties, by way of a decision pursuant to Article 11(3) Decsion of […], to supply [type of internal documents], pending which the merger review time period was suspended from 13 October 2016.

(132)The Parties replied to the second Article 11(3) Decision on 7 November 2016, following which the merger review period was re-started. With its response, the Parties submitted 47 […] documents.

2.3. Conclusion on the procedural issues

(133)To sum up, the Commission considers that the Parties' conduct made the Commission's investigation and assessment of the Transaction more difficult. […].

3. FRAMEWORK OF THE ASSESSMENT

(134) The Commission is aware that other concentrations are currently planned in the agrochemical industry, potentially affecting the same markets where the Parties are active.

(135) In particular on 23 September 2016 (that is, after the notification of the Transaction), ChemChina, an undertaking active among others in crop protection, notified to the 56 Commission its proposed acquisition of Syngenta.At the time of the adoption of this Decision, the Commission's review of ChemChina's acquisition of Syngenta is still ongoing. Also, it has been reported in the press that Bayer reached an agreement to acquire Monsanto, both companies being important players in the agrochemical industry; this planned concentration has not yet been notified to the Commission.

(136) Consistent with its previous practice, the Commission assessed the Transaction according to a priority principle ("first come, first served" approach) based on the 57date of notification.

(137) It should be recalled that assessing the competitive effects of a proposed transaction under the Merger Regulation involves a comparison of the competitive conditions that would result from the notified merger with the conditions that would have prevailed in absence of the merger. The competitive conditions existing at the time of notification constitute, as a general rule, the relevant framework for evaluating the 58 effects of a transaction.However, in some circumstances the Commission may take into account future changes to the market that can reasonably be predicted.

(138) The Commission considers from these principles and the general scheme of the Merger Regulation that a party that is the first to notify a transaction should have it assessed on its own merits as to whether it would significantly impede effective

competition in the internal market or in a substantial part thereof. This first to notify a transaction should therefore be entitled to have its operation decided first (for example, declared compatible with the internal market) within the applicable time limits of the Merger Regulation. It is therefore not necessary or appropriate to take into account future changes to the market conditions resulting from subsequently notified transactions that require approval from the Commission.

(139) Therefore, in the circumstances of this Decision, the Transaction, which was notified to the Commission first, should be assessed in the light of the competitive situation that prevailed at the time of its notification, disregarding the potential changes that may be brought by the proposed ChemChina/Syngenta and Bayer/Monsanto transactions.

SECTION V: CROP PROTECTION

1. BACKGROUND ON THE CROP PROTECTION INDUSTRY

1.1. The crop protection industry: size, evolution and forecast

(140) Crop protection products, also known as pesticides, are used in agriculture in order to protect a crop from other biological organisms (pests) that can negatively affect the crop development either by attacking it or by competing with it for resources. Losses due to pests are estimated to range from 10 to 90%, with an average of 35 to 40%, 59for all potential food and fibre crops.

(141) Crop protection products therefore improve yields and ensure the availability, quality, reliability and price crops to the benefit of farmers and consumers. They constitute an important input for farming. Their use saw a strong increase since the 1940s and the growth of synthetic pesticides.

(142) In 2014, the size of the crop protection and seeds industry taken together was estimated at USD 109 billion globally. By 2025, the size of those industries is 60 expected to grow to USD 150 billion.Growth is expected to be driven by 61 population growth and the rise of the middle class.As the area harvested is not expected to grow, technology will be key, and food safety and security is a critical priority.

(143) Crop protection products can be categorised into herbicides (to fight weeds), insecticides (insects), fungicides (diseases) and others. Herbicides have been the largest segment within crop protection in the last years, while genetically modified ("GM") seeds have been a key development in agriculture globally, as illustrated in Figure 2.

59 International Union of Pure and Applied Chemistry (IUPAC) http://agrochemicals.iupac.org/index.php?option=com sobi2&sobi2Task=sobi2Details&catid=3&sobi2 Id=31.

60 Dow's internal document "Dow Agrosciences – Options, dated July 2015", file name "DAS- 10171052.pdf" (ID6748-19233) slide 3.

61 See for instance DuPont's public document for the Bank of America Merrill Lynch – Global Agriculture and Chemicals Conference, "Delivering sustainable long-term growth in agriculture", 3 March 2016, slide 4.

19

Figure 2 – Evolution of sales of crop protection products and GM seeds 1990-2012

[…]

Source: DuPont's internal document "Herbicide Discovery Targets", file name "Dupont Herbicide DiscoveryTargets Master Feb 14 2014 v2 (2)_CONFIDENTIAL.pdf" (ID1184), slide 4

(144) The global market for conventional crop protection products (excluding sales of herbicide tolerant and insect resistant seeds, as well as non-crop agrochemicals) was 62worth about USD 51 billion in 2015.

(145) The global crop protection market is expected to continue growing at around 2.7% per year for 2015-2020 to exceed USD 58 billion in 2020, largely due to expected 63strong growth in notably Brazil, China, ASEAN, etc.

(146) Globally, cereals (mainly wheat and barley) are the largest crop by cultivated area, 64 followed by maize/corn, rice and soybean. Correspondingly, these crops are major crops in terms of crop protection product sales (Figure 3). Fruits and vegetables also amount for a significant share of crop protection products sales.

Figure 3 – The global crop protection industry

[…]

Source: DuPont's internal document "DuPont Crop Protection Board of Directors Business Review", file name "DOC-000000016.pdf" (ID5273-16)

(147) In the EEA, the market for crop protection products amounts to about 65 USD 10 billion, that is to say 18% of the global sales value. In 2015, the leading markets for crop protection products in the EEA were France (USD 2.5 billion), Germany (USD 2.0 billion), Italy (USD 1.2 billion), Spain (USD 1.0 billion) and the UK (USD 0.8 billion).

(148) The main crops produced in Europe are cereals (35% of total production), maize, fruits and vegetables, vine, oilseed rape and potatoes.

(149) Use of the various categories of crop protection products across Member States varies widely. A study of the European Parliament highlights that "[h]erbicides are the plant protection agents consumed in largest amounts, especially in northern Member States. Fungicides are the second most consumed. Mediterranean countries show the largest consumption share of insecticides and the lowest share of 66 herbicides over the total plant protection agents consumed."The study also notes that crop protection products represented on average between 3 and 15% of farmers' costs over the 2004-2012 time period.

(150) Forecasts from industry experts estimate that the European crop protection market 67(including Russia and Ukraine) will grow 2.6% per year in the period 2015-2020.

62 Phillips McDougall, "Industry Overview – 2015 Market", May 2016.

63 Phillips McDougall, "Industry Overview – 2015 Market", May 2016.

64 Phillips McDougall, "Industry Overview – 2015 Market", May 2016.

65 Phillips McDougall, "Industry Overview – 2014 Market", May 2015.

66 European Parliament study, "Overview of the agricultural inputs sector in the EU", July 2015.

67 Phillips McDougall, "Industry Overview – 2015 Market", May 2016.

20

1.2. Crop protection products: definitions and basic concepts

(151) Depending on the type of organism they target, crop protection products are mainly 68 categorised into herbicides, insecticides, and fungicides. There are also other crop protection products such as plant growth regulators. These products are applied on crops, plants or fields. Some crop protection products can also be applied on seeds to 69protect them from insects and diseases (seed treatments).

(152) The key components of crop protection products are active substances, or active ingredients (hereafter also designated by "AIs"), which produce the desired biological effect (that is, killing the pest or making it inoffensive).

(153) A crop protection AI can typically be classified according to five aspects: the plant(s) to be protected (some AIs are used across several crops), the pest(s) against which it acts (some AIs are used against several pests), the mode of action ("MoA"), the chemical class, and the molecule.

(154) The main players in the agrochemical industry work together in the Resistance Action Committees. For herbicides, the relevant committee is the Herbicide Resistance Action Committee (“HRAC”); for insecticides, the Insecticide Resistance Action Committee (“IRAC”); for fungicides, the Fungicide Resistance Action Committee (“FRAC”).These bodies notably classify crop protection products by AI, chemical family and MoA. They also work on resistance issues.

(155) The MoA describes how a particular molecule acts on a plant. For herbicides, MoAs are classified by HRAC codes (an alphabetical classification, according to the herbicides’ target sites/sites of actions), and by the Weed Science Society of America (WSSA; a numerical system). According to the Resistance Action Committees, there are currently 25 MoAs for herbicides, 29 for insecticides and 11 for fungicides.

(156) The chemical family or chemical class are groupings of molecules. They typically have the same MoA. In some cases different chemical classes can have the same MoA.

(157) Resistance is a consequence of a basic evolutionary selection process in response to the repeated use of the same product(s) with the same MoA, and is also associated with other crop management decisions (for instance, monoculture).

(158) AIs are many times mixed with other AIs with the aim to ensure potency, efficiency, to widen the spectrum of pests targeted, to increase convenience for farmers and to help resistance management.

(159) A formulated product as sold to distributors and farmers is composed of AI(s) mixed with inactive ingredients such as solvents, fillers, and adjuvants. These ingredients aim at making the AIs more stable, effective, or safer or easier to apply. There are various types of final formulations, such as granules, emulsifiable concentrates, etc.

68 Some mixtures exist joining for instance a fungicide and an insecticide, but they are typically not widespread in Europe. 69 In seed treatments, the Parties have limited activities, although DuPont has plans to grow. See for instance DuPont's internal document "DuPont Crop Protection 2016-2018 Business Strategy Review June 24, 2015", file name "DUPONT-VS-00001489.pdf" (ID8008). 70 See the websites of these organisations: http://www.hracglobal.com/, http://www.irac-online.org/, http://www.frac.info/.

21

1.3. Overview of the lifecycle of a crop protection product

(160) The lifecycle of a new AI starts with an R&D company's discovery and development, which requires significant time and financial resources. The average overall costs for the discovery and development of a new agrochemical product brought to the market by an R&D company is estimated by Phillips McDougall at USD 286 million.

(161) Industry sources and market participants highlight that both the cost and lead times have been increasing over the years, allegedly due to a more challenging regulatory environment. The lead time of a global crop protection product discovery and 71development is approximately 11 years.

(162) The discovery process includes the synthesis of candidate molecules, which are then screened to determine the biological activity of the molecule. Molecules are moved forward in the discovery process based notably on their efficacy, toxicological and environmental properties, and fit with company targets. Some products do not succeed and are abandoned.

(163) Patenting of AIs takes place during the discovery stage. Patent strategy is an important consideration for R&D companies, and sometimes differs between companies in terms of both the timing of patent application and the scope of patenting. Companies can apply for and hold patents not only for new AIs, but also for new formulations, new mixtures of AIs, new process technologies (for instance new manufacturing processes that allow for example for higher yields or less waste), or new uses of AIs.

(164) This is also the first moment of some visibility on new products within the industry as competitors screen patents for leads and for defensive measures.

(165) After discovery, an AI enters the development stage.

(166) Regulatory studies are a large part of the development efforts. Field trials represent the largest single cost in the development cycle with approximately [30-40]% of the 72total spend on product development.

(167) After registration is obtained, a company would typically start with some initial preparations/mixtures, to start the roll-out of the AI with later new mixtures, new uses, new countries.

(168) A DuPont internal document estimates the total costs of discovery and development of a compound at USD [150-300] million, of which [30-40]% is spent for discovery, 73 and [60-70]% in development.Discovery takes 3-4 years, and development 5-6 years. The decision to move a molecule to development is thus key.

(169) The lifecycle of an AI continues after registration and commercialisation. The maximisation of revenues of an AI takes various forms, which include developing

71 Phillips McDougall, “The cost of new agrochemical product discovery, Development and Registration in 1995, 2000, 2005-2008, 2010-2014. R&D expenditure in 2014 and expectations for 2019,” a consultancy study for CropLife International, CropLife America and the European Crop Protection Association, March 2016.

72 See DuPont's presentation entitled "DuPont Pioneer & Crop Protection R&D and Commercialization", dated 27 April 2016, slide 10 (ID399).

73 See DuPont's internal document "DuPont Crop Protection Board of Directors Business Review", file name "DOC-000000016.pdf" (ID5273-16).

22

uses for the product (for instance for additional crops or additional effects on the same crops), and creating mixtures with different AIs or other components.

(170) Even after the expiry of patents, R&D companies often succeed to prolong the lifecycle of an AI. They pursue commercial, registration and patenting strategies to avoid commoditisation and generic competition, for instance by offering the AI in mixtures, sometimes with other patented AIs.

(171) The lifecycle for AIs can reach more than 30 years. R&D-based companies carefully manage their portfolio to ensure they have products at different stages of the lifecycle.

Figure 4 – Lifecycle overview

[…]

Source: DuPont's presentation to DG Competition "Product Commercialization in the EEA" (ID1179), 15 June 2016

(172) Depending on the strategic role of the AI in the portfolio of the company, an AI for which related products only bring a limited margin might be divested by an R&D company to a generic company as part of its portfolio management and to maintain a level of margin as expected from shareholders. The R&D company sometimes gets distribution rights back, for instance as Dow did for Oxyfluorfen (divested to Nutrichem in 2015).

1.4. Active ingredients: from discovery to commercialisation

1.4.1. Discovery

(173) The R&D process in crop protection typically starts with the screening of a large number of molecules. Large R&D companies typically screen each year thousands of molecules (for instance more than 60 000 per year for DuPont, and around 160 000 in 2010-2014 for Dow), using high throughput screening methods. These molecules are synthesised from basic chemical entities that can be bought or traded from/with pharmaceutical companies, from academic institutions, etc.

(174) Candidate molecules are subject to a series of biological tests to determine their biological activity. If the molecule has a suitable biological activity, it is a "hit". It must also match other criteria: "it must be able to be patented, possess good toxicological and environmental properties and display good commercial 74 prospects".Therefore the R&D discovery process includes initial toxicological and environmental chemistry testing, as well as an evaluation of commercial prospects.

(175) Dow and DuPont both divide the discovery process into several stages, where some hurdles must be passed each time before moving an AI to the next discovery stage. In light of the resources needed, only a selection of molecules moves forward each year.

(176) Dow and DuPont both have discovery goals or concepts, that is to say that they aim at finding molecules with a certain type of biological activity and safety profile to match discovery concepts designed on the basis of the commercial alternatives of the targeted market and its commercial fit with the existing portfolio.

74 Phillips McDougall, “The cost of new agrochemical product discovery, Development and Registration in 1995, 2000, 2005-2008, 2010-2014. R&D expenditure in 2014 and expectations for 2019,” a consultancy study for CropLife International, CropLife America and the European Crop Protection Association, March 2016.

23

1.4.2. Development

(177) Product development includes further safety testing, testing of formulations using the new AI, biological development to investigate the activity of the AI against different pests and in different environment situations and further assessment of the AI's regulatory profile.

(178) Selecting candidates for the development stage is a significant step in light of the costs involved in this stage. When moving to the development stage, other considerations are taken into account such as the cost of manufacturing and the competitive landscape, notably taking into account patents and the type of molecules on which other crop protection R&D companies are working. "Studies on optimising the manufacturing process for commercial production are subsequently undertaken with the aim of arriving at a suitably cost effective manufacturing process. Another important area of chemistry development is formulation evaluation. This generally involves the testing and optimisation of a variety of formulations of the new crop protection product to ensure that the product can be delivered in a safe and effective 75manner for subsequent field use".

1.4.3. Approval and authorisation

(179) Crop protection products have properties that can endanger human health and the environment since their active components are essentially toxins aimed at harming animals, vegetation, or fungi. In the EEA, the approval and commercialisation of AIs and formulated crop protection products is mainly governed by Regulation (EC) 76No 1107/2009 ("Regulation 1107").

(180) Regulation 1107 sets the rules for (i) the approval of the AI contained in formulated products (referred to as plant protection products in that regulation) and (ii) the authorisation of plant protection products at Zonal and Member State level.

(181) Crop protection products require two types of authorisation before they can be sold to farmers in the EEA. First, the AI must have been authorised by the Commission, upon the advice of European Food Safety Authority ("EFSA"). Second, the crop protection product incorporating the AI (alone or as part of a mixture of AIs) must have been authorised by the competent authority in the Member State(s) where it will be sold. The following recitals describe these two authorisations in turn.

(182) The objective pursued through a Union approval for AIs is to ensure a high level of protection of both human and animal health and the environment. Member States assessing whether a formulated product can be introduced into the market focus on the safe and sustainable use of crop protection products for the production of food, including solutions for resistance management and minor uses (so-called "orphan crops").

(183) An application for the approval of an AI must be submitted by the producer of the AI to a Member State (the Rapporteur Member State), together with a summary and a complete dossier including the information and documents listed in Article 8 of Regulation 1107. The requirements and conditions for approval of an AI are listed in Article 4 of Regulation 1107. The Rapporteur Member State must prepare a draft assessment report relating to the approval of the AI which is sent to EFSA. The EFSA then launches a consultation over the draft assessment and its findings are presented in its conclusions. The Commission then adopts a regulation approving or not approving, as the case may be, the AI.

(184) AIs must be proven to be safe for people’s and animals' health and the environment. This requires, inter alia, an assessment of AI residues in food and feed and of the effects on human and animal health and the environment. Under Regulation 1107, an AI cannot be in principle approved if it is classified as mutagenic, carcinogenic or an endocrine disruptor.

(185) The approval of an AI is valid for the entire Union and, once approved the AI is included on a registry of approved AI. A list of AIs approved in the Union is available online. There are currently more than 400 approved AI for use in the Union.

(186) It takes approximately 2.5 to 3.5 years from the admissibility of an application to the approval of a new AI. However, applicants must start studies for the preparation of the AI's approval dossier several years before they can lodge an application.

(187) The Commission’s first approval period for a new AI does not exceed ten years and may be subject to conditions and restrictions.

(188) Following expiration of the initial authorisation period, the authorisation may be renewed for successive periods of up to 15 years, if the application submitted by a producer of the AI establishes that the approval criteria provided in Article 4 of Regulation 1107 are still satisfied.

(189) The cost of the approval process is approximating USD 33 million and that for renewal is approximately USD 5-8 million.

(190) Once an AI is approved, it can be considered for authorisation and use in formulated products in the Union. Member States take decisions at a national level, within a zonal authorisation process, on whether individual formulated products with specific use recommendations can be authorised. Member States must ensure authorised formulated products comply with harmonised European standards set in Regulation 1107 and reflect any conditions or restrictions for the AI, where appropriate, to the formulated product. In particular, dossiers supporting formulated products must be evaluated in accordance with the Uniform Principles, which are set 77 out in Commission Regulation (EU) No 546/2011 and contain all the documents and information listed at Article 33(2) of Regulation 1107.

(191) Once the examining Member State (zonal Rapporteur Member State) issues an authorisation for a formulated product, the applicant wishing to sell the same formulated product in another Member State(s), within the same "zone" must submit an application for the product to the relevant Member State(s) of such a zone which then would examine and, if appropriate, approve the request.

77 Commission Regulation (EU) No 546/2011 of 10 June 2011 implementing Regulation (EC) No 1107/2009 of the European Parliament and of the Council as regards uniform principles for evaluation and authorisation of plant protection products, OJ L 155, 11.6.2011, p. 127.

25

(192) Regulation 1107 divides the Member States into three zones. For each of the zones, a zonal Rapporteur Member State assesses the application for the registration of a product for the entire zone. Other Member States in the zone shall in principle accept the assessment of the zonal Rapporteur Member State, unless specific environmental or agricultural circumstances require them to conduct their own specific assessment in addition to the zonal Rapporteur Member State’s original assessment. Although intended to facilitate entry from one Member State within a zone to another, the Commission conducted an audit in early 2016, which concluded that the zonal procedure has actually increased barriers to entry by adding an extra layer of dossier review. Indeed, Member States normally conduct their own assessment of formulated products.

(193) A formulated product authorised in one Member State (Member State of origin) may, subject to obtaining a parallel trade permit, be introduced in another Member State (Member State of introduction), if this Member State determines that the formulated product is identical in composition to another formulated product already authorised in its territory (reference product).

(194) The duration of a formulated product authorisation granted in accordance with Regulation 1107 is laid down in the authorisation but shall not exceed one year from the date of expiry of the approval of the AI contained in the formulated product. The holder of the formulated product authorisation can apply for renewal, provided the requirements for the initial authorisation are still met.

(195) Some AIs are classified as "candidates for substitution" in light of their negative properties. This entails that use of such AIs is discouraged since Member States, when examining applications for authorisations of formulated products containing them, will need to consider whether alternative chemical solutions which require less risk mitigation or non-chemical control or prevention methods exist for the pest problems these AIs address and, if they do exist, reject the authorisation of formulated products containing AIs that are candidates for substitution.

(196) Member States can also, under specific circumstances – in essence, the absence of alternative similarly effective solutions –, grant Emergency Use Permits ("EUPs"), which are temporary authorisations allowing use of a formulated product containing an AI which has not been approved at the EEA level.

(197) Overall, it takes approximately 11 years for R&D companies to bring new AIs to market, representing an average cost of USD 286 million, and approximately seven years to obtain the renewal of an AI and subsequent re-registration of a formulated product, for an average cost of USD 10 million.

1.4.4. Patents and other forms of protection

(198) The protection of new AIs and formulated products is primarily ensured by patents.

(199) At the early stages of the research process, agrochemical companies commonly seek 78 patents directed at a chemical genus that includes a new AI. At later stages of the R&D process, a supplementary patent for a composition of matter can be made to protect the identified chemical sub-genus. Patents can also be obtained for some processes ("process patents") or formulations.

78 A taxonomic rank used in the biological classification of living and fossil organisms.

26

(200) In the EEA, a patent offers protection for a period of twenty years from the date of application, irrespective of whether it is exclusively national or based on a European patent. In the EEA there are also supplementary protection certificates. Regulation (EC) No 1610/96 of the European Parliament and of the Council of 23 July 1996 concerning the creation of a supplementary protection certificate for plant protection 79 products ("Regulation 1610/96")allows R&D companies to apply for a supplementary protection certificate before the expiration of the patent. This certificate normally has a validity of five years and grants the same rights as the basic patent to the R&D company. Since it takes on average 10 years from the date of patent application to bring a compound to the market, the holder of a patent followed by a certificate can enjoy up to approximately 15 years of exclusivity from the time its formulated product first obtains authorisation to be placed on the market in the Union.

(201) Data protection is the temporary right of the owner of a test or study report to prevent it from being used for the benefit of another applicant for the purpose of regulatory approval or renewal. When submitting the application for the approval of an AI, the applicant must also hand in a list of any claims for data protection on test and study reports related to the AI.

(202) The protection applies to the data generated through test and studies concerning the AI, safener or synergist, adjuvants and the plant protection product when they are submitted to a Member State by an applicant for authorisation.

(203) The period of data protection is in principle 10 years starting at the date of first authorisation in the Member State, and can be extended to 13 years for specific plant protection products. Renewal of authorisation brings a new period of data protection. Once the period of data protection expires, Member States may rely on existing data for the benefit of other applicants for authorisation, including those of generic players.

1.4.5. Means of commercialisation of AIs other than own products commercialisation

(204) In light of the costs, it is not rare for firms to partner early on in the discovery/development stage to co-develop an AI.

(205) More commonly, commercialisation of an AI can be done via other crop protection companies with the licensing of an AI or the supply of a produced AI ("technical sales"). Reciprocal arrangements can also be reached to fill gaps in the companies' respective portfolios. Typically, due to cost of manufacturing and economies of scale, the company that licenses out/supplies is the one that continues to manufacture the product.

1.4.6. Formulated products

(206) As indicated in recital (190), registration of new final products/new mixtures takes place at the national level. Efforts to commercialise also take place in national markets.

(207) Typically, a R&D company will be in contact with institutes advising farmers that will conduct tests on the new products two to three years before launch. Companies also start communicating to growers for instance one year before launch, for instance

79 OJ L 198, 8.8.1996, p. 30.

27

with field presentations. Advertising spending and branding is significant. Distributors, and in some cases retailers, are also important players for commercialisation. Depending on the Member States, they can play an important role in advising farmers. Distribution contracts between producers of formulated crop protection products and distributors are typically not exclusive. Distributors distribute products from several producers of formulated products (both R&D companies and generic companies). However they can have limited space shelf and make a choice in their offering. Distribution channels will be discussed further in Section V.6.2.2.

(208) R&D continues while formulated products are commercialised, to develop new formulations, mixtures, for instance with different adjuvants, mixture ratios, etc.

(209) When expiry of the approval is near, investing in renewal of authorisation is a business decision, taking into account the costs, whether they can be shared for instance in a taskforce, and the expected revenues.

(210) Generic companies enter if the market opportunity is significant. They often approach the data owner, as generating own data takes time. Development time for a generic is shorter, two to three years. Post patent strategies of crop protection companies R&D companies will be detailed in Section V.6.2.1.

1.5. Overview of main companies active in crop protection

1.5.1. Main agrochemical companies

(211) While some companies offer only crop protection products or only seeds, there is a trend to offer integrated solutions. This trend is largely linked to the growth of GM offers in the past years, mainly outside of the Union.

(212) The top players globally in seeds and crop protection are set out in Figure 5 and Figure 6.

Figure 5 – 2014 Global Top 7 suppliers of Agro Products

Source: Parties' initial Briefing Paper (based on Phillips McDougall, Industry Overview – 2014 Market)

28

Figure 6 – Competitive landscape from Dow's view

[…]

Source: Dow's internal document "Monsanto + Syngenta – Syngenta Seeds Discussion", file name "70. DAS- 00000070-000001.pdf" (ID1327-164), slide 7

(213) Monsanto (US) is the global agriculture and seed leader due to its genetically modified crops resistant to its own best-selling herbicide glyphosate (Round-Up). Its presence in other markets – non-GM for seeds and other chemicals for crop protection – is limited, which explains its relatively small size and relevance in the EEA.

(214) Syngenta (Switzerland) is an integrated crop protection (global number 1) and seed (global number 3) company, with a very large portfolio of products.

(215) Bayer (Germany) is the global number 2 crop protection company with a large portfolio of products. It also has a relatively smaller seed business.

(216) BASF (Germany) is a pure crop protection player, currently the global number 3. It has a large portfolio, with a focus on fungicides.

(217) DuPont/Pioneer is the global number 2 seed player, combined with the global number 6 crop protection business.

(218) Dow AgroSciences ("DAS") combines a relatively small seed business with the global number 4 crop protection business, currently focused on herbicides and insecticides.

(219) Adama, controlled by Chinese state-owned enterprise ChemChina, is the largest "generic" crop protection player, with a large portfolio. It is not active in the discovery of new AIs, but it differentiates itself from pure generic players by developing novel mixtures of old AIs.

(220) Other global players include notably FMC, Nufarm and Sumitomo.

