(1)

On 21 November 2011, the newspaper Le Soir published a series of articles according to which the Walloon Region (3) had been granting financial support to the steel group Duferco (‘Duferco’ or ‘the Duferco Group’) since 2003 without having informed the Commission. According to the Belgian newspaper, in March 2003 the Walloon Region created a new financial holding company, Foreign Strategic Investment Holding (‘FSIH’), a subsidiary of Société Wallonne de Gestion et de Participations (‘SOGEPA’), to invest in companies in the Duferco Group with a registered office outside Belgium or even the EU.

(2)

After receiving this information, by letter dated 29 November 2011 the Commission requested that Belgium send further information concerning the nature of the financial support that the Walloon Region had awarded to the Duferco Group between 2003 and 2011.

(3)

By letter dated 14 December 2011, Belgium requested an extended deadline for responding, which was granted on 16 December 2011. Belgium replied to the Commission's request for information on 4 January 2012. The Commission requested further information by letter dated 29 June 2012. Belgium requested an extension to the deadline for replying, which was granted on 18 July 2012, and replied on 28 September 2012.

(4)

A meeting took place on Commission premises on 6 May 2013 between the Commission and the Belgian authorities and representatives of SOGEPA and FSIH.

(5)

Belgium submitted further information to the Commission on 4 and 10 June 2013 and on 15 and 30 July 2013.

(6)

Another meeting took place on Commission premises on 30 September 2013 between the Commission and the representatives of SOGEPA and FSIH.

(7)

By letter dated 16 October 2013, the Commission informed Belgium of its decision to initiate the procedure provided for in Article 108(2) of the TFEU (the 'formal investigation procedure').

(8)

The Commission's decision to initiate the procedure (‘the opening decision’) was published in the Official Journal of the European Union (4). The Commission invited interested parties to submit their comments on the measures in question.

(9)

The Commission received comments from Belgium on 11 December 2013, after the latter had obtained an extended deadline. The Commission also received comments from interested parties on 23 and 24 June 2014 after also granting them an extended deadline. The Commission sent the aforementioned information to Belgium on 3 and 18 July 2014, giving it the possibility to comment thereon, and received its comments by letters dated 1 and 18 August 2014.

(10)

By letter dated 28 January 2015, the Commission asked Belgium to provide further information. Belgium requested an extension to the deadline for replying, which was granted on 25 February 2015, and submitted its reply to the Commission on 8 April 2015.

(11)

By letter dated 18 May 2015, the Commission requested more additional information. Belgium responded on 2 June 2015.

(12)

A final request for information was sent to Belgium on 15 September 2015. Belgium requested a deadline extension of one month to reply, which was granted on 28 September 2015. Belgium submitted its reply to the Commission by letter dated 23 October 2015.

(13)

The Duferco Group produces and sells steel. It has a presence in some fifty countries around the world. In Europe, the Group is centred in Belgium (approximately 3 000 employees in 2009 at a dozen sites in the Walloon Region) and Italy (approximately 1 300 employees in 2009). It is also active in Switzerland, Luxembourg, France and the Former Yugoslav Republic of Macedonia (FYROM).

(14)

Duferco has been operating in Belgium since 1997, when it took over Forges de Clabecq, which was in financial straits, with support from the Walloon Region (investment of EUR 8,75 million and loan of EUR 13,75 million) (5). In 1999, Duferco bought Usines Gustave Boël in La Louvière from the Hoogovens group, still with support from the Walloon Region (conversion of a claim of EUR 12,5 million into 25 % of the capital and a loan of EUR 25,3 million) (6). In 2002, Duferco took over the hot-phase site in Charleroi, which became Carsid, in cooperation with Usinor, again with support from the Walloon Region (SOGEPA's stake in Carsid's capital was EUR 20 million, reduced to EUR 9 million after the decision to initiate the procedure (7) provided for in Article 6(5) of Commission Decision No 2496/96/ECSC (8)). This procedure was completed on 15 October 2003 (9) with a negative decision from the Commission on the grounds that an investment in a new company cannot be considered pari passu if only the public shareholder bears a new risk, with the private shareholder only transferring a business activity or an existing project to the new company.

(15)

Following this series of acquisitions, Duferco had three main subsidiaries in Belgium: Duferco Clabecq, Duferco La Louvière (‘DLL’) and Carsid.

(16)

In 2006, Duferco formed a strategic partnership with the Russian steel group Novolipetsk (‘NLMK’). This partnership aimed to build on the presence of NLMK in the upstream steel supply chain (supply of raw materials and production of semi-finished products). It took the form of NLMK having a 50 % stake in one of the Duferco Group's holding companies, Steel Invest and Finance (SIF), which was registered in Luxembourg. SIF encompassed many Duferco Group assets, including Duferco Clabecq, DLL and Carsid. The American subsidiary Duferco Farrell Corporation was also affiliated at the end of 2006. The agreement between Duferco and NLMK was ratified on 18 December 2006. NLMK acquired a 50 % stake in SIF, along with an unconditional purchase option on at least one share in SIF.

(17)

The partnership between Duferco and NLMK ended in 2011. The two companies divided SIF's assets. Duferco took back the ‘long products’ sector of DLL, some diversification activities, and Carsid's assets. All of these Belgian assets were combined in a new subsidiary, Duferco Long Products (‘DLP’), a company incorporated under Luxembourg law and created in March 2011.

(18)

In 2012, the Carsid site shut permanently. In 2013, Duferco decided to cease production of ‘long products’ at the La Louvière site.

(19)

The Duferco Group was consolidated through the parent company, Duferco Participations Holding (‘DPH’). The own funds of DPH went from EUR […] (*1) million in 2003 to EUR […] million in 2010.

(20)

DPH was initially established […] but was then relocated […]. DPH itself is held by other companies: […].

(21)

Between 2003 and 2011, FSIH made six contributions to Duferco subsidiaries for a total of EUR 517 million.

(22)

SOGEPA is a limited company with public capital which is wholly owned by the Walloon Region. It implements decisions taken by the Walloon Government relating to interventions in commercial companies and managing those interventions. Furthermore, it is the preferred policy instrument of the Walloon Region in the steel sector. It was formed following the merger in 1999 of the Société Wallonne pour la Sidérurgie (SWS) and the Société pour la gestion de participations de la Région wallonne dans des sociétés commerciales (SOWAGEP).

(23)

According to SOGEPA's website, SOGEPA carries out two types of specific tasks for interventions made at the request of the Walloon Government: (i) analysis and the issuing of an opinion prior to a possible decision by the Walloon Government; (ii) implementing and monitoring decisions taken by the Walloon Government. In the implementation of decisions by the Walloon Government, SOGEPA's departments negotiate and draw up shareholder and/or loan agreements, respecting any conditions that may be specified in the government decision on intervention.

(24)

In principle, the monitoring of interventions involves the presence of SOGEPA and its collaborators on the boards of directors of the companies that benefit from the interventions in the capacity of both administrator and observer. It also involves participation in general meetings and, generally speaking, regular monitoring of the Walloon Region's interests.

(25)

The Walloon Government appoints SOGEPA's directors and also delegates two Government Commissioners to the company.

(26)

On 27 March 2003, SOGEPA formed FSIH with EUR 180 million in capital. According to the deed of incorporation, the object of FSIH is ‘to invest, in any form, in foreign undertakings linked to the steel industry and/or conduct any financial operations with them.’ To this end, FSIH ‘could encourage any collaboration between these undertakings and Walloon undertakings.’

(27)

Members of the board of directors of FSIH are appointed by the general meeting and thus by SOGEPA. A general meeting was held on 27 March 2003, just after FSIH was established, and resulted in the appointment of the administrators and chairman of the board of directors. The appointed chairman was also the chairman of the board of SOGEPA. This double-hatting lasted until 31 March 2013.

(28)

Duferco US Investment Corp. (‘Duferco US’) is a holding company which holds 100 % of a single company, Duferco Farrell Corp. (‘Farrell’). Before the measure under investigation, the Duferco Group owned 99,8 % of Duferco US through one of its industrial subsidiaries, Duferco Industrial Investment (‘DII’).

(29)

Farrell is a steel company established in Pennsylvania (United States). It produces hot and cold rolled steel products.

(30)

On 9 April 2003, FSIH purchased half of the shares, minus one share, in Duferco US for USD 97,22 million (EUR 90 million). DII remained the owner of the other half of the shares, minus one share. The two remaining shares were placed in a special purpose vehicle to hold them in trust for the duration of the partnership between FSIH and DII. On the same day, FSIH acquired a purchase option for EUR 5 million, allowing it to acquire the two shares in this vehicle and therefore take control of Duferco US. However, this option was never exercised.

(31)

Before acquiring, through FSIH, half of the shares minus one share in Duferco US in 2003, SOGEPA had the value of Farrell assessed. According to the British consultancy company Hatch Beddows, the conservative average value of Farrell was USD 330 million. A further study was also made by American consultancy World Steel Dynamics, which estimated the value of Farrell at USD 410 million. The combination of these two analyses gives an average value for Farrell of USD 375 million. Using the monthly average USD-EUR exchange rate for April 2003, this average value is around EUR 345,5 million.

(32)

FSIH received a dividend of USD 2,5 million in 2005 for the financial year ending 30 September 2004. In its comments, Belgium pointed out that the dividend was paid on 23 June 2005. Using the annual average USD-EUR exchange rate for 2005, this dividend is equivalent to some EUR 2,01 million.

(33)

As is the case for many manufacturing industries, Farrell's results were directly influenced by the purchase price of its inputs (slabs). On 15 September 2005, Farrell signed a slab supply contract with Duferco in order to make its results less volatile. This contract stipulated that Duferco had to supply slabs to Farrell at a price contractually set so that Farrell's margin would stay within a given range ([…])

(34)

On 1 January 2006, an agreement was signed between the Duferco Group and the […] group. Under the contract, […] supplied slabs to Duferco at a price calculated as (i) the selling price of Farrell's finished products; (ii) less Farrell's manufacturing costs; (iii) plus a margin. This contract made it possible for the Duferco Group to cover the risk that had been transferred to it from Farrell following the agreement of 15 September 2005.

(35)

In 2006, FSIH decided to review the business case for its stake in Duferco US. To this end, on 14 June 2006 DII and FSIH agreed that DII would have a purchase option on all the shares held by FSIH for EUR 95 million. This option seems to have been granted without remuneration.

(36)

This purchase option was granted in the context of the negotiation of the strategic partnership between Duferco and NLMK. The inclusion of Duferco US in SIF's assets was under discussion.

(37)

On 12 December 2006, after this purchase option was exercised by […], a company in the Duferco Group which was the successor to DII for all the rights and obligations relating to this measure, FSIH withdrew from Duferco US. As a result, the Duferco Group once again became the full owner of Duferco US on 12 December 2006 for USD 125,85 million (EUR 95 million). On 18 December 2006, 50 % of the shares in SIF were transferred to NLMK.

(38)

On 9 April 2003, FSIH acquired from [the company which controlled DPH at the time], with registered office in […], currently in […], around 25 % of DPH's capital as well as a preferred share with a preferential right to dividends. FSIH had previously asked the consultancy Ernst & Young to assess the value of DPH. On the basis of this assessment, the value of 25 % of DPH was between USD […] and […] million, i.e. an average value of USD […] million. The final price of the transaction between FSIH and [the company that controlled DPH at the time] would be USD 86,42 million, or some EUR 80 million using the monthly average USD/EUR exchange rate in April 2003.

(39)

The parties agreed that because FSIH held a preferred share, it would have a preferential right to DPH dividends as follows: (i) payment of EUR 2 million for each of the financial years 2003 and 2004; (ii) payment of EUR 4 million for every subsequent financial year until 2012.

(40)

In January 2005, FSIH therefore received a dividend of EUR 4 million for the 2003 and 2004 financial years. No dividend seems to have been paid for the 2005 financial year. The 2006 financial year had not ended before FSIH resold its stake.

(41)

In 2006 FSIH wanted to withdraw from DPH. In preparation for its exit, FSIH again commissioned Ernst & Young to value its 25 % stake in DPH. Ernst & Young's 2006 report (‘the Ernst & Young 2006 Report’) valued the 25 % stake at between USD […] and […] million, or some EUR […] and […] million using the annual average USD/EUR exchange rate for 2006.

(42)

The parties did not adopt the valuation from the Ernst & Young 2006 Report. On 14 June 2006, FSIH sold its stake for USD […] million, or EUR […] million to [the company that controlled DPH at the time].

(43)

In addition to steel, Duferco had developed diversification activities in Belgium, which in September 2006 were incorporated into a new subsidiary affiliated to SIF: Carsid Développement. In December 2006, several days before SIF came under the partial control of NLMK, the decision was made to remove Carsid Développement from the partnership with the Russian company.

(44)

Duferco thus joined forces with FSIH to take over and develop the assets of Carsid Développement. A new company, […], was created on 29 November 2006 to host these assets. When it was created, its shareholder was […] and its initial share capital was EUR […] million (subscribed but not paid up apart from one share with a nominal value of one euro). However, at its first meeting on 7 December 2006, the board of directors of […] accepted the transfer of the existing share of […] to FSIH. It then accepted the subscription by a subsidiary of the Duferco Group, Duferco Développement, for one share with a nominal value of one euro. The board of directors also decided to change the registered name of the company to Duferco Salvage Investments Holding (‘DSIH’). On 7 December, FSIH and Duferco thus had a new joint subsidiary, DSIH, which was able to host the assets of Carsid Développement.

(45)

On 11 and 14 December 2006, FSIH and Duferco Développement each increased the capital of DSIH by EUR [65-72] million, corresponding to a stake of 50 %, in cash. These funds were used to buy all the shares in Carsid Développement on 14 December 2006 for EUR […] million. Carsid Développement was renamed Duferco Diversification.

(46)

Duferco Diversification then held 100 % of the following three companies: Duferco Environnement, Duferco Immobilière and Marcinelle Energie. Duferco Environnement housed a project to clean up land located in […] and belonging to Duferco Développement (the project known as […]). Duferco Immobilière included a number of property assets previously held by Carsid. Finally, the object of Marcinelle Energie in 2006 was to build a CCGT power station (combined cycle gas turbine) in Charleroi.

Graph 1

Companies involved in Measure 3

Duferco Développement

50 %

50 %

Duferco Immobilière

Marcinelle Energie

Duferco Environnement

Duferco Diversification

FSIH

DSIH

(47)

In 2008, 80 % of the Marcinelle project was sold to the Italian power company ENEL for EUR [30-37] million.

(48)

On 26 March 2010, Duferco Développement subscribed for an additional share in DSIH, which ensured it had majority control of DSIH with 67 001 shares (compared with 67 000 held by FSIH).

(49)

In June 2011, following the end of the partnership between Duferco and NLMK, Duferco and FSIH decided to combine all their Walloon, steel and diversification assets into one new separate subsidiary: DLP. The steel assets resulting from the dismantling of SIF were affiliated to DLP through a subsidiary called Duferco La Louvière Produits Longs (‘DLLPL’, see Measure 6). A series of transactions in summer 2011 also integrated the diversification activities into DLP:

(50)

In 2009, FSIH granted a loan totalling EUR 100 million to […], the parent company of the Duferco Group, for a duration of six years. At the time of the loan, this company was registered in […]. Now it has become the company […], incorporated under the laws of […].

