29.6.2012   

EN

Official Journal of the European Union

L 170/1

(1)

On 3 December 2002, the Belgian authorities notified (3) to the Commission an increase in the capital of De Post-La Poste (hereinafter ‘DPLP’, now ‘bpost’ (4)), the Belgian postal operator. Following a preliminary examination, the Commission decided on 23 July 2003 (5) not to object to the notified measure. The Commission found in its decision that the capital injection, together with other measures in favour of DPLP identified in the preliminary investigation, had not overcompensated DPLP for the net costs incurred in discharging its public service missions from 1992 to 2002.

(2)

The Commission’s decision was challenged before the Court of First Instance (CFI) on 27 November 2003 by Deutsche Post AG and its Belgian subsidiary, DHL International (6). On 10 February 2009, the CFI annulled the Commission’s decision. The CFI concluded that the Commission had decided not to raise objections despite serious difficulties that should have led it to initiate a formal investigation procedure.

(3)

In April 2009, Belgium lodged an appeal (7) (C-148/09) with the Court of Justice (ECJ) against the CFI’s annulment of the Commission’s decision of 23 July 2003.

(4)

On 22 September 2011, the ECJ dismissed Belgium’s appeal and found that the CFI was right to annul the 2003 Commission decision.

(5)

In 2005, a number of Belgian private press distribution companies filed a complaint (8) under antitrust and State aid rules against a press distribution agreement concluded between the Belgian authorities and DPLP. The agreement entrusts DPLP with the public service mission of press distribution throughout Belgium, and fixes tariffs for this service and the compensation DPLP is entitled to receive for the extra costs incurred in fulfilling this mission. The complaint centres on the tariffs for DPLP’s services, which the complainants claim are so low as to effectively exclude any possibility of competition.

(6)

In respect of State aid rules, the Commission rejected the complaint, since the preliminary assessment showed that the net additional costs of the public service had not been overcompensated. However, in May 2009 the complainants contested this preliminary position, submitted additional information and asked the Commission to carry out an in-depth investigation.

(7)

When examining the complaint concerning the compensation for press distribution, the Commission learned that Belgium had entered into a management contract (‘contrat de gestion’) with DPLP in 2005, which had not been notified and provided for the payment of compensation for services of general economic interest (SGEIs). The Commission therefore opened an ex-officio case (9) and sent a request for information to Belgium in April 2007.

(8)

In December 2007, the Belgian authorities replied that they did not regard the SGEI compensation as State aid, since it fulfilled all four conditions set out in the Altmark judgment (10). They subsequently provided information on the fourth management contract, particularly with regard to the cost accounting system used to calculate the compensation paid by the State.

(9)

Subsequently, the measures covered by the ex-officio case and the complaints were included in formal investigation procedure C20/09 opened on 23 July 2009.

(10)

Having regard to the Court of First Instance’s annulment of the Commission’s decision of 23 July 2003 and having examined the information supplied by the Belgian authorities on the measures in question, the Commission decided on 13 July 2009 to initiate the formal investigation procedure laid down in Article 108(2) TFEU (11).

(11)

The Commission decision initiating the procedure was published in the Official Journal of the European Union and the Commission invited interested parties to submit their comments.

(12)

The Belgian authorities submitted comments on 13 August 2009 and made further submissions on 10 September 2009, 12 October 2009, 23 April 2010, 26 April 2010, 19 October 2010, 19 November 2010, 21 December 2010, 18 January 2011, 17 February 2011, 2 March 2011, 29 March 2011, 11 April 2011, 10 May 2011, 27 May 2011, 14 June 2011, 10 August 2011, 26 August 2011, 7 October 2011, 14 December 2011, 19 December 2011, 20 December 2011 and 26 December 2011. They pointed out that they had lodged an appeal on 22 April 2009 against the CFI judgment of 10 February 2009 and provided clarifications regarding the compensation for the public services and other ad hoc measures in favour of DPLP.

(13)

Deutsche Post AG and DHL International NV/SA submitted comments on 28 August 2009.

(14)

The Vlaamse Federatie van Persverspreiders (a Dutch-language association of press distributors, hereinafter ‘VFP’) and Prodipresse (Union des professionnels de la diffusion de la presse, a French-language association of press distributors) submitted comments on 28 August 2009.

(15)

Belgische Distributiedienst NV/Belgique Diffusion SA (a competitor of DPLP in the press distribution market, hereinafter ‘BD’) submitted comments on 28 August 2009.

(16)

The Association Belge des Editeurs de Journaux/Belgische Vereniging van Dagbladuitgevers (Belgian association of newspaper publishers) submitted comments on 28 August 2009.

(17)

The Belgian Post Office was a State agency (‘Régie des Postes’) until 1992, when it was incorporated as a separate legal entity (‘enterprise publique autonome’, autonomous public undertaking). In 2000, its legal status changed to ‘société anonyme de droit public’ (State-owned public limited company).

(18)

DPLP’s operating income in 2010 was EUR 2 318 million and its operating profit was EUR 319 million. At the end of 2010, DPLP employed 33 616 people. In comparison, in 2002, it had 42 000 staff. Staff costs represent 65 % of total operating costs.

(19)

A minority share of 50 % minus one share is held by the private investor CVC Capital Partners. The Belgian State is the majority shareholder.

(20)

DPLP was entrusted with public service missions by the Postal Law (12) and four successive management contracts during the whole period under investigation, i.e. from 1992 to 2010.

(21)

The management contracts stipulate the rules and conditions for carrying out the tasks undertaken by DPLP as part of its public interest activities and fix the financial intervention of the Belgian State.

(22)

Since its incorporation in 1992, DPLP has signed four management contracts with the Belgian State, covering the following periods (13):

(23)

The management contracts lay down the public service missions. The public service missions defined by the Belgian State in the current management contract are as follows (14):

(24)

DPLP was granted public support in the form of several different measures from 1992 to 2010:

(25)

In 1992 the Régie des Postes was transformed by law into an autonomous public undertaking (EPA) called ‘De Post-La Poste’. DPLP, as the legal successor of the Régie des Postes, took over its assets and personnel. The transformation led to a major change in the business model:

(26)

Thus, in 1992 DPLP took over responsibility for 43 747 civil servants and the associated legal protection they enjoyed with regard to job security, remuneration and pension entitlements (Article 33 of the Law of 21 March 1991).

(27)

In 2008 DPLP still employed some 23 500 statutory agents (with a status similar to that of civil servants), representing about two thirds of its total workforce, as set out below.

Figure 1

Employment regimes in DPLP

(28)

Between 1972 and 1996 pensions were paid under a pay-as-you-go scheme. Under this scheme, DPLP reimbursed the State for pensions that the State paid to retired former civil servant employees (17) for pension rights they had accumulated since 1972, when the Belgian Post Office had become a separate accounting entity within the State (Régie des Postes).

(29)

The system of financing pensions remained unchanged until the 1997 pension reform. This reform aligned the first-pillar pension system for DPLP statutory personnel with the private sector system: from 1997 onwards DPLP had to pay a contribution towards pensions to the ONSS (Belgian national social security office) amounting to 8,86 % of salaries.

(30)

In parallel, the State agreed to take over DPLP’s accumulated pension liabilities amounting to EUR 3,8 billion.

(31)

Each contract provided for the payment by the State of annual compensation for discharging the public service missions set out in that contract.

(32)

Under the first three contracts, the compensation was calculated on the basis of the actual difference between the expenditure on public service missions and revenue from them. Under the fourth management contract a fixed amount of compensation is paid, based on the projected difference between the cost borne by DPLP and revenue generated from the use of the public services. This compensation is subject to a ceiling laid down in the management contracts (18).

(33)

Under the current management contract (19), DPLP is entitled to receive compensation from the State for the following public services:

(34)

It should be noted that the universal postal service was excluded from compensation under the last three management contracts (20). Although press distribution can generally be part of the universal service in accordance with Directive 97/67/EC of the European Parliament and the Council of 15 December 1997 on common rules for the development of the internal market of Community postal services and the improvement of quality of service (21), the specific missions (distribution of newspapers and periodicals) to be compensated under the management contracts have been defined by the Belgian authorities according to criteria that distinguish them from the universal postal service. These criteria are: (1) the objectives of the mission, which go beyond universal service objectives, (2) the specific conditions set out in the contracts and (3) the specific provisions on pricing, which differ from those pertaining to universal service pricing.

(35)

Objectives of the mission: the Belgian authorities consider it particularly important to ensure a wide distribution of the written press throughout Belgium in order to enhance the level of information and diversity of opinion, and to foster public participation in the political debate.

(36)

Specific operational conditions are set out in the contract: the specific press distribution service giving rise to compensation imposed on DPLP goes beyond the conventional distribution service included in the universal service. It is subject to a set of stringent quality and performance requirements in terms of time of delivery (e.g. before 07:30), frequency, delivery options for publishers, flexibility for publishers as regards volume, and continuity of service (including a narrow definition of force majeure).

(37)

Specific pricing: the tariffs of the press distribution mission giving rise to compensation are fixed in tripartite negotiations (Belgian State/DPLP/press sector) and, unlike the universal public service tariffs, do not require the approval of the regulator. While the universal public service tariffs are geared to costs, the tariffs of the specific press distribution mission giving rise to compensation are much lower, so that the specific objectives of the mission can be attained.