29

Figure 7 – Company ranking by sales and R&D expenditure

Source: Phillips McDougall

(221) As for crop protection specifically, there are a number of players active in the sale of crop protection products, with different size and business models. The key players globally are set out in Figure 8.

Figure 8 – Crop protection key players

[…]

Source: DuPont's internal document "DuPont Crop Protection Board of Directors Business Review", file name "DOC-000000016.pdf" (ID5273-16)

1.5.2. The Big 5 R&D-integrated players

(222) Five companies are large integrated companies characterised by their scale and their activity at all stages of the value chain (namely discovery, development, mixture/formulation and commercialisation) through large R&D budgets and operations for crop protection. These are Syngenta, Bayer, BASF, Dow and DuPont (thereafter also designated by "Big 5" or "global R&D-integrated players").

(223) Sales by these five R&D-integrated companies account for around 70% of the total EEA crop protection market and 78% of the global crop protection market.

30

(224) These companies are seen as the top players in crop protection, as illustrated by 80documents of the Parties and by other companies.

1.5.3. Other companies active in crop protection R&D

(225) A number of other companies are active in the discovery and development of new AIs. Those firms are discussed in this section and, including the Big 5, are collectively referred to as "R&D players".

(226) Monsanto has significant crop protection revenues, which are however mostly generated by one relatively old blockbuster product (glyphosate Round-Up), a non-selective herbicide, while its research expenditure is almost exclusively focused on seeds.

(227) FMC is a player which used to be active in the discovery and development of new AIs. However it has exited the active R&D discovery approximately ten years ago. Since then FMC launched only few products developed by other companies. Those two companies belong, to some extent, to a group of companies referred to, along the Big 5, as "R&D-integrated players".

(228) A number of other small or mid-sized companies, mostly from Japan, have R&D operations, but do not compete on the same scale as global R&D-integrated companies through global registration and route to market capabilities. The Japanese companies tend to focus on the Japanese market, including relevant products for rice and fruits and vegetables, insecticides and fungicides. They typically do not have a large scale presence in Europe but sometimes cooperate with the Big 5, as well as with generic companies including via some shareholding links.

(229) Isagro, from Italy, has some discovery activity, however mainly using "co-development as a strategy" in light of the "large scale and significant financial resources […] required to engage in discovery and then development, which takes 81about 10-12 years during which no revenues are generated with the product".

80 See for instance DuPont's internal document "DuPont Crop Protection 2013 Strategy Review", file name "M.7932 - Annex DuPont RFI 6 4.4- DCP 2013 Strategy Review 6.20.2013_CONF....pdf" (ID982), 27 June 2013, page 18 "Key Crop Protection Competitors: BASF […]; Syngenta[…]; Bayer CropScience […]; Dow AgroSciences"; Agreed non-confidential minutes of a call with a competitor, 2 June 2016 (ID5800). 81 Agreed non-confidential minutes of a call with a competitor, 6 September 2016 (ID7111).

31

Figure 9 – Companies involved in new crop protection AI research in 2012

Source: Phillips McDougall, "R&D trends for chemical crop protection products and the position of the European Market, A consultancy study undertaken for ECPA", September 2013

1.5.4. "Generic" players

(230) There are a number of generic players for crop protection products with sizeable revenues, led by Adama. None of these, however, has substantially moved to or remained active in the active discovery of AIs.Their activities are largely dependent on access to AIs originally developed by R&D players.

(231) The "generic" players differ to some extent in business model and activities. As noted by a customer, "[t]he different suppliers belong to three categories: (i) R&D players (e.g. Bayer/BASF), who are involved in developing innovative new products/molecules through research; (ii) Generics producers (eg Albaugh, Globachem) who can be very good at bringing generics to market very quickly once the original product goes off-patent, sometimes selling it in different formulations or via mixes as a new co-formulation under their own brands; and (iii) Generic+R&D producers (e.g. Adama, FMC), who are trying to create innovations by mixing older 83(‘off-patent’) ingredients".

(232) As regards downstream sales of crop protection products, it seems that the generics' share has been overall stable in recent years. A DuPont document notes that "[t]he various mergers and product divestments from the majors have enabled the Big 4 generics to grow into larger competitors. However, generic companies as a group have not grown over the last 5 years and are struggling with low profitability".

82 For instance FMC stopped discovery in 2005. Agreed non-confidential minutes of a call with FMC, 8 April 2016 (ID8179). 83 Agreed non-confidential minutes of a call with a crop consultant, 17 March 2016 (ID8246). 84 DuPont's internal document "DuPont Crop Protection 2013 Strategy Review", file name "M.7932 - Annex DuPont RFI 6 4.4- DCP 2013 Strategy Review 6.20.2013_CONF....pdf" (ID982), page 19.

32

1.5.5. The agrochemical industry is characterised by a concentrated shareholders structure, with significant common shareholding

(233) The agrochemical industry is characterised by a significant level of common shareholding, both in terms of the number of shareholders common to several competitors as well as in the level of shares these common shareholders possess 85across the industry.

(234) Using data on "shareholders [of BASF, Bayer, Dow, DuPont, Monsanto and Syngenta] who file Form 13F (report filed by institutional investment managers)", Table 1 lists all shareholders of these firms which hold a total portfolio value in all these companies of more than EUR 1 000 million.

Table 1 – Reported equity holders with shares in any of BASF, Bayer, Dow, DuPont, Monsanto or Syngenta, with a total portfolio value in all these companies of EUR 1 000 million or more

BASF Bayer Dow DuPont Monsanto Syngenta

BlackRock, Inc. (NYSE:BLK)

6.24% (1) 6.37% (1) 5.92% (3) 6.18% (2) 5.67% (3) 5.09% (1)

Capital Research and Management Company 0.85% (9) 3.04% (3) 5.51% (4) 10.28% (1) 8.55% (1) 3.00% (2)

The Vanguard Group, Inc.

2.03% (4) 2.02% (5) 6.28% (2) 6.04% (3) 6.70% (2) 1.97% (5)

State Street Global Advisors, Inc.

1.95% (5) 1.70% (7) 3.92% (5) 4.44% (4) 4.22% (4) 0.45% (19)

Norges Bank Investment Management

2.95% (2) 1.65% (8) 0.47% (27) 0.62% (21) 0.81% (20) 1.77% (7)

Fidelity Investments

0.28% (23) 0.92% (12) 0.91% (12) 2.17% (7) 3.08% (7) 2.02% (4)

Societe Generale Group, Banking Investments - 4.65% (2) -

-

-

-

Berkshire Hathaway Inc. (NYSE:BRK.A)

-

- 6.47% (1) -

-

-

BNY Mellon Asset Management

0.05% (73) 0.48% (23) 1.36% (7) 1.14% (11) 1.38% (12) 2.93% (3)

UBS Asset Management

1.78% (7) 0.65% (16) 0.72% (16) 0.51% (23) 0.58% (27) 1.89% (6)

Lyxor International Asset Management S.A. 1.87% (6) 1.97% (6) -

-

- 0.08% (66)

Sun Life Assurance Company of Canada,

- 2.98% (4) -

-

-

-

Asset Management Arm

Credit Suisse, Investment Banking and

2.26% (3) 0.31% (34) 0.51% (23) 0.11% (83) 0.07% (139) 0.00% (338)

Securities Investments

Deutsche Asset & Wealth Management

1.14% (8) 1.07% (10) 0.36% (38) 0.08% (97) 0.27% (53) 0.31% (29)

T. Rowe Price Group, Inc.

0.12% (38) 0.43% (27) 0.23% (62) 2.19% (6) 0.50% (29) 0.52% (17)

(NasdaqGS:TROW)

Northern Trust Global Investments

0.05% (74) - 1.19% (9) 1.30% (9) 1.42% (11) 0.29% (32)

Franklin Resources, Inc. (NYSE:BEN)

0.64% (11) 0.35% (30) 1.13% (10) 1.13% (12) - 0.00% (351)

Massachusetts Financial Services Company

-

-

- 0.91% (15) 3.28% (5) 0.58% (16)

Teachers Insurance and Annuity Association

of America - College Retirement Equities 0.22% (25) 0.27% (35) 1.07% (11) 0.47% (26) 1.00% (16) 0.37% (26)

Fund

Glenview Capital Management, LLC

-

- 0.83% (13) - 3.22% (6) -

J.P. Morgan Asset Management, Inc.

0.16% (32) 0.58% (20) 0.62% (19) 1.03% (14) 0.05% (168) 0.06% (86)

Geode Capital Management, LLC

0.14% (33) 0.14% (46) 0.81% (14) 0.81% (17) 0.81% (21) 0.15% (46)

85 A more detailed analysis is provided in Annex 5. This annex provides further factual evidence on the significant level of common shareholding in the agrochemical industry and on the involvement of large minority shareholders which, despite some being labelled "passive investors" are in fact "active owners". The annex also reviews the economic literature, both theoretical and empirical, which provides guidance on the effects of common shareholding on competition between firms in industries subject to such feature, and it discusses how the HHI can be modified to account for common shareholding in the industry.

86 Parties' response to the Commission's requests for information RFI 2, question 4 (ID753 and ID1159-257).

33

BASF Bayer Dow DuPont Monsanto Syngenta

Amundi Asset Management

0.56% (14) 1.13% (9) 0.08% (127) 0.04% (148) 0.06% (156) 0.06% (82)

(ENXTPA:AMUN)

Deka Investment GmbH

0.80% (10) 0.90% (13) 0.02% (276) 0.01% (302) 0.03% (199) 0.12% (52)

Trian Fund Management, L.P.

-

-

- 2.32% (5) -

-

Third Point LLC

-

- 2.25% (6) -

-

-

Artisan Partners Limited Partnership

- 0.89% (14) -

-

- 1.04% (10)

Morgan Stanley, Investment Banking and

0.41% (17) - 0.57% (20) 0.58% (22) 0.58% (26) 0.03% (130)

Brokerage Investments

PRIMECAP Management Company

-

- 0.14% (87) 0.05% (125) 2.66% (9) -

Columbia Management Investment Advisers,

0.09% (50) 0.33% (31) 0.37% (36) 0.36% (31) 0.76% (22) 0.20% (41)

LLC

Total

24.59% 32.84% 41.71% 42.78% 45.70% 22.92%

Source: Commission's analysis of Parties' response to Commission's requests for information RFI 2 Note: Shareholders are ranked by decreased order of their portfolio value in all six companies The ranking of each shareholder for each supplier is indicated between brackets.

(235) The 30 shareholders reported account, in total for around 22%-33% of BASF, Bayer and Syngenta, and around 41%-46% of Dow, DuPont and Monsanto. Most of these shareholders hold stocks in more than one those firms, and 17 hold stocks in all six.

(236) Overall, each of BASF, Bayer, Dow, DuPont, Monsanto and Syngenta has a concentrated shareholder structure and a significant level of common shareholders with its competitors.

1.6. Industry trends

1.6.1. Consolidation in the crop protection industry

(237) Historically, the R&D part of the crop protection industry has gone through different waves of consolidation, as illustrated in the non-exhaustive Figure 10. In parallel, the seeds industry has seen consolidation as well.

Figure 10 – Consolidation in the crop protection industry between 1960 and 2008

Source: Agranova

34

(238) Between 1995 and 2012, industry reports estimate that the number of crop protection 87 R&D companies dropped by half.The five leading crop protection players globally have been involved in this consolidation, for instance with the creation of Syngenta 88and Bayer's acquisition of Aventis.

(239) Consolidation also took place in the generic part of the industry in recent years. In 2014 and 2015 Platform Specialty Products Corporation acquired Chemtura AgroSolutions, the Agriphar Group and Arysta LifeScience Limited to form a 89 90vertically integrated agrochemical company.In 2014 FMC acquired Cheminova. Finally, in 2011, Sumitomo Chemical Corp increased its shareholding in Nufarm 91Limited to 21.7%.

(240) Further consolidation is happening in the market currently, with notably the Transaction, the publicly announced offer by ChemChina to acquire Syngenta 92 (notified as Case M.7962 – ChemChina/Syngenta), and the announced offer by Bayer to acquire Monsanto. These will not be considered in the current Decision given the priority rule described in Section IV.3.

1.6.2. Disappearance of AIs, increase of R&D costs and less focus in Europe

(241) In recent years a number of older AIs disappeared from the market due to stricter regulatory requirements as regards toxicology.

(242) These stricter requirements also contributed to a sharply rise of R&D costs for crop protection. According to a Phillips McDougall's report the costs for discovery and development of a new AI has been increasing in the last years, from a total of USD 152 million in 1995 to USD 286 million in 2010-2014.

(243) However, while regulatory costs have increased, expenditure on R&D as a percentage of revenues has decreased over the last twenty years. As a result, innovation output has decreased in the crop protection industry. According to a 93 Phillips McDougall industry report,the average number of AIs introduced per year was 12.3 between 1980-1989, 12.7 between 1990-1999, 10.3 between 2000-2009, and 6.3 between 2010-2013.

(244) There has also been in the recent years a shift in geographical focus, with less emphasis on Europe and more on the rest of the world. A study for the European Parliament highlights reduced investments in crop protection in Europe, decline in the number of patents and expresses concerns that "innovation in this sector will 94 shrink".Meanwhile, there has been strong growth of emerging markets, as illustrated in Figure 11.

87 Phillips McDougall, "R&D trends for chemical crop protection products and the position of the European Market, A consultancy study undertaken for ECPA", September 2013.

88 See Commission Decision in Cases M.1806 – AstraZeneca/Novartis (2000) and M.2547 – Bayer/Aventis Crop Science (2002).

89 http://www.arystalifescience.com/2015release/ArystaLifeScienceacquiredbyPlatform.pdf.

90 http://www.fmccrop.com.au/fmc-to-acquire-cheminova/.

91 https://www.sumitomo-chem.co.jp/english/newsreleases/docs/20110510e.pdf.

92 See http://ec.europa.eu/competition/elojade/isef/case details.cfm?proc code=2 M 7962.

35

Figure 11 – Growth of crop protection markets 2001-2011

(245) Meanwhile, limited solutions are available for so-called "orphan" crops or minor uses. Companies focus on major global crops such as maize, wheat, rice.

1.6.3. New technology developments

(246) First, precision farming appears as a growing trend. It was developed during the 1980s in the US. The recent emergence of cloud storage allowed precision farming to develop. Precision farming agriculture aims at using collected data for farmers' decision-making. Companies provide services to farmers using scientific models that predict the soil and the plants' needs. For instance, precision farming may take into account weather forecasts. A competitor notes that "[p]recision agriculture / data-driven agriculture is a clear industry trend. The digital side of the business becomes more important. Access to tools and information helps farmers react better. Some 95 companies talk of a '4th revolution'." However, it is still unclear which player will emerge as a leader in this area. A majority of respondents to the Crop Protection Stakeholders' questionnaire stated that precision agriculture will play a major role in 96the EEA in the next five to 10 years.

(247) Second, the use of biological crop protection agents is expected to increase to some extent. Biologicals are an area of interest and spending for the industry. However, market participants highlight that biological products still present limited efficacy. Some companies see biologicals as "a trend but of limited weight in the overall crop protection business so far. The absolute percentage of biologicals is still very low." Another player highlighted that "[b]iologicals only play a minor role at present and can only supplement existing crop protection activities. They represent more of a 98strategic investment."

(248) Third, there is a move to integrated solutions, for instance to develop herbicide resistant traits, especially in the US Monsanto's Round-Up uses these methods. Some options also exist using non-GM technology, such as BASF's Clearfield. The industry study notes that "[t]he driving force behind the increase in seeds R&D is the

(248) Third, there is a move to integrated solutions, for instance to develop herbicide resistant traits, especially in the US Monsanto's Round-Up uses these methods. Some options also exist using non-GM technology, such as BASF's Clearfield. The industry study notes that "[t]he driving force behind the increase in seeds R&D is the

95 Agreed non-confidential minutes of a call with a competitor, 12 April 2016 (ID8247). 96 Questionnaire to Crop Protection Stakeholders and Testing Partners (Q3a and Q3b), question 17. 97 Agreed non-confidential minutes of a call with a competitor, 12 April 2016 (ID8247). 98 Agreed non-confidential minutes of a call with a competitor, 5 April 2016 (ID8242).

95 Agreed non-confidential minutes of a call with a competitor, 12 April 2016 (ID8247). 96 Questionnaire to Crop Protection Stakeholders and Testing Partners (Q3a and Q3b), question 17. 97 Agreed non-confidential minutes of a call with a competitor, 12 April 2016 (ID8247). 98 Agreed non-confidential minutes of a call with a competitor, 5 April 2016 (ID8242).

36

36

development of new GM traits, a technology that has only been accepted in the EU in 99the most limited way".

development of new GM traits, a technology that has only been accepted in the EU in 99the most limited way".

(249) Fourth, some new technologies are being developed in the biotech sector, such as RNAi silencing and gene editing.

(249) Fourth, some new technologies are being developed in the biotech sector, such as RNAi silencing and gene editing.

2. CROP PROTECTION BUSINESSES AND PRODUCTS OF THE PARTIES

2. CROP PROTECTION BUSINESSES AND PRODUCTS OF THE PARTIES

2.1. Introduction

2.1. Introduction

(250) DuPont’s agriculture businesses generated in 2014 approximately USD 11.3 billion revenues. DuPont’s Agriculture business comprises the following business areas:

(250) DuPont’s agriculture businesses generated in 2014 approximately USD 11.3 billion revenues. DuPont’s Agriculture business comprises the following business areas:

(1)DuPont Pioneer, active in the development and commercialisation of seeds generating revenues of approximately USD 7.6 billion in 2014, contributing around two-thirds of DuPont’s total Agriculture revenues.

(1)DuPont Pioneer, active in the development and commercialisation of seeds generating revenues of approximately USD 7.6 billion in 2014, contributing around two-thirds of DuPont’s total Agriculture revenues.

(2)DuPont Crop Protection, which provides insect, disease and weed control products as well as seed treatment solutions and accounted for approximately USD 3.7 billion in 2014, contributing the remaining third of DuPont’s Agriculture revenues. These revenues further break down as follows:

(2)DuPont Crop Protection, which provides insect, disease and weed control products as well as seed treatment solutions and accounted for approximately USD 3.7 billion in 2014, contributing the remaining third of DuPont’s Agriculture revenues. These revenues further break down as follows:

(1)Herbicides accounted for approximately 34% of DuPont’s crop protection revenues;

(1)Herbicides accounted for approximately 34% of DuPont’s crop protection revenues;

(2)Insecticides accounted for the largest portion of DuPont’s crop protection sales, namely approximately 44%;

(2)Insecticides accounted for the largest portion of DuPont’s crop protection sales, namely approximately 44%;

(3)Fungicides accounted for approximately 21%; and

(3)Fungicides accounted for approximately 21%; and

(4)Other products (including nematicides and seed treatment products), accounted for 1%.

(4)Other products (including nematicides and seed treatment products), accounted for 1%.

(251) Dow’s agricultural business (Dow AgroSciences, "DAS") generated in 2014 approximately USD 7.3 billion revenues. Dow’s business comprises the following business areas:

(251) Dow’s agricultural business (Dow AgroSciences, "DAS") generated in 2014 approximately USD 7.3 billion revenues. Dow’s business comprises the following business areas:

(1)DAS Seeds, which generated approximately USD 1.6 billion revenues in 2014, contributing approximately 22% to DAS’s total revenues.

(1)DAS Seeds, which generated approximately USD 1.6 billion revenues in 2014, contributing approximately 22% to DAS’s total revenues.

(2)DAS Crop Protection, which includes insect, disease and weed control products, generated approximately USD 5.7 billion revenues in 2014, accounting for approximately 78% of DAS’s revenues. Dow’s crop protection revenues break down as follows:

(2)DAS Crop Protection, which includes insect, disease and weed control products, generated approximately USD 5.7 billion revenues in 2014, accounting for approximately 78% of DAS’s revenues. Dow’s crop protection revenues break down as follows:

(1)Herbicides accounted for the largest portion, approximately 61%, of Dow’s crop protection sales;

(1)Herbicides accounted for the largest portion, approximately 61%, of Dow’s crop protection sales;

(2)Insecticides accounted for approximately 25%;

(2)Insecticides accounted for approximately 25%;

(3)Fungicides accounted for approximately 8%; and

(3)Fungicides accounted for approximately 8%; and

(4)Other products (including soil fumigants, termiticides, and nitrification inhibitors), accounted for approximately 6%.

(4)Other products (including soil fumigants, termiticides, and nitrification inhibitors), accounted for approximately 6%.

99 Phillips McDougall, "R&D trends for chemical crop protection products and the position of the European Market, A consultancy study undertaken for ECPA", September 2013.

99 Phillips McDougall, "R&D trends for chemical crop protection products and the position of the European Market, A consultancy study undertaken for ECPA", September 2013.

37

37

(252) Post-Transaction, the merged entity would become the market leader for many of the segments of the crop protection business as exemplified in Figure 12.

(252) Post-Transaction, the merged entity would become the market leader for many of the segments of the crop protection business as exemplified in Figure 12.

Figure 12 – DuPont and Dow crop protection portfolios

Figure 12 – DuPont and Dow crop protection portfolios

[…]

[…]

Source: DuPont's internal document, file name "DUPONT-CASEM7932-0152373 - STRICTLY CONFIDENTIAL - CONTAINS BUSINESS SECRETS - SUBJECT TO U.S. PRIVILEGE.pdf" (ID7830-34975), slide 13

Source: DuPont's internal document, file name "DUPONT-CASEM7932-0152373 - STRICTLY CONFIDENTIAL - CONTAINS BUSINESS SECRETS - SUBJECT TO U.S. PRIVILEGE.pdf" (ID7830-34975), slide 13

(253) As for the crop protection business in the EEA, in 2015 DuPont had a turnover of approximately USD 730 million and Dow of approximately USD 1.2 billion, with the break down as presented in Figure 13.

(253) As for the crop protection business in the EEA, in 2015 DuPont had a turnover of approximately USD 730 million and Dow of approximately USD 1.2 billion, with the break down as presented in Figure 13.

Figure 13 – Parties' EEA crop protection business

Figure 13 – Parties' EEA crop protection business

Source: Parties' briefing paper

Source: Parties' briefing paper

(254) Both DuPont and Dow had an EBTDA as percentage of revenues of around 21-22% for the crop protection business in 2015. DuPont's gross margin in crop protection as percentage of revenues was 47% while Dow's was 35%.

(254) Both DuPont and Dow had an EBTDA as percentage of revenues of around 21-22% for the crop protection business in 2015. DuPont's gross margin in crop protection as percentage of revenues was 47% while Dow's was 35%.

100 Dow's internal document, file name "DAS-20000001-000828.pdf" (ID7829-9342).

100 Dow's internal document, file name "DAS-20000001-000828.pdf" (ID7829-9342).

(255) In the case of DuPont crop protection business in EMEA, its gross margin was around 56% in 2015 and its projections are for a growth in sales in EMEA of 6.4% per year in the period 2015-2020 (Figure 14).

101 DuPont's internal document "2016 EMEA Regional Execution Plan", file name "DUPONT-CASEM7932-0091090 - STRICTLY CONFIDENTIAL - CONTAINS BUSINESS SECRETS.pdf" (ID6827-29535), slide 20 and 53.

Figure 14 – DuPont's growth projections in the EMEA

(255) In the case of DuPont crop protection business in EMEA, its gross margin was around 56% in 2015 and its projections are for a growth in sales in EMEA of 6.4% 101per year in the period 2015-2020 (Figure 14).

[…]

Figure 14 – DuPont's growth projections in the EMEA

Source: DuPont's internal document "2016 EMEA Regional Execution Plan", file name "DUPONT-CASEM7932-0091090 - STRICTLY CONFIDENTIAL - CONTAINS BUSINESS SECRETS.pdf" (ID6827-29535)

[…]

2.2. Herbicides

Source: DuPont's internal document "2016 EMEA Regional Execution Plan", file name "DUPONT-CASEM7932-0091090 - STRICTLY CONFIDENTIAL - CONTAINS BUSINESS SECRETS.pdf" (ID6827-29535)

(256) In herbicides the Parties have a strong position globally, notably in selective herbicides. Both Parties develop, manufacture and sell a broad range of different

2.2. Herbicides

100 Dow's internal document, file name "DAS-20000001-000828.pdf" (ID7829-9342). 101 DuPont's internal document "2016 EMEA Regional Execution Plan", file name "DUPONT- CASEM7932-0091090 - STRICTLY CONFIDENTIAL - CONTAINS BUSINESS SECRETS.pdf" (ID6827-29535), slide 20 and 53.

(256) In herbicides the Parties have a strong position globally, notably in selective herbicides. Both Parties develop, manufacture and sell a broad range of different

38

herbicides to be applied to all main crop groups across the EEA. While the Parties are strong on products targeting broadleaf weeds, and are notably leaders in cereals, they also have graminicides (targeting grass weeds) and cross-spectrum (also referred to as "broad spectrum") products.

herbicides to be applied to all main crop groups across the EEA. While the Parties are strong on products targeting broadleaf weeds, and are notably leaders in cereals, they also have graminicides (targeting grass weeds) and cross-spectrum (also referred to as "broad spectrum") products.

(257) According to Figure 12, DuPont's overall herbicides sales in 2014 corresponded to USD 1.1 billion, or 7% of the total sales of herbicides by the Big 5. Dow's sales corresponded to USD 3.5 billion or 23% of the total sales of herbicides by the Big 5. Together Dow and DuPont would have 31% of the overall herbicides sales by the Big 5 ahead of any of the other three main players.

(257) According to Figure 12, DuPont's overall herbicides sales in 2014 corresponded to USD 1.1 billion, or 7% of the total sales of herbicides by the Big 5. Dow's sales corresponded to USD 3.5 billion or 23% of the total sales of herbicides by the Big 5. Together Dow and DuPont would have 31% of the overall herbicides sales by the Big 5 ahead of any of the other three main players.

(258) As regards the key crop segments, namely cereals, corn, soybean, rice and specialty (Figure 15), the combined DuPont and Dow would be particularly strong in cereals (with 30% of the global sales), soybeans (with 39% of the global sales) and rice (with 43% of the global sales).

(258) As regards the key crop segments, namely cereals, corn, soybean, rice and specialty (Figure 15), the combined DuPont and Dow would be particularly strong in cereals (with 30% of the global sales), soybeans (with 39% of the global sales) and rice (with 43% of the global sales).