(51)

First, pursuant to the terms of an agreement signed on 4 September 2009, FSIH granted a loan of EUR [30-40] million. The rate applicable to the loan was fixed at the 12-month Euribor rate (1,302 %) plus 75 basis points, i.e. 2,052 % when the agreement was concluded. According to the information presented by Belgium on 4 June 2013, the rate actually adopted was 2,04 %. The principal was redeemable in full on 31 December 2015 or earlier in the event of a change in control of SIF. The loan interest was eligible for capitalisation annually and payable in arrears on the outstanding balance from the sixth year or earlier in the event of a change of control of SIF.

(52)

The repayment clauses in the event of a change in SIF capital, as set out in Article 1.1 of the loan agreement, were introduced to take account of the mutual purchase option clauses negotiated by Duferco and NLMK regarding SIF in an agreement of 1 February 2008. The 2008 agreement provided that this option could be exercised at any time from December 2010 at a price of USD […] million. Under Article 6.1 of the loan agreement between FSIH and [the parent company of the Duferco Group at the time], in the event of a change in control of SIF, the borrower was required to allocate, as a priority, the amounts paid on this occasion towards full repayment of the loan granted by FSIH.

(53)

The loan had two types of guarantees:

(54)

In December 2009, an amendment was added to the loan agreement of 4 September 2009. A new loan of EUR [60-70] million was granted by FSIH to [the parent company of the Duferco Group at the time]. The rate applicable to the loan was also set at the 12-month Euribor rate (1,302 %) plus 75 basis points, i.e. 2,052 % when the agreement was concluded. According to the information presented by Belgium on 4 June 2013, the rate actually adopted was 1,99 % when the loan was released. The same conditions as those of the 4 September agreement apply concerning the repayment of the loan (capital and interest). With regard to guarantees, […] additional shares in DPH were pledged.

(55)

According to Duferco's annual report for 2009, the total value of the guarantee on the two loans was approximately USD […] million. According to Belgium, the value of the guarantee corresponded to this amount if it was based only on the net assets when the loan was granted.

(56)

The loan was repaid early in July 2011 due to the full takeover of SIF by NLMK.

(57)

At the same time, pursuant to the terms of two agreements signed on 29 September and 22 December 2009, FSIH granted a loan of EUR 75 million to SIF. This loan was released in two tranches, EUR 20 million in September 2009 and EUR 55 million in December 2009. The rate applicable to the loan was set at the 12-month Euribor rate (1,302 %) plus 75 basis points, i.e. 2,052 % when the agreement was concluded. According to the information presented by Belgium on 4 June 2013, the rate actually adopted was 1,99 % when the two tranches of the loan were released.

(58)

The capital was repayable […]. The interest was […]. Furthermore, according to Article 5 of the loan agreement, […].

(59)

On 29 June 2011, the parties agreed that given the negative situation of the steel sector worldwide and the change in control of SIF, SIF had to repay the loan early in three payments spread out over 2011, 2012 and 2013. On 30 June 2011, SIF repaid part of the capital, i.e. EUR [15-25] million as well as the interest due on that date, i.e. EUR [1-2] million. The final amendment, of 30 September 2013, provides for early repayment of the outstanding balance, i.e. EUR [55-65] million in principal and EUR [1-2] million in interest. In its response to the opening decision, Belgium informed the Commission that SIF's claim to FSIH had been repaid in full.

(60)

In summer 2011, FSIH injected a total of EUR 100 million into DLP, the investment vehicle under Luxembourg law created to house the parallel investments of FSIH and Duferco in Belgium. This investment was made in two stages.

(61)

First, on 29 June 2011, FSIH granted a convertible loan of EUR 30 million to DLP at a 12-month Euribor rate plus 200 basis points (2,137 % + 2 % = 4,137 %). This loan was concluded for 10 years. The loan did not come with any guarantee given that the aim of the parties was to convert it into capital. The conversion of the loan into capital took place on 23 September 2011.

(62)

On 7 July 2011, through its subsidiary DII, Duferco took part in a capital increase of EUR 101 million in DLP, for 50,26 % of the share capital of DLP.

(63)

On 14 July 2011, FSIH granted EUR 70 million to DLP through a capital increase.

(64)

After its investment, FSIH held 49,74 % of the share capital of DLP.

(65)

The EUR 201 million capital increase (EUR 100 million injected by FSIH and EUR 101 million injected by Duferco) was subscribed with the aim of:

(66)

In December 2012, the decision was made to split the steel activities (which were mostly inactive) and non-steel activities of Duferco in the Walloon Region again. Duferco Belgium kept the steel activities, while the diversification activities were taken by a new DLP subsidiary, Duferco Wallonie. Following this restructuring, the organisation chart of the company was as follows:

Graph 2

Companies involved in Measure 6

[…].

(67)

The Commission points out that its doubts did not concern the conditions under which FSIH had acquired its stake in Duferco US. On the other hand, the Commission doubted whether the conditions under which FSIH had allowed its stake in Duferco US to be terminated were in line with the market economy investor principle. It put forward several arguments in support of its concerns.

(68)

First, the sale price in euros was the same in 2006 as the acquisition price in 2003, which suggests mediocre profitability over the holding period. The Commission was surprised by this, given the positive results obtained by Duferco US between 2003 and 2006, and the consequent increase in its own funds. In this respect, the Commission challenged Belgium's opinion that the return on the stake should be evaluated in US dollars (the functional currency of Duferco US) and not in euros (the functional currency of FSIH).

(69)

Second, the Commission was surprised that no valuation had been made before the option exercise price was set.

(70)

The Commission also questioned the motives behind FSIH's withdrawal, particularly its perception of an economic downturn for the coming years. In this respect, the Commission doubted that a private investor would have agreed to such a long option (almost 10 years), even though they claimed to anticipate an imminent economic downturn. It also noted that the agreements of 2005 and 2006, concluded with Duferco and […] respectively, reduced the exposure of Duferco US to market fluctuations and therefore helped make it more attractive.

(71)

Finally, the Commission was surprised that the purchase option had been granted at no cost when it had a certain economic value for the Duferco Group, which could exercise it to suit its interests.

(72)

The Commission doubted that the conditions under which FSIH had ended its stake in DPH were in line with the market economy investor principle.

(73)

Aside from the Commission's challenging the use of the US dollar to estimate the profitability of the investment in DPH, its main concerns related to the validity of the method for setting the sale price. It was particularly surprised by the difference between the price adopted by the parties and the independent evaluation carried out by Ernst & Young in 2006.

(74)

During the preliminary investigation phase, Belgium had justified this difference by taking into account five haircuts. The Commission questioned whether these haircuts had actually featured in the negotiation between [the company that controlled DPH at the time] and FSIH and, beyond this formal challenge, had reservations about their economic validity.

(75)

Finally, the Commission had concerns about the motives behind FSIH's withdrawal.

(76)

In conclusion, the Commission put forward the idea that, far from acting as a private investor, FSIH had merely sought to recover its initial stake of EUR 80 million, increased only by a dividend of EUR 4 million due for 2005 but which had not yet been distributed.

(77)

The Commission doubted whether the conditions under which FSIH had invested in DSIH were in line with the market economy investor principle.

(78)

In response to Belgium's claim concerning the pari passu nature of the operation, the Commission noted that, at the time of investment, DSIH was already under the control of Duferco and that the assets brought to this new subsidiary were also already held by Duferco (at 50 %, through SIF). Under these conditions, only the new (public) shareholder took on new risks and the operation could not be considered pari passu.

(79)

Concerning the valuation of Duferco Diversification and its subsidiaries:

(80)

Lastly, the Commission had reservations about the consequences for FSIH of the resale of Duferco Diversification to DLP in 2011, given that FSIH held only 49,7 % of DLP, but held all the shares minus one share in DSIH.

(81)

The Commission had reservations about Belgium's claim that the rate adopted for the two loans did not contain any element of aid since it was much higher than the average variable interest rate applied by banks to [the parent company of the Duferco Group at the time] and the SIF group, i.e. 1,31 % and 1,92 % respectively.

(82)

With regard to the loan to [the parent company of the Duferco Group at the time], it was not possible to find private loans with similar characteristics using the information Belgium submitted at the time.

(83)

With regard to the loan to SIF, the Commission doubted that the company could be treated as AAA-rated, and it was surprised at the lack of collateral for this loan.

(84)

Consequently, the Commission was not in a position to definitively rule out the possibility that the loans granted to [the parent company of the Duferco Group at the time] and SIF had included elements of State aid.

(85)

The Commission expressed doubts about the whether the EUR 100 million capital increase in DLP complied with the market economy investor principle.

(86)

The Commission noted that the investment could not be regarded a priori as pari passu since it was made for Duferco in a company already held by the group. Under these conditions, the public and private shareholders did not bear the same risks, because the latter could not be deemed independent in relation to the investment in question.

(87)

Concerning the price of the assets to be acquired by DLP, the Commission had particular concerns about the method used to value DLLPL, Duferco Trebos and the stock acquired from Duferco La Louvière Sales.

(88)

The Commission noted that no framework or guidelines from the Commission concerning the compatibility criteria relating to State aid seemed to apply to the measures under investigation. In terms of the possibility of regional aid in particular, the relevant legislation provided that such aid in the steel sector was not compatible with the internal market.

(89)

With regard to the market-economy investor principle, Belgium considers that, in light of the specific context in which the sale of Duferco US shares was concluded, this transaction cannot be considered not to have been made at market price.

(90)

Belgium notes first that the functional currency of Duferco US, an undertaking established in the United States, is the US dollar pursuant to International Accounting Standard (IAS) 21.9. Consequently, it believes that FSIH's investment and the calculation of its return should be analysed in US dollars and not in euros after conversion. It adds that FSIH took the risk of not hedging against exchange-rate fluctuations because cover was very difficult to obtain, given the nature of the operation and particularly because of the uncertainty surrounding the date of exit from the investment. Belgium confirms that the return on the investment, based on amounts in US dollars, was 6,81 % per year.

(91)

Belgium then disputes that the value of own funds should be taken into account in the valuation of Duferco US. It argues, first, that although an increase in own funds naturally has an impact on the book value of a company, this does not necessarily lead to an increase in its market value. Second, it highlights how very volatile the company's profits were over the years under investigation by the Commission. In this regard, Belgium disagrees with the Commission's claim that the 2003-2006 period had been ‘on the whole a profitable period’ for Duferco US and Farrell.

(92)

Belgium maintains that it is better to use a valuation based on a multiple of EBITDA (earnings before interest, taxes, depreciation, and amortisation) rather than on own funds. It disagrees with the multiple the Commission used on the basis of the independent S&P Capital IQ database. It argues that the average multiple of companies similar to Duferco US (between […] and […]) is below that of the Commission ([…]). Consequently, the valuation of Duferco US is reduced by 49,9 %. According to Belgium, the sale price agreed by the parties (USD […] million) would be significantly higher than the valuation of the stake sold (USD […] million if the adopted multiple were […] and USD […] million if the multiple were […]).

(93)

To refine its valuation process, Belgium then refers to the influence of the supply contracts concluded by Farrell, Duferco and […] in September 2005 and January 2006. It argues that the stabilisation of Farrell's EBITDA resulting from these contracts, leading to a reduced risk/return on investment profile, supported a sale price for FSIH's stake in Duferco US of between USD […] million and USD […] million. It stresses that the sale price that was finally agreed (USD […] million) was significantly higher than these valuations.

(94)

Lastly, Belgium puts forward several reasons as to why FSIH withdrew: (i) the conclusion of the slab supply agreement with […] which, while stabilising the results, still risked limiting the potential gains of Duferco US; (ii) the ongoing negotiations for the sale of 50 % of SIF to NLMK, which envisaged the integration of Duferco US into the partnership with the Russian group; (iii) the feeling FSIH had about a future downturn in the steel sector; (iv) the abundance of liquid assets in the Duferco Group, which created a window of opportunity for FSIH to sell its stake in Duferco US.

(95)

With regard to the price and maturity of the purchase option, Belgium claims that the option price was included in the negotiated exercise price and that the uncertainties surrounding the date of execution of the planned operation between Duferco and NLMK justified for duration of almost 10 years.

(96)

In response to the Commission's first concern, Belgium stresses that it is necessary to value the investment in DPH in US dollars and not in euros given that US dollar is the functional currency of DPH. Using the US dollar as the reference currency would result in an annual return of 8,81 % on the stake in DPH.

(97)

With regard to the Commission's other concerns, Belgium justifies the difference between the sale price and the range in the Ernst & Young 2006 Report by taking into account five haircuts.

(98)

DPH is the parent company of two large divisions: the Trading division (trade) and the Industry division (production). On 29 September 2003, i.e. several months after FSIH acquired a stake in its capital, DPH sold 50 % of the shares in its Trading division (Duferco International Trading Holding — […]) to […]. This sale was made on the basis of the book value of the Trading division in that the sale price was calculated by taking into account half of the own funds of Duferco International Trading Holding, i.e. USD 40,83 million or EUR 36,01 million using the annual average USD-EUR exchange rate for 2003.

(99)

Belgium believes this price is a market price. It therefore had to be taken into account when, around three years later, the whole group was valued. Belgium believes that the Ernst & Young 2006 Report should have taken into account the book value of the Trading division rather than applying its valuation method, i.e. the price/earnings ratio, to the whole Group. A contrario, Belgium does not object to the fact that the Industry division is still valued using the price/earnings ratio.

(100)

Valuing the Trading division at its book value involves applying a haircut of between USD 20 and 31 million on the valuation from the Ernst & Young 2006 Report.

(101)

Finally, Belgium underlines that, even if one wanted to apply the price/earnings ratio method to the whole Group, the ratio adopted in the Ernst & Young 2006 Report does not reflect the risk profile of DPH. Choosing this kind of method does not remove the need to assess the Trading division and the Industry division separately, given that a different ratio should apply to each of them. The Ernst & Young 2006 Report erred in applying a ratio derived from comparable undertakings operating only in one industry to the whole Group (including the Trading division). Belgium believes that trade is a riskier activity than production. Therefore a lower price/earnings ratio should be applied to this type of activity, reflecting investors' reduced appetite for this risk profile (Belgium has provided data taken from S&P Capital IQ to support this conclusion). By not doing so, Ernst & Young overvalued DPH in 2006.

(102)

Belgium maintains that the period selected by Ernst & Young to estimate the profits and the price/earnings ratio to use in the valuation of Duferco US does not adequately reflect the cyclical nature of the steel sector. This period covers only the four years from 2003 to 2006, so it includes three good years (2004, 2005 and 2006) and only one bad year (2003). The parties therefore added 2001 and 2002 to their analysis and established different ratings than those of Ernst & Young for each year.