(38)

The public services mentioned in recital 33 are covered by public service compensation. The public service compensation actually paid in each financial year is as follows:

First management contract

Second management contract

Third management contract

Fourth management contract

(39)

Injections in 1997: DPLP received two non-notified capital injections amounting to EUR 62 million. However, the Belgian authorities subsequently stated that these funds in fact constituted delayed public service compensation, which had not been paid to DPLP on time.

(40)

Injection in 2003: the Belgian authorities injected into DPLP capital of EUR 297,5 million. They claim that this injection was made on market economy investor terms and, therefore, according to the market economy investor principle (MEIP), does not constitute State aid.

(41)

Injection in 2006: the Belgian authorities made a new capital injection of EUR 40 million. They claim that this injection was in line with the MEIP, having been made on terms identical to those applying to a private investor (pari passu).

(42)

Until 31 December 2005, DPLP was exempt from corporate tax.

(43)

Throughout the period under investigation, DPLP was exempt from property tax on the buildings used for public services tasks. The annual advantage varied between EUR […] (23) and EUR […]. If this tax had been paid the cost of the products (both public service and commercial) supplied in these buildings would have been proportionally higher.

(44)

DPLP is also exempt from other indirect and local taxes. The Belgian authorities are not able to estimate the exact value of the advantage but they claim that the measure falls under the de minimis rules.

(45)

The tax exemptions predate the entry into force of the EEC Treaty in 1958 and have continued to exist without material alteration. While the corporate tax exemption was brought to an end on 31 December 2005, the two other measures still apply.

(46)

In 1992 the Belgian State transferred 63 buildings (24) to DPLP free of charge. The buildings were already being used by DPLP although legally they belonged to the State. The total value of the buildings was estimated at EUR 112,2 million (25). These buildings were distinct from and additional to the original assets, which had been separated from the State property when the Post Office became a separate State agency (Régie des Postes) in 1971. The Régie des Postes was already the legal owner of the assets that had been included in its balance sheet. By contrast the additional buildings transferred in 1992 were supposed to compensate for the remaining pension burden on DPLP when it changed its status from a State agency to an autonomous public undertaking.

(47)

The Belgian authorities claim that the transfer of buildings was merely part of an internal reorganisation of State property and did not constitute an advantage. They argue that the buildings had already been used by the Belgian Post Office before the incorporation of DPLP in 1992 (i.e. they had been already used by the State agency).

(48)

DPLP was given the option of requesting a State guarantee for its loans at a premium of 0,25 %.

(49)

In the period under investigation, the guarantee was used only once for a loan of EUR 100 million from the European Investment Bank (EIB). The agreement signed between DPLP and the EIB on 22 December 2004 enabled DPLP to use the EIB credit facility until 10 June 2008. Making use of this facility, DPLP requested EUR 100 million on 8 November 2007, which it received on 5 December 2007. DPLP undertook to reimburse the loan in 11 equal annual instalments of approximately EUR 9,09 million. The first instalment is due on 5 December 2012 and the last on 5 December 2022. Interest is payable quarterly (5 December, 5 March, 5 June, 5 September). The interest rate is calculated on the basis of the three-month Euribor minus 37 basis points.

(50)

On the basis of the Altmark judgment (26), the Commission took the view that the public service compensation paid since 1992, as well as the tax exemptions, capital injections, State guarantee and transfer of buildings referred to above, may have constituted aid within the meaning of Article 107(1) TFEU.

(51)

State aid in the form of compensation for costs linked to universal service obligations may be declared compatible under Article 106(2) TFEU if it is necessary to the operation of the services and does not affect the development of trade to an extent that would be contrary to the interests of the European Union.

(52)

In the 2009 decision initiating the procedure, the Commission expressed doubts about whether the aid measures granted to DPLP were necessary for the fulfilment of the public service missions and proportional to that end.

(53)

As regards the calculation of overcompensation pursuant to Article 106(2) TFEU, the Commission took the view that all revenue and costs linked to the public service obligations had be taken into account in verifying the absence of overcompensation for the public service missions. This calculation would include verifying the soundness of the accounting methodology used to determine the costs attributable to the SGEIs, and taking into account all net costs of the public service, as well as additional aid granted in any form whatsoever.

(54)

In respect of the public financing of civil servants’ pensions, the Commission expressed doubts regarding the extent to which the release from pension obligations placed DPLP in a comparable situation to its competitors as regards social security contributions.

(55)

In line with the approach taken in the Commission Decision on the public financing of pension costs for civil servants working for the French Post Office (27), it was appropriate to examine whether the measure was compatible with Article 107(3)(c) TFEU and, to this effect, to establish the extent to which the social security costs paid by DPLP after it was released from its liabilities were comparable to those of its competitors.

(56)

On 28 August 2009 VFP and Prodipresse submitted their comments concerning the press distribution scheme operated by DPLP.

(57)

In line with their previous complaint, VFP and Prodipresse highlighted the distortion of competition arising from the low tariffs offered by DPLP for press distribution, which made it impossible for other operators to compete in the market.

(58)

According to their submission, DPLP’s prices were fixed in tripartite negotiations between DPLP, publishers and the State. The losses incurred by DPLP as a result of these below-market prices are, they claim, covered by the State as part of the SGEI compensation. While DPLP is the direct beneficiary of this aid, the newspaper publishers also benefit from it indirectly.

(59)

VFP and Prodipresse suspect that DPLP may be overcompensated for the press distribution scheme and may be using the excess compensation to finance its other activities.

(60)

On 28 August 2009, the Association Belge des Editeurs de Journaux/Belgische Vereniging van Dagbladuitgevers submitted comments, in which it underlined the importance, in terms of democracy and equality, of an efficient newspaper distribution system throughout Belgium, at uniform and affordable prices, so that subscribers can receive their newspapers early in the morning.

(61)

On 28 August 2009, Deutsche Post AG and DHL International submitted their comments. The submission highlighted the need to examine the pension relief granted to DPLP, in particular to ensure that it was in line with previous practice, verify the market conformity of the capital injections, check that the allocation of costs complied with the Chronopost judgment, and examine the transfer of buildings.

(62)

On 28 August 2009, Belgische Distributiedienst NV/Belgique Diffusion SA (‘BD’), part of the TNT group, submitted its comments. BD competes against DPLP and particularly its press-distribution subsidiary, Deltamedia, in the market for the distribution of newspapers and periodicals. In line with its earlier complaint, BD underlined DPLP’s below-cost pricing of newspaper distribution. It claimed that the losses incurred by DPLP were financed by public subsidies and that the low prices made it difficult to compete against DPLP.

(63)

BD alleged that DPLP had breached Council Directive 92/50/EEC of 18 June 1992 relating to the coordination of procedures for the award of public service contracts (28), and Article 7 of Directive 97/67/EC. It also mooted the existence of aid within the meaning of Article 107(1) TFEU and questioned the compatibility of compensation for costs linked to SGEIs pursuant to Article 106(2) TFEU.

(64)

BD questioned whether the distribution of newspapers could be qualified as an SGEI, since in most Member States this activity is open to competition and does not involve the postal operator.

(65)

BD also questioned whether a subsidy reserved exclusively for the postal operator was the most appropriate way to support the distribution of newspapers, as opposed to granting aid to publishers, who could then purchase the distribution services on the open market, thereby allowing competition between different providers.

(66)

Even if approval were given for entrusting the distribution of newspapers to DPLP, BD questioned whether the costs were necessary. According to BD, DPLP’s costs are significantly higher than those of other operators, which increases the level of subsidies required. In this regard, the subsidies were neither proportionate nor justified.

(67)

Finally, BD suspects that the costs incurred by DPLP that are borne by the State could include investment costs that are not necessary for supplying the SGEI, since DPLP may have used the infrastructure in place for other services. The marketing campaigns for commercial products may also have been financed through the State subsidy.

(68)

The Belgian authorities claim that, following the approach adopted in the Combus judgment (29), the release from pension liabilities does not constitute State aid since it relieves the Belgian Post Office of an abnormal burden, which would not normally be borne from an undertaking’s budget.

(69)

Such an abnormal burden results from higher payroll costs associated with statutory personnel and operational rigidities, which prevent DPLP from reducing the share of civil servants in its workforce as much as it would like.

(70)

The Belgian authorities identify three sources of rigidity:

(71)

Job security for civil servants: DPLP claims that its statutory personnel have de facto permanent employment contracts. DPLP is unable to terminate the employment of statutory personnel except for serious reasons or consistently poor performance. Even in these circumstances, the dismissal procedure is subject to particularly onerous requirements.

(72)

Moreover, the Belgian Conseil d’État (administrative court which has jurisdiction over statutory personnel whereas labour courts are competent for contractual personnel) has shown a tendency to favour statutory personnel by setting a strict standard for establishing ‘good cause’. As a result, DPLP is unable to dismiss statutory personnel, whether individually or collectively, on economic or technical grounds.