Figure 15 – DuPont and Dow herbicides portfolios

Figure 15 – DuPont and Dow herbicides portfolios

[…]

[…]

Source: DuPont's internal document, file name "DUPONT-CASEM7932-0152373 - STRICTLY CONFIDENTIAL - CONTAINS BUSINESS SECRETS - SUBJECT TO U.S. PRIVILEGE.pdf" (ID7830-34975), slide 14

Source: DuPont's internal document, file name "DUPONT-CASEM7932-0152373 - STRICTLY CONFIDENTIAL - CONTAINS BUSINESS SECRETS - SUBJECT TO U.S. PRIVILEGE.pdf" (ID7830-34975), slide 14

(259) Dow's current portfolio in the EEA consists mainly of products derived from several AIs. Pyroxsulam, fluroxypyr, florasulam, clopyralid, propyzamide account together for 75% of Dow's 2015 herbicides sales in the EEA. Dow is notably rolling out two other products: Arylex (AI halauxifen), a broadleaf herbicide, and Rinskor (AI florpyrauxifen-benzyl), a cross-spectrum herbicide.

(259) Dow's current portfolio in the EEA consists mainly of products derived from several AIs. Pyroxsulam, fluroxypyr, florasulam, clopyralid, propyzamide account together for 75% of Dow's 2015 herbicides sales in the EEA. Dow is notably rolling out two other products: Arylex (AI halauxifen), a broadleaf herbicide, and Rinskor (AI florpyrauxifen-benzyl), a cross-spectrum herbicide.

(260) DuPont's portfolio in the EEA consists mainly of products from the sulfonylurea family ("SUs" family), as well as lenacil, and in-licensed products. Thifensulfuron methyl, tribenuron methyl, nicosulfuron, triflusulfuron methyl, rimsulfuron and flupyrsulfuron account together for […]% of DuPont's herbicides sales in the 102 EEA. DuPont is further conducting development of a cross-spectrum herbicide (SGF45) and has several candidate herbicide molecules in its discovery pipeline.

(260) DuPont's portfolio in the EEA consists mainly of products from the sulfonylurea family ("SUs" family), as well as lenacil, and in-licensed products. Thifensulfuron methyl, tribenuron methyl, nicosulfuron, triflusulfuron methyl, rimsulfuron and flupyrsulfuron account together for […]% of DuPont's herbicides sales in the 102 EEA. DuPont is further conducting development of a cross-spectrum herbicide (SGF45) and has several candidate herbicide molecules in its discovery pipeline.

2.3. Insecticides

2.3. Insecticides

(261) In insecticides, the Parties also have a strong position globally, notably in insecticides targeting lepidoptera, thrips, coleoptera and diptera. Both Parties develop, manufacture and sell a broad range of different insecticides to be applied to all main crop groups across the EEA.

(261) In insecticides, the Parties also have a strong position globally, notably in insecticides targeting lepidoptera, thrips, coleoptera and diptera. Both Parties develop, manufacture and sell a broad range of different insecticides to be applied to all main crop groups across the EEA.

(262) According to Figure 12, DuPont's overall insecticides sales in 2014 corresponded to USD 1.8 billion, or 17.6% of the total sales of insecticides by the Big 5. Dow's sales corresponded to USD 1.4 billion or 13.7% of the total sales of insecticides by the Big 5. Together Dow and DuPont would have 31.3% of the overall insecticides sales by the Big 5 ahead of any of the other three main players.

(262) According to Figure 12, DuPont's overall insecticides sales in 2014 corresponded to USD 1.8 billion, or 17.6% of the total sales of insecticides by the Big 5. Dow's sales corresponded to USD 1.4 billion or 13.7% of the total sales of insecticides by the Big 5. Together Dow and DuPont would have 31.3% of the overall insecticides sales by the Big 5 ahead of any of the other three main players.

(263) As regards the key crop segments globally, namely soybean, rice, cotton, corn and specialty (Figure 16), the combined DuPont and Dow would be particularly strong in

(263) As regards the key crop segments globally, namely soybean, rice, cotton, corn and specialty (Figure 16), the combined DuPont and Dow would be particularly strong in

102 DuPont's internal document "Weed Management Business Update Short Term Outlook 2015-2016", file name "DUPONT-CASEM7932-0067460 - STRICTLY CONFIDENTIAL - CONTAINS BUSINESS SECRETS.pptx" (ID4384-18), slide 4.

102 DuPont's internal document "Weed Management Business Update Short Term Outlook 2015-2016", file name "DUPONT-CASEM7932-0067460 - STRICTLY CONFIDENTIAL - CONTAINS BUSINESS SECRETS.pptx" (ID4384-18), slide 4.

39

39

rice (with 42% of the sales), corn (with 37.5% of the sales) and specialty (with 35% of the sales).

rice (with 42% of the sales), corn (with 37.5% of the sales) and specialty (with 35% of the sales).

Figure 16 – DuPont and Dow insecticide portfolios

Figure 16 – DuPont and Dow insecticide portfolios

[…]

[…]

Source: DuPont's internal document, file name "DUPONT-CASEM7932-0152373 - STRICTLY CONFIDENTIAL - CONTAINS BUSINESS SECRETS - SUBJECT TO U.S. PRIVILEGE.pdf" (ID7830-34975), slide 15

Source: DuPont's internal document, file name "DUPONT-CASEM7932-0152373 - STRICTLY CONFIDENTIAL - CONTAINS BUSINESS SECRETS - SUBJECT TO U.S. PRIVILEGE.pdf" (ID7830-34975), slide 15

(264) DuPont's insecticides portfolio in the EEA mainly consists of two products: Rynaxypyr (AI chlorantraniliprole) and Indoxacarb. These two products currently account for 93% of DuPont's EEA insecticides revenues. Both of these products target lepidoptera. DuPont will shortly launch Cyazypyr in the EEA, which has the same chemical class as Rynaxypyr but has a broader spectrum of activity. DuPont also a number of insecticides in its pipeline, with one molecule targeting sucking insects (EMN08); and the remaining targeting lepidoptera and potentially other pests too.

(264) DuPont's insecticides portfolio in the EEA mainly consists of two products: Rynaxypyr (AI chlorantraniliprole) and Indoxacarb. These two products currently account for 93% of DuPont's EEA insecticides revenues. Both of these products target lepidoptera. DuPont will shortly launch Cyazypyr in the EEA, which has the same chemical class as Rynaxypyr but has a broader spectrum of activity. DuPont also a number of insecticides in its pipeline, with one molecule targeting sucking insects (EMN08); and the remaining targeting lepidoptera and potentially other pests too.

(265) Dow's insecticides portfolio in the EEA currently consists of three products: chlorpyrifos, Spinosad and methoxyfenozide. Chlorpyryifos and Spinosad currently account for 91% of Dow's EEA insecticides revenues. All three AIs target lepidoptera, while Spinosad also targets thrips and chlorpyrifos is a broad spectrum AI that targets a number of additional pests. Dow is currently launching Spinetoram, which mainly targets lepidoptera and thrips. Dow will also soon launch Isoclast, which mainly targets various hemipteran pests. Finally, Dow has a number of insecticides in its pipeline, with one molecule targeting sucking insects ([insecticide pipeline 2]); and the remaining two targeting lepidoptera and potentially other pests too.

(265) Dow's insecticides portfolio in the EEA currently consists of three products: chlorpyrifos, Spinosad and methoxyfenozide. Chlorpyryifos and Spinosad currently account for 91% of Dow's EEA insecticides revenues. All three AIs target lepidoptera, while Spinosad also targets thrips and chlorpyrifos is a broad spectrum AI that targets a number of additional pests. Dow is currently launching Spinetoram, which mainly targets lepidoptera and thrips. Dow will also soon launch Isoclast, which mainly targets various hemipteran pests. Finally, Dow has a number of insecticides in its pipeline, with one molecule targeting sucking insects ([insecticide pipeline 2]); and the remaining two targeting lepidoptera and potentially other pests too.

2.4. Fungicides

2.4. Fungicides

(266) The Parties currently have a relatively weak position in fungicides. In light of this weak position, Dow and DuPont tend to cooperate with other companies to have a larger presence.

(266) The Parties currently have a relatively weak position in fungicides. In light of this weak position, Dow and DuPont tend to cooperate with other companies to have a larger presence.

(267) According to a DuPont's internal document, DuPont's overall fungicides sales in 2014 corresponded to USD 0.8 billion, or 6.6% of the total sales of fungicides by the Big 5. Dow's sales corresponded to USD 0.4 billion or 3.3% of the total sales of fungicides by the Big 5. Together Dow and DuPont would have 10% of the overall fungicides sales by the Big 5.

(267) According to a DuPont's internal document, DuPont's overall fungicides sales in 2014 corresponded to USD 0.8 billion, or 6.6% of the total sales of fungicides by the Big 5. Dow's sales corresponded to USD 0.4 billion or 3.3% of the total sales of fungicides by the Big 5. Together Dow and DuPont would have 10% of the overall fungicides sales by the Big 5.

Figure 17 – DuPont and Dow fungicide portfolios

Figure 17 – DuPont and Dow fungicide portfolios

[…]

[…]

Source: DuPont's internal document, file name "DUPONT-CASEM7932-0152373 - STRICTLY CONFIDENTIAL - CONTAINS BUSINESS SECRETS - SUBJECT TO U.S. PRIVILEGE.pdf" (ID7830-34975), slide 16

Source: DuPont's internal document, file name "DUPONT-CASEM7932-0152373 - STRICTLY CONFIDENTIAL - CONTAINS BUSINESS SECRETS - SUBJECT TO U.S. PRIVILEGE.pdf" (ID7830-34975), slide 16

(268) DuPont's fungicide portfolio in the EEA currently consists of six products (Picoxystrobin, Proquinazid, Penthiopyrad, Cymoxanil, Famoxadone and Carbendazim) and Dow's fungicide portfolio in the EEA currently consists of seven products (Myclobutanil, Quinoxyfen, Meptyldinocap, Tricyclazole, Fenbuconazole, Mancozeb-Dithane and Propiconazole).

(268) DuPont's fungicide portfolio in the EEA currently consists of six products (Picoxystrobin, Proquinazid, Penthiopyrad, Cymoxanil, Famoxadone and Carbendazim) and Dow's fungicide portfolio in the EEA currently consists of seven products (Myclobutanil, Quinoxyfen, Meptyldinocap, Tricyclazole, Fenbuconazole, Mancozeb-Dithane and Propiconazole).

40

40

(269) The Parties have discovery and development pipeline products in fungicides. DuPont has three main products in the pipeline, namely Zorvec, [fungicide pipeline 1] and THQ25. Dow has also three main products in the pipeline, namely Inatreq, XR-659 and XR-481.

(269) The Parties have discovery and development pipeline products in fungicides. DuPont has three main products in the pipeline, namely Zorvec, [fungicide pipeline 1] and THQ25. Dow has also three main products in the pipeline, namely Inatreq, XR-659 and XR-481.

2.5. Nematicides

2.5. Nematicides

(270) The Parties both produce products that are used in the soil to control nematodes. Dow produces the soil fumigant AI 1,3-dichloropropene. It is used on soil before crop seeds are planted to control nematodes and other pests. […].

(270) The Parties both produce products that are used in the soil to control nematodes. Dow produces the soil fumigant AI 1,3-dichloropropene. It is used on soil before crop seeds are planted to control nematodes and other pests. […].

(271) DuPont produces one nematicide AI, oxamyl. Oxamyl was originally developed to control true insects. However, its main usage now is to control nematodes. DuPont is 103also developing a new nematicide, fluazaindolizine (Q8U80) with a new MoA.

(271) DuPont produces one nematicide AI, oxamyl. Oxamyl was originally developed to control true insects. However, its main usage now is to control nematodes. DuPont is 103also developing a new nematicide, fluazaindolizine (Q8U80) with a new MoA.

3. O VERALL STRUCTURE OF THE COMMISSION' S ASSESSMENT OF THE EFFECTS OF

3. O VERALL STRUCTURE OF THE COMMISSION' S ASSESSMENT OF THE EFFECTS OF

THE TRANSACTION ON COMPETITION

THE TRANSACTION ON COMPETITION

3.1. The activities of DuPont and Dow overlap in four respects

3.1. The activities of DuPont and Dow overlap in four respects

(272) As seen in the previous section the activities in crop protection of DuPont and Dow overlap in four respects.

(272) As seen in the previous section the activities in crop protection of DuPont and Dow overlap in four respects.

(273) First, the Parties sell existing overlapping products in a number of European markets. In so far the Transaction may affect product and price competition between existing products of the Parties. In that regard the Commission investigates, in line with paragraphs 24 et seq. of the Commission Guidelines on the assessment of horizontal mergers under the Council Regulation on the control of concentrations between 104 undertakings ("Horizontal Merger Guidelines"), whether the Transaction gives rise to non-coordinated effects, in particular as a result of the creation or strengthening of a dominant position and the elimination of an important competitive constraint.

(273) First, the Parties sell existing overlapping products in a number of European markets. In so far the Transaction may affect product and price competition between existing products of the Parties. In that regard the Commission investigates, in line with paragraphs 24 et seq. of the Commission Guidelines on the assessment of horizontal mergers under the Council Regulation on the control of concentrations between 104 undertakings ("Horizontal Merger Guidelines"), whether the Transaction gives rise to non-coordinated effects, in particular as a result of the creation or strengthening of a dominant position and the elimination of an important competitive constraint.

(274) Second, the Parties intend to launch in the near or relatively near future a number of products in markets where the other Party has existing products or also intends to launch in the near or relatively near future competing products. In so far the Transaction may affect potential product and price competition between forthcoming and existing products or between forthcoming products. Paragraph 58 of the Horizontal Merger Guidelines clarifies that a merger with a potential competitor can have similar effects to mergers between competitors already active on the same market. In that regard the Commission in line with paragraphs 58 et seq. and 24 et seq. also investigates whether the Transaction gives rise to non-coordinated effects, in particular as a result of the creation or strengthening of a dominant position and the elimination of an important competitive constraint.

(274) Second, the Parties intend to launch in the near or relatively near future a number of products in markets where the other Party has existing products or also intends to launch in the near or relatively near future competing products. In so far the Transaction may affect potential product and price competition between forthcoming and existing products or between forthcoming products. Paragraph 58 of the Horizontal Merger Guidelines clarifies that a merger with a potential competitor can have similar effects to mergers between competitors already active on the same market. In that regard the Commission in line with paragraphs 58 et seq. and 24 et seq. also investigates whether the Transaction gives rise to non-coordinated effects, in particular as a result of the creation or strengthening of a dominant position and the elimination of an important competitive constraint.

(275) As regards the overlapping activities described in recitals (273) and (274) the Commission focus its assessment on product and price competition. Product competition may be reduced through a different positioning of the products in order to prevent cannibalisation of each other's' products' sales. Price competition may be reduced in those areas where the products of the Parties would continue to compete head-to-head. It is important to note that product and price competition may be reduced independently of whether the merged entity decides to continue to sell both

(275) As regards the overlapping activities described in recitals (273) and (274) the Commission focus its assessment on product and price competition. Product competition may be reduced through a different positioning of the products in order to prevent cannibalisation of each other's' products' sales. Price competition may be reduced in those areas where the products of the Parties would continue to compete head-to-head. It is important to note that product and price competition may be reduced independently of whether the merged entity decides to continue to sell both

103 Parties' response to the Commission's request for information RFI 27, question 3. 104 OJ C 31, 5.2.2004.

103 Parties' response to the Commission's request for information RFI 27, question 3. 104 OJ C 31, 5.2.2004.

41

41

products or to withdraw one of the products from the market. In the latter case there might however be an additional reduction of competitive pressure on other competitors and thus an even more significant harm to product and price competition in a given market.

products or to withdraw one of the products from the market. In the latter case there might however be an additional reduction of competitive pressure on other competitors and thus an even more significant harm to product and price competition in a given market.

(276) Section V.6 will focus on the effects of the Transaction on product and price competition.

(276) Section V.6 will focus on the effects of the Transaction on product and price competition.

(277) Third, the Parties are pursuing at least partially overlapping important lines of research with similar discovery concepts and pipeline molecules targeting the same product markets in discovery and early development which, if developed and brought to the market, would compete head-to-head against each other. It is also the case that one of the Parties is pursuing important lines of research that will compete in a market where the other Party is an existing or potential supplier. In so far the Transaction may affect innovation competition between the two Parties in the form of discontinuation, deferment or redirection of competing lines of research and early pipeline products.

(277) Third, the Parties are pursuing at least partially overlapping important lines of research with similar discovery concepts and pipeline molecules targeting the same product markets in discovery and early development which, if developed and brought to the market, would compete head-to-head against each other. It is also the case that one of the Parties is pursuing important lines of research that will compete in a market where the other Party is an existing or potential supplier. In so far the Transaction may affect innovation competition between the two Parties in the form of discontinuation, deferment or redirection of competing lines of research and early pipeline products.

(278) Fourth, the Parties are operating two of only a few competing global R&D organisations. The discontinuation of one of those organisations may significantly reduce the overall level of innovation competition and thus product innovation in the crop protection industry.

(278) Fourth, the Parties are operating two of only a few competing global R&D organisations. The discontinuation of one of those organisations may significantly reduce the overall level of innovation competition and thus product innovation in the crop protection industry.

(279) As regards the overlapping activities described in recitals (277) and (278), paragraph 8 of the Horizontal Merger Guidelines clarifies that Union merger control also applies to mergers which diminish innovation or reduce choice of goods. Paragraph 8 also suggests that the parts of the Horizontal Merger Guidelines dealing at first sight only with price competition can be applied mutatis mutandis to innovation competition: "[i]n this notice, the expression 'increased prices' is often used as shorthand for these various ways [including diminished innovation and reduced choice] in which a merger may result in competitive harm".

(279) As regards the overlapping activities described in recitals (277) and (278), paragraph 8 of the Horizontal Merger Guidelines clarifies that Union merger control also applies to mergers which diminish innovation or reduce choice of goods. Paragraph 8 also suggests that the parts of the Horizontal Merger Guidelines dealing at first sight only with price competition can be applied mutatis mutandis to innovation competition: "[i]n this notice, the expression 'increased prices' is often used as shorthand for these various ways [including diminished innovation and reduced choice] in which a merger may result in competitive harm".

(280) Section V.8 will focus on the effects of the Transaction on innovation competition.

(280) Section V.8 will focus on the effects of the Transaction on innovation competition.

(281) There are also potential overlaps as regards the licensing and supply of new AIs, as discussed in section V.7.

(281) There are also potential overlaps as regards the licensing and supply of new AIs, as discussed in section V.7.

3.2. Likely magnitude of the effects

3.2. Likely magnitude of the effects

(282) For the two types of overlaps between the Parties discussed in recitals (273) to (275) the harm to consumers would result mainly from the reduction of competition between the products of the Parties in the respective markets.

(282) For the two types of overlaps between the Parties discussed in recitals (273) to (275) the harm to consumers would result mainly from the reduction of competition between the products of the Parties in the respective markets.

(283) For the two types of overlaps discussed in recitals (277) to (279) the effects on innovation competition, which result mainly in the discontinuation, deferment or redirection of competing lines of research and early pipeline products, would cause an additional long-term harm to consumers in the markets where a reduction of variety would take place. This harm would reinforce the harm resulting from the absence of price and product competition between the Parties in those markets and would be felt repeatedly year after year for any possible future interaction between the Parties. As a consequence, for the latter two types of overlaps the Commission considers likely that the magnitude of harm would be relatively larger.

(283) For the two types of overlaps discussed in recitals (277) to (279) the effects on innovation competition, which result mainly in the discontinuation, deferment or redirection of competing lines of research and early pipeline products, would cause an additional long-term harm to consumers in the markets where a reduction of variety would take place. This harm would reinforce the harm resulting from the absence of price and product competition between the Parties in those markets and would be felt repeatedly year after year for any possible future interaction between the Parties. As a consequence, for the latter two types of overlaps the Commission considers likely that the magnitude of harm would be relatively larger.

42

42

3.3. Depending on how close pipeline products are to the market the Transaction would be likely to affect product or only innovation competition

3.3. Depending on how close pipeline products are to the market the Transaction would be likely to affect product or only innovation competition

(284) As explained in Section 3.1 the second layer focuses on overlaps between currently marketed and forthcoming products, that is to say where one of the Parties intends to launch in the relatively near future a number of products in markets where the other Party has existing products or also intends to launch in the relatively near future competing products.

(284) As explained in Section 3.1 the second layer focuses on overlaps between currently marketed and forthcoming products, that is to say where one of the Parties intends to launch in the relatively near future a number of products in markets where the other Party has existing products or also intends to launch in the relatively near future competing products.

(285) Indeed, paragraph 38 of the Horizontal Merger Guidelines highlights that "effective competition may be significantly impeded by a merger between two important innovators, for instance between two companies with "pipeline" products related to a specific product market. Similarly, a firm with a relatively small market share may nevertheless be an important competitive force if it has promising pipeline products".

(285) Indeed, paragraph 38 of the Horizontal Merger Guidelines highlights that "effective competition may be significantly impeded by a merger between two important innovators, for instance between two companies with "pipeline" products related to a specific product market. Similarly, a firm with a relatively small market share may nevertheless be an important competitive force if it has promising pipeline products".

(286) Similarly, paragraphs 58-60 of the Horizontal Merger Guidelines describe the potential anti-competitive effects of a merger between potential competitors, where "the potential competitor […] already exert[s] a significant constraining influence or there [is] a significant likelihood that it would grow into an effective competitive force [in a relatively short period of time]". These paragraphs point to the significance of costs already incurred or yet to commit as a useful indicator.

(286) Similarly, paragraphs 58-60 of the Horizontal Merger Guidelines describe the potential anti-competitive effects of a merger between potential competitors, where "the potential competitor  already exert[s] a significant constraining influence or there [is] a significant likelihood that it would grow into an effective competitive force [in a relatively short period of time]". These paragraphs point to the significance of costs already incurred or yet to commit as a useful indicator.

(287) However, what constitutes such a significant constraining influence or a significant likelihood that a potential competitor would become an effective competitive force in a relatively short period of time will of course depend on the specific characteristics of each industry. In particular, key factors will be the time necessary to launch a new product on the market, as well as the point in that process where the likelihood of growing into an effective competitive force (and, a fortiori, of effectively entering the market) becomes significant.

(287) However, what constitutes such a significant constraining influence or a significant likelihood that a potential competitor would become an effective competitive force in a relatively short period of time will of course depend on the specific characteristics of each industry. In particular, key factors will be the time necessary to launch a new product on the market, as well as the point in that process where the likelihood of growing into an effective competitive force (and, a fortiori, of effectively entering the market) becomes significant.

(288) The Commission has recently decided several cases in the pharmaceutical sector where it considered both potential competition and innovation competition. In that industry, candidate molecules go through several development stages, starting with pre-clinical trials in laboratories and on animals and later moving on to clinical trials in humans (so called "Phase I", "Phase II" and "Phase III" clinical trials).

(288) The Commission has recently decided several cases in the pharmaceutical sector where it considered both potential competition and innovation competition. In that industry, candidate molecules go through several development stages, starting with pre-clinical trials in laboratories and on animals and later moving on to clinical trials in humans (so called "Phase I", "Phase II" and "Phase III" clinical trials).

(289) As noted in the the Commission Decision COMP/M.1846 – Glaxo Wellcome / Smithkline Beecham, Phase I marks the start of clinical testing on humans, currently some eight to 10 years before a product is marketed, with initial efficacy and toxicology screenings. Statistically, projects in Phase I generally have no more than a 10% chance of reaching the market. Phase II, some four to five years before the product is marketed, involves working out the proper dose for the patient and defining the areas of application. The success of Phase II is generally acknowledged to be approximately 30%. Phase III, starting approximately three years before the product is marketed, involves establishing the product’s effectiveness on larger 105groups of patients. The risk of failure in Phase III is reported to be over 50%.

(289) As noted in the the Commission Decision COMP/M.1846  Glaxo Wellcome / Smithkline Beecham, Phase I marks the start of clinical testing on humans, currently some eight to 10 years before a product is marketed, with initial efficacy and toxicology screenings. Statistically, projects in Phase I generally have no more than a 10% chance of reaching the market. Phase II, some four to five years before the product is marketed, involves working out the proper dose for the patient and defining the areas of application. The success of Phase II is generally acknowledged to be approximately 30%. Phase III, starting approximately three years before the product is marketed, involves establishing the products effectiveness on larger 105groups of patients. The risk of failure in Phase III is reported to be over 50%.

(290) In the Commission Decision COMP/M.7275 – Novartis/GlaxoSmithKline Oncology Business, the Commission concluded that the transaction would lead to a reduction of potential competition for innovative cancer treatments regarding overlaps between a marketed product and a pipeline Phase III product as well as to a reduction of

(290) In the Commission Decision COMP/M.7275  Novartis/GlaxoSmithKline Oncology Business, the Commission concluded that the transaction would lead to a reduction of potential competition for innovative cancer treatments regarding overlaps between a marketed product and a pipeline Phase III product as well as to a reduction of

105 Commission Decision in Case M.1846 – Glaxo Wellcome/Smithkline Beecham (2000), recital 70.

105 Commission Decision in Case M.1846  Glaxo Wellcome/Smithkline Beecham (2000), recital 70.

43

43

innovation competition regarding overlaps between pipeline products at earlier stages 106 of development, namely in Phase I and Phase II.In sum, it concluded to a significant impediment to effective competition on the basis of potential competition where a product had no more than a 50% chance of reaching the market, within approximately three years. It also concluded to such an impediment where two competing projects were, respectively, 8-10 years and 4-5 years from market launch, with only 10% and 30% chances of success.

innovation competition regarding overlaps between pipeline products at earlier stages 106 of development, namely in Phase I and Phase II.In sum, it concluded to a significant impediment to effective competition on the basis of potential competition where a product had no more than a 50% chance of reaching the market, within approximately three years. It also concluded to such an impediment where two competing projects were, respectively, 8-10 years and 4-5 years from market launch, with only 10% and 30% chances of success.

(291) Crop protection is a sector where large investments and significant amounts of time are needed to bring new products to the market. It thus appears to take approximately eight to 10 years to get an AI from its discovery to the launch of formulated products, with an average cost of approximately USD 250-300 million. A crop protection company therefore simply cannot discover and launch a new AI on the market in only one or two years.

(291) Crop protection is a sector where large investments and significant amounts of time are needed to bring new products to the market. It thus appears to take approximately eight to 10 years to get an AI from its discovery to the launch of formulated products, with an average cost of approximately USD 250-300 million. A crop protection company therefore simply cannot discover and launch a new AI on the market in only one or two years.

(292) First of all, the company defines some areas of research or discovery concepts based on the expected market output, thus focusing the research on discovering molecules that can be applied in certain market segments and geographical areas (EEA, for instance).

(292) First of all, the company defines some areas of research or discovery concepts based on the expected market output, thus focusing the research on discovering molecules that can be applied in certain market segments and geographical areas (EEA, for instance).