(103)

The values and weightings adopted by the parties result in a valuation of USD […] million for DPH, (compared with USD […] million in the Ernst & Young reference case). This significant difference in values, to which the effects of other haircuts and readjustments also have to be added, prove that Ernst & Young overvalued DPH in 2006.

(104)

Belgium believes that a haircut reflecting the illiquidity of the shares in DPH held by FSIH should be applied. The factors commonly taken into account to assess the liquidity of shares in a company are as follows:

(105)

In light of these factors, and given the ‘lack of interest in this shareholding from other players’, the parties agreed to apply an illiquidity haircut of 40 %, compared with the 30 % advocated by Ernst & Young.

(106)

Belgium does not feel that it has to further justify the additional haircuts as they fall within the generally applied range and the adopted haircut of 40 % is the result of negotiation between the parties.

(107)

FSIH had negotiated the right to a preferential dividend of EUR 4 million per year for the financial years from 2005 to 2012. Ernst & Young took this right into account in its valuation exercise for 2006, considering that FSIH could hope to receive USD 4 million (11) per year during the seven years until 2012 (i.e. a discounted sum of USD 26,3 million).

(108)

Ernst & Young did not add this discounted sum to its valuation as such. The consultant applied a haircut of 6,2 % to it, taking the view that the risk of non-payment of this preferential dividend was comparable to the default risk of a bond with a BBB rating. The Ernst & Young 2006 Report thus added USD 22,2 million to the value obtained for 25 % of the shares in DPH.

(109)

According to Belgium, the parties took the view at the time that this haircut of 6,2 % was not enough to reflect the risk of non-payment of the preferential dividend.

(110)

Belgium puts forward two main arguments to support this. The first relates to the volatility of DPH's results: in the event of insufficient profits or cash flow, DPH would have had to forego payment of the dividend. In this regard, Belgium stresses that payment of the preferential dividend was at the company's discretion. The second argument relates to the organisation of the Duferco Group: given that the company owing the dividends is a holding company (DPH), the profits at its level comprise solely the consolidation of the profits generated by its subsidiaries. Now, for various reasons (particularly tax rules, financial commitments and contractual obligations), it was not always easy, or indeed possible, for profits from the subsidiaries to flow up to the holding company.

(111)

All of these factors were sources of real uncertainty concerning the future payment of the preferential dividend from DPH. Belgium feels that the haircut applied therefore had to be higher than the haircut applied for future yields on bonds with a BBB rating. The parties agreed to a figure of 14 %, i.e. a haircut of USD 4 million in relation to the Ernst & Young valuation.

(112)

Haircut 5, in reality, is the choice not to apply a premium proposed by the Ernst & Young 2006 Report. In this report, the consultant observed that FSIH, through the shareholder agreement, has special rights allowing it to influence certain strategic decisions of the Group. The Ernst & Young 2006 Report therefore takes the view that a control premium of between 0 % and 10 % of the value of the shareholding should be added.

(113)

Belgium believes the parties chose not to adopt such a control premium because the special rights it relates to did not actually have any market value in the context of a sale to a third party or to [the company that controlled DPH at the time]. According to Belgium, in the event of a sale to a third-party shareholder, [the company that controlled DPH at the time] could oppose the transfer of rights laid down in the DPH shareholder agreement. Consequently, [the company that controlled DPH at the time] did not have to pay a certain amount to recover full powers.

(114)

Belgium points out that the parties agreed that the special rights attached to the FSIH shares under the shareholders' agreement would be terminated in the event of one of them selling their shares to a third party. If it is indisputable that the special rights provided for in the shareholders' agreement conferred on FSIH the power, at the very least, to influence the management of DPH and, in this regard, these rights had a certain value when FSIH acquired its stake in DPH, the fact remains that FSIH could not freely transfer the special rights it held to a third party. Therefore, these rights had no market value.

(115)

Belgium puts forward two more factors which did not lead to application of a haircut but which it wanted to mention so that the Commission could see the different negotiating points that the parties discussed in order to establish the final price.

(116)

The first factor relates to the 2005 agreement between DPH and Farrell which provided for the transfer of certain profits to Farrell. Because of this agreement, the Ernst & Young 2006 Report applied a haircut of USD 5,5 million to the reported earnings of each year taken into account. The parties did not increase this haircut.

(117)

The second factor relates to a supply agreement with the company […]. This agreement provided for the purchase of a substantial quantity of semi-finished products (slabs) at a price corresponding to the cost of production. For DPH, this agreement represented both a risk of loss and a potential gain, depending on global steel prices. This agreement set out obligations for a capital injection and significant penalties in the event of withdrawal and was therefore a potential cost risk for DPH. The Ernst & Young 2006 Report did not quantify this risk and, according to the information Belgium sent to the Commission, neither did the parties.

(118)

Lastly, Belgium points out the motives behind FSIH's withdrawal: the cyclical nature of the sector, entry and influence of new private investors in the Duferco Group ([…] and NLMK), and availability of liquidity in the Duferco Group. In this regard, it stresses that the joint venture with NLMK showed the will of Duferco to gradually exit industrial steel production with a view to turning to new activities such as site rehabilitation and developing energy projects. FSIH pursued this trend by withdrawing from DPH and investing in the Group's new diversification projects (see Measure 3).

(119)

Belgium justifies the price of EUR […] million by the total of various factors.

(120)

Carsid Développement, which became Duferco Diversification, inherited EUR [30-37] million in trade receivables towards another company in the Group (Duferco La Louvière Sales).

(121)

Since […] 2002, the Duferco Group had been making efforts to develop an extensive site redevelopment project. Although the land is in the hands of Duferco Développement, it is the Carsid Développement subsidiary (now Duferco Diversification) that is responsible for developing this project.

(122)

Belgium points out that on the day FSIH decided to approve acquisition of shares in DSIH, i.e. 5 December 2006, it had already been agreed that Duferco Développement, which had no human resources, would use Duferco Diversification to go ahead with work to prepare and clean up the […] site. With regard to the remuneration of Duferco Diversification for the services provided within this project, it had already been agreed that Duferco Diversification would receive half the profits expected on the sale price of the land belonging to Duferco Développement, which would keep the other half.

(123)

Belgium admits that, given the complexity and ambition of the property project, the formal conclusion of a contract between Duferco Diversification and Duferco Développement took a long time. Certain aspects still had to be determined, on the basis of the conclusions of the study syndicate established by ministerial decree of 18 January 2007, i.e. one month after FSIH acquired its stake in DSIH.

(124)

The remuneration of Duferco Diversification within the […] project was finally formalised in an agreement concluded on 4 February 2009 between Duferco Développement, the owner of the land, and Duferco Diversification. Belgium argues that, in accordance with what had been agreed at the outset, the agreement concluded on 2009 provides that the profit from the sale of the cleaned-up land would be shared equally between Duferco Développement and Duferco Diversification. This agreement from 2009 therefore merely formalises an existing agreement between Duferco Développement and Carsid Développement from 2006. Sot it was entirely justified for FSIH to take this project into account in the process of valuing Carsid Développement when it invested in DSIH.

(125)

Furthermore, Belgium stresses that as early as 2006, FSIH based its investment on a valuation of the project. It considers that the expected profit of the […] project was conservatively evaluated in 2006 at EUR […] million.

(126)

This amount corresponded to the difference between the estimated cost of cleaning up the […] site (EUR […] million according to an expert report from 2006) and the value of the land held by Duferco Développement (already estimated, according to Belgium, at EUR[…] million — a value confirmed by the consultancy GALTIER in January 2010).

(127)

Belgium therefore takes the view that FSIH behaved conservatively and prudently by asking that the parties retain a profit of EUR […] million which, in accordance with their agreements, would be shared equally between Duferco Développement and Duferco Diversification (i.e. EUR […] million to be taken into account by DSIH).

(128)

At the end of the internal reorganisations at Duferco, the […] project was transferred to Duferco Wallonie on 19 December 2012 for EUR […] million. This amount corresponded to the expected profit on the project, which was reassessed in 2012 according to new reports.

(129)

Belgium does not give further indications on whether or not the […] project has been implemented yet or on its commercial success or failure.

(130)

Some of the property owned by Duferco La Louvière was transferred to Duferco Immobilière, a subsidiary of Duferco Diversification. Its value, confirmed by a consultant, is assessed at between EUR [14-21] and [15-22] million.

(131)

In 2006, Marcinelle Energie, a subsidiary of Duferco Diversification, planned to construct a CCGT power station in Charleroi. FSIH relied on the evaluation made by the consultancy Bain (‘Bain’) in July 2006 to value this project.

(132)

According to Bain, the overall value of the project could be estimated at EUR […] million: (i) EUR […] million in net present value for the power station as such and (ii) EUR […] million if 100 % of the CO2 allowances were allocated to the project.

(133)

An intermediate valuation proposed by Bain was based on (i) a net present value of EUR […] million (based on the assumption that Duferco would not join forces with a power company) and (ii) an allocation of only 50 % of the CO2 allowances for EUR […] million. The valuation of the CO2 allowances was then reduced again, taking into consideration the risk of a reduction in the trading price of the allowances. The value that FSIH finally accepted for the CO2 allowances was EUR […] million.

(134)

On the basis of this report, FSIH could legitimately consider that investing in the construction of a CCGT would bring a net present value of EUR […] million (EUR […] million). However, the Bain report suggested that an association with an experienced power company would make it possible to extract greater value, which would bring the net present value of the project to EUR […] million.

(135)

According to Belgium, the median overall value of the Marcinelle Energie project as conservatively estimated on the basis of the Bain report was therefore EUR […] million.

(136)

In light of this valuation, FSIH started negotiations on the valuation of the Marcinelle Energie project, based on exploratory contacts made by the Duferco Group with various players in the energy sector to form a partnership. As shown by the exchanges with the key players in the energy sector — including EDF, Electrabel, SPE, NUON and ENEL — the Marcinelle Energy project received a lot of interest and hinted at major opportunities.

(137)

Nevertheless, given that amounts of only EUR […] million were mentioned during these exploratory contacts for the acquisition of 80 % of the shares in the Marcinelle Energy project, FSIH requested and obtained a reduction of EUR […] million in the median value of the project according to Bain, taking into consideration the amounts referred to in these informal contacts.

(138)

Ultimately, a value of EUR […] million was adopted as the value of 80 % of the project. FSIH and Duferco agreed to value the remaining 20 % of the project at 20 % of the median value of the project as per the Bain report, i.e. EUR […] million (20 % of EUR […] million). A total value of EUR […] million was therefore adopted for the Marcinelle Energy project.

(139)

Belgium thus maintains that, in December 2006, when FSIH decided to acquire shares in DSIH alongside Duferco Développement, the value of the CCGT project of Marcinelle Energie was the subject of negotiations that resulted in the parties adopting a markedly lower valuation than the median value of EUR […] million from the Bain report. Ultimately, the partners adopted an amount of EUR […] million. By doing so, FSIH ensured that a very prudent calculation method was used.

(140)

It is clear from the above that ultimately, the parties adopted a prudent valuation, much lower than the maximum profitability that the project was likely to generate. It would therefore be fundamentally incorrect to criticise FSIH for having allegedly invested ‘an amount equivalent to the expected return on the project’ (paragraph 161 of the Commission's opening decision).

(141)

However, Belgium admits that, according to the Bain report, taking the value of CO2 emission allowances into account required the power station to be operational very quickly, and in any case during 2009. If the project were to overrun, it would no longer be able to take advantage of the possibility of receiving the available CO2 emission allowances.

(142)

In June 2008, two years after these discussions, Duferco Diversification sold 80 % of the Marcinelle Energie project to the Italian power company ENEL for EUR [30-37] million. This sale came with a sale option enabling Duferco Diversification to re-sell its remaining shares ([…]) to ENEL one year after the power station began operating at a price of EUR […] million, increased by repayment of possible losses resulting from an agreement to supply at cost price (off-take right).

(143)

However, the energy market suffered another strong downturn starting in 2010 and the profitability of CCGT power stations plummeted. This did not prevent the power station from entering into service on 31 March 2012.

(144)

Consequently, at the start of 2013, Duferco Diversification (which became Duferco Belgium) exercised its exit right for EUR […] million, increased by EUR […] million in repayment of losses recorded on the agreement to supply at cost price, EUR […] million by way of a rebate on the price of gas and EUR […] million in ‘revaluations and miscellaneous interest’ (without further details).

(145)

Ultimately, the Marcinelle Energie project was therefore sold for a total of EUR [70-78] million. Nevertheless, as Belgium points out, investments (EUR […] million) and losses (approximately EUR […] million) made between 2006 and 2013 should be deducted from this amount, resulting in a non-discounted net profit of approximately EUR [40-46] million.

(146)

These results are less than the forecasts outlined in 2006. Belgium justifies this by the complete turnaround of the gas market from 2010, noting that the ENEL group also lost money on this project.

(147)

On the basis of the above four factors, in 2006 FSIH and the Duferco Group concluded that the value of Duferco Diversification was between EUR […] million and EUR […] million. Ultimately, the value of EUR […] million was to be adopted, half of which was to be provided by FSIH.

(148)

Belgium argues that this investment in DSIH should be considered an investment that is pari passu with Duferco Développement. Given that FSIH and Duferco Développement provided the same amount, they would have the same interest and would be subject to the same risks. In this regard, Belgium points out that the DSIH did not belong to the Duferco Group before the investment by FSIH, given that it was established on 29 November 2006 by a third party (under the name […]) before being transferred to FSIH and Duferco Développement as an ad hoc vehicle for their joint investments in December 2006. For these reasons, FSIH's intervention complies with the market-economy investor principle.

(149)

As a reminder, on 8 July 2011 DSIH sold its stake in Duferco Diversification to DLP for the same amount, i.e. EUR […] million. However, DLP did not pay this amount; Duferco Diversification kept a claim of EUR […] million in its balance in relation to this.

(150)

With regard to the loan to [the parent company of the Duferco Group at the time], Belgium disputes any elements of aid. It points out that the average variable interest rate on bank loans held by the various subsidiaries of [the parent company of the Duferco Group at the time] in 2009 was [1,1-1,55] %, which is lower than the loans on the rates under investigation.

(151)

Belgium maintains that this point of comparison is valid because the loans taken into account to calculate it are comparable to those granted by FSIH. In this it disagrees with the Commission, which does not consider these loans relevant for a comparative analysis.

(152)

According to Belgium, the Commission is wrong to take into consideration such criteria as the maturity of the loan, the amount lent or the nature of the loan to decide whether private loans are comparable to those of FSIH. This approach goes against the Communication from the Commission on the revision of the method for setting the reference and discount rates (the ‘2008 Reference Rate Communication’) (12), which explicitly states that ‘the margin is largely independent of the maturity of the loan’.

(153)

Belgium illustrates the lack of correlation between the maturity, amount and nature, on the one hand, and the margin, on the other, by analysing the loans granted to Duferco by commercial banks. The loans from banking institutions on the highest amounts have interest rates that vary between [0,9-1,35] % and [1,7-2,2] % and are often lower than the interest rates on loans for small amounts (up to [2,25-2,75] % for a loan of EUR […] million or even [2,3-2,75] % for a loan of EUR […] million). Rates on loans for working capital fluctuate between [0,75-1,15] % and [2,75-3,25] % while rates on loans for capital expenditure (CapEx) fluctuate between [1,5-2] % and [2,75-3,25] %.