(73)

Accordingly, termination of employment is not a management tool for adjusting staffing to actual operational needs. Also, the impossibility of dismissal for economic reasons significantly limits DPLP’s leverage to address such issues as inadequate mobility and above-market salaries and benefits of statutory personnel. Furthermore, statutory personnel are not entitled to transfer automatically to the public administration, nor can DPLP impose such a transfer.

(74)

This means that a reduction in the number of DPLP’s statutory personnel is essentially limited to voluntary departures, which are insignificant in number, and to natural wastage.

(75)

No internal mobility: in addition to its inability to reduce the number of statutory personnel in areas where the demand for their services has diminished, DPLP’s ability to reallocate statutory personnel within its organisation according to business needs is subject to severe constraints.

(76)

On account of factors such as language group, qualifications, seniority level and functional classification, DPLP is largely unable to move statutory personnel from areas in which staffing needs are diminishing, e.g. as a result of the decline in the volume of the traditional mail business, to areas which are understaffed, inter alia, because of retirement or increased demand.

(77)

In other words, as a result of the rigidities of their status, DPLP is largely unable to reallocate statutory personnel according to its needs or to address a mismatch between overstaffed and understaffed activity segments. Paradoxically, to meet needs in understaffed areas, DPLP needs to hire interim or contractual personnel, while keeping on its payroll excess statutory personnel in overstaffed areas.

(78)

No access to temporary lay-offs: in addition, DPLP cannot temporarily lay off statutory personnel like private-sector operators can in case of shortage of work.

(79)

Higher payroll costs: when the Régie des Postes became an autonomous public undertaking in 1992 and was renamed DPLP, the statutory personnel had significantly higher salaries and benefits than contractual personnel in the private sector. Since 1992, automatic seniority-linked salary increases and a strong union bargaining position have helped to keep salaries and benefits at their comparatively higher levels.

(80)

In addition to higher salaries and benefits compared with contractual personnel, statutory employees are entitled to between 4 and 11 additional days of paid leave (30), giving rise to an additional cost to DPLP. These features of statutory personnel have made it difficult for DPLP to manage its unit staff cost, which is substantially higher than that of contractual employees in the private sector.

(81)

Since taking over the statutory personnel of the Régie des Postes in 1992, the DPLP management has recognised that these staff, with their significantly higher payroll costs and operational rigidities, represented a structural handicap that was increasingly difficult to bear as postal market liberalisation progressed. Accordingly, DPLP stopped hiring statutory workers in 1998. Attrition, predominantly through retirement, then reduced these staff in absolute numbers to 23 538 in 2008.

(82)

As the reduction in the number of statutory personnel is primarily dependent on the natural attrition rate in the context of the existing age pyramid, its progress has been slow and is likely to remain so until 2030, long after full liberalisation of the postal market, as portrayed below.

Figure 2

Evolution of DPLP’s statutory workforce

(83)

The Belgian authorities therefore consider that DPLP’s statutory personnel represents a serious operational and cost handicap vis-à-vis private operators, and that the State has relieved DPLP of an abnormal burden not borne by its competitors.

(84)

They claim that the measure does not constitute State aid and, therefore, the approach adopted by the Commission in the decision initiating the procedure, in conformity with previous pension scheme cases (EDF (31), La Poste (France) (32), Royal Mail (33)), is not necessary in the case of DPLP.

(85)

However, the Belgian authorities claim that, if such a method were applied to DPLP, it would lead to the conclusion that the share of social security costs, as a percentage of wage costs, borne by DPLP is the same as that of its competitors.

(86)

The Belgian authorities submitted DPLP’s accounting data (34) and described the principles of its cost accounting system.

(87)

The Belgian authorities claim, first of all, that the annual public service compensation does not constitute aid since it complies with the Altmark judgment.

(88)

Secondly, the Belgian authorities argue that, if the public service compensation were aid, it would be compatible since it does not overcompensate DPLP for the net costs it incurs when discharging its public service missions. They claim that the calculations they have presented show that DPLP is undercompensated for the costs it incurs.

(89)

Lastly, the Belgian authorities point out that some of the measures constitute existing aid. See section on existing aid (recitals 224 et seq.).

(90)

In a letter to the Commission dated 12 August 2009, the Belgian authorities indicate that the 1997 capital injections totalling EUR 62 million should be considered deferred compensation for public service obligations.

(91)

According to the Belgian authorities, the 2003 and 2006 capital increases were both made on the basis of a strategic plan (35) whose ultimate goal was to enhance the company’s competitiveness and profitability in the context of the progressive liberalisation of the postal market.

(92)

The strategic plan had two main objectives:

(93)

2003 capital injection: according to the Belgian authorities, the 2003 capital increase was aimed at strengthening DPLP’s equity capital with regard to the major financing commitments it had undertaken, in particular for the construction of four new sorting centres and the upgrade of an existing one, the optimisation of mail distribution routes and automation.

(94)

Moreover, the Belgian authorities argue that the funds injected also covered personnel costs arising from a reduction in the number of employees as a result of an early retirement scheme, a reduction in working time and an increase in the salaries of new postal workers.

(95)

These measures offered a reasonable assurance of increasing profitability over the period 2003-2007 to the levels expected by private investors. Thus, when it decided to make an equity investment, the Belgian State as the sole shareholder in DPLP in 2003 was guided by the prospect of profitability, in full compliance with the MEIP.

(96)

On 5 August 2009 the Belgian authorities further argued that the profitability forecasts showed that the capital injection was in line with the MEIP, because the expected level of return was above what a private investor would have required. These forecasts were not unrealistic since actual profitability between 2003 and 2008 had been consistently above predicted profitability.

(97)

The Belgian authorities have clarified that there is no link between the 2003 capital injection and the 2005 cancellation of receivables from the State in DPLP’s accounts. The cancellation, which took place at a different time, was a normal accounting operation, since the State had clearly indicated that those receivables were not going to be paid to DPLP. The final amounts cancelled in the 2005 balance sheet clean-up were different from the amounts of capital injection in 2003 and accumulated losses at the end of 2002.

(98)

2006 capital injection: the Belgian authorities claim that this injection complied with the MEIP because it was made under the same terms as applied to a private investor (pari passu), namely Post Invest Europe SA, which invested EUR 300 million. New shares were issued to both the State and the private investor. As a result, the new investor obtained 50 % minus one share, while the State retained 50 % plus one share (36).

(99)

In order to determine the price of the new shares issued, the investment bank hired by the Belgian authorities estimated DPLP’s total value at between EUR […] and EUR […]. The State finally opted for a valuation of EUR […] for the entire company before the capital injections. The Belgian authorities acknowledge that DPLP’s subsequent performance has exceeded the assumptions made at the time of the valuation.

(100)

According to the Belgian authorities, the implementation of the strategic plan made DPLP one of the most efficient operators in Europe. It achieved significant improvements in profitability, efficiency (37) and quality, while maintaining prices to customers at an affordable level.

(101)

In their submission of 12 August 2009, the Belgian authorities acknowledge that DPLP was not subject to a corporate tax regime, but to a tax on profits (i.e. a tax on legal persons) until 27 December 2005. They nonetheless claim that during the period 2003-2005, DPLP did not benefit from this tax regime because, if it had been subject to corporate tax, it would have been able to carry over past losses, and would not have had to pay corporate tax during that period.

(102)

In addition it is claimed that DPLP had to pay taxes on profits in the order of EUR 2 million during the period 2003-2005.

(103)

From the fiscal year 2006, DPLP paid corporate taxes without being able to carry over past losses. The Belgian authorities estimate the foregone tax credit up to that time at EUR 51,87 million.

(104)

For these reasons, they conclude that DPLP did not obtain a net benefit from the corporate tax exemption before 2006.

(105)

The Belgian authorities specify that DPLP is exempt from property tax on the revenue from property used to supply a public interest service.

(106)

The Belgian authorities have estimated the net advantage enjoyed by DPLP and argue that it is only a fraction of the additional costs DPLP incurs as a result of its obligation to maintain a postal network with a strictly defined territorial density, which is not compensated in any other way.

(107)

The Belgian authorities explain that DPLP benefits from certain exemptions from indirect and local taxes. They argue that the net effect of these exemptions is de minimis.

(108)

In their submission of 12 August 2009 the Belgian authorities confirm that a number of buildings were transferred to DPLP when it was transformed into an autonomous public undertaking. These were buildings that DPLP had been using since 1971 and were necessary for discharging its postal service mission.

(109)

The Belgian authorities argue that the transfer of these buildings did not confer any advantage on DPLP since it was already using them for its public service mission. In other words, it was merely a reorganisation of public property.

(110)

In their submission of 12 August 2009 the Belgian authorities claim that DPLP does not automatically benefit from a State guarantee for loans. If DPLP requested such a guarantee, the Belgian authorities would have to give their approval and DPLP would have to pay an annual premium of 0,25 % to the State Treasury.

(111)

During the period under consideration, DPLP only once resorted to a State guarantee.

(112)

The Belgian authorities consider a premium of 0,25 % appropriate given DPLP’s financial position and the remote possibility of its failing to repay. Moreover, they estimate that the market rate for a similar loan without the State guarantee would have been up to 40 basis points higher.