(293) As explained in Sections V.1.3 and V.1.4, selected candidate molecules – in fact, several analogs of a base molecule at these early stages – go through the discovery process, where they are first screened and optimised for efficacy and intended uses. Tests are also run to assess toxicology and their environmental impact in order to prevent later failures. Patents are typically filed.

(293) As explained in Sections V.1.3 and V.1.4, selected candidate molecules  in fact, several analogs of a base molecule at these early stages  go through the discovery process, where they are first screened and optimised for efficacy and intended uses. Tests are also run to assess toxicology and their environmental impact in order to prevent later failures. Patents are typically filed.

(294) Once enough data has been gathered to the satisfaction of the company, candidate molecules are moved to the development stage, where further data is gathered for regulatory submissions for the companies to be able to market their products. These trials will be made both on AIs and on formulated products, including mixtures of several AIs. The business case will also be gradually refined, with specific target crops, pests and geographies, precise revenue estimates as well as marketing and differentiation strategies.

(294) Once enough data has been gathered to the satisfaction of the company, candidate molecules are moved to the development stage, where further data is gathered for regulatory submissions for the companies to be able to market their products. These trials will be made both on AIs and on formulated products, including mixtures of several AIs. The business case will also be gradually refined, with specific target crops, pests and geographies, precise revenue estimates as well as marketing and differentiation strategies.

(295) Development entails strong financial commitments from crop protection companies, since many of the expensive studies required for regulatory approval can apparently not be halted. Because development is so costly, these studies will typically only be launched where there is relative certainty that the AI will eventually be successfully launched – albeit perhaps with a slightly modified formulation or marketing scope reducing market prospects somewhat, for example use or dose limitations, in light of the full results of the studies and of the approval and authorisation process. It is thus statistically unlikely that a product in development will eventually not be launched.

(295) Development entails strong financial commitments from crop protection companies, since many of the expensive studies required for regulatory approval can apparently not be halted. Because development is so costly, these studies will typically only be launched where there is relative certainty that the AI will eventually be successfully launched  albeit perhaps with a slightly modified formulation or marketing scope reducing market prospects somewhat, for example use or dose limitations, in light of the full results of the studies and of the approval and authorisation process. It is thus statistically unlikely that a product in development will eventually not be launched.

(296) Finally, companies will apply for registration of their products, a process taking several years during which they may be called upon to provide additional data.

(296) Finally, companies will apply for registration of their products, a process taking several years during which they may be called upon to provide additional data.

(297) Accordingly, the likelihood of success is very different in crop protection than it is in the pharmaceutical industry. It gradually increases in the discovery phase, from a relatively low level in the early stages – albeit higher than for pharmaceutical candidates in the pre-clinical phase – to higher levels in the later stages (see Figure 18). Moreover, because the main toxicology and environmental screenings are

(297) Accordingly, the likelihood of success is very different in crop protection than it is in the pharmaceutical industry. It gradually increases in the discovery phase, from a relatively low level in the early stages  albeit higher than for pharmaceutical candidates in the pre-clinical phase  to higher levels in the later stages (see Figure 18). Moreover, because the main toxicology and environmental screenings are

106 Commission Decision in Case M.7275 – Novartis/GlaxoSmithKline Oncology Business (2015), recital 90.

106 Commission Decision in Case M.7275  Novartis/GlaxoSmithKline Oncology Business (2015), recital 90.

44

44

increasingly done as early as possible in discovery, the likelihood of success becomes very high as soon as the molecule is moved into development, with an 80 to 90% chance of reaching the market.

increasingly done as early as possible in discovery, the likelihood of success becomes very high as soon as the molecule is moved into development, with an 80 to 90% chance of reaching the market.

(298) In spite of this very high likelihood of success, market launch is still many years away because the regulatory constraints for registration – particularly in the EEA – make the requisite studies long, resulting in a prolonged development phase. The processing of applications for registration also takes a significant number of years.

(298) In spite of this very high likelihood of success, market launch is still many years away because the regulatory constraints for registration  particularly in the EEA  make the requisite studies long, resulting in a prolonged development phase. The processing of applications for registration also takes a significant number of years.

(299) Hence, pipeline projects in crop protection will have a significant likelihood of reaching the market much earlier than similar pharmaceutical projects. By way of illustration, projects approximately six to eight years away from launch will have no more than a 10% to, at best, 30% chance of launch in the pharmaceutical sector, whereas similar projects in crop protection will have an 80 to 90% chance. Conversely, pharmaceutical projects will only reach an 80 to 90% likelihood of success once the results of the Phase III trials have been assessed and decisions are made regarding launch, which will be fairly close to effective product launch.

(299) Hence, pipeline projects in crop protection will have a significant likelihood of reaching the market much earlier than similar pharmaceutical projects. By way of illustration, projects approximately six to eight years away from launch will have no more than a 10% to, at best, 30% chance of launch in the pharmaceutical sector, whereas similar projects in crop protection will have an 80 to 90% chance. Conversely, pharmaceutical projects will only reach an 80 to 90% likelihood of success once the results of the Phase III trials have been assessed and decisions are made regarding launch, which will be fairly close to effective product launch.

(300) In parallel, this long development and registration lead time means that detailed testing of mixture concepts and sales plans, including relatively detailed roll-out and business plans by crop and geography, will be prepared already six to eight years before effective launch.

(300) In parallel, this long development and registration lead time means that detailed testing of mixture concepts and sales plans, including relatively detailed roll-out and business plans by crop and geography, will be prepared already six to eight years before effective launch.

(301) This is made possible by the fact that crop protection companies have some visibility on competitor pipelines through publicly available information such as patent applications – the subject-matter of which they typically manufacture and test – filings for ISO names, and investor presentations. With this close monitoring of competing pipelines, crop protection companies are able to somewhat reliably assess the competitive dynamics of the market for the coming decade. In fact, they are compelled to do so in order to assess the revenue potentials of their own pipeline products, and make business and investment decisions on that basis. Because discovering and developing new products takes a long time, these anticipations are very unlikely to be unexpectedly disrupted.

(301) This is made possible by the fact that crop protection companies have some visibility on competitor pipelines through publicly available information such as patent applications  the subject-matter of which they typically manufacture and test  filings for ISO names, and investor presentations. With this close monitoring of competing pipelines, crop protection companies are able to somewhat reliably assess the competitive dynamics of the market for the coming decade. In fact, they are compelled to do so in order to assess the revenue potentials of their own pipeline products, and make business and investment decisions on that basis. Because discovering and developing new products takes a long time, these anticipations are very unlikely to be unexpectedly disrupted.

(302) The Commission thus concludes that, in crop protection, the assessment of potential competition should take into account AIs which have entered or are about to enter the development stage (as defined in Section V.1.4.2), and accordingly have a significant likelihood of being launched on the market in spite of this launch in Europe sometimes being six to eight years away.

(302) The Commission thus concludes that, in crop protection, the assessment of potential competition should take into account AIs which have entered or are about to enter the development stage (as defined in Section V.1.4.2), and accordingly have a significant likelihood of being launched on the market in spite of this launch in Europe sometimes being six to eight years away.

4. CROP PROTECTION MARKET DEFINITION PRINCIPLES

4. CROP PROTECTION MARKET DEFINITION PRINCIPLES

4.1. Introduction

4.1. Introduction

(303) Agrochemical companies mainly sell formulated products to crop protection distributors or cooperatives of farmers and seed treatment products to seed suppliers (which are very often the same as for crop protection).

(303) Agrochemical companies mainly sell formulated products to crop protection distributors or cooperatives of farmers and seed treatment products to seed suppliers (which are very often the same as for crop protection).

(304) In addition, R&D agrochemical companies sell their technology on the upstream market through the licensing of AIs (which may include IP rights, data,

45

registration, etc.) to competing crop protection players. This can also take the form of a supply of AIs which encompasses a transfer of technology. The receiving party may then use those AIs as an input to produce their own formulated products benefiting from the licensor's/supplier's proprietary technology (for example IP rights, data, etc.). These sales are to be distinguished from the bulk supply of off-

107

patent(generic) AIs when it does not involve any transfer of technology (that is to say no IP, data package, etc.).

(305) Finally, it is also necessary to consider the innovation efforts of R&D agrochemical companies to discover and develop new AIs which will then feed into the sales of AIs and of formulated products.

4.2. Crop protection formulated product markets

4.2.1. Relevant product markets

(306) According to the Commission notice on the definition of relevant market for the purposes of Community competition law (the "Market Definition Notice"), a relevant product market comprises all those products which are regarded as interchangeable or substitutable by the consumer, by reason of the products' characteristics, their prices and their intended use (paragraph 7). According to paragraph 13 of the Market Definition Notice, "firms are subject to three main sources or competitive constraints: demand substitutability, supply substitutability and potential competition".

(307) The main criteria for the definition of the relevant market is demand side substitutability according to which "the range of products which are viewed as substitutes by the consumer" should be in the same relevant market (paragraph 15 of the Market Definition Notice).

(308) Crop protection formulated products consist of finished products that are mixtures of active and inert ingredients (such as solvents, fillers, and adjuvants) ready to be applied for their respective purpose. These latter ingredients aim at making the AIs more stable, effective, or safer or easier to apply. The farmers are the ultimate consumers of formulated products.

(309) The Parties argue in their response to the Statement of Objections that the segmentation of markets for crop protection formulated product by crop-pest combinations is too narrow from a demand-side and supply-side perspective and

108

suggest to consider instead ranges of pests and crops.This is because while products are differentiated as regards driving pests, all of them cover a range of pests, and the farmer will take that into account when deciding between different products.

(310) The Commission considers instead that farmers buy a formulated crop protection product to address their particular needs. They will choose based on the crop, pest(s), timing, etc., they want to target. In line with the Market Definition Notice, this implies a narrow relevant product dimension consisting of a crop/pest combination. In fact, in general, a formulated product that applies to a given crop/pest combination

107 "Off-patent" in this context means AIs which do not enjoy any more patent, data or other types of protection.

46

is not substitutable from the farmer point of view with another product that applies to a different crop/pest combination.

(311) Moreover, all formulated products have a label that typically indicates AIs, formulation, permitted use crops, pests targeted, options for compatible tank mixing, crops that can be sown after a crop treated with the product, safety instructions, etc. As explained by a technical institute for herbicides "products are authorized for a

109

specific use, that is to say for a given type of seeds and a precise crop". This

110

implies that off-label use or misuse by market participants is typically uncommon and potentially subject to fines.

111(312) Based on recitals (306) to (311) and according to previous Commission decisions, from a demand-side perspective, formulated products can be distinguished notably by target crop, target pest, and other specific features as elaborated in Sections V.6.3 to V.6.6 more for herbicides, insecticides, nematicides and fungicides.

(313) Supply side substitutability may be considered for product market definition when "suppliers are able to switch production to the relevant products and market them in the short term without incurring significant additional costs or risks in response to small and permanent changes in relative prices" (paragraph 20 of the Market Definition Notice). According to paragraph 14 of the Market Definition Notice "the competitive constraints arising from supply side substitutability other than those described in paragraphs 20 to 23 […] are in general less immediate and in any case require an analysis of additional factors. As a result such constraints are taken into account at the assessment stage of competition analysis".

(314) As explained in Section V.1.4, the discovery and production of a new formulated product involves high costs (above USD 250 million) and a long period of time (above 10 years). This implies that a supplier producing a given formulated product targeting a given crop/pest combination cannot start competing in the short term and without incurring in significant costs for another crop/pest combination, if the label of the product does not allow for this use.

(315) Internal documents of the Parties indicate that they look at their formulated products as targeting markets based on crop/pest segmentations. For instance, Dow's internal

112

documents […].

(316) The Commission notes that some crop protection products of the Parties are registered for several crop/pest combinations. However, this does not imply that the relevant product market should include all those combinations since products targeting several crop/pest combinations only partially overlap among them.

109 Agreed non-confidential minutes of a call with an institute, 13 September 2016 (ID8557). Courtesy translation from French: "les produits sont autorisés pour un "usage" particulier, c'est-à-dire sur un type de mauvaises herbes et pour une culture précis".

110 Agreed non-confidential minutes of a call with a competitor, 6 September 2016 (ID7383). Similarly a Dow regulatory filing document states that they "do not foresee misuse of spinetoram in Europe in vegetables and ornamental crops as growers are highly knowledgeable and responsible users of pesticides". Dow's internal document, "Zonal Biological Assessment Dossier – EU Central Zone", file name "DAS-10201807.docx" (ID6696-15304).

111 See for instance Commission Decision in Cases M.1806 – AstraZeneca/Novartis (2000) and M.2547 – Bayer/Aventis Crop Science (2002).

112 Parties' response to Article 11(3) decision of 3 November 2016.

47

(317) For instance, consider the case of an herbicide A targeting cereals and oilseed rape, and an herbicide B targeting cereals and corn. Those two formulated products partially overlap for cereals. If one includes on the same relevant market all products targeting cereals, oilseed rape and corn, one would be incorrectly assuming that formulated product B would be an option for a farmer with a need for oilseed rape. In other words, by widening the relevant market to include several crops one would be overestimating the competitive pressure of formulated products which cannot be applied for some of those crops.

(318) In spite of relevant product markets corresponding to a crop/pest combination, from a practical point of view, it will be impossible to assess so many narrow markets. Therefore, in order to conduct a meaningful assessment, the Commission will, in Sections V.6.3 to V.6.6, consider sometimes groupings of crop/pest combinations. Such groupings can be seen in some of the Parties internal documents, as well as in reports from specialised consultancies such as Phillips McDougall. However, these aggregate groupings may hide relevant product markets where the Parties shares are higher than the grouping shares.

(319) Based on recitals (306) to (318), the Commission considers that the relevant product market for formulated products correspond to segmentations by crop/pest combination. Throughout the Decision the Commission will try to identify these markets and to make the assessment at that level. For the sake of convenience grouping shares will also be presented, bearing in mind their limitations.

4.2.2. Relevant geographic markets

(320) In the past, the Commission assessed the formulated products markets as national in scope, since their authorisation is still regulated at the national leveland each Member State remains responsible for maintaining specific national data requirements.

113

(321) The Parties, however, claim that the situation has evolved significantly over the past 15 years and that markets are now Union-wide or at least regional, notably for the following four reasons.

(322) First, they argue that conditions of competition do not substantially differ across the EEA and that the main large, international manufacturers of crop protection products are active throughout the Union/EEA.

(323) Second, they claim that products are principally identical in all countries of the Union. Parallel imports are customary and meet minimal barriers today.

(324) Third, they submit that distributors and co-operatives on the demand side have created pan-European alliances. Examples include 'Novafield', the alliance of European distributors that covers 12 EEA Member States, and 'Euro-A-Pro', the pan-European joint venture.

(325) Fourth, the Parties point out that Regulation 1107 recognises the existence of three geographic zones (North, Centre and South) in the Union, which have comparable agricultural, plant health and environmental conditions.

113 See Commission Decision in Cases M.2547 – Bayer/Aventis Crop Science (2002), recital 23-27; M.1806 – AstraZeneca/Novartis (2000), recital 82-83; and M.1932 – BASF/American Cyanamid (AHP) (2000), recital 29-31.

114 Form CO, part B.I, paragraph 85.

48

(326) Nonetheless, in line with the Commission’s previous practice, the Parties have provided market data and a competitive assessment at the national level.

(327) Contrary to the arguments of the Parties, the Commission's in depth investigation does not confirm that the downstream markets for formulated products are wider than national. This is notably for the following reasons.

(328) First, the Parties' direct customers on these markets are distributors, some of which may be active in several Member States. However, these distributors' portfolio of products is adapted to varying demands of their customers in different Member States. Customers' needs, habits, and preferences appear to be dependent on geography and differ across EEA countries. Producers' design, brand products as well as their prices vary according to country. Customers of the Parties who replied to the Commission's questionnaires indicated that the relevant geographic market is national in scope, with a majority of respondents purchasing crop protection products

115

at the local or national levels.

(329) Second, data on prices submitted by the Parties indicates national differences in price levels and evolutions for the same products.

(330) Third, authorisations have remained national. In practice, the zonal system described in Regulation 1107 has not been implemented widely yet. This is indicated in some

116

internal documents of the Parties,as well as underlined by the European Crop

117

Protection Association ("ECPA").Market participants also underlined the limited application of the zonal system. For instance, "[s]ometimes formulations are authorised in other Member States but, despite theoretical mutual recognition, not

118

authorised in the UK".

(331) One competitor also stated that "geographic markets remain national, due to the regulatory process. This is because although registration for active ingredients is done at EU level, what counts is the registration at national level for the formulated products, which is done by crop and type of pest. While in theory there is a system of zones, the concept of mutual recognition of products has a negligible effect in practice. Parallel imports are also negligible (less than 1%). There are also differences in climate. Distribution channels vary by country, and customers are typically organised at national level due to language and the need for customer proximity. Countries also vary in their regulatory approach, for instance the

119

Netherlands place significant emphasis on ground water protection".Parallel trade exists but is limited, as will be detailed in Section V.6.2.3.

(332) Based on recitals (320) to (331), the Commission considers that the markets for formulated products are national in their geographic scope.

115 Questionnaire to Crop Protection Customers (Q1), question 57.

116 See for instance DuPont's internal document "EMEA R&D Towntalk Stine", September 2015, file name "DOC-000000199.pdf" (ID1329-199), and Dow's internal document "Presentation to CPMF – Inatreq 2015 Update", December 2015, slide 36, file name "Case M.7932 DowDuPont_Key Issues (RD)_Presentation_June 9.pdf" (ID1034).

117 See ECPA position paper ECPA Position on the future revision of Regulations 1107/2009 and 396/2005, July 2015.

118 Agreed non-confidential minutes of a call with a farmers' association, 18 March 2016 (ID8254). See also Questionnaire to Crop Protection Competitors (Q2), question 28, "Mutual recognition does not work well because of lack of uniformed document exchanges between member’s states" (ID4027).

119 Agreed non-confidential minutes of a call with a competitor, 5 April 2016 (ID8248).

49

4.3. Technology markets

(333) As explained in recital (304), R&D companies may sell their technology to competing crop protection players through the sale or licensing of AIs (which may include IP rights, data, registration, etc.). These "sales" can also take the form of swaps between R&D companies through which one company obtains access to a given third-party AIs in exchange of giving access to its own AIs.

120 (334) In past decisions,the Commission found that each AI constitutes a separate product market, although recently it left open whether for off-patent ingredients broader markets exist including all AIs within the same class of molecules.The Parties argue that for the purposes of the Transaction the exact market definition can be left open.

121

(335) The Commission investigation confirmed that there is active trading of AIs between crop protection players. Hence, the Parties have so-called "technical sales" of various AIs to other companies. Also, the majority of the relevant respondents to the Commission questionnaires stated that they (cross-)license or purchase AIs from other companies.

122

(336) The relevant actors on the demand side for the sale and licensing of AIs are different from those in downstream markets for formulated products. Customers of these products are other producers of crop protection products. Also the relevant actors on the supply side are often different. There are fewer upstream developers and suppliers of AIs than downstream producers of finished formulated products. It is also possible to distinguish those suppliers (originators) who license or sell molecules they brought to the market, as is the case for the global integrated crop protection companies, from chemical companies that in some instances have manufacturing abilities of off-patent AIs.

(337) In line with the Commission's Guidelines on the application of Article 101 of the Treaty on the Functioning of the European Union to technology transfer

123

agreements(the "Technology Transfer Guidelines") the Commission considers that the Parties are active and compete on upstream technology markets.

(338) The Technology Transfer Guidelines define technology as an input, which is integrated either into a product or a production process. As such, technology right licensing can affect competition both upstream in input markets and downstream in output markets: "[f]or instance, an agreement between two parties which sell competing products downstream and which also cross license technology rights relating to the production of these products upstream may restrict competition on the downstream goods or services market concerned. The cross licensing may also restrict competition on the upstream market for technology and possibly also on other upstream input markets. For the purposes of assessing the competitive effects

120 See Commission Decision in Cases M.6141 – CNAC/Koor Industries/Makhteshim Agan Industries (2011) and and M.2547 – Bayer/Aventis Crop Science (2002).

121 See Commission Decision in Case M.6141 – CNAC/Koor Industries/Makhteshim Agan Industries (2011).

122 Questionnaire to Crop Protection Competitors (Q2a and Q2b), question 63.

123 OJ C 89, 28.3.2014, pages 3-50.

50

of licence agreements it may therefore be necessary to define the relevant product

124

market(s) as well as the relevant technology market(s)."

(339) In line with the assessment in the Article 6(1)(c) Decision, however, the Commission finds that it is possible to draw a distinction between the licensing and sale of AIs (including swap) involving technology from the mere bulk supply of AIs which, besides being off-patent, do not involve any transfer of technology.

(340) As regards geographic markets, the Parties submit that the geographic scope for AI markets is at least EEA-wide, if not worldwide.

(341) Because the Transaction would not raise competition concerns at the level of the technology markets or the supply of AIs, the exact scope of the product and geographic market for the sale and licensing of AIs can be left open.

4.4. Innovation spaces

4.4.1. Product dimension

(342) As explained in recital (304), R&D companies engage in innovation efforts to discover and develop new AIs.

(343) In the US, the antitrust agencies have used the concept of innovation markets. The recent proposal for "Antitrust Guidelines for the Licensing of Intellectual Property" issued by the DoJ and FTC (Federal Trade Commission), distinguishes between

125

goods markets, technology markets and research and development market.

(344) Goods markets correspond to the markets where the final product is sold. In the current case it corresponds to formulated products.

(345) Technology markets "consist of the intellectual property that is licensed (the ‘licensed technology’) and its close substitutes—that is, the technologies or goods that are close enough substitutes to constrain significantly the exercise of market power with respect to the intellectual property that is licensed".This is similar to the concept of technology market defined in Section V.4.3.

126

(346) Research and development markets "consists of the assets comprising research and development related to the identification of a commercializable product, or directed to particular new or improved goods or processes, and the close substitutes for that research and development".

127

(347) According to the Technology Transfer Guidelines, "innovation is a source of potential competition which must be taken into account when assessing the impact of the agreement on product markets and technology markets. In a limited number of cases, however, it may be useful and necessary to also analyse the effects on competition in innovation separately. This is particularly the case where the agreement affects innovation aiming at creating new products and where it is possible at an early stage to identify research and development poles. In such cases it can be analysed whether after the agreement there will be a sufficient number of

124 Technology Transfer Guidelines, paragraph 20. 125 DoJ and FTC, "Antitrust Guidelines for the Licensing of Intellectual Property", 12 August 2016. 126 DoJ and FTC, "Antitrust Guidelines for the Licensing of Intellectual Property", 12 August 2016. 127 DoJ and FTC, "Antitrust Guidelines for the Licensing of Intellectual Property", 12 August 2016.

51

competing research and development poles left for effective competition in innovation to be maintained".

128

(348) When considering both the downstream product markets and the upstream technology markets, innovation should not be understood as a market on its own right, but as an input activity for both the upstream technology markets and the

∗

downstream technologymarkets. This however does not prevent the Commission to assess the impact of the Transaction at the level of innovation efforts by the Parties and its competitors.

(349) First, the assessment of innovation competition requires the identification of those companies which, at an industry level, do have the assets and capabilities to discover and develop new products which, as a result of the R&D effort, can be brought to the market.

(350) Secondly, it is also relevant to identify and analyse those spaces in which innovation competition occurs in the crop protection industry. The R&D players do not innovate for all the product markets composing the entire crop protection industry at the same time. They also do not innovate randomly without targeting specific spaces within that industry. When setting up their innovation capabilities and conducting their research R&D players have specific discovery targets (the strong focus at early stages on specific target pest and crops is testified by a number of internal documents detailing the Parties' innovation targets in very specific terms, see also Section V.8.6.1 for a more detailed explanation and examples).

(351) A given discovery target is based on lead crops and lead pests and may thus comprise AIs that can be used in several downstream formulated product markets (for example chewing Lepidopteran insecticides, broadleaf herbicides). The spaces where innovation competition takes place are thus broader than an individual downstream crop protection market, but are nonetheless small. In fact, in light of increasing regulatory hurdles, which require crop protection products to be ever more selective, the innovation spaces in the crop protection industry are getting ever smaller: the innovation output tends to be confined to ever narrower spaces from which it is more difficult to adapt the innovation to other purposes.

(352) In conclusion, in order to assess innovation competition, the Commission will both consider metrics of innovation taking place at industry level, as well as innovation taking place in spaces consisting of groupings of crop/pest combinations (as will be defined specifically for the areas where the Parties overlap in Section V.8.8).

4.4.2. Geographic dimension

(353) On the one hand, AIs are typically developed globally for sales in several regions, depending on the crops targeted. R&D companies also pursue global strategies for patents.

(354) During the Commission's market investigation, market participants overall indicated that innovation in the crop protection industry takes place on a global level. Respondents to the Commission's questionnaires notably indicated that (i) major

128 Technology Transfer Guidelines, paragraph 27. ∗ Should read: product.

52

industry issues are of global importance (for instance whiteflies, moths, thrips), and

129

(ii) new solutions are predominantly developed by international companies.

(355) On the other hand, because innovation is developed aiming at innovation spaces that can be defined by groupings of crop/pest combinations, and given the specificities of different regions of the world as regards the main crop and pests, the competitive environment may differ significantly among regions like EEA, North America, Asia, South America, etc.

(356) For instance, some companies, like the Japanese, primarily seek herbicides, fungicides, and insecticides for rice and other crops present in their geographic area and only occasionally these molecules have also application in European agriculture.

(357) Weather and soil conditions also differ from region to region and thus companies that want to have a strong presence in a given region need to have assets for testing in those regions.

(358) Moreover, there are differences in the regulatory requirements of different regions of the world. In particular, in the Union the registration requirements are tougher than in most of the other regions.

(359) As explained in Section V.1.4.3, according to Union legislation, crop protection products must undergo two regulatory steps before they can be placed on the market. In the first step, the AI must be approved by the Union, which implies the obligation to submit extensive test data. In the second step, formulated plant protection products that contain the AI (or a mix of several AIs) must be authorised in each Member State where they will be employed.

(360) According to the Parties "a number of regional factors tend to impact the focus of companies’ innovation efforts, relating to the key crops grown in a region, the corresponding pest, weed and disease control requirements, local environmental conditions, the level of control offered by existing products on the market, and local

130

regulatory requirements".

(361) Based on recitals (353) to (360), the Commission considers the innovation spaces to be global, with strong differentiation between the different regions, or at least EEA-wide.

5. M ARKET SHARES : SOURCES AND METHODOLOGIES

5.1. Market shares relevant for product competition

(362) The Commission illustrates its assessment of competition at the formulated product level with shares computed at the level of crop/pest combination groupings, at the national level. The Commission also uses shares for these groupings at the EEA level for context, as being informative of the strength of market players at the level of their portfolio of AIs, as well as shares of R&D suppliers, as being informative of the relative strength of suppliers bringing new AIs in the market.