(154)

According to Belgium, in order to rule out any element of aid in the granting of a public loan, nothing imposes the obligation to establish the existence of an identical loan from a private bank. Neither the EU State aid rules nor case-law require public loans to be identical to private loans in terms of maturity, amount and nature to rule out any element of aid.

(155)

Concerning the rating of [the parent company of the Duferco Group at the time], Belgium infers from the rates charged for the different loans granted to Duferco by banking institutions that it could claim an AAA rating. It based this on the rates that the 2008 Reference Rate Communication associated with companies rated A to AAA for a high level of collateral in 2009: 2,37 % in September 2009 and 2,05 % in December 2009.

(156)

Belgium argues that the collateral associated with the FSIH loan is extremely high in relation to the 2008 Reference Rate Communication. It follows from DPH's very solid finances that the pledge on 15 % of its shares is excellent collateral. Furthermore, the quasi-certain nature of the claim of EUR […] million to be allocated as a priority to repay the loan in the event of a change of control of SIF makes this collateral very strong.

(157)

With regard to the loans granted to the SIF group, the comparison shows that, once again, there is only a limited correlation at most between the criteria put forward by the Commission (maturity, amount and nature) and the rate of remuneration on these loans.

(158)

The comparison also establishes that the remuneration on the FSIH loan is close to the remuneration on the different loans granted to SIF by private financial institutions. With an interest rate of 1,99 %, its remuneration is slightly higher than the average rate on loans to the SIF Group, which is [1,65-2,15] %.

(159)

Taking into account only the long-term loans contracted by SIF provides a range that is even lower than the average of [1,65-2,15] %, i.e. a range between [1-1,5] % and [1,65-2,15] %.

(160)

Belgium believes that the lack of collateral for the loan awarded to SIF makes it necessary to add a margin of [20-60] basis points to this range. This is justified by the rating category ‘strong’, under the terms of the 2008 Reference Rate Communication, which SIF could claim. This increase leads to a range of [1,5-2] % to […] % which the loan granted by FSIH fits into easily.

(161)

In the absence of a public rating, Belgium accounts for SIF's ‘strong’ rating by the fact that the interest rates applied by private financial institutions correspond to (or are less than) the interest rates laid down by the 2008 Reference Rate Communication for financially sound companies (AAA to A rating).

(162)

According to Belgium, FSIH's investment in DLP was made pari passu with that of DII given that: (i) the amounts invested by the private partner and the public partner are practically identical; (ii) DLP did not belong to the Duferco Group before this investment but was established specifically to serve as a vehicle to house the respective investments of FSIH and DII, and (iii) the goal of the investments was not to buy back assets that already belonged to the Duferco Group (but to SIF).

(163)

Belgium points out that the fact that FSIH's investment was, in the first instance, granted in the form of a convertible loan (and was quickly converted) is irrelevant. On the date the convertible loan was granted, FSIH and DII had at the same time concluded a shareholders' agreement through which they undertook to increase the capital of DLP by EUR 70 million and EUR 101 million respectively. The loan of EUR 30 million should therefore be investigated as an element of the capital increase of EUR 201 million decided pari passu by FSIH and DII.

(164)

With regard to the purchase of DLLPL shares, the Commission is reminded that the Walloon Region had been a shareholder in DLL since 1999 through SOGEPA and in this respect was very familiar with its activities and financial situation. This explains the fact that, in the context of DLP's investment in DLLPL, the parties did not deem it necessary to produce a business plan in due form. FSIH knew perfectly well what it was acquiring. Belgium further believes that FSIH, when making its investment, could be reassured by the prudent nature of the business plan since the forecasts were perfectly in line with the actual figures that were already available. The same applies to Duferco Trebos.

(165)

With regard to the purchase of the stock from Duferco La Louvière Sales, Belgium argues that although it is correct that the investments by FSIH and DII in DLP allowed the latter to buy back a substantial stock of products with a value of EUR [50-70] million (sales contract of 30 June 2011), it should be noted that the seller of these products was by no means ‘another Duferco company’. Indeed, at the time, Duferco La Louvière Sales was, and still is, a company controlled by SIF. After the negotiations, the stocks were bought at their book value although their actual value was much higher.

(166)

As regards the investment of EUR [30-40] million in DLLPL, separate business plans had been drawn up for different projects or investment scenarios. The parties ultimately opted for an investment of EUR [30-40] million in DLLPL, which became Duferco Belgium. This investment was the subject of only one business plan, which Belgium submitted to the Commission. The other business plans to which Belgium refers became irrelevant, given that the projects or scenarios set out were not ultimately adopted by the parties.

(167)

By way of introduction, it is worth pointing out that […], the holding company of the Duferco Group, was the successor to […] and to the companies […] and […], the former parent companies of DPH.

(168)

[…] stresses first that it agrees completely with Belgium's observations.

(169)

[…] then recalls the motives leading FSIH to invest in Duferco US. The acquisition of a stake in Farrell was related to the desire of FSIH to control the site so that Duferco did not abandon the Belgian site in favour of the American site. The Walloon Region wanted to protect itself against the risk of potentially diverging interests within Duferco being used against it. Additionally, FSIH's investment in Farrell was also intended to improve cooperation between the Belgian and American production sites in order to optimise exchanges of information and possible synergies between these sites.

(170)

[…] considers it important to emphasise that the initiative behind the withdrawal of FSIH lies with Duferco. In February 2006, Duferco had signed a letter of intent with NLMK with a view to selling a stake of 50 % in SIF to NLMK. However, one of the preconditions was that all the shares in Duferco US should be held by SIF, which therefore implied a buyback of the shares held by FSIH. Duferco then suggested to FSIH a purchase option to be able to meet the demands of NLMK while still maintaining the necessary flexibility, given the uncertainty at that time concerning the implementation of the transaction. At the same, the buyback offer by Duferco was interesting in the context of the end of the steel cycle expected for the end of 2006, which made the investment less attractive.

(171)

[…] confirms Belgium's observations to the effect that the opportunity to have a prior competitive tender procedure for potential buyers was not relevant in light of the purpose of the sale: to sell of a minority shareholding in a private company with a non-integrated production site as target. Although the situation did not necessitate a prior competitive tender procedure, […] maintains that the sale was still made at a market price.

(172)

[…] points out that the functional currency of Duferco US is the US dollar. Consequently, except for the internal needs of FSIH, it is not relevant that the purchase price was paid in euros.

(173)

To supplement the valuation exercises carried out by the Walloon Region, […] hired KPMG to value FSIH's 49,9 % stake in Duferco US on 14 June 2006, the day the purchase option was granted, using the fair value method.

(174)

KPMG considered that FSIH's stake in Duferco US should be valued as a stake which does not confer control and which is non-negotiable, while applying a control premium, given that the shareholders' agreement of 9 April 2003 gave FSIH a significant influence over Duferco US beyond the influence held by a minority shareholder.

(175)

KPMG indicated that industrial cycles in the steel industry generally last between three and seven years. KPMG noted that even though the steel industry was in good shape in 2006, the indicators clearly showed major risks, given the weakness of the domestic automotive industry and the possibility of a correction in the market for non-residential construction.

(176)

Essentially assuming cycles of three and five years as well as a valuation on an EBITDA multiple from comparable companies, KPMG estimated the fair value of the 49,9 % stake in Duferco US at USD […] million on 14 June 2006.

(177)

[…] then expresses its opinion on the discussion on whether or not to take into account the value of own funds to value Duferco US. It confirms that a valuation based on own funds, as envisaged by the Commission in its opening decision, is not the most appropriate. However, it notes that even if own funds were taken into account, the conclusion would be that the transaction was carried out at a market price. To prove this, it draws the Commission's attention to what it considers to be a comparable transaction: the sale to NLMK of a 50 % stake in SIF (accompanied by an unconditional sale option on a least one SIF share). The price applied in this transaction from 2006 corresponds to a multiple of own funds of the company which is practically identical to the multiple of own funds represented by the sale by FSIH of its stake in Duferco US (in both cases a multiple of almost 2). As there was no indication that the transaction concerning SIF had not been agreed under market conditions, this comparison confirms that taking into account own funds would also result in the transaction being considered to have been made at a market price.

(178)

Concerning the duration of the purchase option, […] considers that it falls within the context of the concurrent negotiation of a possible sale by Duferco of 50 % of the capital of SIF to NLMK. NLMK had to draw up a first draft of the share transfer agreement and a shareholders' agreement within 10 days of signing the letter of intent. This letter of intent itself was only valid until 1 October 2006 at the latest. The intention was therefore for Duferco to either use or not use the purchase option in the weeks or months after it was granted by FSIH. There was therefore no real reason for the parties having agreed that the option would run until 31 December 2015.

(179)

KPMG assessed the fair value of the purchase option granted to Duferco in June 2006. KPMG agrees with Duferco's position that the expected term of the option of four months should be taken into account because the option was neither transferable nor assignable, which markedly decreased its value. KPMG takes the view that the fair value of the purchase option was USD […] million.

(180)

On the basis of the contractual duration, the fair value is USD […] million. If such an amount were taken into account, added to the fair value of FSIH's stake in Duferco US, the sale price would be USD […] million, which is significantly lower than the price paid by [the parent company of the Duferco Group at the time].

(181)

According to […], FSIH's investment in DPH had two purposes: on the one hand, participation in the capital, representation on the board of directors, access to strategic information and documents and, on the other hand, attractive and secured financial returns (preferential dividend).

(182)

The main reason for FSIH exiting DPH is the sale of the Belgian production sites to NLMK. Duferco would lose its status as a strategic partner for the Walloon Region in Belgium and the production sites of Clabecq and La Louvière would no longer be directly affected by decisions by Duferco's Trading division.

(183)

[…] asserts that DPH is a company whose functional currency the US dollar and therefore the return on any investment in such a company must be assessed in US dollars. FSIH's investment in DPH thus generated a return of 8,81 % per year.

(184)

[…] commissioned KPMG to analyse the valuation made by Ernst & Young in March 2006. With regard to the valuation methods, KPMG notes that Ernst & Young does not specify the type of value for the valuation (market value, investment value or fair value). Ernst & Young mainly applied the market approach by comparable companies. KPMG does not give any preference to the method based on comparable companies in relation to the method based on comparable transactions. KPMG did not use the net asset value as a basis because that method ignores future gains and operational risks.

(185)

KPMG concludes that, in light of its analysis of the cyclical nature of the steel industry, periods of between four and six years should be taken into account with equal weighting for each year to value DPH. KPMG agrees with the parties' analysis, which takes into account the years 2001 and 2002 in the valuation of DPH in order to adequately reflect the cyclical nature of the steel industry and apply an equal weighting to each year.

(186)

KPMG confirms the relevance and level of the adjustments made by Ernst & Young for the valuation of DPH in light of non-recurring events. KPMG also made adjustments to the 2001 and 2002, which Ernst & Young did not take into account.

(187)

KPMG examined the list of comparable companies selected by Ernst & Young. KMPG was not able to find certain multiples included in the Ernst & Young 2006 Report in the database used by Ernst & Young (Bloomberg). Furthermore, KPMG deemed it appropriate not to include certain companies in the list of comparable companies provided by Ernst & Young. KPMG established its own list of comparable companies, which shows that the median price/earnings ratios for 2003 to 2006 are lower than those used by Ernst & Young.

(188)

KPMG was not able to observe a sufficient number of listed companies that are active only in the steel trade. It is not possible to make a separate valuation of the Trading division. Nevertheless, KPMG confirms that it would have been appropriate to make a separate valuation of the trade and production activities. With a higher risk and more volatile results, the trade activities offer a higher return to investors, which results in a lower valuation of the activities. A lower multiple should therefore be used for the trading activity. […] believes it is relevant to refer to the sale of 50 % of […] in September 2003, for which the agreed price was close to the value of own funds.

(189)

With regard to off-balance-sheet items, KPMG confirms that the amount of a company's own funds should be adjusted to take into account off-balance-sheet liabilities if these are non-recurring. Unlike Ernst & Young, KPMG believes that only these items should be deducted from the estimated value of the company's own funds, regardless of the valuation method.

(190)

For two of the off-balance-sheet items, KPMG deemed it appropriate to reconsider the amount of the adjustment applied by Ernst & Young. This concerned the 'French and Belgian guarantees' that Ernst & Young had evaluated at USD […] million in 2006, whereas according to KPMG they amounted to USD […] million. These guarantees were adjusted on the basis of the risk they contained. This resulted in a total amount of USD […] million compared with USD […] million used by Ernst & Young.

(191)

KPMG confirms the relevance and the level of the illiquidity haircut (30 %) applied by Ernst & Young. A rebate of 12 % in the framework of the valuation based on comparable transactions would have been equally relevant.

(192)

KPMG feels it was appropriate to apply a control premium of 5 % in the event of a valuation on the basis of comparable companies, given that this kind of method generates a valuation corresponding to a minority shareholding while FSIH's stake allows significant influence over DPH. KPMG believes that a haircut of 20 % should be applied in the event of a valuation on the basis of comparable transactions, given that this method results in a valuation with a controlling interest.

(193)

With regard to the return on preference shares, KPMG notes that Ernst & Young was mistaken in its valuation of preferential dividends for 2006-2012 in basing its calculation on annual dividends of USD 4 million, whereas the amount was EUR 4 million. KPMG also takes the view that the discount Ernst & Young applied given the risk of default on payment (6,2 %) falls within the identified acceptable range, between 1,8 % and 16,4 %. The discount of 14 % applied by the parties is also considered acceptable since it is in the value range.

(194)

These adjustments result in a valuation of the shareholding in DPH that fluctuates between USD […] million and USD […] million, or i.e. a median value of EUR […] million, which is in line with the sale price in June 2006 of EUR […] million. […] concludes that there can be no question of State aid to [the company that controlled DPH at the time].

(195)

DSIH confirms the arguments put forward by Belgium to defend the pari passu nature of FSIH's investment. DSIH claims that the conditions of the transaction were identical for Duferco Développement and FSIH. The two shareholders agreed jointly to take a new risk in investing in DSIH in order to carry out the diversification projects of Duferco Diversification. The two shareholders benefit from proportional distribution based on their participation in the capital, potential dividends and the financial opportunities of projects developed by their joint subsidiaries. According to DSIH, the private intervention of Duferco Développement is economically very substantial and the amount of the investment is identical to that of FSIH.

(196)

DSIH asserts that it did not belong to Duferco Développement before the simultaneous investments. DSIH is a company that was established by a third party shortly before these investments (29 November 2006) and then acquired by the two partners to facilitate their investments. DSIH further maintains that the projects acquired by DSIH did not belong to the Duferco Group either before the investment by FSIH, although they were housed by SIF in draft form.