(113)

In their submission of 12 October 2009, the Belgian authorities argue that only the comments by BD and FVP/Prodipresse contain specific critical elements. These elements mainly concern the early distribution of newspapers and periodicals.

(114)

The Belgian authorities argue that, contrary to BD’s remarks, the award of non-reserved postal services (such as press distribution) does not fall within the scope of EU public procurement rules (Directive 2004/18/EC of the European Parliament and of the Council (38)).

(115)

In addition, the Belgian authorities argue that the distribution of newspapers and periodicals, which constitutes an SGEI, does not violate EU provisions on the liberalisation of the postal market.

(116)

The Belgian authorities also point out that the award decision was not arbitrary. They claim that DPLP is the only company that fulfils the conditions and obligations imposed on the provider of such services.

(117)

The Belgian authorities argue that, contrary to BD’s claims, during the period 2002-2008 DPLP did not benefit from overcompensation for the costs linked to the supply of its newspaper and periodical distribution service.

(118)

The Belgian authorities do not agree with the claim made by FVP/Prodipresse that DPLP exaggerated the investments and additional costs related to increasing national early morning coverage from 82 % to 100 %. Since the 18 % concerned mainly remote areas, DPLP’s personnel costs increased disproportionately and it had to acquire long-haul vehicles with large depreciation costs.

(119)

Lastly, the Belgian authorities argue that the costs of the publicity campaign have not been taken into account in the calculation of the State compensation which DPLP received for early morning press distribution.

(120)

BD claims that DPLP is manifestly cost inefficient. The Belgian authorities contest this argument by emphasising that it does not fully take into account the considerable additional costs related to early press delivery and overlooks major efforts by DPLP to control its costs and increase its productivity.

(121)

As BD affirms, Deltamedia’s costs are lower than those of DPLP. According to the Belgian authorities this is because of the following factors:

(122)

In their submission of 21 December 2010, the Belgian authorities claim that certain of the measures under investigation constitute existing aid.

(123)

Firstly, the Belgian authorities claim that the tax exemptions enjoyed by DPLP predate Belgium’s accession to the European Union and therefore constitute existing aid.

(124)

Secondly, the Belgian authorities argue that the annual subsidy paid to DPLP dates from before 1999, i.e. before the liberalisation of the postal market and therefore constitutes existing aid. Similarly, the capital injections made before 1999, if they constituted State aid, should not be considered existing aid.

(125)

Thirdly, on account of the 10-year limitation period, the Belgian authorities claim that all measures taken before 13 July 1999 (10 years before the opening of the current formal investigation) constitute existing aid because an earlier CFI annulment would have annulled also all previous investigative acts of the Commission. However, they argue that, because the first investigative act of the Commission took place in December 2002, measures before December 1992 should be left outside the scope of the investigation. This concerns the transfer of buildings and the first annual subsidy paid in 1992.

(126)

Article 14 of Directive 97/67/EC as amended by Directive 2008/6/EC of the European Parliament and of the Council of 20 February 2008 amending Directive 97/67/EC with regard to the full accomplishment of the internal market of Community postal services (39) (the third postal Directive) requires postal operators to have a cost accounting system which clearly distinguishes between services and products that are part of the universal service and those which are not. The accounting system must assign direct costs directly to the products and must allocate common costs directly or indirectly, and in particular allocate appropriately those common costs that are necessary for both universal and non-universal services. DPLP’s cost accounting is an activity-based costing (ABC) methodology, which allocates all operational costs, up to the level of earnings before interest and tax (EBIT), to the various products, in line with the fully distributed costing (FDC) method. The scope of the cost accounting includes all operating costs and all organisational units of DPLP. In 2009 commercial products made up […]% of DPLP’s turnover, while public service products accounted for […]%.

(127)

The cost accounting methodology has three layers: resources, activities and products. The aim is to calculate the share of costs attributable to each product.

(128)

Costs are divided into 398 different types and 3 954 responsibility centres. For accounting purposes, these are further grouped into 34 resource pools and 108 code groups. There are 986 different activities.

(129)

The costs of the resources are allocated to activities depending on their nature: ‘direct resources’ are directly associated with a specific activity, while ‘indirect resources’ contribute to several activities and need to be allocated using allocation keys that reflect the degree to which the resource is used by each activity. Finally the global contribution to the activity includes the overheads that cannot be allocated to specific activities, so they are attributed evenly to all products on a cost-plus basis.

(130)

The 986 different activities constitute the building blocks of each product’s value chain. A share of the costs of an activity is allocated to products in accordance with the usage of the activity in each product. Costs of non-operational activities are first allocated to the operational activities which contribute directly to the products.

(131)

At the last stage of the procedure, the costs of the activities are finally allocated to each product in accordance with an allocation key, which is typically based on volume. In other words, all products that use an activity bear a share of the costs of that activity proportional to their volume of usage of that activity.

(132)

The Commission considers that the cost accounting method currently used by DPLP is sound as it correctly allocates all costs with a sufficient level of adequacy. The Commission notes, however, that, as DPLP has developed the methodology over time, it can be assumed that the accounting data from the early part of the period under investigation, although globally reliable, do not have the same level of detail as the data from the current accounting system.

(133)

In March 2011 the Belgian authorities presented to the Commission a retroactive modification of DPLP’s cost accounting methodology with a view to adjusting the allocation of certain costs between commercial activities and public service activities. The modification contemplated by the Belgian authorities concerns the last stage of cost allocation, when products are charged costs in proportion to their volume of usage of each activity.

(134)

The modification alters the method of allocating operating costs to products that DPLP has used up to now. Under such a revision, some costs linked to commercial products would be shifted to the public service sphere, thereby increasing the cost base of public services. As a result, the public services would be allocated on average EUR […] per year of the costs currently allocated to commercial products. The other parts of the cost accounting methodology would remain unchanged.

(135)

The Belgian authorities claim that the costs of some activities are by their nature fixed at a certain level, which is determined by public service needs. In other words, the costs of such activities used for public service purposes would remain unchanged even if no commercial products used those activities. To the extent that a commercial product has not caused the activity, even if it uses the activity, the product should not bear the operational cost of that activity.

(136)

While the Belgian authorities argue that the revised methodology is compatible with Directive 97/67/EC, the Chronopost judgment (40) and the Community framework for State aid in the form of public service compensation (41) (hereinafter ‘SGEI Framework’) because it allocates to products both directly attributable costs and a share of common costs, and better reflects the real costs of the public service obligations, the Commission considers that there are formal, economic and legal grounds not to accept the new claims.

(137)

From a formal point of view, the proposed methodology is not actually being used in DPLP’s internal accounting, rather it is an ex-post modification of past data. The revised method has not been approved by the regulator as a basis for tariff-setting, nor is it being used by the State to calculate the SGEI compensation. The revised method therefore appears to be a theoretical depiction produced for the purposes of this State aid procedure. From an economic point of view, it cannot be assumed, in terms of internal transfer pricing, that the commercial products would not need to bear any share of the costs of the resources they use (e.g. under this method the commercial products would not contribute to the costs of delivery rounds by postmen even though those postmen also deliver commercial products). Finally, in legal terms, such a change does not appear to comply with Article 14(3)(b)(iv) of Directive 97/67/EC, the Chronopost judgment (42) or the 2005 SGEI Framework (43), which all require commercial activities to be charged an appropriate share of common costs.

(138)

Consequently, the Commission will not rely in its assessment on the modified accounts, but will instead use the cost accounting data validated, in accordance with Directive 97/67/EC, by the Belgian institute of postal and telecoms services (IBPT/BIPT, the Belgian postal regulator (44)), which were submitted previously and are based on the cost accounting method officially used by DPLP.

(139)

Article 107(1) TFEU states: ‘Save as otherwise provided in the Treaties, 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’.

(140)

In order to constitute State aid within the meaning of Article 107(1) TFEU, four cumulative conditions must be met. The measure must:

(141)

Following the approach adopted in the Combus judgment (45), the Belgian authorities claim that the pension liabilities constituted an abnormal burden, which competitors were not required to finance and, therefore, the relief from the State does not constitute aid.

(142)

Their main argument is that, because of the rigidities of the status of statutory personnel, it was not possible for DPLP to reduce the share of this category of personnel among its workforce even though it was much more costly than a hypothetical workforce consisting of employees with less favourable contracts.

(143)

In general, it should first be recalled that the Combus judgment has not been confirmed by the Court of Justice. On the contrary, the Court’s case law contradicts the assumption that compensation for a structural disadvantage would rule out any classification as aid. For instance, the Court has consistently ruled that the existence of aid is to be assessed in relation to the effects and not the causes or objectives of state intervention (46). It has also maintained that the concept of aid covers advantages granted by the public authorities that, in various forms, mitigate the charges normally included in the budget of an undertaking (47).

(144)

It has also indicated that the costs linked to employee pay naturally place a burden on the budgets of undertakings, irrespective of whether or not those costs stem from legal obligations or collective agreements (48). In this connection, it has ruled that State measures aimed at compensating for additional costs cannot exclude them from being qualified as aid (49).