129 Questionnaire to Crop Protection Stakeholders and Testing Partners (Q3a and Q3b), question 34. 130 Form CO.

53

5.1.1. The Parties' markets shares provided in the Form CO

(363) As from the pre-notification stage and during part of the Phase I investigation, the Commission and the Parties had numerous and extensive discussions regarding the data and the methodology used by the Parties for the computation of the market shares in the Form CO, as well as regarding the various results obtained through this process. In that context, the Commission issued at least five pre-notification and five post-notification requests for information partly or entirely devoted to issues related to market shares.

131

(364) The methodology note attached to the Form CO provided limited information on the Parties' data and methodology used for the computation of markets shares provided in the Form CO, in particular in light of the importance of market shares in a case with, potentially, hundreds of affected relevant (downstream) markets for formulated products.

132

(365) In responding to the Commission's request to "provide an updated methodology note […] explaining in detail the preparation of all global data and indicating sources for their respective data point, year, unit",the Parties indicate that "the Parties do not consider it necessary to provide a long description of this incredibly simple and easy-to-understand methodology".The Commission finds it apparent from [details on communication regarding market share methodology].

133

134

(366) Overall, the Commission was not [assessment why the Commission considers that it had required information on market share and methodology only 17 days after the Transaction was notified to the Commission].

135

(367) [Details on market share methodology].

(368) In light of the limited time available in a Phase I investigation and of the elements described in recitals (363) to (367), in particular given the difficulties faced by the Commission, despite its efforts, to be provided with sufficient level of information on the data selection and on the methodology used by the Parties, the Commission considered necessary to compute market shares for formulated products.

(369) The Commission notes that, on 19 August 2016, at a time at which the Parties were aware of the Commission's data and methodology used for computing market shares, several Dow's employees involved in the Commission's procedure had the following exchange. One stated: "[Information regarding discussions on market share methodology]." To which another one responded (emphasis in the original email):

136,137

"[Information regarding discussions on market share methodology]."

131 For more details, see Annex 3, Section 2.1.

132 For more details, see Annex 3, Sections 2.2, 2.3 and 2.4.

133 Commission's request for information dated 6 July 2016, question 2, reiterating question 2 of the Commission's request for information RFI 16.

134 Parties' response to the Commission's request for information dated 6 July 2016, question 2 (reiterating question 2 of RFI 16), dated 11 July 2016 at 1:01 (ID5025).

135 For more details, see Annex 3, Section 2.5.

136 Dow's internal document, entitled "RE: Orion -- Insecticides & Nematicides", file name "DAS-USPRIV-40030819.pdf" (ID7081-2531).

137 Agrowin is a database used internally by Dow and DuPont and which is also used by the Commission, in part for the purpose of computing market shares for downstream formulated products.

54

5.1.2. Data used by the Commission for the computation of markets shares on the (downstream) relevant markets for formulated products

(370) For the purpose of the calculation of market shares and related figures, such as concentration measures, for the purpose of the assessment of the present Transaction on the (downstream) relevant markets for formulated products, the Commission relied on databases provided by a third party, Agrobase-Logigram,through its product called Agrowin.

138

(371) The Commission understands that Agrowin is widely used and recognised in the agrochemical industry. The Parties are both clients of Agrowin and both use it internally in ordinary course of business for the purpose of estimating market size and their competitors' and their own market position.

139

(372) Agrowin defines itself as the "crop protection and seeds use database", which compiles information on the use of pesticides and seeds from multiple sources: "Agrowin is better labelled as 'use data' rather than 'sales data'. Indeed, Agrowin tracks quantities and prices of products used by farmers, which only partly correspond to the products sold by suppliers." In a nutshell, the added value of Agrowin is to standardise the primary data obtained from panel companies and to provide its users with a standardised set of variables valid across the various primary data.

5.1.3. The Commission's markets shares and concentration measures used in the Decision for the purpose of the assessment of the present Transaction on the (downstream) relevant markets for formulated products

(373) As explained in Section V.4.2.1 detailing market definition principles for relevant crop protection formulated product markets, the Commission considers that, in general, the relevant product markets for formulated products correspond to segmentation by crop/pest combinations. Nevertheless, the Commission decided, in order to be in a position to conduct a meaningful assessment of the Transaction, to group crop/pest combinations into "markets".

(374) The market shares and the concentration measures computed by the Commission therefore reflect such grouping and use different methodologies to allocate sales value to each of these groupings. Whenever no data are available from Agrowin, the Commission's assessment relies on figures provided by the Parties.

(375) The Commission uses market shares in this Decision for the purpose of assessing the likely effects of the Transaction on downstream formulated products (at the level of market groupings) though, because of the grouping of markets, these market shares are meant to provide an informative approximation at an aggregated level of relevant markets.

(376) Moreover, as shown in Section V.1.5, suppliers present in a given relevant market can be divided in different categories, among which R&D suppliers, which have innovated and continue to innovate by identifying new AIs, and generic suppliers, which copy inventions brought by R&D players when their related patent protections expire.

(377) While, in some (downstream) relevant markets, generic suppliers can supply a significant share of the volumes sold in that market, these suppliers will not bring any new products, that is products with new AIs, but will mainly provide farmers with combination of already invented AIs. As a consequence, the presence and relative strength of R&D players provide contextual elements to be factored in the Commission's assessment of a specific relevant (downstream) market.

(378) Finally, the Commission also calculates several concentration measures, including the Herfindahl-Hirschman index ("HHI").

5.2. Market shares relevant for innovation competition

(379) As regards innovation competition the Commission will focus on measures of innovation output.

(380) In an Advocacy Paper, the Parties propose to calculate shares on innovation and assess the R&D companies' strength based on R&D input measures, such as R&D expenditure, and output measures, such as patent applications count and number of AIs launched. This approach is further proposed in the report of one of the Parties' experts.

(381) In this report, the Parties' expert also discussed why the direct metrics proposed would be superior to the Commission's methodology as set out in the Article 6(1)(c) Decision.

(382) On R&D expenditure, the report acknowledges that this is an input, but because there is uncertainty in innovation, looking strictly at ex post measures of output can be misleading.

(383) As regards output measures, the report indicates that the number of patent applications and the number of new AIs introduced are both more direct measures of innovation than downstream figures used by the Commission, and contend they remain more direct and relevant than the preliminary approach retained by the Commission in its Article 6(1)(c) Decision.

(384) As regards R&D expenditure, the Commission maintains that such a measure is an input measure, which does not account for a number of factors that are relevant for the assessment.

(385) The Parties' expert contends that, for instance, the Phillips McDougall report has available data for the top 35 firms, and other firms than leading innovators collectively have a billion dollar worth of crop protection expenditure. As it will further be explained in Section V.8.6.3, however, not all these firms have comparable capabilities nor do they necessarily focus on R&D aimed at discovery or development of new AIs. Also generic companies often report figures for an R&D budget, focused for instance on new formulations of existing AIs. Moreover, output measures are better suited to describe relevant capabilities and expertise of crop protection companies and their track record of bringing innovation in crop protection markets.

(386) In the Sections V.5.2.1 and V.5.2.2, the Commission will discuss which output measures it regards as most representative of the companies' position and strength at innovation level.

5.2.1. The "patent shares" used by the Commission in the Decision for the purpose of the assessment of the present Transaction on innovation competition

(387) One of the relevant activities within crop protection R&D is, as explained in Section V.1.4.1, the discovery of new molecules (AIs), which are normally patented by the discovering company. While different companies have different patenting strategies, the analysis of the patent portfolio of crop protection companies can be a metric to assess their strength at the discovery level.

(388) The Parties suggest the use of R&D spend and patent counting to measure the competitive strength of firms involved in research for crop protection. As regards the R&D spend, the Commission considers that this measure captures, to a certain extent, the R&D effort of a company but that it may not provide insight into the output of the R&D organisation. As regards the patent counting, a purely quantitative count of patents does not take into account that some companies do not patent as extensively as others and in particular file patents at a later stage of the research which are generally of better qualities, and does not reflect the significant variation in qualitative significance of these patents. Actually, it is well-established in the economic literature that simple patent counts are not informative about innovative output. Simple patent counts are rather associated with the input side of the innovative process, primarily with contemporaneous R&D expenditures.

(389) Instead, the economic literature suggests using patent citations to measure the importance of a patent, that is to say the quality or value of a patent. This amounts to count the number of times each patent has been cited in subsequent patents to compute, on that basis, a citation-based index as a measure of innovative output. One important finding of the economic literature is that citation-based indexes are informative on the technological importance of patents.

(390) For that purpose, the Commission uses several data sources, submitted by the Parties at the request of the Commission: patent data submitted by Dow, which track patents related to crop protection published by industry competitors over the period 1990-2015, and DuPont, for its own patents. Dow keeps a single database of competitor crop protection patent filings.

(391) Moreover, the Commission used data and metrics from a service provider, PatentSight. PatentSight defines itself as "providing reliable and relevant benchmarks of patent portfolios to key decision makers in technology companies. To derive reliable insights it is key to measure actual patent strength rather than mere patent filings. This is achieved by the Patent Asset Index™ methodology [… which] considers both the quantity and the quality of patents in a portfolio."

(392) PatentSight is used by Dow in its internal analyses of patents in the crop protection industry. It is recognised in the industry as a reference as regards patents. For example, PatentSight's website displays the following quote from the Executive Vice President and CTO of Dow: "[t]he Patent Asset Index provides an accurate view of the impact and efficiency of an enterprise’s investment in innovation."

(393) The citation-based indicator is based essentially on the number of worldwide citations received from later patents, as reported and adjusted for age, patent office practices and technology fields by PatentSight. Citations to a given patent come from any subsequent patents, including from patents owned by the same firm as the one holding the cited patent (internal citations, also called self-citations).

(394) The difficulty in interpreting internal citations is that they increase mechanically with the size of the patent portfolio because the more patents a firm has, the higher is the probability that a citation from a new patent will be given by a patent that it already owns. Therefore, firms with a larger portfolio size have mechanically a larger number of internal citations, regardless of whether internal citations are indicative of the value of a patent. The link between internal citations and the quality of a patent is likely to be weakened with the size of the patent portfolio due to this mechanical effect.

(395) On that basis, the Commission considers external citations to measure the quality of a patent and uses total citations, that is citations including internal and external citations, as a sensitivity analysis.

5.2.2. The "new active ingredients" shares used by the Commission in the Decision for the purpose of the assessment of the present Transaction on innovation competition

(396) In addition, the Parties suggest that the number of new AI launches is a good measure of innovation output. For instance, in a presentation provided at a meeting held on 19 July, the Parties provided an overview of launches of new AIs in the period 2005-2016, showing a number of other players launching new AIs.

(397) The Commission maintains that the total number of new AIs may not account for their quality and commercial significance.

(398) AIs are likely to differ greatly in their quality, as suggested also by the analysis of patent data (see Annex 1). In order to measure for the effective significance of the R&D output of crop protection companies, one must use a measure which may keep into account the commercial success of the AIs launched by crop protection companies.

(399) Gauging the commercial success of AIs allows controlling for an important factor affecting the development capabilities of crop protection companies, that is their capability to develop an AI on a large scale and to distribute it to enable its commercial success on the market.

(400) The Parties themselves seem to implicitly acknowledge that the commercial success of an AI is an important factor when assessing the relevance of the innovation effort of crop protection companies. For instance, the Parties also propose an analysis of the strength of a number of crop protection companies' innovation effort as measured by the turnover generated by the AIs they launched on the market ("blockbuster AIs").

(401) Moreover, in case of an eventual reduction of innovation by a company introducing AIs with a large scale, this could hardly be compensated by the introduction of new AIs without any commercial relevance in the market for formulated products. Thus weighting competitors innovation activity (that is the number of AIs) by the commercial success of the AIs launched in recent years is an important element to measure their ability to counteract any reduction on innovation by the Parties.

(402) Therefore, on the basis of information submitted by the Parties at the request of the Commission, which contain information on AIs introduced since 1995, the Commission provides an assessment of the innovation output by weighing the importance of the new AIs introduced by R&D players according to the 2015 turnover generated by each AI in the global market as well as in Europe.

6. COMPETITIVE ASSESSMENT: PRODUCT AND PRICE COMPETITION

6.1. Test under the Merger Regulation and the Horizontal Merger Guidelines and theory of harm in this case

6.1.1. Legal basis

(403) In this section the Commission will focus on the assessment of the effects of the Transaction on product and price competition as opposed to innovation competition.

(404) Article 2 of the Merger Regulation stipulates that "[a] concentration which would significantly impede effective competition, in the common market or in a substantial part of it, in particular as a result of the creation or strengthening of a dominant position, shall be declared incompatible with the common market." In its appraisal,

the Commission is required by the Merger Regulation to take into account, among others, the need to maintain effective competition in light of the structure of the markets concerned, the market position of the undertakings concerned and their economic and financial power, as well as the development of technical and economic progress provided that it is to consumers' advantage and does not form an obstacle to competition.

(405) Recital (25) of the Merger Regulation clarifies that the language of Article 2 is meant to encompass the appraisal of the effects of concentrations in oligopolistic markets, and in particular those that may significantly impede effective competition by the elimination of important competitive constraints that the merging parties had exerted upon each other as well as by a reduction of the competitive pressure on the remaining competitors.

(406) Recital (28) of the Merger Regulation clarifies that the Commission may publish guidance aimed at providing a sound economic framework for the assessment of concentrations, with a view to determining whether or not they may be declared compatible with the internal market.

(407) In this context, the Horizontal Merger Guidelines provide further guidance on the underpinning concepts of the Commission's assessment.

(408) The Transaction consists of the merger of two crop protection R&D companies active in a large number of crop protection product markets in the EEA, with significant combined market shares in several of these markets.

6.1.2. Focus on non-coordinated effects

(409) Horizontal mergers, such as the Transaction, may significantly impede effective competition within the meaning of Article 2(3) of the Merger Regulation in two ways – through non-coordinated and/or coordinated effects. This Decision focusses on non-coordinated effects, characterised as "eliminating important competitive constraints on one or more firms, which consequently would have increased market power, without resorting to coordinated behaviour".

(410) The "most direct effect" of such a merger will "be the loss of competition between the merging firms", in this case the loss of competition between Dow and DuPont.

(411) The Horizontal Merger Guidelines indicate that such a horizontal merger can also lead to the reduction of competitive pressure on other competitors: "[n]on-merging firms in the same market can also benefit from the reduction of competitive pressure that results from the merger, since the merging firms' price increase may switch some demand to the rival firms, which, in turn, may find it profitable to increase their prices."

(412) Paragraph 25 of the Horizontal Merger Guidelines distinguishes two ways in which non-coordinated effects result in a significant impediment of effective competition.

(413) Firstly, "[g]enerally, a merger giving rise to such non-coordinated effects would significantly impede competition by creating or strengthening the dominant position of a single firm".

(414) Secondly, "mergers in oligopolistic markets involving the elimination of important competitive constraints that the merging parties previously exerted upon each other together with the reduction of competitive pressure on the remaining competitors may […] also result in a significant impediment of competition."

(415) These two alternatives are not mutually exclusive, since a merger may both eliminate important competitive constraints and in addition and at the same time create or strengthen a dominant position of a single firm.

(416) In this Decision, many product and geographic markets are at stake when looking at crop protection product markets in the EEA, and various situations arise in various markets with the Transaction. Depending on the situation, the Transaction will be assessed under the first or second or both alternative(s) in Sections V.6.3 to V.6.6.

(417) In some instances, a forthcoming product, just launched or close to be launched, will not appear in the current existing market shares, for instance Dow's Arylex herbicide, which is expected to be a successful product on the market. In these cases the Commission will examine the potential competition exerted by these products.

(418) In fact, paragraph 58 of the Horizontal Merger Guidelines indicates that "[c]oncentrations where an undertaking already active on a relevant market merges with a potential competitor in this market can have similar anti-competitive effects to mergers between two undertakings already active on the same relevant market and, thus, significantly impede effective competition, in particular through the creation or the strengthening of a dominant position".

(419) Paragraph 60 of the Horizontal Merger Guidelines describes two basic conditions for a merger with a potential competitor to have significant anti-competitive effects: "[f]irst, the potential competitor must already exert a significant constraining influence or there must be a significant likelihood that it would grow into an effective competitive force. Evidence that a potential competitor has plans to enter a market in a significant way could help the Commission to reach such a conclusion. Second, there must not be a sufficient number of other potential competitors, which could maintain sufficient competitive pressure after the merger".

6.1.3. Factors taken into account by the Commission for the assessment of whether significant non coordinated effects are likely

(420) Paragraphs 27 to 38 of the Horizontal Merger Guidelines list a number of factors which may significantly impede effective competition by influencing the ability and incentives of the merging parties to increase prices or affect other parameters of competition. Paragraph 26 of the Horizontal Merger Guidelines states that (i) taken separately these factors do not have to be decisive, (ii) not all these factors need to be present for such effects to be likely and (iii) the list should not be considered as exhaustive.

(421) First, the Commission will present the Parties' market shares in the various crop protection product markets.

(422) According to paragraph 27 of the Horizontal Merger Guidelines, "the larger the market share, the more likely a firm is to possess market power". And the larger the addition of market share, the more likely it is that a merger will lead to a significant increase in market power.

(423) As will be seen in Section V.6, the Parties have very large combined market shares in several crop protection product markets. The Transaction also leads to sizeable

market share additions. Given the very large pre-existing and/or combined market shares in some of the markets, the Commission will examine whether one of the Parties has a degree of market power pre-Transaction and whether that market power will be enhanced as a result of the Transaction.

(424) Second, in line with paragraphs 25 and 28 to 30 Horizontal Merger Guidelines the Commission will examine whether the Transaction eliminates a significant and close competitor.

(425) Under paragraphs 25 and 28 to 30 of the Horizontal Merger Guidelines, and in line with Commission practice and economic theory, it is sufficient for competitive harm to arise that the merging parties are significant and close competitors. They do not necessarily have to be the closest competitors. This is all the more true in cases where the merger creates or strengthens a new market leader with large or very large market shares.

(426) The assessment will examine the degree of substitutability between the Parties' various crop protection products, and the importance of rivalry between them for competition in the market. "[t]he higher the degree of substitutability between the merging firms' products, the more likely it is that the merging firms will raise prices significantly."

(427) As will be discussed in Section V.6, the investigation shows that Dow and DuPont are close competitors in many markets where they have overlapping products.

(428) Third, according to paragraph 31 of the Horizontal Merger Guidelines "[c]ustomers of the merging parties may have difficulties switching to other suppliers because there are few alternative suppliers. … The merger may affect these customers' ability to protect themselves against price increases."

(429) This Decision will assess the availability of effective alternatives, taking into account differing demand characteristics and product offers in the various crop protection product markets.

(430) As explained in more detail in Section V.6.2 the Commission's main findings are that generics do not constitute a strong competitive constraint on the Parties and parallel trade is limited in terms of countries, product range and volumes involved. Moreover, for many markets the number of effective available alternatives is currently reduced and some of these alternatives may even be taken off the market in the near future.

(431) Fifth, according to paragraph 38 of the Horizontal Merger Guidelines: "effective competition may be significantly impeded by a merger between two important innovators, for instance between two companies with ‘pipeline’ products related to a specific product market. Similarly, a firm with a relatively small market share may nevertheless be an important competitive force if it has promising pipeline products."

(432) This is particularly the case of markets where the Parties have products just launched or about to be launched that are soon expected to conquer a significant market share from rivals.

(433) The relatively high degree of closeness of competition between the Parties together with the very high barriers to entry, large combined market shares, and limited number of alternatives are important drivers of the likelihood of anticompetitive price effects in this case.

6.1.4. Structure of the assessment

(434) In light of the foregoing, the Commission's assessment will be structured as follows. In Section V.6.2 the Commission will start by analysing the general arguments of the Parties applicable to all crop protection products, namely (i) whether generics are a strong competitive constraint on the Parties (Section V.6.2.1); (ii) whether significant buyer power exists to compensate for any potential added market power from the Parties (Section V.6.2.2); (iii) whether parallel trade is an added competitive constraint, which both compensates for any potential added market power from the Parties and broadens the geographic scope of product markets beyond single countries (see Section V.6.2.3) and (iv) whether the switching data shows that the Parties are not close competitors (see Section V.6.2.4 as well as Annex 6). Section V.6.3 will assess the impact of the Transaction as regards the horizontal overlaps in herbicides. Section V.6.4 will assess the impact of the Transaction as regards the horizontal overlaps in insecticides. Section V.6.5 will assess the impact of the Transaction as regards the horizontal overlaps in nematicides. Section V.6.6 will assess the impact of the Transaction as regards the horizontal overlaps in fungicides.

6.2. General arguments applicable to all crop protection products

(435) In contesting the existence of horizontal concerns in the crop protection markets, the Parties make a number of similar arguments for each type of product, namely that:

(1)generic products are a strong competitive constraint on the Parties;

(2)significant buyer power exists to compensate for any potential added market power from the Parties;

(3)parallel trade is an added competitive constraint, which both compensates for any potential added market power from the Parties and broadens the geographic scope of product markets beyond single countries; and

(4)the switching analysis shows that the Parties' customers mostly switch to generic suppliers or other branded manufacturers rather than between the Parties.

(436) Since these arguments are made at a general level applicable to all areas of crop protection (herbicides, insecticides, nematicides and fungicides), they are addressed in the following recitals as a general part of the competitive assessment.

6.2.1. Generic players are only a partial and often not significant constraint for R&D players

(437) As explained in Section V.1.5, undertakings predominantly active in the sale of products containing off-patent AIs can follow different business models. The level of competitive constraint such undertakings entail for R&D players can therefore vary significantly, and will depend on the specific market segment in which they are active. Nevertheless, it is possible to draw some general conclusions applicable across crop protection which relate to the common features of the industry and its different types of players.

(438) As explained in Section V.1.4.4, when their AIs reach the end of patent protection, R&D players are at risk of generic players being able to enter the market with copies of their products. The Commission fully acknowledges this competitive threat and does not deny that generics are viewed as a threat by the Parties. However, the investigation has shown that R&D players can address and manage the competitive constraint posed by generic players, thereby probably strongly reducing the actual threat they pose on prices and market shares. Because R&D players react and innovate when faced with the mere threat of generic competition, they appear able to avoid any significant competition by generic players.

(439) In particular, the Parties submitted two economic studies which do not appear to consider the defence strategies that branded manufacturers use to limit competition from generics, such as: (i) the use of patents on formulated products/mixtures, (ii) market segmentation strategies with the use of mixtures (sometimes by combining off-patent AIs with on-patent AIs) to differentiate their product offerings from generic manufacturers, and (iii) cost improvement strategies to maintain their variable costs below generic manufacturers' so that generic manufacturers are less competitive on a variable cost basis.

6.2.1.1. Parties' arguments

(440) First, in the Form CO, the Parties provided two economic studies by RBB Economics aimed at assessing the competitive constraint exerted by generic suppliers.

(441) The studies acknowledge that branded manufacturers can respond to generic entry by reducing prices and/or by introducing new and innovative mixtures products that bring added value to farmers.

(442) In the Parties' view, the Commission essentially argues that the Parties’ reactions to generic entry insulate them from any competitive constraint exerted by generic suppliers, and considers generic manufacturers to be only distant competitors to branded manufacturers.

(443) In response to the Commission’s criticism, the Parties explain that generic suppliers still represent their closest competitive constraint. In their view, both the economic analysis and the internal documents show that their competitive initiatives in recent years were primarily driven by competitive pressure from generic suppliers, not by competitive pressure from each other.

(444) Second, the Parties argue that stable market shares do not per se justify the conclusion that generics are no competitive constraint. On the contrary, stable market shares could be a sign of rigorous competition.

(445) Third, the Parties argue that internal documents point to generics as their closest competitors and a constant threat, forcing other R&D players to develop defence strategies to defend their market position.

162 For an overview of "post patent strategy [for indoxacarb] in EMEA" (notably France), in particular use of data protection, secondary patents, segmentation/differentiation with mixtures, replacement with newer AIs and even withdrawal of product authorisations to slow generic registrations, see DuPont's internal document, file name "DUPONT-CASEM7932-0081998 - STRICTLY CONFIDENTIAL - CONTAINS BUSINESS SECRETS.pptx" (ID6827-020443) (also Annex DuPont 2269, file name "M7932_Annex DuPont 2269 Indoxacarb post patent strategy in EMEA Oct.05 2015_CONFIDENTIAL.pdf" (ID1332-00619)).

163 See also Annex 2.

(446) By contrast, the Commission would allegedly focus on a very small number of internal documents, quoted out of context and which would not support the Commission's conclusions.

(447) Fourth, in the Parties' view, the Commission has failed to take into account the bulk of the evidence they submitted showing that generic players fully compete with equivalent quality, and face the same regulatory constraints as R&D players.

(448) Fifth, the Parties argue that, while a "majority of respondents said that generics are not able to solve all farmer needs", that also applies to branded players since no individual company can solve all farmers’ needs. Moreover, such broad statements would not indicate that generics do not constrain the Parties in specific relevant markets. They conclude that, on balance, the key documents made available to them show that responses to the investigation were quite positive.

6.2.1.2. Generic players cannot offer competitive products for the parts of the market which require novel AIs

(449) As explained in Section V.1.5.4, generic players do not have their own discovery capabilities where they would be able to invent new AIs. They generally limit their activities to manufacturing and/or selling products containing off-patent AIs, sometimes in mixtures. Therefore, generic suppliers cannot offer new AIs to satisfy unmet needs or needs that are not currently sufficiently addressed.

(450) Moreover, R&D players do not usually share novel and impactful technologies with generic players. In fact, they typically refuse to license their most recent and valuable AIs to generic players in order to avoid generic competition and preserve monopoly profits for as long as possible. A generic company in this respect indicated that "[g]eneric companies […] cannot compete in that direction because they would need a license from the patent holder to create similar copy products." They will sometimes grant one or more generic players access to one or more of their most recent AIs if they consider it profitable, for example to address new markets where they have limited presence and reach. They will also do this to maintain high production volumes and, therefore, costs, a key competitive advantage which will be further detailed in the following recitals.

(451) Generic players usually only compete with off-patent technology, and the Commission's investigation suggests that generics are not able to solve all farmer needs. Some respondents noted that generic products were older, not able to tackle new technical problems, and that generic companies did not offer adequate consultancy services. In particular, a customer stated that "generic products usually contain old active ingredients with many restrictions" and another that "generic products do not cover every need and in this case there is not appropriate consultancy". A third customer stated that "[n]ew technical problems are hardly solved by old molecules".