(197)

With regard to the […] project, DSIH points out that the Walloon Region, through SOGEPA, was perfectly familiar with the situation of the […] site, given that it had intervened with the Duferco Group in the takeover of […], which was then bankrupt. The Walloon Region was therefore a shareholder in […] and held several seats on the board of directors. The Walloon Region was also involved in the […] project from the outset and had access to all the information available in the context of the project launch, in particular exactly the same information that was available to private investors.

(198)

DSIH relies on an ex post evaluation of the project which proved the validity of the valuation of the project as initially made and indicated very attractive profits.

(199)

On the Marcinelle project, DSIH states that the lower than expected profits were not due to the partners investing too much but to factors beyond the control of the parties. First, in January 2008 the Commission announced a proposal to withdraw the free allocation of CO2 emission allowances, which had a great impact on the valuation of the project before the close of negotiations with ENEL. Second, the electricity market experienced a complete and unforeseeable downturn in 2010, with a fall in the prices of electricity, gas and CO2.

(200)

Duferco Diversification came out of this crisis well by selling Marcinelle Energie just before this happened. The company sold 80 % of Marcinelle Energie at a reasonable price and was able to secure the remaining […] by a sale option negotiated with ENEL.

(201)

[…] notes that in order to apply the market economy investor criterion, various factors should be examined, primarily the applicable interest rate and the collateral required to guarantee the loan.

(202)

With regard to the collateral, […] points out that Duferco had a very sound financial situation, as shown by the terms of the loans granted by banking institutions to the Duferco Group, with rates comparable to those offered to AAA-rated companies according to the Commission's reference rates. Additionally, [the parent company of the Duferco Group at the time] was required to allocate the funds from the future sale of DII's stake in SIF to NLMK as a priority to the early repayment of the loan from FSIH. The existence of a future, but almost certain, claim of USD […] million on NLMK represented a liquid and easily payable guarantee. This loan is similar to an advance on a future claim. FSIH also benefited from a pledge on 15 % of the shares in DPH, a very financially sound company (no loss reported up to 2009), for an approximate value of USD […] million. Lastly, the 12-month Euribor reference rate guaranteed a stable rate for a period of 12 months, which is an attractive feature in a time of crisis and falling rates. Therefore, given these guarantees, the risk for FSIH was minimal.

(203)

The rates applied to the Duferco Group by private banks varied between [0,65-1,15] % and [2,75-3,25] % with an average of [1,1-1,55] %, i.e. rates which are in line with those of FSIH. […] does not think that loans that do not have exactly the same maturity, the same amount, the same nature or that have different base rates can be excluded from the comparison. Only the collateral and the interest rates applied should be taken into account.

(204)

The amounts of the loans presented by Belgium are generally lower than those of the FSIH loan since Belgium only presented loans taken out in euros, which represent only a small part of the total amount. […] was interested in multi-currency loans covering significantly higher amounts. These data show that Duferco had access to credit lines of a very high amount and that certain lending was in the form of long-term loans (10 years). The rates on these loans varied between [0,5-1] % and [2,2-2,65] % with a weighted average of [1,6-2,05] %.

(205)

[…] thus concludes that, in light of the strong collateral enjoyed by FSIH, the interest rates are perfectly in line with market rates.

(206)

According to DLP, since the outset the parties had anticipated that the loan of EUR 30 million would be likely to be converted at short notice, which was the case in September 2011. FSIH's initial intention was to make a capital investment of EUR 100 million. DLP believes the convertible loan should be analysed together with the capital injection and be treated as such.

(207)

DLP argues that the interventions by DII and FSIH are concurrent, since DII participated in a capital injection in DLP on 7 July 2011, i.e. several days before the capital injection by FSIH and after the convertible loan was granted. It also argues that the conditions of the operation are identical, with the two partners having invested almost identical amounts. The investments were therefore made under the same conditions since FSIH and DII bear the same risk and enjoy proportional distribution of dividends and increase in the value of their subsidiary.

(208)

DLP states that the company did not belong to the Duferco Group at the time of FSIH's investment but was a company that was newly established specifically to house the joint investments of the two partners. The fact that these investments were used in part to buy the assets of SIF is irrelevant given that the sale price was based on expert reports. Consequently, DLP concludes that the investment made by FSIH in DLP is pari passu with DII.

(209)

DLP also points out that SOGEPA had been a shareholder in DLL since 1999 and was therefore very familiar with the activities and financial situation of the company as regards the long products sector. Through SOGEPA, the Walloon Region had sound knowledge of the market conditions, the operating sites and their environment. DLL also indicates that there is only one business plan for each of the companies Duferco Trebos and DLLPL. FSIH was therefore able to make a fully informed decision and had a level of information identical to what a private investor would have required in an identical situation.

(210)

According to DLP, the financial results show that the EBITDA achieved after three months was in line with the EBITDA forecast in the business plan for a period of six months. When it made its investment, FSIH could therefore be reassured as to the prudent nature of the business plan.

(211)

The Commission points out that the comments by SIF refer only to the measure it benefits from, namely Measure 5.

(212)

SIF considers that its financial situation was sound when the loan was granted and that the guarantees at the time of the loan agreement were more than enough to repay the entire loan from FSIH. Moreover, SIF also believes that FSIH had no doubts about the profitability of the loan transaction given SIF's investment and development policies. SIF's business plan for the Belgian sites, particularly for 2007-2010, contained plans for major investments promising significant growth potential of the group in Belgium.

(213)

SIF also considers that its ownership by two major players in the steel sector was taken into account in the risk analysis linked to the loan transaction by FSIH. NLMK granted a guarantee to FSIH in the event of a change of control. In granting a loan of EUR 75 million to SIF at a rate of 2,05 % (margin of 75 basis points) while the National Bank of Belgium rate for new loans was 1,63 %, FSIH did not take an unnecessary risk in relation to a private operator.

(214)

SIF points out that applying the market economy investor criterion does not require any comparison with loans that are identical in all aspects to the loan granted by FSIH. SIF also points out that the Commission states that the type of operator involved, the nature of the transaction and the relevant markets should be taken into account. In light of this clarification, two private loans granted to Duferco Clabecq, a subsidiary of SIF, should be taken into account for the comparison with the loan granted by FSIH. The analysis of these loans shows that, during the period when FSIH granted the loan, SIF was able to obtain credits from private financial institutions and that the conditions of the loan granted by FSIH do not differ substantially from the market. The interest rates applied by credit institutions for these two loans [1-1,5] % (Sumitomo loan) and [1,2-1,7] % (Rabobank loan) are lower than that granted by FSIH (2,05 %). It follows that the subjective risk inherent in the loan granted by FSIH is reflected in the applied interest rate, the 12-month Euribor rate, which makes the loan more secure for the lender in that it guarantees a stable rate for 12 months.

(215)

Moreover, SIF takes the view that the lack of collateral can in no way constitute an advantage. If no collateral was required, this is because its financial strength, its affiliation with the NLMK and Duferco Groups and […] were enough to reassure FSIH that SIF was able to repay the loan.

(216)

SIF also notes that a margin of 75 points is consistent with the banking practice at the time. The Sumitomo loan shows that, for a normal level of collateral, the margin was [10-50] points, i.e. less than that indicated in the 2008 Reference Rates Communication, for a good rating (100 points) or even a strong rating (75 points). Likewise, for the Rabobank loan the margin is [45-85] points for a normal level of collateral. It follows that if one considers that SIF had a good rating, a margin of 75 points (below 100 points) for a 10-year loan is in line with the practices of banks, which considerably tightened their margins at the time.

(217)

SIF concludes that the loan granted to it by FSIH corresponds to market conditions and did not give it any advantage.

(218)

However, SIF points out that if the loan were to constitute aid, such aid could be compatible with the common market under the Commission's Communication on the Temporary Community framework for State aid measures to support access to finance in the current financial and economic crisis (‘the 2009 Communication’) (13). To assess the compatibility of subsidised interest rates with the rules on State aid, the 2009 Communication establishes a reduced rate specific to the crisis period. SIF notes that all the conditions laid down by the 2009 Communication are met: the contract was concluded on 29 September 2009 and SIF was not in difficulty on 1 July 2008. Therefore, if the loan did constitute aid, the compatibility of the interest rate on the loan granted to SIF by FSIH should be assessed in light of the reduced rate.

(219)

[…]'s observations require no specific comments by Belgium. They confirm that the decisions were made in accordance with an industrial strategy that corresponded to market conditions.

(220)

SIF's observations require no specific comments by Belgium. Belgium also declares that it is at the Commission's disposal to carry out an audit should the loan granted to SIF be considered State aid.

(221)

Article 107(1) TFEU lays down that ‘any aid granted by a Member State or through State resources in any form whatsoever which distorts or threatens to distort competition by favouring certain undertakings or the production of certain goods shall, in so far as it affects trade between Member States, be incompatible with the internal market’. This provision sets out the conditions under which a measure by a Member State could be considered State aid.

(222)

First, the measures in question must have a State origin in the sense that they must involve State resources and be imputable to the State. Second, the Commission must ascertain whether the State resources confer an advantage. Third, the measures must be selective (e.g. apply to a company, an economic sector or a geographical area) rather than a generally applicable measure. Fourth, the measures must be likely to distort competition and affect trade between Member States.

(223)

For the six measures in question the Commission will first examine the criteria of State origin, selectivity and the impact on competition and intra-Community trade. It will then analyse the economic advantage criterion for each of the measures.

(224)

The Commission notes that Belgium did not dispute the State origin of the financing, its selectivity or its impact on competition and trade between Member States. Consequently, for these criteria the Commission will limit itself to confirming the analysis made in the opening decision. Belgium contests only the condition concerning the existence of an economic advantage.

(225)

The Commission points out that ‘for advantages to be capable of being categorised as aid within the meaning of Article 87(1) EC, they must, first, be granted directly or indirectly through State resources […] and, second, be imputable to the State […]’ (14).

(226)

The body at the origin of the measures under investigation is the financial holding company FSIH. FSIH, a company whose object is intervention in steel companies based abroad, is a subsidiary wholly owned by SOGEPA, which itself is wholly owned by the Walloon Region. The funds available to FSIH are therefore under the control of the Walloon Region and therefore constitute State resources.

(227)

The description of SOGEPA and FSIH demonstrates that the role of the public authorities is key in the interventions of SOGEPA and FSIH. SOGEPA acts at the request of the Walloon Government. Article 3(1) of its statutes states that ‘[The object of the company is to carry out all tasks entrusted to it by the Walloon Government, …. In this context, it implements decisions to intervene in commercial companies taken by the Walloon Government and manages the holdings, obligations, advances or interests that the Walloon Region or itself may have in such companies.’ On the subject of steel, any decision by SOGEPA that goes beyond simple portfolio management also requires agreement from the Ministers for the Economy and for the Budget of the Walloon Region. In this respect, the articles that appeared in the newspaper Le Soir on 21 November 2011 report the reactions of the Walloon Ministers for the economy at the time. These individuals clearly bear the responsibility of the Walloon Government for FSIH's interventions.

(228)

In conclusion, the Commission considers that the measures under investigation meet the criteria of State origin and are imputable to the Walloon Region.

(229)

The measures under investigation concern only the companies and subsidiaries of the Duferco Group. They are therefore selective.

(230)

The Duferco Group is active on the steel market. The interventions by FSIH allowed the Duferco Group to maintain a stronger competitive position than it would have had in the absence of these interventions. Moreover, within the EU, there are a number of other active operators on the steel market and steel products are traded between a number of Member States. The advantage conferred by FSIH's interventions on a market open to competition affected this trade. Indeed, where State financial aid strengthens the position of an undertaking as compared with other undertakings competing in intra-Community trade, the latter must be regarded as affected by that aid' (15).

(231)

Consequently, in line with the case-law of the Court of Justice of the European Union (16), the measures are likely to distort competition and affect trade between Member States.

(232)

The economic and financial transactions carried out by public bodies do not confer an advantage on the recipient, and therefore do not constitute State aid, if they are carried under normal market conditions (17). In other words, it is necessary to consider whether, under similar circumstances, a private investor operating under normal market conditions would have made the same intervention. If this is not the case, the recipient undertaking has received an economic advantage which it would not have obtained under normal market conditions, placing it in a more favourable position than that of its competitors (18).

(233)

The Commission will examine the six measures granted to the Duferco Group using this criterion.

(234)

The Commission takes the view that although the analysis of the reasons for FSIH's withdrawal is not enough to prove the imprudent nature of the investment, it still sheds precious light on the measure in question.

(235)

Belgium justifies FSIH's withdrawal from Duferco US by: (i) the feeling FSIH had about a future downturn in the steel sector; (ii) the conclusion of the slab supply agreement with […] which, while stabilising the results, still risked limiting the potential gains of Duferco US; (iii) the abundance of liquid assets in the Duferco Group, which created a window of opportunity for FSIH to sell its stake in Duferco US; (iv) the ongoing negotiations for the sale of 50 % of SIF to NLMK, which envisaged the integration of Duferco US into the partnership with the latter.

(236)

With regard to the first reason, the Commission notes that Belgium did not consider this to be very important in its response of 8 April 2015. The Commission had asked Belgium to produce any study, prior to the measure, on the prospects for the steel market for the next five years. Belgium produced an overview of the state of the American economy taken from The National Economic Review on the first eight months of 2006 and a study of the steel sector dated 10 April 2006 by Crédit Suisse. The first two studies are general overviews of American growth in the first two quarters; they show that GDP growth had slowed due to the decline in the housing sector but do not predict recession in the short term. The second study, which directly concerns the steel market, remains cautious as to the evolution of this market in the United States:

(237)

On reading these studies, the Commission notes that Belgium was correct in limiting the importance that the perception of the steel market had in FSIH's decision to withdraw. The economic prospects at the time of FSIH's withdrawal seem insufficient to present this withdrawal as a real opportunity.

(238)

With regard to the second reason, the Commission notes that the implementation of the supply contracts led to significant improvements in Farrell's results from the 2006 financial year. Net sales almost doubled; the operating result moved from a loss of USD […] million to a positive result of USD […] million; the net result moved from a loss of USD […] million to a positive result of USD […] million. Given that the 2006 financial year closed on 30 September, FSIH could not have been unaware of the improvement in June 2006.

(239)

Since FSIH was not a hedge fund, the Commission is not convinced that an increase in and stabilisation of revenue at a predictable and comfortable level were valid reasons for withdrawing.

(240)

The Commission has no objections to the third reason, but considers that it should be qualified in relation to the real reason behind the withdrawal of FSIH: the cooperation between Duferco and NLMK.

(241)

It appears that the initiative to withdraw came from Duferco. As […] indicated in its comments to the Commission (19), the sale by FSIH of its stake in Duferco US was an essential condition for NLMK to acquire a stake in the capital of SIF.

‘NLMK had in particular set as a precondition to the entire operation that all the shares in Duferco US be controlled by SIF, which therefore implied a buyback of the shares held by FSIH.’

In light of the above, FSIH was undoubtedly in a position of strength to negotiate the sale of its stake in Duferco. In this context, the fact that it did not, however, seek to recover more than its initial investment of EUR 95 million indicates that FSIH did not behave as a prudent investor.