(145)

In addition, it can be observed that the competitive background in which Combus was operating differed from that of DPLP. Combus had to conduct its transport business on a commercial basis and operate under market conditions of competition comparable to those for private bus companies. After a tendering procedure, public transport companies divest their bus transport operations to private and public undertakings. Under the tendering rules, the contracts are awarded to the ‘economically most advantageous bid’, irrespective of the private or public nature of the tenderer. For the whole period under assessment, DPLP had a wide-ranging monopoly, in which economic constraints operated completely differently. In addition, unlike in Combus, in the present case Belgium has not adopted any legislative provision aimed at abolishing or adapting the special employment arrangements for civil servants working for DPLP or the way in which they acquire pension rights.

(146)

Therefore, the Commission considers that the factual differences between the Combus case and the case at issue justify a different conclusion in the present case.

(147)

In order to ascertain whether the measures under scrutiny contain elements of State aid, it needs to be determined whether they confer an economic advantage on DPLP in that they allow it to avoid costs that would normally have had to be borne by its own financial resources and have thus prevented market forces from producing their normal effect. Aid consists in the mitigation of charges normally included in an undertaking’s budget, taking account of the nature or general scheme of the system of charges in question. Conversely, it could potentially be possible to define a concept of ‘special charge’, which would consist in an additional charge over and above those normal charges. The withdrawal of such a special charge by way of a legislative provision would not grant any advantage to the beneficiary and would not, therefore, constitute State aid.

(148)

In line with the Court’s case law on selectivity analysis, which involves a comparison with a reference framework in order to determine whether or not differential treatment of certain undertakings and products is in conformity with the ‘nature or general scheme of the system’, the question as to whether a charge is ‘normal’ or ‘special’ in what is, from a structural viewpoint, a normal market situation has to be based on a reference framework or comparator with a view to identifying undertakings which would be in a legal and factual situation that is comparable in the light of the objective pursued by the measures in question.

(149)

However, it does not appear possible to identify an exogenous comparator that would make it possible to define a ‘normal’ contribution for undertakings in a legal and factual situation comparable to that of DPLP in the light of the objective pursued by the measure under review. In particular DPLP’s competitors are private-law companies operating on competitive markets, whereas DPLP benefited from a statutory monopoly during the period under assessment before full liberalisation in 2011.

(150)

In determining the existence of an advantage within the meaning of Article 107(1) of the Treaty, the Commission should, therefore, examine DPLP’s situation and compare social security contributions before and after the pension relief.

(151)

Following the law of 6 July 1971 and until the pension reform of 1997, DPLP bore all pension costs and social liabilities for its civil servant employees. In these circumstances, the Commission takes the view that these costs are part of the normal costs DPLP had to meet from its own resources, taking into consideration its legal and factual position in the Belgian postal market.

(152)

Since the 1997 pension relief under review relieved DPLP of costs that would normally have had to be financed from its own financial resources, the measures in question confer on the operator an advantage within the meaning of Article 107(1) of the Treaty.

(153)

The advantage in question is selective since it concerns only DPLP.

(154)

The pension relief was financed by the State, which took over DPLP’s pension responsibilities.

(155)

Since the State pays the pension of DPLP retired personnel directly from the State budget, State resources are clearly involved.

(156)

The measures described above are capable of affecting trade between Member States, given that DPLP operates in markets that are open to competition, such as parcel services, and where there is significant cross-border trading. In the parcel market DPLP faces competition from undertakings with activities in other Member States, such as UPS, FedEx and TNT Express. DPLP also operates in the press distribution market, a market in which companies from other Member States are active or can become active. Parcel distribution has never been entirely reserved for DPLP, and other undertakings already provided some parcel delivery services (50) and other specific postal services at the beginning of the 1990s (51). DPLP is also active in the financial services sector (post office current and savings accounts and payment services), where it competes with operators offering financial products, such as banks and financial operators. The post office current and savings accounts, which are used for both payment and savings purposes, are in competition with bank current accounts and savings products. The banking sector was already open to competition before 1992 and has been characterised by extensive trade between Member States. Banks from different Member States were already operating in Belgium before 1992, when there were more than 70 foreign financial institutions in the country (52).

(157)

In the light of the foregoing it is clear that any State measure conferring on DPLP an economic advantage may affect trade between Member States.

(158)

For the reasons explained in recitals 141 to 157, the Commission considers that the pension relief constitutes aid within the meaning of Article 107(1) TFEU.

(159)

Under the 1997 pension reform DPLP’s refund to the State of annual net pension costs was replaced by a contribution in full discharge of liabilities aligning the pension costs borne by DPLP with those of its competitors. Without this reform, the level of these pension costs would, in the subsequent years, have continued to rise significantly, notably as a result of the deterioration in the ratio of civil servants in work (contributors) to those not in work (recipients) following DPLP’s decision to halt recruitment of civil servants.

(160)

As the derogations provided for in Articles 107(2) and 107(3)(a)(b) TFEU clearly do not apply and as Belgium has not invoked Article 106(2) TFEU as justification for the compatibility of the pension relief granted to DPLP, the Commission will examine whether the pension relief can be declared compatible pursuant to Article 107(3)(c) TFEU, which states that aid to facilitate the development of certain economic activities or of certain economic areas may be declared compatible with the common market where it does not adversely affect trading conditions to an extent contrary to the common interest.

(161)

The Commission will first analyse the extent to which the relief of the full pension cost fulfils an objective of common interest and is necessary to that end before investigating its proportionality and carrying out a balancing test of the positive and negative effects of the measure.

(162)

EU law has promoted even greater liberalisation of the postal sector since 1992 with the adoption of the green paper, which later led to the adoption of the three directives: (i) Directive 97/67/EC, (ii) Directive 2002/39/EC of the European Parliament and of the Council of 10 June 2002 amending Directive 97/67/EC with regard to the further opening to competition of Community postal services (53), and (iii) Directive 2008/6/EC. It is also acknowledged that the liberalisation of postal markets at EU level has an important role to play in the Lisbon strategy for growth and employment.

(163)

Therefore, it is important, besides safeguarding the provision of basic postal services as reflected in the concept of the universal service obligation, to create a level playing field between the incumbent, existing competitors and new entrants.

(164)

In the La Poste (France) Decision, the Commission already recognised that creating a level playing field between the incumbent and its competitors with regard to social contributions is a key condition for fair competition in the postal sector. The Commission declared the relief of the obligations imposed on La Poste to finance its civil servants’ pensions compatible with the internal market because the public subsidies guaranteed an equal rate of social contributions by La Poste and its competitors. The Commission intends to apply the same logic in the present case as it applied in chapter 6.3 of that Decision, which must be considered applicable in the present case in so far as it is relevant.

(165)

The Commission considers that the 1992 transformation of DPLP was necessary to promote competition in the postal market because it put DPLP and its competitors on an equal footing as regards legal form (and thereby eliminated the unlimited guarantee that DPLP had previously enjoyed as a public entity).

(166)

However, neither the Belgian State nor the Régie des Postes had accumulated reserves matching the accumulated pension liabilities. When DPLP was incorporated in 1992, it continued to be responsible for pension obligations, but no corresponding assets were transferred to its accounts to cover those obligations.

(167)

As a result, the payment of pensions would gradually have become an unsustainable burden on DPLP, since the retirement age of many civil servants was approaching. Even though measures (improved productivity, gradual personnel reduction as a result of natural wastage, pay indexation) would have been taken to improve this situation, they would not have been sufficient to balance DPLP’s finances in the long term (see Figure 3 below).

(168)

This is in particular because the transformation into a private-law company also meant that DPLP stopped employing civil servants and only hired contractual employees (subject to a general employment contract and labour law requirements). Consequently, after 1997, it would have had to finance from its revenues both the pension expenses for retired civil servants and social contributions linked to pensions for its active workforce of contractual employees.

(169)

DPLP’s cash flow would have turned increasingly negative year after year, as the increasing pension payments would have exceeded revenues generated from its activities (see table below).

Figure 3

Simulations of DPLP’s cash flow conducted by the Belgian authorities in 1996

(170)

It cannot be denied that DPLP would have faced significantly higher costs if it had had to finance pensions for all retired civil servants from its own resources after 1997. It is evident that, without any aid, the change of legal form from public administration to private-law company would have been hampered.

(171)

Furthermore, complete refinancing of pension costs through increased regulated letter revenues would have led to very high letter prices and perhaps substantially decreased the number of letters sent to an extent that would have endangered the provision of the universal letter services.

(172)

The Commission also takes the view that the measures in question are tailored to the objective of common interest pursued. No other instrument could have been more effective. In the sectors with an SGEI, public service compensation could indeed have been granted, but such an approach would not be sustainable over the long term because of the specific structural nature of the problem.

(173)

Considering the objectives of liberalisation of the postal markets, notably safeguarding the provision of high quality and affordable universal postal services, and considering the objective of promoting fair competition, the pension relief constituted a necessary measure.

(174)

The La Poste Decision set out the Commission’s proportionality assessment of aid measures that provide relief from pension costs for incumbents in the postal sector who continue to employ civil servants at terms which were agreed during the monopoly period. The proportionality assessment must be carried out with regard to the establishment of a level playing field on the markets which are open to competition (e.g. mail, parcels, and financial services). The incumbent must be subject to the same rate of social security contributions as its competitors.