164 Parties' response to the first Letter of Facts, paragraphs 5-9.

165 Agreed non-confidential minutes of a call with a competitor, 15 July 2016 (ID8259).

166 Questionnaire to Crop Protection Customers (Q1), question 72; Questionnaire to Crop Protection Competitors (Q2), question 44; Questionnaire to Crop Protection Stakeholders (Q3), question 38.

167 Questionnaire to Crop Protection Customers (Q1), question 72 (ID2896, ID9332 and ID9448). Courtesy translation from French: "[l]es nouveaux problèmes techniques sont difficilement résolus avec des molécules anciennes."

(452) This is particularly relevant for products, such as insecticides and fungicides, where resistance to older products is already strong and continues to develop. For insecticides and fungicides, new products, but most importantly new AIs and – where possible – new MoAs, appear to be very much needed. Because generic players do not have access to these latest products, they may not be able to compete for a large part of the market, likely the most profitable.

(453) Some market participants also noted that generic products could be of a lower quality, for example regarding formulations. A stakeholder noted that "[f]or the same AI (fungicide), the generic product are often less efficient than a product coming from Syngenta, Bayer, BASF, Dow and DuPont. It's surely due to the formulation of the product". A customer similarly stated that "we do not have enough trust in the adjuvants of generic products. Many doubts on effectiveness and mixtures".

168 Questionnaire to Crop Protection Stakeholders (Q3a), question 37.1 (ID1913).

169 Questionnaire to Crop Protection Customers (Q1), question 72.1 (ID9541). Courtesy translation from Spanish: "[n]o tengo suficiente confianza en los coadyuvantes de los productos genéricos. Muchas dudas de efectividad y mezclas con otros cultivos".

(454) Furthermore, it appears from the investigation that the perceived difference in quality influences farmer choice. For instance, one market participant explained that "a specific brand of pesticide may be of importance even when generics for an active ingredient are available, because the generic product may not have been technically proven. Even if the active ingredient is the same, the efficacy of the generic product might be worse, because of a different carrier agent or a different solvent used for delivery. The main factor that farmers consider in choosing a specific product is efficacy. Price considerations are secondary." Moreover, a distributor added that, due to uncertainty regarding effectiveness of generic products, "a distributor cannot be sure that he will be able to sell them".

170 Agreed non-confidential minutes of a call with a customer, 17 March 2016 (ID8246).

171 Agreed non-confidential minutes of a call with a customer, 21 March 2016 (ID8263).

(455) However, other respondents seem to highlight that some (older) groups of products are commoditised and are typically purchased from generic players. Some market participants have even explained that the leading generic players – Adama in particular, with strong formulation capabilities – can offer very good products, sometimes better than products from R&D players.

172 Agreed non-confidential minutes of a call with a customer, 21 March 2016 (ID8263).

(456) Overall, generic players appear likely to constitute a potential competitive constraint only for some parts of the overall crop protection market.

6.2.1.3. R&D players use a number of legal and practical tactics to prolong the legal or de facto exclusivity for their AIs and products

(457) In the EEA, generic players face a number of hurdles which effectively delay or restrict their apparent possibilities of acquiring market shares and exercising competitive pressure on R&D players.

(A) Patent and data protection periods are long

(458) The initial patent on an AI is granted for a 20 year term from the date of application, which can be extended by a further five years if a supplementary protection certificate is requested. Since it takes on average 10 years from the date of patent application to bring a compound to the market, the holder of a patent followed by a certificate can enjoy up to approximately 15 years of exclusivity from the time its formulated product first obtains authorisation to be placed on the market in the Union.

(459) In parallel, the data needed to support an application for an AI is protected – independently from patent protection – for a period of 10 years (for new AIs) starting from the date of first authorisation of a product containing the AI. Therefore, depending on the date on which a product is first authorised, protection on the data can expire years after the compound patent and certificate on the AI. In practice, absent an agreement with the patent holder, any generic player wishing to bring to the market a new AI has a forced choice between waiting for the expiration of the protection of the data package, which delays its entry into the market, or generating new data itself, a process that is both expensive and time consuming and would likely propend for the former option.

(B) Regulatory entry barriers are high

(460) The EEA regulatory framework for the approval of AIs and the authorisation of formulated products is generally considered to be the most strict,and accordingly the most lengthy and costly, in the world. In fact, many AIs never make it to the EEA market because of these complexities and costs. DuPont explicitly considers that the complexity of the Union regulatory environment constitutes a "barrier for generics".

(461) Generic players wishing to launch products in the EEA must comply with registration obligations. In so doing, they have no significant advantage compared to R&D players at that stage. In particular, they must generate or purchase the same data for regulatory authorities.

(462) The Parties argue that R&D players have the large disadvantage of having incurred significant R&D costs to launch new products, which generics are able to easily copy. However, although some rare exceptions may not be excluded, R&D players typically recoup their R&D investment costs by the time of patent expiry, which the Parties acknowledge, and which generic players do not enjoy.

(463) Moreover, generic players seem to have several disadvantages compared to R&D players when dealing with the Union regulatory system.They encounter

173 The data protection period is 2.5 years for existing AIs.

174 Form CO, part D, page 33; agreed non-confidential minutes of a call with a competitor, 6 September 2016 (ID7111); agreed non-confidential minutes of a call with a competitor, 9 September 2016 (ID7972).

175 DuPont's internal document Annex DuPont 166, file name "DOC-000000166.pdf" (ID1329-166), page 37. See also the agreed non-confidential minutes of a call with a competitor, 15 July 2016 (ID8259).

176 Agreed non-confidential minutes of a call with a competitor, 13 September 2016 (ID7128); agreed non-confidential minutes of a call with a farmers' association, 18 March 2016 (ID8254); agreed non-confidential minutes of a call with a competitor, 15 July 2016 (ID8259); agreed non-confidential minutes of a call with a competitor, 7 April 2016 (ID8260).

177 Agreed non-confidential minutes of a call with a competitor, 7 April 2016 (ID8260).

178 As explained by the Parties in response to the Commission's request for information RFI 7 (ID1001), question 52, a recent General Court decision may lead to some changes in this regard, which would make generic entry somewhat easier from a regulatory standpoint. For the moment, it appears that, although there formally is a so-called Bolar exception in crop protection, which would enable generic companies to start preparing entry before patent or data protection expiry, the unavailability of the chemical specifications of the AIs (particularly impurities) makes this exception relatively inoperative (compare agreed non-confidential minutes of a call with a competitor, 15 July 2016 (ID8259) and the Parties' response to the Commission's request for information RFI 7 (ID1001), question 52).

179 Agreed non-confidential minutes of a call with a competitor, 15 July 2016 (ID8259).

180 Agreed non-confidential minutes of a call with a competitor, 7 April 2016 (ID8260).

(464) Difficulties in determining the key characteristics of the AIs and products they wish to copy, which entails delays and added costs. They also lack the technical expertise with the specific AIs coming off-patent. In particular, their production processes (and costs) will usually not be as efficient as those of the R&D player, as will be detailed in the following recitals.

(465) For instance, a generic player explained that it considered launching an off-patent insecticide in the EEA, but ultimately gave up in light of the costs and complexities of registration in multiple crops and countries in relation to the market opportunity. Conversely, it did launch a cereal herbicide, a product used on fewer crops in a limited number of countries but with larger individual market opportunities, where technical complexities were more limited.

(466) Significantly, in its comments of 4 January 2017 on the Statement of Objections Finchimica, the Parties' competitor and an interested third party, provides a detailed technical explanation strengthening the Commission's finding on the limited competitive constraint constituted by generic players. It focuses particularly on the use R&D players make of the European regulatory framework to unduly protect their business from generic competition, contrary to what would be the acknowledged objectives of the applicable legislation. In Finchimica's own words: "[t]he Commission observes that the R&D players in the industry have numerous means to counter the generic threat, that is, mainly by means of market segmentation via mixtures with patented active substances or developing mixtures with other old active substances. However, it misses a subtle and more insidious factor to their success in warding off generic competition. R&D firms are able to exploit regulatory barriers for generic entry and continuation in the market"(emphasis in original).

(467) Indeed, Finchimica confirms that R&D players may strategically decide to withdraw older products to make generic entry impossible or more difficult, as explained in recital (484): "[t]he regulatory barriers represent a very significant facet to the protection afforded to the Parties and the R&D industry. The development of mixtures not only serves to segment the market into costlier smaller access channels, but it also creates added regulatory hurdles for generics to overcome. Where a mixture product is itself patented or a mixing partner is patented, generic competition cannot follow without a license to the IP right. This is not disputed and the generic industry has no argument with this strategy, per se. However, such strategies are also accompanied by tactical withdrawal of old and 'unprotected' reference products and other technical manipulations within their replacement product re-authorization processes, as part of a strategy to extend data protection and foreclose access to unprotected regulatory data requirements under very

184 complex rules set out in Regulation (EC) No 1107/2009 ("Regulation")"(emphasis added).

(468) More importantly, as explained in recital (459), access to regulatory data appears to be a strong barrier to generic entry: "[i]t is the uncertainty embodied in access to unprotected regulatory data requirements and the exclusion from data sharing consortia in order to comply with the continuous cycle of regulatory re-approval and re-authorization that are the key barriers to generic competition in the EU. […] Indeed, the AIR Data protection period shall vary from Member State to Member State according to how the R&D companies play the game and furthermore, according to how the Member State Competent Authorities (CAs) cope with the burden of product re-authorisation procedures."

(469) It also appears that the way specific national regulatory authorities will apply the relevant legislation may vary and in some cases constitute a further significant barrier for generics above and beyond the legislation per se: "[f]urthermore, some Member States and their competent authorities are known to act harshly in the operation of their pesticide product authorization schemes, to the direct disadvantage of generic competition. German farmers, in the second largest crop protection product market in the EU, are widely understood to enjoy very little effective generic competition. The reasons are directly attributed to the regulatory barriers particular to Germany, created by the authorities themselves in their own manner of operating the harmonized pesticide product framework regulation" (emphasis added).

(470) Specifically, delays in the review by national regulatory authorities of renewals of existing authorisations facilitate the extension of data protection by R&D players to 10 or even 13 years: "[w]hat is the impact of such extensions? Obviously, it facilitates the implementation of the Ten Year Trick by the R&D companies. Any period of extension that delays the re-newal of an existing product favours the successful application for a new product authorisation to be achieved in advance of the old one, hence deriving a period of 10 years data protection for the AIR Data instead of 30 months."

(471) Moreover, Finchimica confirms dissimilar treatment between, one the one hand, products from R&D players currently on the market and, on the other hand, new generic copies of these which generics are looking to launch. In sum, generics entrants face considerable difficulties in getting their products authorised, whereas R&D players appear to obtain long extensions to keep their current products on the market beyond expiry of the original authorisation, sometimes in spite on new scientific data: "[n]ew entrants, generics, seeking to apply for registration of products similar to those already existing in the market, face a variety of administrative policies which effectively lock out new entrants as the renewal of an active substance approaches. […] As such, new entrants are locked out of the process for filing for new product authorisations between 12-18 months in advance of the active substance renewal of approval program and further locked out until they reproduce the AIR Data themselves or wait for the AIR Data to become unprotected."

188 Submission on behalf of Finchimica S.p.A. of 4 January 2017 in response to the Statement of Objections, page 3.

(471) Finchimica thus concludes that "the Regulation is a poorly drafted piece of legislation that has the unintended consequence that R&D companies are effectively shielded from any meaningful competition from generics in the European pesticides market(s) to the detriment of consumers/farmers alike (having to ultimately pay higher prices while at same time having reduced choice of products in the market)"(emphasis added).

189 Submission on behalf of Finchimica S.p.A. of 4 January 2017 in response to the Statement of Objections, page 7.

6.2.1.3. R&D players may also use secondary patents and patent thickets

(472) To protect their revenues and margins once generic entry becomes possible, R&D players expand their products' use into new crops and pests (with the accompanying patent protection), either as straight products or in mixtures. Because these new uses can be patent protected, generics will again not be able to compete.

(473) R&D players may also use process or formulation patents, which bring potential generic entrants to find alternative production processes or formulations for their products. While this can generally be achieved by generic players given the many chemical routes to synthesise a given molecule – provided the R&D player has not managed to patent all economically viable alternatives – it, however, requires considerable time and investments, that add an additional layer of complexity and cost for such generic players to already difficult and capital intensive business processes. A generic company confirmed that R&D players use "always subsequent follow-up patents, process patents that are filed by the originator of the molecule in subsequent years in order to block alternative means to produce the active ingredients".

(474) Some of these so-called "secondary" patents (as opposed to the patent on the compound) may be filed with the main purpose of creating a patent "thicket" to block competitors (including generics) from entering the market with similar products, not to genuinely protect a useful invention for the R&D player, which may in practice not at all use the patented production process or formulation. In this respect a generic competitor indicated that R&D players "patent various different types and means of presentation of the active ingredient formulations in order to block the generic industry from entering even after the original active substance patent has expired".

188 Submission on behalf of Finchimica S.p.A. of 4 January 2017 in response to the Statement of Objections, page 3.

189 Submission on behalf of Finchimica S.p.A. of 4 January 2017 in response to the Statement of Objections, page 7.

190 Agreed non-confidential minutes of a call with a competitor, 15 July 2016 (ID8259).

191 Agreed non-confidential minutes of a call with a competitor, 15 July 2016 (ID8259).

6.2.1.4. R&D players may also rely on production know-how to achieve de facto continued protection of exclusivity

(475) In addition to the possibility of obtaining new patents, R&D players may also choose not to patent their production know-how. This will be especially relevant for molecules which are particularly difficult to produce or require some specific expertise unavailable to all players. R&D players in such a case would anticipate that keeping this knowledge secret rather than making it public by protecting it with patents would ultimately provide longer protection than patents themselves because of the time needed for third parties to discover and acquire the requisite know-how.

(476) As an illustration, Dow appears to have a very high level of expertise in fermentation products, which only a limited number of players may be able to copy even once all patents will have expired. [Internal assessment].

192 Dow's internal document "XR-659 Pre-Development Advancement Presentation", file name "Annex Dow RFI 59a_Q 14 - XR-659 CPMF Presentation_FINAL (Nov 28).pdf" (ID9408), page 16.

(477) In sum, R&D players appear to have a number of strategies to delay generic entry. If generics ever do decide to launch a product in the EEA and are successful, they will in any event enter the market no sooner than one to two years after patent expiry, as the Parties acknowledge in their internal documents. This provides an additional period for R&D players to reap the benefits of their inventions without any competitive constraint from generic players. In a context of spreading resistance, this delay may also in particular circumstances prove sufficient to significantly reduce the value of launching a generic version of a given AI.

6.2.1.4. R&D players use differentiation and segmentation strategies as a defence against generics

(478) R&D players typically launch new AIs as straight products. However, when approaching patent expiry, as a defence strategy R&D players appear to engage in market segmentation and differentiation strategies consisting in creating specific new mixtures: "innovating and developing differentiated products (especially, but not only, formulated products based on AI mixtures or co-packs)". For DuPont, "the Big 3 also renew their portfolio mainly from introducing new actives in combination with existing products (herbicides and fungicides".

192 Dow's internal document "XR-659 Pre-Development Advancement Presentation", file name "Annex Dow RFI 59a_Q 14 - XR-659 CPMF Presentation_FINAL (Nov 28).pdf" (ID9408), page 16.

(479) Indeed, a generic player confirms that "[o]ne of the strategies of the non-generic R&D companies is to stop selling straight products containing off-patent molecules and to develop mixtures of those same molecules with patented active substances. This way they re-segment the market so that there are fewer straight off-patent products on the market. The market is thus redefined in mixtures with patented products. Generic companies like […] cannot compete in that direction because they would need a license from the patent holder to create similar copy products."

(480) Figure 19 clearly shows this strategy to switch the market towards mixtures rather than straight products, which will favour R&D players as opposed to generic players.

Figure 19 – Strategy of R&D players to switch the market towards mixtures rather than straight products

[…]

Source: DuPont's internal document Annex DuPont 1045, file name "DOC-000001249.pdf" (ID1329-01249)

(481) Another generic player also confirms that "multinational innovation is moving towards innovative mixtures rather than new active ingredients in solo formulations. Indeed, tribenuron is no longer sold by DuPont as a solo formulation, only in mixture with other SUs".

198

(482) These mixtures may (when they provide some innovative step or include a patented mixture partner) or may not be patented. In both cases, they limit the ability of generic players to compete effectively. The Commission's investigation shows that these mixtures move customers from large straight product markets to smaller, differentiated, mixture markets.

(483) Generics may be able to target the former straight products, but face considerable difficulties in attempting to address the latter differentiated ones, even where there is no patent protection. In essence, an overall addressable market can be split into several smaller markets, targeted by a specific product/mixture offering, which taken individually can be too small for generic entry to be profitable, with the exception of a limited number of countries/segments. Figure 20 illustrates Dow's strategy of using market segmentation – at least in part – as a protection against generics.

Figure 20 – Spinosad generic defence strategy

[…]

Source: Dow's internal document "Value Growth Molecules 2014 Plan and 2014-2018 BVM Guidelines", file name "Spinosad 2014 BVM Strategy Planning Final .pdf" (ID7081-00378), page 1

196 DuPont's internal document Annex DuPont 982, file name "DOC-000001186" (ID1329-01186), page 19.

197 Agreed non-confidential minutes of a call with a competitor, 15 July 2016 (ID8259).

198 Agreed non-confidential minutes of a call with a competitor, 13 September 2016 (ID7128).

199 However, the level of innovation to reach appears to be fairly low according to some market players. Agreed non-confidential minutes of a call with a competitor, 15 July 2016 (ID8259).

200 For other examples of internal documents referring to mixtures as a generic defence strategy, please see Dow's internal documents, file name "DAS-00000629-000001.pdf" (ID1073), file name "DAS-00000619-000001.pdf" (ID1056-00052) and file name "DAS-00000101-000001.pdf" (ID455-00102), pages 3, 10, 28 and 44.

72

(484) In addition, when patent expiry approaches, R&D players may also attempt to switch their customers from an existing AI to newer patented AIs which are unavailable for molecule competition by generics. The replacement of Dow's spinosad with Spinetoram in Figure 20 and Figure 21 provides a good illustration of such "antigeneric" strategies.

Figure 21 – Spinosad key strategic assumptions

[…]

Source: Dow's internal document "2012 Plan / SVM / Supply Review, Growth Insecticides, December 2011", file name "RFI 14 - Dow annexes - Case M 7932 Dow DuPont - RFI 14 - Dow annexes - DAS-00000711-000001" (ID3987-00100), page 4

(485) With this segmentation and differentiation strategy using mixtures, R&D players seem able to limit generic penetration to only a relatively small part of the crop protection markets and to preserve their own aggregate sales and margins by focusing on a large number of small markets, as shown in Figure 22. Moreover, R&D players may also move into new uses, new crops and new geographies, where generic players are not present yet.

Figure 22 – Cereal portfolio strategy

[…]

Source: Dow's internal document, file name "DAS-00000118-000001.pdf" (ID00455-00119), page 10

(486) In addition, brand importance in the EEA appears to vary with geography. For instance, German growers seem to favour (German) branded products. On the contrary, growers in other markets (Southern or Eastern Europe) appear more likely to use – cheaper – generic products. For instance, an Italian distributor explains that "[i]f there was a price difference of 5-10% between a generic and branded product, a farmer may choose the more expensive branded product. If there is a big price difference farmers may turn to generics because of the difficult financial situation of the agriculture sector".

(487) It thus appears that brand value and quality plays a significant role in at least some EEA countries.

6.2.1.5. R&D-integrated players typically enjoy superior economies of scale and production

(488) R&D-integrated players have an in-depth knowledge of their AIs and products. When generics threaten to enter the market, they have had decades to optimise production processes and costs. Due to their broader market presence (broad portfolio and distribution network, out-licensing, technical sales), they also have a much larger scale than (smaller) generic players, especially at patent expiry, and thereby lower costs. Overall, even years after generic entry, R&D-integrated players may have lower production costs. This is of course a key competitive advantage, especially for commoditised products.

201 See, for instance, the agreed non-confidential minutes of a call with a competitor, 9 September 2016 (ID7972).

73

(489) For example, for clopyralid production, Dow considers that it "has better utilization of assets/raw materials and therefore lower variable cost" than a Chinese generic, with a slightly lower total cost. Similarly, for fluroxypyr, Dow appears to have significantly lower total and variable costs compared to Chinese and Indian generics. The document also shows decreasing production costs over time, as a competitive advantage against generics.

(490) More generally, because of their smaller scale and sales, generics have more limited financial resources from which to finance multiple simultaneous registrations and distributions, which are required for a quick return on investment and large profits. Dow explains that "[s]mall companies focus on generics as soon as patents are off. They struggle to survive due to lack of critical mass and profitability". Diachem, a generic competitor, confirms that its "limited resources enable it to prepare only a small number of regulatory dossiers at the same time."

(491) Generic players will therefore likely only target the markets that are more likely to be profitable for them. They will rationally target the largest crop protection markets which they can access. They will likely not be able to easily build market reach, and avoid costly multiple registrations (across small crops but also across countries).

(492) In the EEA, it seems that they will thus likely focus on row crops (cereals in particular) rather than specialty crops, and avoid very fragmented/differentiated (with mixtures) markets like fungicides, where the addressable segments are unlikely to make entry profitable.

(493) A generic player confirms that "[t]he time and costs to obtain a registration varies depending on the type of product and the number of crops that are registered. Particularly for vegetables fungicides (and insecticides), numerous crops have to be registered individually, and when each requires field trials, it can be an enormous task". These smaller, niche, markets are the heart of European agriculture.

(494) Moreover, the market investigation suggests that generic players are usually active in only a few countries within the EEA. Only a large generic player like Adama (and perhaps FMC) appears to be able to have a market reach comparable to that of an R&D-integrated player.

(495) On the contrary, R&D-integrated players' market reach – in terms of distribution capabilities and expertise, product portfolio and brand value – is much greater. In DuPont's words – assumedly describing Bayer, BASF and Syngenta – "the breadth of portfolio of the Big 3 allows them to quickly take advantage of new market opportunities".

204 Dow's internal document, file name "DAS-00000118-000001.pdf" (ID00455-00119), page 27.

205 Dow's internal document, file name "DAS-00000118-000001.pdf" (ID00455-00119), pages 53, 54 and 56.

206 Agreed non-confidential minutes of a call with a competitor, 13 September 2016 (ID7128); agreed non-confidential minutes of a call with a competitor, 9 September 2016 (ID7972). In this regard, Adama's "strong range across major crop segments" makes it an atypical generic player, and the largest (DuPont's internal document Annex DuPont 166, file name "DOC-000000166.pdf" (ID1329-166), slide 9).

207 Dow's internal document, file name "DAS-10160458.pdf " (ID6696-3350).

208 Agreed non-confidential minutes of a call with a competitor, 13 September 2016 (ID7128).

209 Parties' response to the Commission's request for information RFI 56 (ID8089), paragraph 1.2.

210 Agreed non-confidential minutes of a call with a competitor, 15 July 2016 (ID8259).

211 Agreed non-confidential minutes of a call with a competitor, 13 September 2016 (ID7128).

212 DuPont's internal document Annex DuPont 982, file name "DOC-000001186" (ID1329-01186), page 19.

74

(496) For instance, a generic player cited the difficulties of competing against R&D-integrated players: "[i]t can be difficult to compete against R&D players in the off-patent market, because R&D players have different cost structures, having already recouped their investment during the patent exclusivity phase. Furthermore, there are economies of scale in production. A generic player typically needs to buy access to the data from the R&D companies: either they physically share the data or they provide a so-called letter of access, which can be provided to the regulatory authority".

6.2.1.6. Generic players are typically left with portfolios of low margin AIs discarded by R&D-integrated players

(497) Although R&D-integrated players appear to have a number of strategies to limit the competitive impact of generic players, some older products ultimately become commoditised. When that happens and gross margins reach predefined levels considered too low (typically around 30-40%), R&D-integrated players may sell their AIs to other players, usually the generic players, which are accordingly kept at a safe distance from the R&D-integrated players.

(498) In so doing, R&D-integrated players curate their portfolio compositions to preserve the high margin products, thus avoiding head-to-head competition with generic players on commoditised molecules. DuPont sees this as a strategic focus for Dow: to "improve the overall profitability of the CP business by phasing out of older generic products with low margins". It seems that, contrary to generic players, R&D-integrated players' business model is one of value, not volume.

6.2.1.7. The case study presented by the Parties on Dow's herbicides indicates that R&D-integrated players have means to defend their margins

(499) As is discussed in Section V.6.2.1.9, the Commission's analysis seems to be fully borne out by the examples of generic threat identified in Dow and DuPont internal documents by the Parties. These examples appear largely limited to cereal herbicides, a very large market where no new MoAs have been introduced in decades. Indeed, Bayer confirms that "generics are also successful in areas where there is limited innovation, such as herbicides (Nufarm, for example, is strong in herbicides) and insecticides."

(500) The examples provided by the Parties do not seem to point to significant effective competition by generic players on all crop protection markets in the EEA.

213 DuPont's internal document Annex DuPont 982, file name "DOC-000001186" (ID1329-01186), page 19.

214 Agreed non-confidential minutes of a call with a competitor, 7 April 2016 (ID8260).

215 Agreed non-confidential minutes of a call with a competitor, 27 September 2016 (ID8562). Agreed non-confidential minutes of a call with a competitor, 13 September 2016 (ID7128). Agreed non-confidential minutes of a call with a competitor, 15 July 2016 (ID8259).

216 DuPont's internal document Annex DuPont 982, file name "DOC-000001186" (ID1329-01186), page 18.

217 Parties' response to the Statement of Objections, paragraph 649; Parties' response to the Article 6(1)(c) Decision, paragraphs 55-65. See also the Parties' response to the first Letter of Facts, paragraph 10.

218 Agreed non-confidential minutes of a call with a competitor, 5 April 2016 (ID8248).

75

6.2.1.8. Internal documents confirm that the market share of generic players is stable, that they are seen as struggling and that generic defence works well

(501) While internal documents show that the Parties do refer to competitive pressure from generics, the investigation suggests that the Parties are still be able to successfully limit generic impact even after patent protection by adopting a differentiated positioning on the market through mixtures and brands.

(502) For instance, DuPont notes about Dow: "Dow 'has a largely generic range and pasture and vegetation management herbicide portfolio that retains high brand value. Dow has maintained its share and pricing despite the attempted inroads by generic companies through an elaborate product segmentation strategy that has created a large number of products for every conceivable use within this market space."

(503) Figure 23 is another illustration of Dow's successful strategy to limit generic competition, partly by moving the market to mixtures.