(242)

Faced with such an assumption, Belgium responded in its comments that FSIH's investment should be analysed in the functional currency of Duferco US (US dollar) and not that of FSIH (euro). In US dollars, the sale price obtained by FSIH (USD […] million) gave it a decent return of 6,81 % per year. The Commission does not agree with this position for two reasons. First, it believes that the only currency that makes sense for FSIH, as the manager of a fund denominated in euros, is the euro. Second, and more fundamentally, it points out that this debate has no impact on the only important question, which is whether the price of USD […] million (or equivalently, EUR […] million) corresponds to the market price that a private investor could have obtained under similar circumstances. Having duly put this into context, the Commission would like to reply to this question.

(243)

When the sale of FSIH's stake in Duferco US was being negotiated, the Commission first finds that FSIH did not set up an open, transparent and unconditional tender procedure because, according to FSIH, it ‘was difficult to envisage turning to a third-party buyer for the sale of a minority shareholding in a (unlisted) private company’ (20). The examination of the reasons for FSIH withdrawing clearly discredits this argument: if FSIH did not seek bidders other than Duferco, it was to please the latter, which had a pressing need for FSIH's shares to conclude the agreement with NLMK.

(244)

However, the Commission does not systematically require the setting up of a tender procedure (21) for a sale to be deemed to be in line with the market price. On the other hand, in that case it recommends seeking independent studies to assess the market price.

(245)

FSIH did not commission any report of this nature. In this respect, Belgium considers that preparing a valuation report was as pointless as a call for tenders ‘given the active and effective involvement of FSIH in Farrell since it acquired its stake in the capital in 2003 and its excellent knowledge of the steel sector.’ (22)

(246)

The Commission therefore notes that the two methods it generally uses to determine a market price, i.e. a call for tenders or an independent report, were not used by the Walloon Region.

(247)

When a Member State argues that an economic transaction meets the market economy investor criterion, it must provide evidence demonstrating that the decision to proceed with the transaction was made, at the time, on the basis of economic evaluations comparable to those which, under similar circumstances, a rational private operator would have had carried out in order to determine the resulting economic advantage.

(248)

The above findings show what follows from the EDF judgment (23):

(249)

However, the Commission notes that the valuation methods produced by Belgium and by […] are studies carried out after the measure in question was implemented. They were carried out solely to respond to the Commission's concerns in the opening decision, i.e. after the events in question.

(250)

In this regard, the Commission notes in the observations by […] (24) that, while the latter had asked KPMG to reconstruct the valuation of Duferco US ex post KPMG had to discard the income approach since there was no business plan prepared by Duferco US at the time.

(251)

The Commission on three occasions requested any previous document indicating the existence of work preceding FSIH's decision (25). Belgium replied that, since 2003, when FSIH acquired its stake in the capital of Duferco US, three managers representing FSIH were on the boards of directors of Duferco US and Farrell and were involved in the management of these companies. Numerous reports were exchanged between the shareholders of Duferco US (26). Belgium also submitted two examples of reports dated August 2004 and March 2005. These reports are the records of several pages of two visits to Farrell made by representatives of FSIH. The Commission considers that they do not make it possible in any way to establish the valuation of FSIH's stake in Duferco US. Belgium believes that these steel-industry experts were able to evaluate the company's situation themselves (27). However, the Commission notes that Belgium has not produced any memo, correspondence or other documents recording the exchanges the steel-industry experts might have had with the Duferco Group to determine the price of FSIH's stake in Duferco US.

(252)

In light of these facts, the reasons for FSIH's withdrawal and the case-law in the EDF judgment, the Commission concludes that FSIH did not act as a private investor in a market economy when it sold its shares in Duferco US.

(253)

The Commission has demonstrated that, by not acting as a private investor, FSIH granted an economic advantage to DII that it would not have obtained under normal market conditions.

(254)

In the absence of evidence from Belgium that an ex-ante evaluation of Duferco US resulted in the sale price that was actually paid, the Commission undertook its own evaluation of the valuation of 49,99 % of Duferco US (stake held by FSIH).

(255)

First, the Commission refuses to completely rule out the own funds valuation method that was used in the opening decision. Nevertheless, and to take account of Belgium's comments concerning the intrinsic limits of this method, the Commission will use it only for verification.

Table 1

Own funds and net results of Duferco US (USD thousand)

Table 2

Own funds and net results of Farrell (USD thousand)

(256)

These data simply show that a valuation method based on the value of own funds unquestionably results in a higher value in 2006 than in 2003.

(257)

However, the Commission, estimated the value of FSIH's stake in Duferco US using the method recommended by Belgium, i.e. the EBITDA multiple. This method requires an estimate of (i) the EBITDA for 2006 (the year of the sale); and (ii) a multiple to be applied to it on the basis of comparable companies.

(258)

Belgium advocated using an EBITDA of between EUR 25 and 42 million for 2006. According to the consolidated accounts of Duferco US at 30 September 2006 (28), the EBITDA was in fact USD 52,25 million (operating result: 45,97 + depreciation and amortisation: 6,28). The bulk of this good result should have already been known on 14 June 2006 since there were only two and a half months before the end of the financial year. The Commission therefore considers that the range advocated by Belgium is unfairly low. It also notes that the ex post study carried out by KPMG uses a ‘forecast’ EBITDA for the 2006 financial year of USD 42 million. The combination of these different amounts leads the Commission to maintain that an EBITDA of USD 47,12 million (29) could realistically have been anticipated in 2006.

(259)

The Commission's estimate is supported by the following information. On 12 June 2006, in a memo (30) to its board of directors meeting on 14 June, FSIH predicted a stabilised result of USD 20 to 25 million for 2006 and subsequent years. This confirms, first, that FSIH had a certain view of the company's financial situation. But this mainly provides a good indication of the EBITDA level that could be expected ex ante. The accounting and financial reconstruction carried out by KPMG in 2014 concludes that a net result of EUR 20,3 million goes hand in hand with an EBITDA of EUR 42,3 million. If we accept this EBITDA/net result ratio of 2,09 and apply it to a provisional net result of EUR 22,5 million (arithmetic mean of the range predicted by FSIH in its memo to the board of directors), one obtains a forecast EBITDA for 2006 of EUR 47 million, which is very close to the EUR 47,12 million estimated by the Commission in the preceding recital.

(260)

With regard to the multiple to be applied, the Commission accepts the multiple taken from the methodology used by Hatch Beddows when it valued Duferco US in 2003, i.e. a multiple of […]. This study is the only ex-ante analysis that was available at the time of sale.

(261)

It follows that the value of FSIH's stake in Duferco US is USD […] million (31). The amount of aid is the difference between valuation and the sale price of USD […] million set in 2006, i.e. USD 15,24 million.

(262)

The value of the purchase option, i.e. USD […] million as estimated ex-post by KPMG, should be added to this amount. The total amount of aid received by DII is therefore USD 15,34 million, or in principle EUR 11,58 million (32).

(263)

The Commission observes that the amount of aid estimated in this way confirms that the sale price of FSIH's stake in Duferco US does not correspond to what a private investor operating under normal market conditions could have expected.

(264)

In its comments, Belgium used the following factors to explain why FSIH withdrew from DPH: the cyclical nature of the sector, the entry and influence of new private investors in the Duferco Group ([…] and NLMK) and the availability of liquidity in the Duferco Group.

(265)

Although this last factor is not in doubt, the Commission questions, on the other hand, whether the first two factors played a decisive role. Neither of these reasons figures in the minutes of the meeting of the board of directors of FSIH on 14 June 2006 (33) which dealt with the sale of its shares in DPH. However the minutes report another, more pressing, motive: an urgent need for liquid assets on the part of FSIH.

‘At the end of 2005, FSIH was faced with a significant need for cash to finance urgent projects.

At the same time, the Duferco Group had the opportunity to sell its industrial unit located in Russia.

FSIH took advantage of this opportunity to ask [the company that controlled DPH at the time] to buy its shares in DPH.’

(266)

It should be noted that FSIH had received EUR 180 million from SOGEPA when it was established. At the end of 2005, this amount was almost entirely tied up: EUR 95 million had been invested in Duferco US and EUR 80 million in DPH. Yet in 2006, FSIH wanted to support other projects which were no doubt regarded as greater priorities from the point of view of the Walloon public authorities. These projects included the refurbishing of Carsid's blast furnace and coke plant, ‘a strategic project for the future of the steel industry in the Charleroi basin [and] that will most certainly fall under the scope of SOGEPA’ (34).

(267)

In light of these reasons, it seems that FSIH decided to transfer its stake in DPH for mainly political and social considerations: FSIH had to find the cash to be able to continue to support the steel industry in Wallonia. By acting this way, FSIH's intention was not to maximise the value of its stake. Indeed, the Walloon company was put in a weak position to negotiate its exit from DPH with Duferco.

(268)

In this context, [the company that controlled DPH at the time] was free to make an aggressive takeover bid. This is when FSIH commissioned an independent study to validate the bid from [the company that controlled DPH at the time], as shown in the minutes of the meeting of the FSIH board of directors of 14 June 2006 (35):

‘To validate the proposed price for the transaction, FSIH had the Group valued by Ernst & Young’.

(269)

When a public body sells an asset to a private person, it is necessary to determine whether the sale price is equivalent to a market price in that it corresponds to what the buyer could have obtained under market conditions. To do this the public body must use either an open, transparent and unconditional call for tenders or, failing that, an independent report.

(270)

The Commission does not deny that FSIH behaved as a private investor when it commissioned the independent study from Ernst & Young. Nevertheless, it notes that, despite requests along these lines (36), Belgium was unable to provide proof that the conclusions of the study were actually taken into account during negotiations between the parties. On the contrary, it seems that in its haste to sell its stake, FSIH left it to [the company that controlled DPH at the time] to propose a price, which was accepted without the information in the subsequent report being used in the negotiation. Therefore, the Commission maintains that the Ernst & Young 2006 Report was not used in the way of a private investor in a market economy would have done and served merely to make a negotiation that was already practically over look like a market transaction.

(271)

In light of the above, it seems that the haircuts, which were the subject of lengthy argument by Belgium, were not discussed, or at least not sufficiently discussed, as such with [the company that controlled DPH at that time] during the negotiations. On the contrary, they appear to be an ex-post structuring of justifications partially relied on internally by FSIH to convince itself of the merits of the price proposed by [the company that controlled DPH at the time].

(272)

The Commission also notes that, however questionable Belgium feels the results of the Ernst & Young 2006 Report to be, the Walloon Region did not deem it necessary to seek a second expert opinion as it did in 2003 when it invested in Duferco US. Such technical and important discussions as those relied on by Belgium on the haircuts allegedly negotiated between FSIH and [the company that controlled DPH at the time] warranted a second opinion in any event.

(273)

For these reasons, the Commission does not feel it has to comment on the haircuts point by point, given that FSIH's behaviour was not, in any case, that of a private investor in a market economy. The Commission concludes that FSIH, by selling its stake for USD […] million, gave [the company that controlled DPH at the time] an economic advantage that it would not have been able to obtain under normal market conditions.

(274)

Since the only independent report from the time indicated a market price of between USD […] and […] million, the aid to [the company that controlled DPH at the time] that is present in this measure is the difference between the lower end of the range and the actual price of the transaction, i.e. USD 25,58 million (or EUR 20,36 million (37)).

(275)

A transaction is considered to have complied with the market economy investor principle if it can be proven that it was made pari passu between the public and private investors.

(276)

Belgium maintains that FSIH's investment in DSIH is pari passu with that of Duferco Développement, given that (i) the risks are the same for FSIH and Duferco Développement (as DSIH did not belong to either the Duferco Group or FSIH before their joint acquisition of a stake), (ii) the amounts contributed are identical and (iii) the interventions were concurrent.

(277)

[…] confirms the arguments put forward by Belgium.

(278)

The Commission does not dispute that the transactions were made at the same time. However, it refutes Belgium's first two arguments and therefore does not consider FSIH's acquisition of a stake to be pari passu with that of Duferco Développement.

(279)

Although DSIH was formally established outside the Duferco Group, the Commission points out that DSIH served as a simple vehicle for Duferco to acquire, jointly with FSIH, a company, Carsid Développement in which it held all the capital within the framework of SIF. For the record, Duferco was the sole shareholder in SIF until 18 December 2012, when NLMK purchased 50 % of the Luxembourg holding company.

(280)

Consequently, the Duferco Group paid EUR […] million to acquire Carsid Développement through DSIH and received EUR […] million from the sale of Carsid Développement through its subsidiary SIF. This transaction, which in fact is a simple transfer of subsidiaries, is profitable for the Duferco Group while being a cash payment of EUR […] million for FSIH.

(281)

The Commission concludes that, to acquire an equal stake, FSIH contributed EUR […] million in cash to a subsidiary of the Duferco Group, while the latter received EUR […] million net without taking on any new risk. Under these conditions, it cannot be argued that this investment was pari passu.

(282)

Beyond the non-pari passu nature of the transaction, the Commission sets out the reasons why it concluded unequivocally that the transaction did not comply with the market economy investor principle. These reasons relate to the way in which FSIH evaluated its acquisition of a stake in DSIH, in particular (i) the assets and liabilities held by Carsid Développement and (ii) the valuation of the […] project.

(283)

Carsid Développement was established in 2006 out of the split-off from its parent company, Carsid, which contributed a number of assets in kind to the newly established company. This contribution was evaluated by KPMG on 14 November 2006 (38). The KPMG report (KPMG Report 2006) describes a contribution of assets of EUR […] million, over 99 % of which was made up of trade receivables towards Duferco La Louvière Sales, and a contribution of liabilities of a similar amount (EUR […] million) corresponding to a provision for environmental risks intended to cover the costs of cleaning up the Carsid site that had not yet been incurred.

(284)

FSIH had taken the EUR […] million of trade receivables into account in its valuation of Carsid Développement (and therefore of DSIH) (39). However, and despite a request from the Commission to this end (40), Belgium did not provide convincing proof that the liabilities of Carsid Développement had been taken into account in FSIH's valuation of DSIH. Belgium claims that these liabilities did not actually correspond to any legal obligation, given that the regulations on contaminated soil applicable in the Walloon Region only date from 2008. A clean-up project would have had to be undertaken only with a view to a subsequent valuation of the land. In this last scenario, the clean-up costs would have been fully compensated by the value of the cleaned-up land, which the parties allegedly established as EUR […] million. There was therefore no reason to quantify the environmental provision that was none the less included in the accounts. The Commission notes that Belgium did not provide it with any proof of these discussions or any report confirming the value of the cleaned-up land. It therefore concludes that this liability was not sufficiently taken into account in FSIH's valuation of Carsid Développement.

(285)

With regard to the […] project, the region claims that FSIH had based its analysis on objective cost items set out in an independent study commissioned by Duferco in 2006 from a company specialising in the redevelopment of brownfield and landfill sites (SPAQuE) and on evaluations of the market value of decontaminated land.