(175)

Given that DPLP’s contractual personnel is legally in the same situation as employees of any private company, while the statutory personnel has a status similar to civil servants, a comparison of the social contributions paid by DPLP for each category of employees makes it possible to verify whether, following the implementation of the pension relief measure, DPLP meets the condition of being subject to the same rate of social security contributions as its competitors.

(176)

A detailed comparison of the social security costs paid by DPLP for its statutory personnel and its contractual personnel is summarised in the following table:

(177)

As this table shows, the rate of DPLP’s contributions to the Belgian social security system (nominal percentage of wage costs) is considerably lower than that of private operators ([…] % against […] %).

(178)

However, when the costs directly borne by the employer are included, in order to compare the actual social contribution rates, the two percentages are almost equal and even slightly higher for DPLP statutory personnel ([…] % against […] %).

(179)

It can also be noted that the contributions paid by employees are essentially equal in both cases ([…] % against […] %).

(180)

The lower nominal rate of social security contributions paid by DPLP for statutory personnel is because of the exemption from certain social security contributions that are not applicable to employees with civil servant status (e.g. occupational illness, unemployment, industrial accident). However, DPLP itself directly supports the cost of the exempted contributions on account of the special status of its statutory personnel. This approach results in high direct costs for DPLP.

(181)

The employee contribution, which should also be taken into account as employees have to pay this from the salary paid to them by the employer (therefore it is also a cost for the employer), is slightly higher for statutory personnel than for contractual personnel.

(182)

The Belgian authorities and DPLP have provided information showing the overall stability of direct contributions over time.

(183)

Although the direct costs borne directly by DPLP ([…] %) are not stricto sensu equivalent to social contributions, it should be taken into account that these costs are not voluntary, but instead stem directly from the particular status of the statutory employees and were therefore imposed on DPLP. It seems therefore logical to treat them as equivalent to other compulsory social security costs in terms of their origin and their purpose.

(184)

Consequently, these direct costs should be taken into account when calculating DPLP’s actual social contributions. The resulting total social security contribution rate would be slightly higher for DPLP statutory personnel ([…] %) than for the personnel of private companies ([…] %).

(185)

The reform of the pension arrangements applicable to DPLP was an important stage in adapting the undertaking to progressive liberalisation and with a view to the full liberalisation of the Belgian postal market, which plays an important role in the Lisbon strategy for growth and employment.

(186)

In the La Poste Decision, the Commission stated that the liberalisation of the postal sector could be made more difficult if plans to reform pension schemes (comparable to that implemented in Belgium) were not approved.

(187)

The Commission also considers that the measures in question, even though they do not concern special arrangements for the benefits paid to pensioners, introduce sustainability into a financing mechanism that past developments had rendered obsolete. The Commission considers that the measures are, therefore, to be seen in the more general framework of pension system reform in Member States, which has the support of both the Council and the Commission.

(188)

The negative effects of the measure appear limited since, given DPLP’s history and its activities, it is evident that a large proportion of its pension liabilities relate to activities in the reserved sector, where the distortion of competition is by nature limited.

(189)

Even if the measure in question could, in theory, enable DPLP to retain a dominant position, the Commission regards this as a low risk since the measures simply align the contributions paid by DPLP with those paid by its competitors.

(190)

It follows from the foregoing observations that the negative effects of the aid granted to DPLP have been limited in comparison with the positive effects of the measure.

(191)

Since the measures are limited to what is strictly necessary to establish a level playing field for social security contributions and put an end to a distortion of competition that was a disadvantage for DPLP, they do not affect trading conditions to an extent contrary to the common interest.

(192)

Consequently, the Commission is of the opinion that the 1997 pension waiver can therefore be declared to be compatible aid within the meaning of Article 107(3)(c) TFEU since DPLP is not in a better situation than other operators as regards the social security costs of its statutory personnel.

(193)

In respect of the compensation for public service costs, the Altmark judgment (56) lays down the criteria public measures must fulfil in order to be regarded as compensation for public service obligations and not State aid within the meaning of Article 107(1) TFEU.

(194)

In the present case, these criteria, in particular the fourth criterion, are not fulfilled: DPLP was not chosen by means of a public procurement procedure. Moreover, the Belgian authorities have not proved that the level of compensation was determined on the basis of an analysis of the costs which a typical, well-run undertaking with adequate means would have incurred in discharging the public service obligations entrusted to it, nor is there any indication or information available to the Commission that would suggest that the fourth criterion is met.

(195)

Since the annual compensation benefited only one undertaking and provided additional funds to that undertaking, the measure conferred a selective advantage on DPLP.

(196)

1997 capital injections: the Belgian authorities have acknowledged that the 1997 capital injections of EUR 62 million were in fact a delayed payment of SGEI compensation for 1996. Therefore, they should be assessed on the same basis as the annual compensation. Since not all the Altmark conditions are fulfilled, the measure conferred a selective advantage on DPLP.

(197)

2003 capital injection: the Belgian authorities have claimed that the 2003 capital injection (EUR 297,5 million) complied with the MEIP and therefore did not constitute State aid. According to the Belgian authorities, that capital increase aimed at strengthening DPLP’s equity capital for the construction and upgrade of the sorting centres, the optimisation of mail distribution routes and automation. The funds injected were also intended to cover personnel costs arising from a reduction in the number of employees through an early retirement scheme, a reduction in working time and an increase in the salaries of new postal workers.

(198)

In order to prove that the capital injection complied with the MEIP, the Belgian authorities carried out profitability forecasts as part of their strategic plan, which showed a profitability level above the cost of capital that would be required by a private investor.

(199)

The projections and profit calculation carried out by the Belgian authorities are future-oriented, very detailed and in line with the method that would be used by a private investor:

Figure 4

Ex-ante profit projections for 2003-2007 by the Belgian authorities

[…]

Source: Belgian authorities.

Figure 5

Ex-ante internal rate of return calculation by the Belgian authorities for the most likely scenario

[…]

Source: Belgian authorities.

(200)

It can also be noted that the robustness of the projections has been globally validated by the financial results achieved in subsequent years (see Figure 6). As a consequence of this increase in profitability, DPLP started to distribute dividends from 2007 onwards.

Figure 6

Ex-post profits 2003-2007

[…]

Source: Belgian authorities.

(201)

The performance over a 10-year period is also in line with the 2003 forecasts since the ex-post internal rate of return should amount to […] %, which is above the […] % forecast in 2003.

(202)

Although the profitability of the Belgian State’s investment seems to be sufficient to justify its compliance with the MEIP, it should be noted that the projected revenues used for the assessment of that profitability included State aid and, in particular, the annual compensation granted to DPLP.

(203)

However, these State aid measures have clearly distinct aims and have no chronological link with the 2003 capital injection.

(204)

Moreover, the potential uncertainty for a private investor that may have arisen from the non-notified aid in favour of DPLP was mitigated by the formal authorisation to proceed with the capital injection granted by the Commission in its 2003 decision. At the time the investment was made, any private investor faced with good prospects of a return would have probably taken the risk of making the investment following the adoption of a positive Commission decision, even if that decision was not yet final. A private investor would not have waited several years for a judgment from the General Court and the ECJ before taking an investment decision. Therefore, the subsequent annulment of the Commission decision is an event that was not foreseeable at the time of the investment, and must have appeared very unlikely as Commission decisions are considered legal until they are annulled.

(205)

The Commission therefore takes the view that the 2003 capital injection can be regarded as complying with the MEIP and does not constitute State aid within the meaning of Article 107(1) TFEU.

(206)

2006 capital injection: the Belgian authorities have claimed that the 2006 capital injection (EUR 40 million) was made on MEIP terms and therefore did not constitute State aid.

(207)

The 2006 capital injection was made on the same terms as those applying to a private investor: the increase in DPLP’s capital was subscribed to by the State (EUR 40 million) and by Post Invest Europe (EUR 300 million) at an equivalent price per share (pari passu). At the time, Post Invest Europe was an external investor, which had no connection with DPLP or the Belgian State.

(208)

In addition, the issue price of the shares was set properly. In order to determine the price of the new shares, the investment bank hired by the Belgian authorities estimated DPLP’s total value at between EUR […] and EUR […]. The State finally opted for a valuation of EUR […] for the entire company before the capital injections. The assumptions regarding the level of return and risk based on a joint business plan were guided as such by considerations of long-term profit optimisation. DPLP’s subsequent performance following the 2006 capital increase (see below) also confirms the robustness of the assumptions and parameters taken into account at the time of the investment decision.

Figure 7

DPLP’s performance from 2006 to 2010

[…]

Source: Belgian authorities.

(209)

The Commission considers that the 2006 capital injection of EUR 40 million clearly corresponds to the behaviour of a market economy investor and therefore does not constitute State aid.

(210)

The corporate tax exemption, while likely to confer an advantage on DPLP by possibly reducing the amount of tax it would otherwise have paid, did not produce any concrete advantage because, from 1992 to 2005, when the tax exemption was abolished, DPLP’s cumulated result after tax was negative. While DPLP’s net result in some years was positive, the tax credit carried over from one year to the next would have resulted in no tax payments even if DPLP had been liable for corporate tax. Consequently, this exemption is not examined in any more detail in this Decision.