Figure 23 – Fluroxypyr regulatory summary

[…]

Source: Dow's internal document, file name "DAS-00000118-000001.pdf" (ID00455-00119), page 59

(504) As a consequence of these "anti-generic" measures, the overall market share of generic players has remained stable over recent years – from 2008 to 2012 in Figure 24 – and internal documents of the Parties highlight this stability: "generic companies as a group have not grown over the last 5 years and are struggling with low profitability." This marks a significant difference from the neighbouring pharmaceutical industry, where the share of generic players has increased over the last years, largely under pressure from health systems to limit expenditures.

Figure 24 – Market share evolution, generics vs majors

[…]

Source: DuPont's internal document Annex DuPont 982, file name "DOC-000001186" (ID1329-01186), page 17

(505) Figure 25 illustrates DuPont's post patent strategy in the United Kingdom, largely used to limit the competitive constraint exercised by generic players.

Figure 25 – DuPont's UK post patent strategy

[…]

Source: DuPont's internal document Annex DuPont 1048, file name "DOC-000001252.pdf" (ID1329-01252), page 2

219 For example, DuPont mentions the "preparation of generic picoxystrobin defense, even if beyond 5 yr", and the "early adoption of defense activities" against generics. See DuPont's internal document Annex DuPont 3251, file name "DOC-000000006.pdf " (ID4384-00006), pages 5 and 9.

220 DuPont's internal document Annex DuPont 541, file name "DOC-000000541.pdf" (ID1329-000541), page 1.

221 DuPont's internal document Annex DuPont 982, file name "DOC-000001186" (ID1329-01186), pages 17 and 19. This is also quoted by the Parties in their response to the Article 6(1)(c) Decision with emphasis on the preceding sentence (not cited here).

76

(506) Similarly, and in spite of the generic constraint, the Parties appear to have been able to preserve high market shares in crop protection markets, thereby further confirming how limited the generic constraint truly is. For example, one of the document states that "Clopyralid and Florasulam, in spite of generic threats, are still very solid and very profitable products".

(507) It follows that the Parties and other R&D-integrated players are likely to still be able to extract relatively high gross margins from off-patent molecules, sometimes decades after patent expiry, as shown in Figure 26.

Figure 26 – DuPont's SU post patent strategy

[…]

Source: DuPont's internal document Annex DuPont 982, file name "DOC-000001186" (ID1329-01186), page 45

(508) Dow's tricyclazole provides another good illustration. It has been off-patent for decades, with generic entry. In spite of this, Dow appears to still retain high gross margins ([80-90]% in Southern Europe), high market shares (see Section V.6.6), and to contemplate a strategy to "[extract internal document]". In fact, in Spain where a number of other players are present, Dow appears to be able to maintain a farmer price for tricyclazole [50-60]% higher than generics.

(509) Other market players also appear to confirm the limited competitive impact of generics. For example, a competitor highlighted that it "does not consider that the generics will increase their market share in the future. There are significant regulatory hurdles and these companies do not have the necessary expertise to overcome these. The innovation hurdle is also very high, for example Nufarm cannot enter in fungicides."

(510) Similarly, R&D-integrated players generally seem to look mainly at the other R&D-integrated players as their competition. Smaller players like generic producers, but also small companies active in discovery, are monitored and taken into account when planning product development and commercialisation. However, they are apparently taken into account only to the level of competitive constraint they seem to constitute for the larger players, that is as second- or third-tier players, with limited product offering, market reach and market power.

(511) In conclusion, it appears that generic players are unable to exercise significant competitive constraints on R&D players like Dow and DuPont.

6.2.1.9. The economic analyses submitted by the Parties do not support their general claims related to the competitive constraints brought by generic suppliers

(512) The economic analyses submitted by the Parties, if properly read, do not point to a different conclusion than the Commission's own on the basis of the available evidence. The Parties have submitted several economic analyses(hereafter collectively referred to "Submissions on generics") that aimed at assessing the competitive constraint brought by generics products on their own products when they lose patent protection. Most of the time, the underlying data and methodology were not provided along these submissions, as requested by DG Competition Best Practices for the submission of economic evidence, but as the result of requests for information.

(513) These Submissions on generics focus on two Dow active ingredients, fluroxypyr and florasulam, because (i) "DuPont does not have historical transaction data covering a sufficiently long period of time to assess the impact of generic entry" and (ii) these two AIs are the only ones from Dow which lost patent protection across several countries in the EEA during the last 15 years.

(514) These economic studies provide a qualitative analysis of the evolution over time of volume, price and margin, in light of market events such as patent expiry and "generic entry". In other words, the Parties aim at assessing graphically the causality of market events, such as patent expiry or generic entry, by solely identifying whether downward changes in price or margin occurred in similar periods than these events.

(515) On the basis of the Submissions on generics, the Parties claim that:

(a)Generic entry is widespread after patent expiry;

(b)Straight products experience significant price drops as a result of the competitive constraint brought by the introduction of generic products;

(c)Mixtures also experience significant price and margin drops as a result of the competitive constraint brought by the introduction of generic products, in particular due to the possibility of tank-mix by farmers;

(d)Dow reacts to the competitive constraint on price and margins exerted by the introduction of generic products by introducing new mixtures around the patent expiry date.

(516) The Commission made a thorough analysis of the Parties' Submissions on generics and is unable to validate the Parties' claim on the competitive constraints brought by generic suppliers, with particular respect to florasulam and fluroxypyr, for several reasons.

(517) First, the Submissions on generics suffer from important limitations. For example, the Submissions are based on the analysis of price of an AI per kilogram of this AI, while the internal document of the Parties indicate that customers do care about the cost of treatment per hectare, and not about the price per kilogram of AI.

(518) Another important issue is that entry by generics is mischaracterised and branded manufacturers, such as BASF or Syngenta, are often reported as generic entrants. Such mischaracterisation is frequent and is at odds with the Parties' own assessment of the agrochemical industry supply side in which BASF and Syngenta are not identified as generic suppliers but as branded manufacturers.

(519) Moreover, the analysis amounts solely to the graphical identification of downward changes in price or margin occurring in similar periods than patent expiry or generic entry, and do not attempt to control for the heterogeneity of the data that could be caused, for example, by Dow's change in cost structure, market trends, country characteristics or product characteristics.

(525) The Parties claim that “mixture are a manifestation of the competitive constraint exerted by generic suppliers”,that is that the constraints brought by generic suppliers induce R&D suppliers to bring new mixtures on the market. The Commission notes that this argument is not relevant for the assessment of downstream competition for formulated products, which occurs between existing or forthcoming mixtures and products. To that respect, mixtures brought to the market by branded manufacturers are patent protected and, consequently, cannot be copied by generic products, and the Parties' Submissions do not take into account the effective market segmentation strategy branded manufacturers put in place with mixtures to differentiate from generic manufacturers. Such successful strategy suggests that the competitive pressure from generic companies is limited. The Commission also notes that the Parties’ argument is not relevant for the innovation competition assessed in this Decision, which focuses on the discovery of new AIs and not the creation of mixtures made of known AIs.

240 Changes in market power are proxied by change in Lerner Index, that is changes over time of the margin expressed as a share of price. On the Lerner index, see, for example, Tirole (1998), The theory of industrial organisation, MIT press, section 1.1, page 66.

241 Parties’ response to the Statement of Objections, annex 5, entitled “Dow/DuPont: Response to the Commission’s analysis of closeness”, dated 21 December 2016, section 2.1.

80

(526) Taking all these elements in consideration, the Commission concludes that the data provided by the Parties with their Submissions on generics do not allow to reach the Parties' claims, in particular that branded manufacturers are significantly constrained by patent expiry or generic entry and that generic products are close competitors to products of branded manufacturers.

6.2.1.10. Conclusion

(527) In conclusion, the Commission's investigation show that, while generic players may be a threat to R&D players, the latter are likely to implement highly effective strategies which would strongly limit the competitive constraints generic players could exercise on R&D players.

6.2.2. Distribution channel appears to act more as a barrier to entry and expansion for smaller competitors than as a competitive constraint to the Parties

6.2.2.1. Parties' arguments

(528) In the Form CO, the Parties argued that they "face substantial countervailing bargaining power by their sophisticated customers, namely distributors and agricultural cooperatives". Moreover, they consider that distributors "constrain the manufacturers’ pricing" because they "multi-source a wide range of products from different suppliers and have a strong influence on the farmers’ choice".

6.2.2.2. The Commission's assessment

(529) Against such claims, the Commission provisionally considers that countervailing purchasing power of distributors would not be sufficiently important to exercise constraints on suppliers. This is for the following reasons.

(530) First, certain EEA Member States such as Italy and Spain have a very fragmented distribution network in which the five main distributors account for less than 20% of total crop protection products sold.In such countries distributors' bargaining power is likely limited as they individually represent only a fraction of the total sales that agrochemical companies make.

(531) Second, distribution is rather concentrated in a range of EEA Member States, including Austria, Belgium, the Czech Republic, Germany, and the UK where the five main distributors represent over 50% of total crop protection products sold. However, even where distribution networks are concentrated it is unlikely that distributors would exercise significant countervailing buyer power in their negotiations with crop protection companies. The following considerations support this conclusion.

(532) In the first place, distributors' profitability is influenced by rebates granted by their main suppliers. As one competitor explained "[l]oyalty programmes are to be seen as considerable price reductions and incentive programmes that many distributors depend on as an additional income source".The Commission's investigation

242 See, among others, Form CO, part B.I, paragraphs 285 and 286.

243 Parties' response to the Commission's request for information RFI 47, file name "M7932 Annex DowDuPont_RFI47_7_1.XSLX" (ID7150).

244 Parties' response to the Commission's request for information RFI 47, file name "M7932 Annex DowDuPont_RFI47_7_1.XSLX" (ID7150).

245 Agreed non-confidential minutes of a call with a competitor, 23 September 2016 (ID8870).

81

246shows that rebates can be very substantial and in some instances exceed 30%. Moreover, rebates (including rebates offered by the Parties) are dependent in part on volumes of purchases but also on a range of other criteria, such as marketing, support, range, targets, which provide a certain margin of discretion to crop protection companies on whether and to what extent rebates are granted. It must also be noted, that crop protection companies tend to offer loyalty inducing rebates such as retroactive rebates, target rebates and discretionary rebates.On balance, rebates appear to be a tool for crop protection companies to encourage distributors' loyalty and constrain their ability to source crop protection products from a broad range of suppliers.

(533) In the second place, the Commission's investigation indicates that distributors are unlikely to be materially impacted by price increases applied by crop protection companies because they are often in a position to pass on such price increases to the final customers. For instance, one competitor indicated that "it is understood that whenever the agrochemical companies increase the price of their products, this increase will be reflected on the final price for costumer. However, this rule is not applied in case of decreasing pricing as only the distributors will profit from it. There is a considerable delay in the passing on of lower prices to farmers. Therefore, distributors are not significantly affected by price increases implemented by the large agrochemical companies. If prices go up it is the final customers that suffers and not the distributors" moreover this competitor pointed out that "[t]he distributors' margins have been increasing in recent years".Therefore, distributors' incentives to counter price increases by crop protection companies is likely limited.

(534) In the third place, distributors face constraints in the number of products from different crop protection producers they can offer. This is mainly due to the need for distributors to streamline their procurement and inventory management and to limit the financial risks of carrying a large and diversified inventory. As one competitor explained "[e]ach product line (SKU) creates cost of logistics from catalogue to warehouse and delivery. This indirectly restricts the number of products offered". Indeed, distributors have an incentive to focus their procurement on suppliers offering a large number of high-volume products. In this respect, one competitor pointed out that "to have big volumes [distributors] have the tendency to limit the n° of products on their catalogue".Moreover, such a sourcing strategy facilitates the achievement of volume and loyalty rebates from suppliers. Therefore, while distributors have in principle an interest in offering alternative options to their customers in practice they tend to source a limited number of products, often two or three, for each crop/pest combination. In the words of one competitor "some distributors will choose to carry a limited number of products/ brands to address a specific customer need (ie 2-3 seed brands for a given species; one branded and one generic offer for a given chemical / target)".Another competitor indicated that distributors "limit the number of products mainly to limit the difficulties in purchase, traceability and storage management. Therefore most distributors have only 3-4 products per segment in their range".In turn, this limits distributors' ability to multisource and to leverage the threat of switching vis-à-vis their suppliers.

252 Competitor's response to the Commission's request for information to competitors on distribution (ID6523).

253 Competitor's response to the Commission's request for information to competitors on distribution (ID6874).

(535) In the fourth place, as the Parties argue, the merged entity would not be in a unique position to leverage the range of products it would offer post-Transaction to win sales with distributors. Indeed, Adama, BASF, Bayer, and Syngenta hold portfolios that are similar in scope, if not larger, than that of the merged entity. However, the combined effect of the strength of the main crop protection companies in the distribution channel, thanks to their ability to offer rebates, the range of their products, the presence of their catalogues of must-have products, and their ability to generate farmers' demand thanks to their presence in the territory of the individual EEA Member States, restricts the ability for smaller generic competitors to win sales with distributors. One competitor in this respect indicated that "[b]ig suppliers with a dominant portfolio are able to push their product catalogue through and limit smaller suppliers in their vertical market access"while another pointed out that "[l]arger multinationals with aggressive, high value rebate schemes have the ability to shut out smaller competitors by offering rebates on aggregate purchases of a variety of products. As a result, smaller competitors who cannot match the breadth of product offerings may have difficulty competing for distributor shelf space".Therefore, the distribution channel appears to act more as a barrier to entry and expansion for smaller competitors than as a countervailing factor to the Parties' and their main competitors' market position.

254 Competitor's response to the Commission's request for information to competitors on distribution (ID6647).

255 Competitor's response to the Commission's request for information to competitors on distribution (ID6869).

6.2.3. Parallel trade appears limited in terms of countries, product range and volumes involved

6.2.3.1. Parties' arguments

(536) The Parties submit that the constraint represented by parallel imports on their competitive behaviour should be assessed on a country-by-country and market-by-market basis. Moreover, the Parties point out that, in Germany in certain segments parallel imports can be higher that the average figure (4-5%) indicated by the Commission in its Article 6(1)(c) Decision.

6.2.3.2. The Commission's assessment

(537) The Commission considers that the competitive constraint posed by parallel import would not be material to exercise constraints on the combined market power of the Parties. This is for the following reasons.

(538) First, while the share of crop protection products that is imported in each individual EEA Member State can vary it must be noted that the overall legal and commercial

252 Competitor's response to the Commission's request for information to competitors on distribution (ID6523).

253 Competitor's response to the Commission's request for information to competitors on distribution (ID6874).

254 Competitor's response to the Commission's request for information to competitors on distribution (ID6647).

255 Competitor's response to the Commission's request for information to competitors on distribution (ID6869).

83

framework in which parallel trade takes place is rather homogeneous across the EEA. Indeed, respondents to the Crop Protection Parallel Traders questionnaire reported a decrease in volume due to the tightening of conditions for legal parallel trade of crop protection products in the Union. In this respect a parallel trader reports that Regulation 1107 "led to a partitioning-off the market and a reinforcement of an existing oligopoly".Another submitted that national authorities in some countries are becoming increasingly restrictive to protect the domestic industry. For instance, in Slovakia, although parallel trade formally exists, it is hardly possible because of formal obstacles.

(539) Second, while some customers highlighted that "[p]arallel trade is a growing area", other respondents highlighted a limited scope for parallel trade and some technical and regulatory issues. For instance, a customer explained that it is "hesitant to engage in parallel trade, meaning the import of products from other EU countries to exploit price differences. It is technically and legally difficult. There have been some issue with products coming e.g. from France, the Netherlands or Poland. Labelling is not always correct. Parallel trade is a very small percentage of [company's] business, because [company] will have a good and reliable image."

(540) Third, for respondents to the Commission's Crop Protection Customers questionnaire, only a small number of distributors commercialise products obtained from other Member States via parallel imports. Distributors that engage in parallel imports tend to buy their products from neighbouring countries (for example for Germany from Poland, for Romania from Bulgaria etc.).

(541) Fourth, the indicative overview of companies registered in the EEA to trade their AIs between Member States submitted by the Parties highlights that only some products are concerned. It also concerns registration, and not actual sales.

6.2.4. The switching analysis submitted by the Parties does not allow the Commission to validate the claim of lack of closeness of competition between Dow and DuPont

(542) In the Form CO, the Parties report the results of a switching analysis which intends to show that the Parties' customers mostly switch to generic suppliers or other branded manufacturers rather than between the Parties. This analysis is further developed in their response to the Article 6(1)(c) Decision and in their response to the Statement of Objections. The underlying data and methodology were provided in response to the Commission's request for information RFI 6. This section refers to those three submissions as "Submissions on switching".

84

(543) The Commission has assessed the Parties' data, methodology and arguments in Annex 6. The Commission’s findings on such submissions can be summarised as follows.

6.2.4.1. Parties' data and methodology

(544) The Parties base their Submissions on switching on sales data of each of Dow and DuPont to all their customers for a selection of their products in the following categories: broadleaf herbicides, cross-spectrum herbicides and insecticides. For each category, the Parties selected their most important AIs and, on that basis, all products with one of these AIs as the primary AI. The data show the yearly amount of a particular product a customer bought from each of the Parties in a given year (in 2013, 2014 and 2015) for each EEA country, except for Iceland and for Liechtenstein. The analysis is performed only on 20 countries (referred to as "EEA-20" in this section).

(545) Since neither Dow nor DuPont track wins/losses of customers on a systematic basis, the Parties claim that no better data are available to them for the purpose of that exercise.

(546) For the purpose of their analysis, the Parties define, in each product category, a customer group called "potential switchers" by identifying customers for whom, first, the value of sales of products in this category from either Party fell by more than a given threshold, 25% in their base analysis, and for whom, second, the value of sales of products in this product did not fall for both Parties at the same time. Customers who reduced their purchases from both Parties are excluded from the analysis on the basis that such scenario could reflect a reduction in customer's total demand rather than a switch of the customer from one supplier to the other.

85

(547) For the customers identified as "potential switchers", the Parties compute the proportion of sales which switched from one Party to the other as follows. The results are then aggregated both at national and at EEA-20 level. The proportion of sales switched from one Party to the other Party is calculated by dividing the potential switchers' sales lost from one Party to the other Party, with the total sales loss of potential switchers.

(548) For the sake of the argument, it is assumed that a customer decreased its purchases from Dow by more than 25% between 2013 and 2014. If, in the same period, this customer did not increase its purchases from DuPont, then all the sales lost by Dow are flagged as resulting from a switch from Dow to other competitors. If, on the contrary, this customer increased its purchases from DuPont, then all the sales gained by DuPont are flagged as resulting from a switch from Dow to DuPont. If, on top of that, this customer increased its purchases from DuPont by less that it decreased its purchases from Dow, then, the difference between the sales lost by Dow and the sales gained by DuPont is flagged as resulting from a switch from Dow to other competitors.

(549) In response to the Article 6(1)(c) Decision, the Parties also provide two variations of the initial analysis in which potential switchers are defined by customers for whom the value of sales in a product category from either Party fell by more than either 0% or 50% (rather than 25%) but did not fall for both Parties at the same time.

(550) Finally, in the response to the Statement of Objections, the Parties make a supplementary analysis in which (i) potential switchers include customers for whom the value of sales in a product category from either Party fell by more than a given threshold between 0% or 50% but did not fall for both Parties at the same time, and (ii) potential customers also include, to some extent, customers for whom the value of sales in a product category increased from both Parties or decreased from both Parties. The switching rates for these customers are computed following the rule used in stylised examples provided by the Commission in its Statement of Objections for the purpose of showing the strong limitations of the data provided by the Parties.

6.2.4.2. Parties' arguments

(551) The Parties argue that their selection of "potential switchers" is conservative in the sense that it excludes instances of possible switching away from the Parties to other firms, as some of these changes in purchases might also be considered as demand reductions.

(552) On a EEA-20 level and with the 25% threshold, the Parties' analysis shows that the value of customer sales switched between the Parties is low for both broadleaf and cross-spectrum herbicides: in broadleaf herbicides the proportion is below 10% for both Parties, while in cross-spectrum herbicides the proportion is even lower (below 1%) for switching from Dow to DuPont and approximately 16% for switching from DuPont to Dow. On insecticides, the Parties' analysis shows a relatively low proportion of switching for both directions, around 10%. The Parties interpret the results of their analysis as evidence that there is low switching and that the Parties are therefore not close competitors.

278 Form CO, part B.I, paragraph 189.

279 The Parties also provide a similar analysis with potential switchers defined by different thresholds, namely potential switchers being customers for whom the value of sales in a product category from either Party fell by more than 0% or 50% (rather than 25%) but did not fall for both Parties at the same time.

(553) In their supplementary analysis, the Parties’ own computations result in a switching rate for broadleaf herbicides around 15% from Dow to DuPont and around 21% from DuPont to Dow. For cross spectrum herbicides, the switching rate lies between 15% and 25% from Dow to DuPont and between 26% and 34% from DuPont to Dow. For insecticides, according to the Parties, the switching rate is between 15% and 19% from Dow to DuPont and between 23% and 27% from DuPont to Dow.

6.2.4.3. Commission's assessment

(554) The Commission considers that the switching analysis submitted by the Parties does not allow the Commission to validate their claim of lack of closeness of competition between Dow and DuPont, for the following reasons.

(555) First, the Commission notes that the data provided by the Parties suffer from many limitations. In particular it does not include sales from competitors, and it does not provide any indication of customer specific characteristics or demand specific characteristics, for example whether the demand has increased or decreased over time. This in itself limits the ability to identify precisely customers who have switched purchases between Dow and DuPont and other suppliers.

(556) In their response to the Article 6(1)(c) Decision, the Parties essentially agree that a clear differentiation between switching and demand reductions is not possible. Nevertheless, they argue that large changes in the quantity bought by one of the Parties are likely to reflect switching while small changes are more likely to reflect overall fluctuations in the market environment. In their response to the Statement of Objections, the Parties do not address the fundamental criticism that such data are not suited for a meaningful switching analysis.

280 It is worth noting that the Parties misrepresent the Statement of Objections main contention against their switching analysis. Contrary to what Annex 5 of the response to the Statement of Objections mention in its first paragraph page 18, the main contention of the Statement of Objection against their switching analysis is not related to the exclusion of common customers, but the main contention is that the data provided by the Parties suffer from many limitations, so that they are not suited for a meaningful switching analysis.

(557) Second, as a consequence of this lack of certainty on which customers do switch purchases between Dow and DuPont and other suppliers, the Parties make assumptions which exclude from their analysis a large proportion of customers and sales. Common customers that increase or decrease their purchases from both Parties represent a significant share of the Parties' sales, which could include significant switching between the Parties. All these customers are excluded from the analysis and only partially incorporated in the supplementary analysis.

(558) Here also, the Parties agree with the Commission's observations in the Article 6(1)(c) Decision that situations in which customers decrease or increase their sales from both Parties simultaneously could include occurrences of switching of sales between the Parties. Nevertheless, they mention that in the absence of specific information on the actual behaviour of these customers, it is not possible to incorporate these customers in the switching analysis.

(559) The Parties also argue that excluding those customers do not bias the results towards less switching between the Parties since these sales could also include switches to other firms, and customers for which purchases from both Parties fell would be less likely to have switched between the Parties than other customers' categories included in the Parties' analysis. The Commission provides stylised examples showing that, in fact, the Parties' claims cannot be sustained. Moreover, the Parties’ supplementary analysis, which partly includes customers that increase or decrease their purchases from both Parties, results in significantly higher switching rates than their initial analysis.

(560) Third, the Parties also include in their switching analysis customers who never bought from one of them. On the one hand, the fact that these customers never bought from one of the Parties suggest that this supplier was not present in these customers' choice set, and as a consequence they did not have the possibility to switch purchase to this supplier. On the other hand, the switching analysis assumes that any decrease in the purchase from the supplying Party is switched away to other competitors, whereas such decrease may just be due to a change in demand. The Parties' approach therefore mechanically underestimates the switching rate between the Parties.

(561) In their response to the Statement of Objections, the Parties argue that excluding these customers “disregards relevant customers that do not view the Parties’ products as substitutes and therefore biases upwards the switching rates between the Parties presented in tables 4, 5 and 6 of the [Statement of Objections].” The lack of methodological information provided by the Parties in relation to these empirical exercises makes it difficult to the Commission to address that argument. The mere existence of such customers might related to errors in handling the data, or difficulty in cleaning and matching Dow’s and DuPont’s customers. It could as well result from the fact that the analysis proposed by the Parties mixes products that belong to different (downstream) relevant markets for formulated products.

281 Parties’ response to the Statement of Objections, annex 5, entitled […] (ID10040), section 3.2.

(562) Fourth, the analysis provided by the Parties suffers from significant volatility of switching rates, depending on the period considered. Such volatility typically requires more attention and, absent robustness checks, casts doubts on whether any inference on closeness can be made from the Parties' analysis.

(563) Fifth, if anything, the Commission's assessment of the data provided by the Parties in support of their Submissions on switching provides indication of higher switching rates between Dow and DuPont than those provided in their Submissions. The Parties’ supplementary analysis also results in significantly higher switching rates than their initial analysis.

(564) Sixth, and without prejudice to the Commission’s assessment of the relevance of the Parties’ switching analysis, the Commission notes that, given the levels of the Parties’ margins for their crop protection products and given the levels of switching rates (even as computed by the Parties), standard techniques used for measuring likely price effects of merger would lead to the anticipation of significant price increases.

(565) Seventh, the limitations of the data provided by the Parties are such that the Parties' switching analyses need to make several critical assumptions which are not correct as soon as, for example, the customers' demand increase or decrease over time.

(566) Finally, and more generally, the Commission is of the view that it is not possible to perform a meaningful switching analysis when, as it is the case in the Submissions on switching, the only data available are sales data from both Parties to their customers, with no understanding of sales made by other suppliers, customers’ characteristics or markets’ characteristics and evolution.

6.3. Herbicides

6.3.1. Introduction to herbicides

6.3.1.1. Herbicide use in agriculture

(567) Herbicides are crop protection products aimed at preventing or controlling weeds which compete with crops for water, light, nutrients and space. Since the Parties' 282 activities in non-selective herbicidesare limited, the main focus of this Decision is on selective herbicides, which are herbicides designed to eliminate only the weeds while leaving intact the crop to which they are applied.

(568) Selective herbicides are usually intended for use in specific crops, such as rice; some are registered for use in several crops, for instance cereals and oilseed rape. The type of crop is thus the key element when considering the application of selective herbicides.

(569) There are, however, thousands of weeds in the EEA. An AI will typically target either grasses or broadleaf weeds, but it can also be effective against a broader range of weeds from both weed categories. The target weed is another distinguishing element for selective herbicides.