(286)

The Commission acknowledges the relevance of the cost items, which were established at the time by an independent expert. On the other hand, it notes that no independent ex-ante study examined the market value of the decontaminated land. Belgium justifies the figure of EUR […] million by a study by the expert Galtier from 2010, i.e. four years after the measure under examination.

(287)

The Commission notes that the independent study by SPAQuE in 2006 did propose an evaluation of the revenue that could be generated from the sale of the decontaminated land (41). The revenue did not exceed EUR […] million in any of the scenarios envisaged by SPAQuE. On the basis of the data from this time, it therefore seems impossible to justify the value of EUR […] million paid by FSIH.

(288)

Lastly, and even if one were to accept the projection of EUR […] million defended by Belgium, the Commission notes that it is not usual, for a private investor in a market economy, to value a project using the undiscounted total of the costs and expected profits. A private investor in a market economy does not attach value in the same way to profits that will not materialise until the distant future and expenditure that must be incurred soon.

(289)

The Commission notes that such errors of assessment would not have been made if a complete and unified valuation of Carsid Développement had been carried out at the time. The Commission therefore notes that FSIH did not, ex ante, undertake an economic evaluation of the type required by a private investor in a market economy.

(290)

In light of the above and of the EDF case-law already referred to in recital 248, the Commission concludes that FSIH did not behave as a private investor in a market economy and thereby conferred an economic advantage on DSIH under conditions that were not market conditions.

(291)

Given that the sale by DSIH of its activities gave rise only to a claim, which has not been collected and is today held by Duferco Wallonie, the Commission considers the amount of aid granted to DSIH to be equal to the total amount of FSIH's acquisition of a stake, i.e. EUR […] million.

(292)

The Commission considers that in order to assess whether a loan includes an element of aid, it is necessary to determine whether the beneficiary company could have obtained the amounts in question under similar conditions on the capital market (42). To assess this, the Commission may adopt a comparative approach and evaluate the loans in question with reference to market transactions that it considers comparable. This is the framework in which the Commission asked Belgium to indicate whether [the parent company of the Duferco Group at the time] and SIF had obtained comparable loans in 2009 through financial institutions other than FSIH and, if so, under what conditions.

(293)

Belgium presented the Commission with tables setting out a number of loans granted to [the parent company of the Duferco Group at the time] and to SIF.

(294)

With regard to [the parent company of the Duferco Group at the time], Belgium pointed out that, as a non-operational holding company, [the parent company of the Duferco Group at the time] had never needed to resort to bank loans, so it is impossible to compare these two rates with the rates actually applied by banks to this particular company. For [the parent company of the Duferco Group at the time], the comparators provided by Belgium therefore all concern subsidiaries of the company.

(295)

For [the parent company of the Duferco Group at the time], as for SIF, the Commission had already noted in its opening decision (43) that the tables included loans which were not comparable with those in question here: different maturities (generally shorter than those of the two loans in question), different amounts (much lower than the two loans), different nature (the tables include, for example, renewable one year loans and overdraft facilities) and, lastly, the different nature of the collateral (mortgages on buildings or other tangible assets).

(296)

Belgium criticised the Commission for not having sufficiently justified its refusal to treat the loans presented as valid comparators. The analysis criteria used by the Commission were not admissible since, according to Belgium (44), the interest rate is not a function of (i) the maturity of the loan; or (ii) the nature of the loan; or (iii) the amount of the loan.

(297)

Contrary to the claims by Belgium, the Commission notes that there is a common understanding in the financial sphere that the maturity of a loan undoubtedly has an influence on determining its remuneration. The European Central Bank represented the relationship between the rate of remuneration of a loan and its maturity in a graph (45). The curve of this graph shows very clearly that remuneration increases with the maturity of the loan, in particular for the first 10 years. This observation is also found in the case-law of the General Court of the European Union. In the Arbel Fauvet case (46), the General Court confirmed the Commission's analysis that the conditions of remuneration of a short term credit and a long term credit cannot be similar:

‘A current account overdraft granted by a private bank is a very short-term credit facility, unlike the repayable advances at issue, which have a maturity of three years, and the overdraft is not, therefore, subject to the same risk analyses by creditors. The fact that a debtor can obtain short-term credit is therefore not sufficient to assess whether it could obtain a longer-term loan, the repayment of which will depend on the debtor's ability to survive.’

(298)

The Commission considers that the nature of the loan also influences its remuneration. The risk of relating to the financing project for working capital requirements cannot be remunerated in the same way as the risk of a capital investment project. Since the level of risk is different, depending on the nature of the project to be financed, its remuneration must be adapted as a result.

(299)

For these reasons, the Commission does not accept Belgium's objection and confirms its refusal to apply a comparative approach on the basis of the loans provided by Belgium.

(300)

After the Commission refused to examine the loans provided by Belgium, […] and SIF wished to bring other loans to the Commission's attention.

(301)

[…] therefore produced a table setting out 19 multi-currency loans for amounts higher than those listed by Belgium. Of these 19 loans, the Commission notes that only one was granted in 2009. In this instance, it concerns a payment facility that could be mobilised for durations of the order of one month (the FSIH loan was concluded for six years). Most of the other loans were granted between 2003 and 2008, i.e. before the financial crisis began and at a time when the steel market was still growing. Given these substantial differences, the Commission takes the view that the 19 loans presented by […] are not good comparators for the FSIH loan.

(302)

With regard to SIF, the Commission does not consider the two loans brought to its attention to be relevant. Admittedly they are two long-term loans. However, both these loans had significant collateral: (i) collateral from DPH and NLMK; (ii) for the Rabobank loan: […]; (iii) for the Sumitomo loan: a guarantee from […]. The loan granted by FSIH does not have any guarantee. Furthermore, the loans presented by SIF were […], while the loan granted by FSIH is […]. It follows from this information that the loan granted by FSIH was riskier than those mentioned by SIF. Therefore these loans cannot be used as comparators.

(303)

In light of the above, the Commission concludes that the examples of loans supplied by Belgium […] and SIF are not comparable with the two loans under investigation. The comparative method is therefore not conclusive in this instance.

(304)

In cases where it is difficult or impossible to identify comparable transactions on the market, the Commission has drawn up indicators that serve to determine whether loans contain any aid. For loans, the methodology for calculating a reference rate that would serve as the market price is described in the 2008 Reference Rate Communication. This methodology is based on two basic parameters: the interest margin and the level of collateral.

(305)

The Commission would first apply this methodology to the loan granted to [the parent company of the Duferco Group at the time].

(306)

Belgium argues that [the parent company of the Duferco Group at the time] could justify an AAA rating at the time the loan was granted by FSIH. On the basis of the table setting out some forty loans in euros held by Duferco in 2009 (47), Belgium argues that the margins applied by commercial banks to Duferco reflect an AAA rating with high collateral, according to the 2008 Reference Rate Communication. If the Communication accurately reflects the market, the conclusion would have to be that Duferco could claim an AAA rating.

(307)

The Commission rejects this argument. The loans set out in the table provided by Belgium present a margin generally ranging from 70 to 150 basis points. A significant number of these loans were negotiated at 100 basis points. Since Belgium seems to claim that Duferco furnished its loans with high collateral, the 2008 Reference Rate Communication indicates that a margin of 100 basis points corresponds to a BB rating for this level of collateral. Therefore, Belgium's reasoning results in [the parent company of the Duferco Group at the time] receiving a BB rating at most.

(308)

This estimate of the financial rating of [the parent company of the Duferco Group at the time] is confirmed by a comparison with the ratings of the main steel groups worldwide in 2009. First, no global steel group, regardless of the rating agency, was rated AAA in 2009. Belgium does not contest this finding. Of the 10 or so leading global steel groups rated in 2009 (48), the Commission notes that only two of them were rated A and A-, three were in the BBB category and four were in the BB category. The Commission concludes that the Duferco Group would have found it difficult to justify a rating equivalent to AAA in 2009 and that the BB rating resulting from the approach proposed by Belgium seems to be in line with that of a number of other steel groups — it is even the most frequent rating in 2009.

(309)

With regard to the value of the collateral, the Commission does not agree with Belgium that the collateral on the loan granted to [the parent company of the Duferco Group at the time] was very strong. As a reminder, the collateral consists of a pledge on 15 % of the shares in DPH and an assurance that it will be repaid as a priority on maturity of a future, but almost certain, claim of USD […] million.

(310)

With regard to the pledge on 15 % of shares in DPH, the Commission notes that DPH, registered in […], is a unlisted company whose shares are, therefore, by definition not very liquid. The value of collateral increases in relation to its liquidity in the event of a default of the lender, i.e. the possibility of selling it in the short term without losing value. Furthermore, as for all ordinary shares, the value of DPH shares remains variable, which is problematic for the purpose of guaranteeing a long-term loan to be repaid in full on maturity. Lastly, Belgium argues that the Commission had signalled DPH's financial soundness in recital 148 of the opening decision (increase in own funds and net result). The Commission points out that this finding applied to the period from 2003 to 2006, not to 2009. Duferco's annual report from 2009, on the contrary, underlines the drastic fall in sales, by almost half in relation to 2008, and a consolidated loss of around USD […] million. Consequently, the Commission does not consider this collateral can be classified as normal, but should rather be classified as low.

(311)

With regard to the claim of USD […] million, the Commission points out that, under Article 3 of the loan agreement, [the parent company of the Duferco Group at the time] had to repay the loan on 31 December 2015, or earlier in the event of a change of control of SIF in accordance with the agreement of 1 September 2008 concluded between DII and NLMK. In the latter case, Article 6.1 of the agreement specifies that [the parent company of the Duferco Group at the time] undertook to allocate as a priority, before any other payment or repayment of any debts, the amounts necessary for the repayment due under the loan agreement. Article 6.2 also stipulates that [the parent company of the Duferco Group at the time] undertakes that DII will delay the exercise of its sale option until 18 December 2012 and/or until the investments mentioned in the annex are made.

(312)

The Commission points out that the value of collateral increases in line with its capacity to be realised. In this instance, Duferco's claim cannot be considered certain at the time the loan agreement was signed since DII had undertaken not to exercise its option before 18 December 2012. Until that date, in the event of the default of [the parent company of the Duferco Group at the time], the collateral had no value. Nevertheless, the Commission recognises that given the context of the time, it was highly probable that Duferco would withdraw from SIF. Its probability of realisation and thereby its quality increased over time, all the more so because Duferco had undertaken to prioritise the claim which became certain on the repayment of the loan granted by FSIH. Consequently, the Commission is minded to agree with Belgium's position on classifying the level of collateral as normal.

(313)

The Commission concludes that a rating of BB should be applied to [the parent company of the Duferco Group at the time] and that the level of collateral on the loans in question should be classified as normal at best. On the basis of the 2008 Reference Rate Communication, the margin should have been 220 basis points — a much higher level than the 75 basis points agreed by FSIH. An economic advantage was therefore conferred on SIF under conditions that were not market conditions.

(314)

Second, the Commission applies the 2008 Reference Rate Communication to the loan granted to SIF.

(315)

The Commission observes that, like Duferco, the joint venture SIF did not have a public rating. In the absence of a rating or other information enabling approximation of the financial rating of SIF, the Commission had no other choice than to use the financial rating of the parent companies as a prudent basis.

(316)

It therefore notes that NLMK's parent company was rated BBB- by Standard & Poor's, Ba1 by Moody's and BB+ by Fitch Ratings (ratings taken from NLMK's annual report 2009). According to the 2008 Reference Rate Communication, these ratings correspond to the ‘satisfactory rating’ (BB). The second parent company, Duferco, was also rated BB by the estimate adopted by the Commission in recital 313.

(317)

As a subsidiary of the two companies (NLMK and Duferco) rated BB according to the Commission's estimates, and taking into account the particularly difficult macroeconomic context of 2009 in the steel sector, the Commission takes the view that SIF's rating could prudently and reasonably be classified as BB.

(318)

The Commission also notes that the loan was granted without collateral. SIF objects that it had the capacity to provide its own guarantee, even that of NLMK or Duferco: ‘If FSIH did not ask for any guarantee from it, that is because its financial situation at the time, the fact that it belonged to the NLMK/Duferco Groups and the guarantee offered by NLMK in the event of a change of control were such as to reassure FSIH as to its capacity to repay the loan (49).’ SIF insists that it was financially sound in 2008. However, the loan granted by FSIH was for [7-12] years with repayment on maturity. Given these two characteristics, the Commission considers that a private lender would have required sufficient and proportional guarantees to remunerate its risk over [7-12] years and would not have been satisfied with only the financial results of SIF for the 2008 financial year to waive any collateral. The Commission concludes that the level of collateral on the loan is weak.

(319)

The estimate of SIF's financial rating (BB) and of the level of collateral (weak) attached to the loan from FSIH lead the Commission to take the view that a remuneration of 400 basis points would have been required under the 2008 Reference Rate Communication. This remuneration is much higher than the 75 basis points negotiated between SIF and FSIH. Since SIF received an economic advantage under conditions that were not market conditions, the Commission finds that there is aid.

(320)

In light of the above, the Commission takes the view that no private lender would have agreed to grant EUR 100 million and EUR 75 million respectively to [the parent company of the Duferco Group at the time] and to SIF under the conditions agreed to by FSIH. FSIH therefore conferred an economic advantage on the two beneficiaries of the loans by not acting as a private lender, which places the beneficiaries in a more favourable situation than their competitors.

(321)

The amount of aid results from the difference between the rates actually applied (1,99 %) and the rates under application of the 2008 Reference Rate Communication, i.e. 3,502 % (12-month Euribor + 220 basis points) for [the parent company of the Duferco Group at the time] and 4,302 % (12-month Euribor + 400 basis points) for SIF, applied to the period from the day the loans were granted to the day they were repaid early. Thus, the loan to [the parent company of the Duferco Group at the time] was signed on 4 September 2009 for an amount of EUR 30 million, a duration of six years in principle and effective rate of 2,04 % at the time of disbursement. The amount was then increased by EUR 70 million with an effective rate of 1,99 % at the time of disbursement. The early repayment was made in July 2011. The amount of the loan to SIF was EUR 75 million, released in tranches of EUR 20 million (September 2009) and EUR 55 million (December 2009) with an effective interest rate of 1,99 % for both tranches. Early repayment was agreed by tranches of principal and interest to be paid in June 2011, then in 2012 and 2013.

(322)

On the basis of a simplified discounting calculation, by way of illustration, the amount of aid is therefore approximately EUR 2,08 million for the loan to [the parent company of the Duferco Group at the time] and EUR 10,41 million for the loan to SIF (50).

(323)

A transaction is deemed to have complied with the market economy investor principle if it can be proven that was made pari passu between public and private investors.

(324)

Belgium argues that FSIH's stake in the capital is pari passu with that of DII because the invested amounts are practically identical, DLP did not belong to the Duferco Group at the time of FSIH's investment and the purpose of these investments was to buy shares belonging to SIF and not to the Duferco Group.

(325)

[…] confirms these observations and adds that DII did not possess any financial document other than those presented by Belgium with a view to examining the business case for the investments.