(211)

The property tax and local tax exemptions, by contrast, place DPLP in a more favourable position than that of other undertakings by relieving it of obligations it would have normally have had to bear. These tax exemptions therefore confer a selective advantage on DPLP.

(212)

The Belgian authorities have argued that the transfer of buildings was part of an internal reorganisation of State property since at the time DPLP was fully State-owned. However, the transfer of buildings made available to DPLP new assets which had previously belonged to the State. DPLP had already been using the buildings while they belonged to the State, but was not the legal owner of them. The transfer of the legal title of ownership gave DPLP, an undertaking separate from the State, property rights over the buildings. DPLP therefore obtained additional resources compared with the situation before the transfer. The measure therefore conferred a selective advantage on DPLP.

(213)

Without the State guarantee, DPLP could not have obtained a loan on the same terms. The Belgian authorities have estimated that market rates would have been up to 40 basis points higher, whereas the premium paid to the State was 25 basis points.

(214)

In accordance with the Commission communication on the application of Articles 87 and 88 of the EC Treaty to State aid in the form of guarantees (57) (paragraph 3.2(d)), ‘risk-carrying should normally be remunerated by an appropriate premium on the guaranteed or counter-guaranteed amount. When the price paid for the guarantee is at least as high as the corresponding guarantee premium benchmark that can be found on the financial markets, the guarantee does not contain aid. If no corresponding guarantee premium benchmark can be found on the financial markets, the total financial cost of the guaranteed loan, including the interest rate of the loan and the guarantee premium, has to be compared to the market price of a similar non-guaranteed loan’. As the premium does not appear to be a normal market rate, the State guarantee allowed DPLP to obtain funding at a cost below what it would have normally had to bear.

(215)

The measure therefore conferred a selective advantage on DPLP, which amounts to the difference between the interest rate DPLP would have had to pay in the absence of the State guarantee and the actual interest rate plus the premium paid for the guarantee. Therefore the advantage obtained was 15 basis points on the capital outstanding each year. The advantage can therefore be estimated at EUR 1,5 million over the duration of the loan.

(216)

The annual SGEI compensation was financed from the national budget, as specified in the successive management contracts between DPLP and the Belgian State. According to the information provided by Belgium, the 1997 capital injections are de facto SGEI compensation and are therefore included in this part of the Decision.

(217)

The tax exemptions have decreased the State’s tax revenue by reducing the income it would have received in the absence of the tax exemptions. They involve, therefore, a transfer of State resources in the form of a loss of tax revenue for the Belgian State.

(218)

The buildings transferred to DPLP were previously State property. The State surrendered ownership to DPLP, thereby reducing its own assets and increasing those of DPLP. The measure therefore constitutes a transfer of State resources.

(219)

The loan guarantee was granted by the State, which assumed the associated risks from its own budget without a consideration, and therefore constitutes a State resource in the form of foregone revenue.

(220)

Revenues from mail distribution services do not constitute State resources following the Preussen Elektra judgment (58) since the prices paid by users are financed from their own funds, which are outside the State’s control.

(221)

Nevertheless, for the purposes of the compatibility assessment, these revenues are directly linked to the provision of a service of general interest in the postal sector and must therefore be taken into account in accordance with point 17 of the SIEG Framework and Article 7 of Directive 97/67/EC for determining the necessity and proportionality of the compensation.

(222)

The measures are capable of affecting trade between Member States for the reasons indicated in recital 156 above.

(223)

For the reasons explained in recitals 193 to 222, the following measures constitute State aid within the meaning of Article 107(1) TFEU:

(224)

Some of the measures under assessment may be considered existing aid within the meaning of Article 108(1) TFEU.

(225)

The exemptions from corporate tax and property tax benefited DPLP when it was still a State administration before 1971 and predate Belgium’s EEC membership. They have not been modified since. These measures therefore constitute existing aid.

(226)

The corporate tax exemption was abolished in 2005, whereas the property tax exemption still applies.

(227)

The Belgian authorities have argued that all the measures dating from before the liberalisation of postal markets in 1999 should be considered existing aid. However, the measures in questions benefited all the activities of DPLP, which was active in competitive markets (such as parcel distribution and financial services) well before the entry into force of Directive 97/67/EC in 1999. Consequently, according to the Alzetta judgment (59), such aid cannot be considered existing aid because of the evolution of the internal market: it can be considered existing aid to a given beneficiary to the extent that the beneficiary was active solely on markets not open to competition when the aid was put into effect.

(228)

This claim cannot therefore be accepted by the Commission.

(229)

The claim that the 10-year limitation period should be counted from the opening of the current formal investigation in 2009 cannot be accepted either. The General Court’s 2009 judgment did not annul all the preceding investigative acts, but only implied that the Commission should have opened a formal investigation as a result of them (60).

(230)

By contrast, as the Commission’s first investigative act occurred on 23 December 2002, when it sent its first request for information to the Belgian authorities, the measures put into effect before 23 December 1992 are to be considered existing aid.

(231)

The compatibility of the SGEI compensation is assessed under Article 106(2) TFEU and the SGEI Framework. The SGEI Framework is applicable as of its entry into force on 29 November 2005. Since the Framework is largely a codification of rules that already existed before it entered into force, it will also be referred to when assessing the measures that precede its entry into force.

(232)

In their submissions, the Belgian authorities and DPLP strongly opposed the Commission carrying out an overall analysis of all aid granted to DPLP over the entire period, in line with the approach it adopted previously for assessing other cases where aid was illegally paid in the past, as it assumes that the aid financed without distinction all the public service missions entrusted to the undertaking. The Belgian authorities and DPLP have contended that in the case at issue, unlike in others, the management contracts they entered into strictly delimit the scope of DPLP’s aided activities and the amounts of aid for which they were eligible (61).

(233)

The Commission considers that the relevant management contracts rigorously define the respective rights and obligations of the State and the aid beneficiary by setting the parameters for calculating the compensation to which the beneficiary is entitled for the specific activities listed and for the periods of time indicated. Accordingly, the Commission will base its assessment on the provisions laid down in the various management contracts.

(234)

The SGEIs for which the compatibility of aid is to be checked are defined in the acts of entrustment established by the Member State. The public service missions to be discharged by DPLP are entrusted to it in the successive administrative acts approving the management contracts signed between DPLP and the State approximately every five years. These contracts develop and complement the provisions of national postal law.

(235)

Each successive management contract constitutes a separate act of entrustment, which describes the public missions to be discharged by DPLP, fixes the duration of the entrustment and the parameters for calculating the compensation to be paid by the State. The management contracts specify the activities included in the calculation of the amount of financial compensation granted by the State.

(236)

The compatibility of the aid measures must be considered in the light of Article 106(2) TFEU since the compensation was granted to finance the costs incurred in discharging the public service missions. The compensation is compatible to the extent that it does not overcompensate DPLP for the extra costs incurred in carrying out the public service missions entrusted to it under the management contracts, including a reasonable profit. Any overcompensation will have to be recovered.

(237)

The scope of the assessment of overcompensation includes those SGEIs for which compensation is provided for in the clauses of the successive management contracts.

(238)

It is important to note that the provisions of the successive management contracts differ to some extent. Consequently, overcompensation must be calculated on the basis of the specific parameters provided for in each contract.

(239)

In the first management contract (1992-1996), the SGEI compensation covers all public services. The provisions of this contract do not apportion the compensation between the different public services according to whether or not they are eligible for compensation under the provisions of the contract.

(240)

Accordingly, the compensation provided for in the contract has to be understood to cover the additional costs of discharging all the public service missions without distinction. Checking potential overcompensation in this period will therefore encompass all the public service measures: the universal postal service (USO), press distribution and other missions entrusted to DPLP.

(241)

By contrast, under the clauses of the second, third and fourth management contracts, the aided activities are limited to certain SGEI missions specified therein. The SGEI missions entitled to be compensated are: the distribution of newspapers and periodicals, and other public services listed in recital 33. The clauses determining the scope of the aided activities are contained in the management contracts: Article 7 of the second management contract, Article 15 of the third management contract, and Articles 12 and 13 of the fourth management contract.

(242)

The universal postal service, by contrast, does not receive any compensation under the management contracts, with the exception of the first management contract.

(243)

The costs and revenues of the universal postal service are therefore not included in the calculation of overcompensation, except, where relevant, for the excess profits from the reserved area (62).

(244)

On account of the differences in the public service missions entrusted, as referred to in recitals 239-243, the potential overcompensation is calculated differently for the first management contract from the three subsequent management contracts.

(245)

In addition to the coverage of costs incurred in discharging the public service obligations, the SGEI Framework also allows for compensation for a reasonable rate of return (63).

(246)

Paragraph 18 of the SGEI Framework specifies that the reasonable profit ‘should be taken to mean a rate of return on own capital that takes account of the risk, or absence of risk, incurred by the undertaking by virtue of the intervention by the Member State, particularly if the latter grants exclusive or special rights. This rate must normally not exceed the average rate for the sector concerned in recent years. In sectors where there is no undertaking comparable to the undertaking entrusted with the operation of the service of general economic interest, a comparison may be made with undertakings situated in other Member States, or if necessary, in other sectors, provided that the particular characteristics of each sector are taken into account. In determining what amounts to a reasonable profit, the Member State may introduce incentive criteria relating, among other things, to the quality of service provided and gains in productive efficiency.’