(570) In the EEA, from a demand-side perspective, farmers usually face at least 10 different varieties of weed that they seek to eliminate, which fall into the two main categories of broadleaf weeds and grass weeds. This requires looking at specific weeds when examining the demand for herbicides. Certain weeds are more decisive for farmers' choices than others as they appear in higher numbers and may therefore have a considerable impact on yields. Those key weeds vary among crops and regions. Typical examples of key broadleaf weeds in the EEA include Galium aparine (cleavers), Stellaria media (chickweed) and Rumex (dock). Key grass weeds include Avena fatua (wild oat), Echinochloa crus-galli (barnyard grass) and Lolium multiflorum (annual ryegrass).

282 See recital (628).

(571) When farmers seek to eliminate both weed categories, they can either buy separate products for grass weeds (graminicides) and broadleaf weeds (broadleaf herbicides) or buy ready-made products that are effective against a range of weeds from both categories ('cross-spectrum' or 'broad-spectrum' herbicides). Different products may also have different levels of efficacy on certain weeds compared to others.

(572) Formulated herbicide products can either contain a single AI or a mixture of two or more AIs. The AI determines the core characteristics and efficacy of the formulated product. In herbicides, mixtures of several AIs are more common than in other crop protection denominations such as insecticides.

(573) When choosing a selective herbicide, farmers also need to consider at what time during the growing season and at what point in the crop's lifecycle they need to apply a product. For instance, in cereals, certain herbicides are applied in autumn while others are used in spring. Whether autumn application is required depends on the crop type, namely winter crops or spring crops, and the specific time of sowing: early, mid- or late autumn. The timing of application is further limited to a specific period in the plant's lifecycle: before sowing (pre-plant herbicides), after sowing but before germination (pre-emergence herbicides) or after germination (post-emergence herbicides).

(574) One problem facing herbicide use in crop growing is the development of resistance in weeds. This occurs when the same herbicide or various herbicides with the same MoA are used repeatedly. Resistant mutant biotypes survive, resulting in an increasing number of plants until a majority of plants may become resistant. After a weed has become resistant to a certain MoA, it may become resistant to all products with the same MoA and all products in the same chemical class.

(575) From a supply-side perspective, herbicide AIs are to be classified into MoA groups and chemical classes. MoA groups consist of herbicides that use the same mechanism to kill weeds, from absorption to death. There are currently over 20 MoA groups in herbicides. Chemical classes of herbicides group together AIs which present chemical similarities in terms of their molecular structure and physical properties. Members of the same chemical class all have the same MoA but one MoA group will cover several chemical classes.

(576) Because weed resistance limits the lifecycle of herbicides, herbicide manufacturers recommend rotation as a means to control the development of herbicide-resistant weeds. Farmers are advised to rotate herbicides annually and during the growing season. In the EEA the European Herbicide Resistance Action Committee (EHRAC) promotes the development of resistance management strategies in order to maintain efficacy and reduce the risk of resistance. Other bodies provide guidance to farmers, including, depending on the country, institutes, crop consultants, distributors, cooperatives and crop protection companies. Based on these recommendations, farmers decide which products they will use to suit their specific circumstances.

(577) Herbicides require two types of authorisation before they can be sold to farmers in the EEA. First, the AI must be authorised by the Commission, on the advice of EFSA. AIs that are given Union-wide authorisation are included in Annex I of the Authorisations Directive (Directive 91/414/EC). Second, the crop protection product incorporating the AI (alone or as part of a mixture of AIs) must be authorised by the competent authority in the Member State(s) where it will be sold. For more information on those authorisations, see Section V.1.4.3.

90

(578) In the early stages of the research process, agrochemical companies typically seek to 283 register patents covering a chemical genus that includes a new AI. Later on in the research process, supplementary patents for composition of matter can be registered to protect the chemical sub-genus identified. Patents can also be registered for processes ("process patents") or formulations. In the EEA, patents offer protection for a period of 20 years from the date of application, regardless of whether they are exclusively national or based on a European patent. The patent protection period in the EEA can also be extended through supplementary protection certificates. For more information on patents and other forms of intellectual property protection see Section V.1.4.4.

6.3.1.2. Current turnover figures

(579) Total herbicide sales amount to about USD 25 billion globally, and around USD 4 billion in the EEA. Selective herbicides represent the large majority of herbicide sales in the EEA, at around USD 3.4 billion (USD 19 billion market 284globally) in 2015.

(580) Cereals, corn, oilseed rape, fruits and vegetables and beets are the main crops for herbicides in the EEA. Cereals accounted for 35% of herbicides sales in the EEA in 2015 (USD 1.5 billion), corn 20% (USD 800 million), oilseed rape 12% (USD 490 million), fruits and vegetables 12% (USD 480 million) and beets 7% (USD 277 million). Rice represented 1.5% (USD 60 million), pasture between 0.3% and 0.5% (estimates range between USD 11 million and USD 20 million) and 285sunflowers 5% (USD 200 million).

(581) In this Decision, in light of the Parties' activities in the EEA, the relevant crops for the competitive assessment on selective herbicides are cereals, rice, pasture, oilseed rape, sunflowers, beets and corn.

6.3.2. Product portfolios of the Parties

(582) The Parties have strong product portfolios globally, notably in selective 286herbicides.

(583) The Parties have very limited activities in non-selective herbicides, which do not give rise to any affected markets in the EEA. The Parties‘ sales of non-selective 287 herbicides are small: USD 5.2 millionin the EEA in 2015 for Dow (EEA market share: 2%) and less than USD 200 000 in 2015 for DuPont, which were limited to 288 Greece (EEA market share: 0%).The Decision will therefore review only the Parties’ portfolios of selective herbicides, as they are the only products giving rise to affected markets.

283 A taxonomic rank used in the biological classification of living and fossil organisms.

284 Parties' response to the Commission's request for information RFI 29, question 3.1.

285 Parties' response to the Commission's request for information RFI 8 and follow-up requests, and Agrowin data.

286 Form CO and estimates of the Parties (Parties' response to the Commission's request for information RFI 8 and follow-up requests).

287 Estimate of the Parties (Parties' response to the Commission's request for information RFI 29, question 3.1).

288 Form CO, part B.I, page 14; and Form CO, Annex B.1.6.2.

91

(584) Both Parties develop, manufacture and sell a broad range of different herbicides to be applied to all main crops across the EEA, including beets, cereals, corn, cotton, fruits and nuts, grapes/vines, oilseed rape, plantation (tobacco), pasture, potatoes, rice, vegetables and sunflowers. Both companies also have the technology to develop selective herbicides that can be applied across crops.

(585) While the Parties are strong in products targeting broadleaf weeds, and are notably 289 leaders in cereals,they also have graminicides and cross-spectrum products in their portfolios of selective herbicides.

(586) Sections V.6.3.2.1 and V.6.3.2.2 describe the herbicide portfolios of Dow and DuPont respectively, including current and pipeline products in the EEA. AI names will be mainly used, as brand names tend to vary by country. When sales are given by AI for mixture products, the sales figure is usually allocated to the "lead AI" in the mixture (typically the most recent of the constituent AIs).

(587) The Parties' selective herbicides for cereals, rice, pasture, oilseed rape, sunflowers, beets and corn will be examined further in the relevant crop sections.

6.3.2.1. Dow

(588) Herbicides have traditionally represented the majority of Dow's agrochemical 290 business (61.4% in 2014). Growth of this portfolio is driven by strong demand for recently introduced AIs, notably penoxsulam, pyroxsulam and aminopyralid.

(589) The largest market for Dow in Europe is France with 2014 sales estimated at USD 195 million, followed by Germany (USD 125 million), Spain (USD 60 million), Italy (USD 55 million) and the United Kingdom 291 (USD 50 million).Key Dow herbicides for the European market include pyroxsulam, fluroxypyr, florasulam, clopyralid and propyzamide, which together accounted for 75% of Dow's herbicides sales in the EEA in 2015.

(590) The company itself recognises its strength in this segment: a 2015 Dow internal document notes that Dow is "#2 Cereal Herbicide Company – all products - Global" 292and "#1 in the Broadleaf weed herbicide segment".

289 "Dow is the leading company globally for products controlling broadleaf weeds in cereals while DuPont is number two." See DuPont's internal email "Strategic Considerations regarding Ag Consolidation BoD prep DuPont (Pioneer)", July 2015, file name: "DOC-000000039.pdf" (ID1329-39).

290 Phillips McDougall – AgriService, Companies Section, Part 1 The Leading 15 Companies – 2014 Market Situation, page 202.

291 Phillips McDougall – AgriService, Companies Section, Part 1 The Leading 15 Companies – 2014 Market Situation, page 201.

292 Dow's internal document "Cereals and Broadleaf Crops Herbicide Portfolio Technology Review", 21 September 2015, file name "DAS-00000118-000001.pdf" (ID445-119), slide 9.

92

Table 2 – Dow's herbicides portfolio in the EEA

Overall EEA sales Main category ofStatus in Chemical in EUR AI name relevant weedsthe EEA class million crop(s) 293 (2015)targeted

Comments

Patent expires in Sold in Cross- Triazolopyri 2022, data Pyroxsulam Cereals 123 the EEA spectrummidines protection in Union until 2024

Sold in Cereals, Florasulam the EEAMaize

Triazolopyri BLW 74.6 midine

Pyridine carboxylic acid

Sold in Cereals, Fluroxypyr the EEAMaize

BLW

70.0

Oilseed rape, Sold inCereals, Clopyralid the EEABeets, Pasture

Pyridine carboxylic acid

BLW

58.5

Oilseed rape, Sold inCereals, Grapes, Cross- Propyzamide Benzamide 43.0 the EEAFruits,Spectrum Vegetables

Rice, Sold in Cross- TriazolopyriPatent expires in Penoxsulam Cereals, 28.0 the EEA Spectrummidine 2017 Fruits

Non-crop, Sold in Triclopyr Cereals, the EEA Grassland

Pyridine carboxylic acid

BLW

24.5

Maize, Phenoxy- Sold inCereals, BLWcarboxylic- 16.0 the EEAFruits, Nuts, acid Pasture, Rice

2,4-D

Oilseed rape, Sold in Aminopyralid Cereals, the EEA Pasture, Rice

Pyridine carboxylic acid

Patent expires in 15.0 2021

BLW

Cereals, Sold in Isoxaben Perennial the EEA crops

BLW Benzamide 12.6

Pasture, Sold inCereals, Picloram the EEAMaize, Oilseed rape

Pyridine carboxylic acid

BLW

9

Oilseed rape, soybeans, Sold in Haloxyfop-P fruits and the EEA vegetable, cotton, beets

Aryloxyphen Grassesoxypropiona 6.6 tes

293 "BLW" stands for broadleaf weeds.

93

Overall EEA sales Main category ofStatus in Chemical in EUR AI name relevant weedsthe EEA class million crop(s) 293 (2015)targeted

Comments

Aryloxyphen Grassesoxypropiona 5.2 te

Sold in Cyhalofop the EEA

Rice

Pyridine carboxylic acid

Sold in Cereals, Phenoxies the EEAMaize

BLW

4.0

Fruits, Sold in Cross- Dinitroanilin Oryzalin Vegetables, 3.5 the EEA Spectrum es Vines

Launch in 2016/2017 for cereals; in 2019 for OSR; in 2021 for sunflowers; in 2020 for grassland; in 2019 for orchards

Launch in Cereals, Arylex2016 to OSR, (halauxifen)2021 insunflowers, the EEA etc

Arylpicolina BLW te

Launch in Italy sometime in 2021 and in the rest of the EEA in 2022

Rinskor Launch as Cross- Arylpicolina (florpyrauxifenRice, Pasture from 2021 Spectrum te -benzyl)

Fruits, PPO RecentlyVegetables, Cross- Diphenylethe 9.7 divestedPlantationSpectrum Nutrichem r crops

Recently Glyphosate Vines Cross- spectrum Divested in 2016 to 10.7 Nutrichem

EPSP

Recently Peanuts, Cross- Dinitroanilin Ethalfluralin divestedSunflowerSpectrum es

Divested in 2015 to Gowan Company

Recently Fruits, Cross- DinitroanilinDivested in 2015 to Benfluralin 4.5 divestedVegetablesSpectrum es Gowan Company

Recently Cereals, Trifluralin divestedCotton

Dinitroanilin es

Divested in 2015 to Gowan Company

Phenoxy carboxylic acid

rd 3 party AI

1.3

BLW

rd 3 party AI

Chlorotoluron

Cross- spectrum

Urea

rd 3 party AI

Diflufenican

Pyridinecarb oxamide

rd 3 party AI

Mesotrione

BLW Triketone

94

Overall EEA sales Main category ofStatus in Chemical in EUR AI name relevant weedsthe EEA class million crop(s) 293 (2015)targeted

Comments

rd 3 party AI

Metazachlor

Cross- Chloroaceta s spectrum mide

rd 3 party AI

Oryzalin

Cross- Dinitroanilin spectrum es

rd 3 party AI

Pendimethalin

Cross- Dinitroanilin spectrum es

Source: Parties' submissions

(A) Existing products

(591) The offering of selective herbicides in Dow’s portfolio can be grouped by chemical class and MoA, as is usually the case in the agrochemical industry (see 294recital (574)).

(592) Dow offers products in the chemical class of pyridine carboxylic acids, which are part of the synthetic auxin MoA group. They account for a large part of Dow’s agrochemical portfolio, with worldwide sales in excess of USD 1 billion. Dow’s leading pyridine herbicides are picloram and fluroxypyr, with global sales of around USD 250 million in 2014. Clopyralid, aminopyralid and triclopyr also belong to this chemical class.

(593) Dow’s portfolio also includes herbicides in the chemical class of triazolopyrimidines, which act by inhibition of acetolactate synthase (ALS). This group makes up a large part of Dow's herbicide portfolio, with global sales totalling approximately USD 900 million in 2014. Dow's main AIs in this class are penoxsulam, pyroxsulam and florasulam, which all have large sales globally.

(594) Dow is one of the leading manufacturers of phenoxy carboxylic acid herbicides worldwide, which also belong to the synthetic auxin MoA group. Its worldwide sales of this product group reached approximately USD 255 million in 2014. The range is led by 2,4-D, which is widely used both in crop and non-crop situations, followed by MCPA and a number of other analogues.

(595) In addition, Dow produces herbicides in the aryloxyphenoxypropionate chemical class, which is part of the acetyl CoA carboxylase inhibitor (ACCase) MoA group. This range includes the graminicide cyhalofop, which was introduced in 1996 for use in rice to control grass weeds, including barnyard grass. Sales in 2014 are estimated to have reached USD 155 million worldwide, with results benefiting from both co-formulations and tank-mix usage.

(596) Another member of the aryloxyphenoxypropionate chemical class is Dow graminicide haloxyfop, which was first introduced in 1986 and is sold for post-emergence grass weed control on soybean, oilseed rape, fruit & vegetables, cotton

294 See the product portfolios of the leading 15 agrochemical companies discussed in Phillips McDougall, – AgriService, Companies Section, Part 1 The Leading 15 Companies – 2014 Market Situation.

295 Phillips McDougall, – AgriService, Companies Section, Part 1 The Leading 15 Companies – 2014 Market, page 204.

95

and sugarbeet. Global sales in 2014 are estimated to have reached 296USD 100 million.

(B) Forthcoming products

(597) Dow has recently developed the new chemical class of arylpicolinates, of which Arylex (AI halauxifen-methyl) is the first member. It is a new broadleaf auxinic herbicide specialised in hard-to-control weeds worldwide with utility in multiple crops and uses. It is intended to target primarily Galium, as well as other weeds such as Conyza, Chenopodium and Papaver in cereals, oilseed rape, sunflowers, pasture, olive trees and citrus. Arylex is expected to be a key development for Dow's cereal herbicide portfolio.Dow has recently obtained the first commercial authorisations of Arylex-based products for use in cereals in the Union. It plans to continue the launch of Arylex-containing products in the EEA in 2016-2017 in cereals and from 2019 to 2021 in other crops (oilseed rape, orchards, pasture, sunflowers). In most markets Arylex will be sold in combination with other Dow herbicides, such as […]. It is designed to rejuvenate existing Dow products by developing new markets and segments and to protect and regain market positions from generics through 298 mixtures.Dow estimates that total sales of Arylex will generate USD […] million 299revenue at maturity.

(598) Another member of the new chemical class of arylpicolinates is Dow’s late-development product Rinskor (florpyrauxifen-benzyl), a cross-spectrum herbicide for the post-emergent control of grasses, broadleaf weeds and sedges, including resistant species, primarily in rice but also in pasture. It has a synthetic auxin MoA, which is new in rice crops. Formulated products based on this molecule are not yet registered in the Union. Dow applied for Union registration of Rinskor in March 2016 and expects approval in 2021. It aims to launch a Rinskor-formulated product in Italy sometime in 2021 and in the rest of the EEA in 2022. Dow estimates that Rinskor will generate revenues of USD […], which will help Dow grow its [internal 300 assessment] .

301(C) Early pipeline products

302 303(599) Dow has a [information about herbicide pipeline 1].

304 305 306(600) In addition, Dow has a [information about herbicide pipeline 2].

296 Phillips McDougall – AgriService, Companies Section, Part 1 The Leading 15 Companies – 2014 Market Situation, page 210.

297 Dow's internal document "Cereals and Broadleaf Crops Portfolio Technology Review", dated 21 September 2015, file name "DAS-00000118-000001.pdf" (ID445-119), slide 7.

298 Dow's internal document "GBL 2014 Plan-SVM Review-Supply Cereals&BLC", file name "DAS-00000400-000001", 2014 (ID729-000009), slide 5.

299 Dow's internal document "PMF Arylex Active 2016 Update", 20 October 2016 (ID8833-92), file name "M7932 Annex Dow RFI 59_14.2 - cPMF Pipeline ARYLEX Active for 2016 (Nov 4).pdf", slide 4.

300 Dow's internal document "CP PMF RINSKOR PROJECT UPDATE 2015", December 2015, file name "40 .DAS-00000040-000001.pdf", slide 5.

301 The Commission points out that notwithstanding its description of certain discovery products in this section, it has conducted its final competitive assessment on the sole basis of existing products and forthcoming, that is to say development or near-development products.

302 Dow's internal document "Crop Protection Discovery Pipeline" (ID1328-131), 10 December 2015, file name "DOC-000000131.pdf".

303 Dow's internal document "Confidential--CP_Discovery_Pipeline_Dec_2015_Final", 10 December 2015 (ID6696-29505), file name: DAS-10216008.pptx.

96

(601) For more information on Dow's early pipeline research in herbicides see Sections V.8.8.1.3. to V.8.8.1.5

(D) Divestitures

(602) According to the information provided by the Parties, Dow has recently made some divestitures in its herbicides portfolio. However, Dow maintains links in relation to certain divested molecules. One divestiture was to Gowan, namely the DNA portfolio, a group of molecules including trifluralin, benfluralin, ethalfluralin. Internal documents indicate that the rationale for the divestment was "[b]usiness is 307 highly commoditized with low profitability".However, while the Parties submitted that "[t]here are no more remaining links with Gowan", internal documents indicate 308that Dow would distribute and toll formulate for several months.

(603) Another divestiture, also in 2015, was oxyfluorfen to Nutrichem. Internal documents indicate that the rationale for the divestment was that Oxyfluorfen is a forty year old generic molecule that currently has reasonable margins, with a potential for margin erosion and increased regulatory costs. Dow has limited resources that are better focused on new patent-protected technologies and actives that carry higher margins. 309 Dow will retain ownership of the oxyfluorfen mixture products.However, internal 310documents also indicate that Dow would "enter into a 5 year Supply Agreement", 311and that oxyfluorfen is still seen as part of the portfolio in 2016.

6.3.2.2. DuPont

(604) Herbicides were traditionally the most significant sector for DuPont's agrochemical operations, supported by its strong portfolio of sulfonylurea (SU) products. However, the herbicide business of DuPont has more recently been replaced in relative importance by the insecticide products, currently representing almost 45% of 312DuPont’s total agrochemical sales.

Dow's internal document, "Confidential--CP_Discovery_Pipeline_Dec_2015_Final", 10 December 2015 (ID6696-29505), file name "DAS-10216008.pptx". Dow's internal document "Bisphosphonate Herbicide Project CPDT meeting", 12 February 2016 (ID6143-057697), file name "DAS-10057100.pptx", slide 15. Dow email "DAS-10038726.msg" – reply to RFI 37. Dow's internal document "Mergers and Acquisitions - Dow AgroSciences Members Committee'', dated 18 September, 2015, file name "10. DAS-00000532-000001.pdf", slide 4. Dow's internal document "Mergers and Acquisitions - Dow AgroSciences Members Committee'', dated 11 December 2015, file name "DAS-USPRIV-40032182", slide 5. Dow's internal document "Divestiture Transaction Proposal", 4 November 2015, file name "Annex Dow RFI9 27.2.pdf". Dow's internal document "Mergers and Acquisitions - Dow AgroSciences Members Committee'', 11 December 2015,, file name "DAS-USPRIV-40032182", slide 4. Dow's internal document, "SVM - 10 YR Plan Review – Rice, Trees & Vines", May 2016 (ID742), file name "M7932_Annex 43a.6_CONFIDENTIAL.xlsx". Phillips McDougall – AgriService, Companies Section, Part 1 The Leading 15 Companies – 2014 Market Situation, page 240.

97

Table 3 – DuPont's herbicides portfolio in the EEA

Overall EEA sales Main Sales status category of Chemical in EUR AI name relevant in the EEA weeds class million crop(s) targeted (2015)

Comments

Cereals, Thifensulfuro Sold in theSoybeans, n EEAMaize, Pasture

BLW Sulfonylurea 25

Cereals, Tribenuron- Sold in theSunflower, BLW Sulfonylurea 37 m EEAPasture, Rice

Cereals, Triflusulfuro Sold in theMaize, n-methyl EEASoybeans, Beets, Rice

BLW Sulfonylurea 25

Maize, Sold in thePotato, Rimsulfuron EEA Fruits, Sunflower

Grasses Sulfonylurea 19

Flupyrsulfuro Sold in the Cereals, Cross- Sulfonylurea 17 n EEAPasturespectrum

Sold in the Maize, Nicosulfuron EEA Fruits

Grasses Sulfonylurea 16

Cereals, Sold in the Metsulfuron Rice, OSR, BLW Sulfonylurea 11 EEA Pasture

Beets, Sold in thePotatoes, Lenacil EEA fruits, vegetables

BLW

Uracil

8.5

Ethametsulfu Sold in the Oilseed ron-methyl EEA rape

Union registration BLW Sulfonylurea 4.5 process ongoing

Chlorsulfuro Sold in the Cereals, Cross- Sulfonylurea n EEApasturespectrum

8

DuPont did not invest in the renewal of Annex I (expires in 2021)

Cross- Sulfonylurea spectrum

Sold in the Azimsulfuron EEA

1

Rice

Aminocyclop Not in EEA Pasture yrachlor

Recently Bensulfurondivested in the EEA

Cross- spectrum

Rice

Sourced from Nissan

rd Quizolafop 3 party AI

98

Overall EEA sales Main Sales status category of Chemical in EUR AI name relevant in the EEA weeds class million crop(s) targeted (2015)

Comments

Sourced from Syngenta

Only in mixture with nicosulfuron and rimsulfuron or bromoxynil /nicosulfuron /rimsulfuron

rd Mesotrione 3 party AI

BLW Triketone

4

Sourced from BASF

Only in mixture with rimsulfuron and/or nicosulfuron

rd Dicamba 3 party AI

BLW Benzoic acid 15

Amidosulfur rd 3 party AI on

BLW Sulfonylurea

Benzothiadia BLW zinone

rd Bentazone 3 party AI

Bromoxynil- rd 3 party AI octanoate

BLW

Nitrile

Cross- Triazolone spectrum

Carfentrazon rd 3 party AI e-e

Aryloxyphen oxy- Grasses propionate ‘FOPs’

rd Clodinafop 3 party AI

Isoxazolidin BLW one

rd Clomazone 3 party AI

rd Diclofop-m 3 party AI

Grasses

rd Diflufenican 3 party AI

BLW

Triazolopyri BLW midine

Sourced from Dow

rd Florasulam 3 party AI

Cross- spectrum

rd Fluometuron 3 party AI

Urea

Pyridine carboxylic acid

Sourced from Dow

rd Fluroxypyr 3 party AI

BLW

Cross- Imidazolino spectrum ne

rd Imazamox 3 party AI

Phenoxy- BLWcarboxylic- acid

rd MCPA 3 party AI

Phenoxy- BLWcarboxylic- acid

rd MCPP-P 3 party AI

99

Overall EEA sales Main Sales status category of Chemical in EUR AI name relevant in the EEA weeds class million crop(s) targeted (2015)

Comments

Chloroaceta Grasses mide

rd Pethoxamid 3 party AI

Aryloxyphen oxy- Grasses propionate‘F OPs’

Propaquizafo rd 3 party AI p

Terbuthylazi rd 3 party AI ne

BLW

Triazine

Source: Parties' submissions

(A) Existing products

(605) The offering of selective herbicides in DuPont’s portfolio can also be grouped by chemical class and MoA. DuPont's portfolio in the EEA consists mainly of products derived from 11 AIs of the SU family, as well as lenacil, and in-licensed products. The SUs thifensulfuron methyl, tribenuron methyl, nicosulfuron, triflusulfuron methyl, rimsulfuron and flupyrsulfuron account together for 75% of DuPont's 2015 herbicides sales in the EEA. Other DuPont SUs include metsulfuron, triflusulfuron, azimsulfuron, ethametsulfuron and chlorsulfuron.

(606) DuPont introduced the sulfonylurea chemical class in the early 1980s, enabling the company to compete successfully in a broad range of crop and non-crop applications. SUs are part of the ALS (acetolactate synthase) Inhibitors MoA group. This product group has recently become less important to DuPont. However, despite resistance issues, pressure from generic competition and newly developed competitive products DuPont’s SU product sales experienced modest growth overall in recent years. SU sales grew between 2010 and 2013 and sales in 2014 came to USD 828 million, accounting for 67% of DuPont's total herbicide sales.

(607) DuPont has not introduced any new SU products since 1997. However, the product range has been extended by new mixture formulations, blends, and new technology based on this chemistry, addressing specific niches in the market place.

(608) DuPont has an ongoing 'SUs renewal strategy'for its SU portfolio, notably with new mixtures and formulations. Some mixtures include Dow products, such as the RXR49 SU/florasulam mixture.The sulfonylurea products were initially offered in dry formulations, but the range has recently been extended with liquid formulations (liquid SUs). This new delivery form makes it easier for the products to be used in co-formulations for weed resistance management.