(326)

The Commission takes the view that, like DSIH (Measure 3), DLP is merely a vehicle used by the interested parties, FSIH and DII, to give a pari passu appearance to Duferco's takeover of assets that already belonged to it in part. DLP was set up in March 2011 by FSIH and DII as an ad hoc vehicle to accommodate Duferco's Belgian assets. This concerns mainly assets resulting from the dismantling of SIF (of which Duferco held 50 % at the time) and the assets held jointly with FSIH through DSIH.

(327)

Although the shell company DLP was formally set up outside the Duferco Group, the Commission takes the view that the simultaneity and the series of transfers of stakes between the different subsidiaries of the Duferco Group between June and July 2011 (51) show that DLP had depended on Duferco since it was established.

(328)

Moreover, and more fundamentally, the Commission notes that the asset situation of the Duferco Group and the underlying risks hardly changed at all following the measure under investigation. Duferco holds through DLP (i) 50,3 % of the steel assets it already held previously at 50 % through SIF and (ii) 50,3 % of the diversification assets it already held previously at 50 % through DSIH. Conversely, the asset situation of FSIH evolved significantly following the measure under investigation. FSIH acquired 49,7 % of substantial steel-industry assets located in La Louvière and Trebos which had not previously belonged to it.

(329)

In terms of financial commitments, this situation for the Duferco Group can be reflected as follows:

(330)

The net financial contribution of the Duferco Group for its acquisition of a 50,3 % stake in the capital of DLP was EUR […] million. This amount is out of proportion with the EUR 99,7 million (52) agreed by FSIH to become a 49,7 % shareholder in DLP.

(331)

Under these conditions it cannot be argued that this investment was pari passu.

(332)

The Commission sets out the reasons, beyond the non-pari passu nature of the transaction, why it concluded that the transaction did not comply with the market economy investor principle. These reasons relate to how FSIH evaluated its stake in the capital of DLP.

(333)

Before going into the detail of these evaluations, the Commission refutes an objection made by […]. According to […], the fact that FSIH did not have access to other documents than DII (private investor) at the time is enough to prove the prudent nature of FSIH's investment.

(334)

For the reasons explained in recital 329, the transaction under investigation was financially almost neutral for the Duferco Group. Therefore the Group did not have any incentive to produce detailed studies and reports: the price of the assets bought by DLP had only a small impact on its cash flow. This was not the case for FSIH, which had to pay 49,7 % of the price set for these assets. Consequently, it was for FSIH and not Duferco to make every possible effort to evaluate the assets at their fair value. The fact that Duferco did not have access to documents other than those produced by Belgium does not, in any case, prove that FSIH behaved as a private investor in a market economy.

(335)

With regard to the content of these evaluations, the Commission points out that four items had to be valued: (i) the shares in DLLPL; (ii) the shares in Duferco Trebos; (iii) the stocks bought from Duferco La Louvière Sales; and (iv) the project justifying an investment of EUR [30-40] million in DLLPL.

(336)

The Commission takes the view that the Galtier report of May 2011, submitted by Belgium, does not constitute a thorough valuation exercise for DLLPL. It merely refers to an estimate of the fair value of tangible assets at the La Louvière site and does not deal with any liabilities or any business plan.

(337)

However, the expert Galtier had informed the Commission, through Belgium, that it had based its estimate on a business plan in due form. The summary version of this business plan, provided by Belgium (53), is a short document that does not contain any date or refer to any assumptions. The Commission did not receive the full version of the business plan, if it ever existed.

(338)

Belgium also put forward the same argument concerning the purchase price of the shares in Duferco Trebos, i.e. that the document ‘Value in use of the tangible and intangible assets of Trebos’ of May 2011 is actually a combination of studies and reports that includes a two-page preliminary business plan drawn up by Duferco (54). The completion date of this document is not specified. The objective of the 2011 report was to evaluate the assets' value in use and to identify the value of customers in the business assets. The accounts were not checked and the company was not audited. No overall valuation of Duferco Trebos as a whole was provided.

(339)

Regarding these two acquisitions, Belgium also acknowledges that ‘the parties did not deem it necessary to produce a business plan in due form. FSIH knew perfectly well what it was acquiring.’ The Commission considers knowledge of a company to be useful but insufficient for an investor whose decision to invest depends on expected future profitability.

(340)

With regard to the evaluation of the stocks of Duferco La Louvière Sales, a company controlled by SIF, Belgium provided a document that it believes supports the market value of the stocks being higher than their book value, which was adopted to set the sale price. Again, the Commission notes that the completion date of this document is not indicated.

(341)

Lastly, with regard to the investment of EUR [30-40] million in DLLPL, the Commission had, in its opening decision, asked Belgium to produce all the studies and business plans concerning this investment. Belgium confirmed that, ultimately, a single business plan had justified the investment of EUR [30-40] million; the other plans referred to had become irrelevant as they concerned other projects that were not adopted. Belgium therefore provided the Commission with an undated, one-page plan with a table of EBITDA up to 2016 with no explanation or assumption to support the numbers presented.

(342)

In the Commission's view, the other documents relied on by Belgium do not make up for this document's shortcomings. The document entitled ‘A plan for the future of the steel industry in La Louvière’ from February 2011 details the strategy of continuing activities involving long products and flat products. One page is devoted to the development of DLLPL over the previous five years. The document ‘Long Products Marketing Plan 2011-2013’ is limited to a study of the relevant sector.

(343)

It follows from the above that FSIH's investment cannot be considered to be in line with the market economy investor principle. The Commission notes that DLP allowed FSIH to again support steel activity in Wallonia without any justification by an economic analysis in line with the market economy investor principle.

(344)

It follows from the above that the investment cannot be considered pari passu and in line with the market economy investor principle. The amount of aid received by DLP is therefore the amount of the capital increase, i.e. EUR 100 million.

(345)

After having analysed each measure individually, the Commission took the view that it was necessary to carry out a more general analysis of FSIH's investment strategy since it was set up. It considers that these measures are inextricably linked to each other and that they work towards a single objective: to support and sustain the Duferco Group's activities in Wallonia, through indirect investments in the Group's offshore subsidiaries in order to circumvent the rules prohibiting aid to the steel sector in the EU. The beneficiaries of FSIH's interventions were Duferco's parent company or subsidiaries of the group in the United States (Measure 1), […] (Measure 2), Ireland (Measure 3), […] (Measure 4) and […] (Measures 5 and 6) to finance projects that were carried out in Wallonia, in the Walloon subsidiaries of the group active in the steel sector, then in site decontamination, while the group gradually ceased its steel activities in Wallonia.

(346)

This analysis is supported by the simultaneity of or short period of time between certain events.

(347)

Duferco has been operating in Belgium since 1997, when it took over Forges de Clabecq, which were in financial straits, with support from the Walloon Region and the backing of the Commission (investment of EUR 8,75 million and loan of EUR 13,75 million) (55). The Commission notes that this takeover followed the negative decision by the Commission regarding Forges de Clabecq (56): the Belgian State's attempt to rescue the same steelworks was judged incompatible with the common market in December 1996. The company was therefore declared bankrupt in January 1997.

(348)

There followed a series of takeovers of Walloon steel sites by the Duferco Group, still with the financial support of the Walloon Region through SOGEPA, then FSIH from 2003. In 1999, Duferco acquired Usines Gustave Boël in La Louvière from Hoogovens, with support from the Walloon Region (conversion of a claim of EUR 12,5 million into 25 % of the capital and a loan of EUR 25,3 million) (57). Then, following the announcement by the Chairman of the Usinor-Sacilor group in February 2001 that he intended to close down Cockerill Sambre's hot-rolling line in Charleroi, talks were entered into between Usinor-Cockerill Sambre, the Duferco Group and SOGEPA primarily with a view to setting up a joint venture to produce slabs using Cockerill Sambre's existing mill in Charleroi together with the plant owned by Duferco Clabecq. In 2002, Duferco took over the hot-phase site in Charleroi, which became Carsid, in cooperation with Usinor, again with support from the Walloon Region (SOGEPA's stake in Carsid's capital was EUR 20 million, reduced to EUR 9 million after the decision to initiate the procedure provided for in Article 6(5) of Decision No 2496/96/ECSC).

(349)

The Commission notes that FSIH was created in March 2003, i.e. during the Commission's investigations following the adoption of the decision to initiate the formal investigation procedure on 3 April 2002 on the financial participation by the Walloon Region in the company Carsid. This procedure was completed on 15 October 2003 (58) with a negative decision from the Commission on the grounds that an investment in a new company could not be considered pari passu where only the public shareholder bears a new risk, with the private shareholder only transferring a business activity or existing project to the new company. The Commission also notes that, in many cases, the interventions by the Walloon Region in Duferco's favour made it possible to delay adjustments that were socially difficult but economically necessary in the steel sector in Wallonia. This behaviour, motivated by regional and social considerations of safeguarding employment to the detriment of streamlining and modernising the steel industry, cannot be considered to be that of a private investor.

(350)

Furthermore, the Commission observes that the investment behaviour of FSIH does not correspond to that of a prudent market investor. In this case, FSIH's behaviour should be compared to the behaviour that an investment fund manager would adopt. Taken as a whole, the investment policy followed by FSIH, as highlighted by this procedure, is not likely to be followed by such a market operator. A non-speculative investment fund must diversify the risks in its investment portfolio in order to ensure its longevity. While taking into account the desired performance and profit horizon, the diversification of assets in the portfolio geographically and by sector reduces the volatility of the assets and therefore the investment risk of the portfolio as a whole. This diversification does not necessarily have to be extreme, but in this case FSIH concentrated all its investments in a single company.

(351)

Indeed, by investing exclusively in the Duferco Group, FSIH's exposure to risk was at its greatest since the return on investment for FSIH depended on the success or failure of a single company. It follows, overall, that the investment behaviour of FSIH is an additional indication that its operations do not satisfy the market economy investor test.

(352)

In light of the above, the Commission concludes that none of the examined measures complies with the market economy investor principle. The conditions of the interventions in each case conferred an advantage on the different entities concerned in the Duferco Group. They therefore constitute State aid within the meaning of Article 107(1) TFEU.

(353)

The total amount of aid is therefore EUR 211,43 million, in principle.

(354)

The prohibition on State aid is neither absolute nor unconditional. In particular, paragraphs 2 and 3 of Article 107 TFEU constitute legal bases allowing some aid measures to be considered compatible with the internal market.

(355)

In its response to the Opening Decision, Belgium did not put forward any reason relating to the compatibility of the measures under investigation.

(356)

The Commission observes that the derogations in Article 107(2) TFEU are clearly inapplicable. Of the derogations in paragraph 3 of this Article, only points (a) and (c) could prove useful. Point (a) stipulates that aid to promote the economic development of areas where the standard of living is abnormally low or where there is serious underemployment may be considered to be compatible with the internal market. Point (c) states that aid to facilitate the development of certain economic activities or of certain economic areas, where such aid does not adversely affect trading conditions to an extent contrary to the common interest may also be considered to be compatible with the internal market.

(357)

However, no framework or guidelines from the Commission concerning the compatibility criteria relating to State aid seem to be applicable.

(358)

Since the ECSC Treaty expired on 23 July 2002, the Commission has adopted various guidelines and communications to maintain the prohibition on regional aid for investment and rescue and restructuring aid to the steel industry.

(359)

Thus, paragraph 8 of the Guidelines on National Regional Aid for 2007-2013 (59), the period during which FSIH granted the aid under investigation, explicitly excludes regional aid to the steel industry from their scope and confirms that it is not compatible with the common market.

(360)

Likewise, the Communication on aid for the steel sector (60), which was in force until 31 December 2009, states that rescue and restructuring aid for firms in difficulty in the steel sector is not compatible with the common market. Only aid to cover payments payable by steel firms to workers made redundant or accepting early retirement and aid to steel firms which permanently cease production of steel products is authorised, under certain conditions. The purpose of the aid granted by FSIH does not correspond to the two aforementioned aid categories. On the contrary, it was used to finance investments.

(361)

With regard to possible compatibility of the aid granted to SIF under the 2009 Communication, the Commission takes the view that the Communication does not apply to SIF. Point 4.4.2 of the 2009 Communication covers companies having difficulties in finding finance in the market circumstances between December 2008 and December 2010. It does not seem from SIF's comments that it had this type of difficulty. The Commission therefore considers that SIF did not meet the eligibility conditions laid down by the 2009 Communication and that the aid cannot be deemed compatible with the internal market.

(362)

Finally, point 18 of the Community Guidelines on State aid for rescuing and restructuring firms in difficulty (61), which were in force until 30 July 2014, confirms that the steel sector is excluded from their scope.

(363)

In light of the above, the Commission concludes that the six aid measures under investigation are not compatible with the internal market.

(364)

In accordance with the TFEU and the settled case-law of the Court of Justice, the Commission is competent, when it has found that aid is incompatible with the internal market, to decide that the State concerned must abolish or alter it (62). The Court has also held on a number of occasions that the obligation to abolish aid incompatible with the internal market which a Commission decision imposes on a Member State has as its purpose to re-establish the previously existing situation (63).

(365)

In this context, the Court has established that this aim is achieved once the recipient has repaid the amounts granted by way of unlawful aid, thus forfeiting the advantage which it had enjoyed over its competitors on the market, and the situation prior to payment of the aid is restored (64).

(366)

Article 16(1) of Council Regulation (EU) 2015/1589 (65) lays down: ‘Where negative decisions are taken in cases of unlawful aid, the Commission shall decide that the Member State concerned shall take all necessary measures to recover the aid from the beneficiary …’.

(367)

Consequently, since the measures in question were implemented in breach of Article 108 TFEU and must be considered aid that is unlawful and incompatible with the internal market, it must be recovered in order to re-establish the situation that existed on the market before the aid was granted.

(368)

In this case, the aid was granted to the following companies: DII (Measure 1), [the company that controlled DPH at the time] (Measure 2), DSIH (Measure 3), [the parent company of the Duferco Group at the time] — which became […] — (Measure 4), SIF (Measure 5) and DLP (Measure 6), all subsidiaries of the Duferco Group in different countries. These companies or their possible legal successors must be required to repay the aid that was unduly received.

(369)

Recovery must also cover the period during which the beneficiaries received an advantage, i.e. between the time when the aid was made available to each beneficiary and when it was actually recovered, having regard to the criteria for defining an economic advantage and the calculation principles defined in Sections 7.1.4.1 to 7.1.4.5. In the context of Belgium's obligation of loyal cooperation within the framework of the recovery procedure, the amount of aid granted to [the parent company of the Duferco Group at the time] — which became […] — and to SIF will have to be established more precisely during the procedure, on the basis of information to be supplied by Belgium and taking into account, in particular, for the aid in the form of loans, the actual dates of payments or repayments as well as all other relevant circumstance reported by Belgium. In each case, the amounts to be recovered should include interest due until recovery takes place.

(370)

The Commission finds that Belgium has unlawfully implemented the aid measures in question in breach of Article 108(3) of the TFEU. The Commission considers all the measures under investigation to be State aid incompatible with the internal market,