(247)

The appropriate level of reasonable profit is thus to be chosen taking into account a number of parameters, in particular the average rate for the sector concerned, the risk associated with the undertaking’s activities and, where applicable, incentive criteria relating to the company concerned.

(248)

In order to establish a level of reasonable profit, the Commission has based itself on several expert studies, which are set out in more detail below. Based on the studies and on the level of risk of each management contract, the Commission has determined a range of reasonable levels of profit for DPLP for mail delivery, on the one hand, and press distribution, on the other, for each of the four management contracts. These studies all attempt to identify sets of companies which are sufficiently comparable in terms of activities and risk profile to the postal incumbent in question and to establish a benchmark profit level for the latter on the basis of the observed levels of profit of the comparator firms.

(249)

As regards risk in particular, the Commission wishes to underline that DPLP, during the entire period under investigation, has faced little or no competition in the vast majority of its operations since it either enjoyed a legally protected monopoly (this applies to most mail services), or a dominant position characterised by very high market shares and barriers to entry through the subsidisation of below-cost prices (e.g. press distribution). In the Commission’s opinion, this aspect has to be borne in mind in all analyses involving benchmarking with companies that, while providing similar services to DPLP, had to operate in a competitive environment.

(250)

Furthermore, the extent to which a postal incumbent is exposed to risk depends essentially on the characteristics of the regulatory mechanism established by the public authority.

(251)

In the context of mail delivery, a sector in which DPLP’s USO was mainly financed through stamp prices (64), the Commission considers that it makes a substantial difference in terms of risk whether the stamp prices were, to a greater or lesser degree, frequently adjusted by the regulator (e.g. at the start of each year or simply whenever necessary) so as to cover the net costs of the USO on an ongoing basis, or whether stamp prices (65) were fixed in advance for a long period (e.g. four to five years) as part of a long-term contract (66). The first type of price regulation tends to be rather low risk for the firm in question, in so far as stamp prices are adjusted on a relatively frequent basis so as to maintain a constant financial equilibrium between revenues and costs. This type of regulation protects the company’s profit margins against uncertainty. The second type involves more risk as it exposes the operator to the consequences of variations in costs and revenues over a long period.

(252)

Likewise, in the context of press distribution activities (USO financed through financial compensation), it makes a substantial difference whether the compensation is based on actual net costs (i.e. net costs actually incurred) or whether it is based on projected net costs over a long period. For instance, if the compensation takes the form of a fixed lump-sum payment covering expected net costs over a long period, it leaves the operator exposed to all variations in revenues and costs. In such a case, the operator faces a higher degree of risk than when it is compensated in full on the basis of ex-post costs and revenues.

(253)

Where the contract between the operator and the public authority is such that the public service is connected with a certain degree of risk, for instance because it leaves the incumbent exposed to significant variations in revenues and costs, the reasonable profit allowed to the incumbent should be commensurate with the corresponding level of risk involved. By contrast, where the contract between the incumbent and the public authority is such that the public service is not connected with substantial risk, the reasonable profit allowed should be low as well.

(254)

The Commission would also point out that the degree of risk incurred by the company and the extent to which the company is entitled to compensation for efficiency gains are closely intertwined. For instance, a public authority can set in advance a fixed level of compensation which anticipates and incorporates the efficiency gains that the undertaking can be expected to make over the lifetime of the entrustment act. By exposing the company to all variations in costs and revenues, the contract gives it incentives to improve performance and become more cost efficient. At the same time, it results in more risk for the company. To some extent, therefore, risk and efficiency incentives go hand in hand. Likewise, an ex-ante price cap over a long period normally provides stronger incentives for efficiency improvements, but is also more risky for the company involved.

(255)

In order to address the question of a reasonable rate of profit for DPLP, the Commission bases itself on several expert studies:

(256)

WIK study: in the context of Case C-36/07, a State aid procedure relating to the German postal operator, Deutsche Post AG (hereinafter ‘DP’), the Commission asked WIK Consult to carry out a study on the appropriate level of reasonable profit for postal incumbents (67) (hereinafter ‘WIK study’). The relevant aspects of the WIK study were shared with Belgium on 25 March 2010 and 7 February 2011. The Commission received comments from Belgium on the WIK study on 23 April 2010 and 1 March 2011.

(257)

Deloitte study: on 23 April 2010, the Commission received a study by Deloitte (68) (hereinafter ‘Deloitte study’) setting out its views on the level of reasonable profit to be used for DP’s universal services. This study was sent to Belgium on 9 December 2011. The Commission received comments from Belgium on the Deloitte study on 26 December 2011.

(258)

CRA study: on 1 March 2011, Belgium submitted an expert study in response to the WIK study by Charles Rivers Associates (69) (hereinafter ‘CRA’).

(259)

In this study, WIK establishes a reasonable profit benchmark for DP. Specifically, WIK first performs a function and risk analysis of DP and concludes that DP predominantly provides routine functions and faces limited entrepreneurial risk (70). On the basis of this analysis, WIK selects appropriate comparators, i.e. determines the composition of the benchmark groups from which the reasonable profit benchmark is derived. Overall, WIK points out that, conceptually, the adopted benchmark procedure follows the standard methodology used for profitability benchmarks in transfer pricing (71). Such a methodology is applied not only in national transfer pricing regulations but also in the OECD Transfer Pricing Guidelines (hereinafter ‘OECD Guidelines’) (72).

(260)

WIK considers several measures of profitability, distinguishing between the internal rate of return (IRR), which relies on cash flows and is deemed the conceptually correct measure of the profitability of an activity, and other measures derived from accounting data (rates of return), such as the return on capital employed (ROCE), the return on equity (ROE), the return on assets (ROA) and the return on sales (ROS) (73). As set out in Chapter 5 of the study, WIK proposes using the ROS as a profitability indicator. One advantage of this measure is that it only depends on accounting profit (EBIT (74)) and sales data, which are both easily observable in the company’s accounts. In addition, the use of the ROS avoids the valuation and attribution of assets between different services, which would be necessary for a capital-based benchmark, such as the ROE and the ROCE, and which is normally a difficult exercise. WIK also points out that a benchmark based on the ROS is advisable if the comparator companies and the benchmarked company belong to the same branch/industry sector so as to ensure that the structure of EBIT and sales is homogeneous.

(261)

Regarding the composition of the benchmark group, WIK concludes that there is a trade-off between the comparability of companies in the sample and the size of the sample. WIK argues against including European postal incumbents in the sample because their prices have been regulated in a market which has been foreclosed to competition in most countries over the entire period under investigation. For the same reason, the number of companies in direct competition with the postal incumbents is rather limited. Mostly, these are smaller companies. In order to obtain a balanced set of comparable companies, WIK includes both the direct competitors on national markets and big multinational players, such as large international express parcel companies (e.g. UPS, FedEx, DHL), which run a more capital-intensive and riskier business than a postal incumbent on a protected national market.

(262)

WIK addresses the trade-off between comparability and sample size by considering three different benchmark groups: (i) benchmark group I, or the narrow parcel sector, including companies that are primarily active as parcel service providers (10 companies, considered highly comparable with universal postal service providers); (ii) benchmark group II, or the broader parcel sector, including companies whose main activity might focus on other services, such as courier services (26 companies with a somewhat lower level of comparability); and (iii) benchmark group III, including also companies from the land transportation sector (1 163 companies with lower comparability) (75). This approach, according to WIK, reinforces the statistical validity of the final benchmark.

(263)

One of the selection criteria applied to all benchmark groups is the size of the company; only companies with revenues above EUR 3 million are considered (76). As regards the time span of the data, reliable data were available only for the period 1998-2007, while backward interpolation was used for the period 1990-1997 (77).

(264)

Furthermore, WIK follows the analysis adopted in transfer pricing by relying on the concept of the interquartile range (78) to reduce the effect of outliers and to establish a level of reasonable profit.

(265)

WIK obtains the following results for the period 1998-2007:

(266)

WIK concludes that the broader parcel sector, or benchmark group II, is the best benchmark because the results obtained are relatively stable over time compared with the alternative benchmark (narrow parcel sector) and because this sector is functionally closer to letter and parcel activities than the other alternative benchmark (land transport sector).

(267)

A further adjustment is made to obtain an upper limit for the reasonable profit for DP on the basis of the analysis of DP’s entrepreneurial risk relative to the benchmark group. On this basis WIK proposes the median benchmark II ROS of 3,48 % as an upper limit for a reasonable profit for DP. WIK chooses the median instead of the 75th percentile as the upper value of reasonable profit to reflect the fact that the companies in the sample bear higher risk than postal incumbents since they do not benefit from special rights.

(268)

The Belgian authorities considered the methodology used in the WIK report to be flawed and submitted several documents (letters, position papers and reports) on 23 April 2010, 1 March 2011, 14 June 2011, 5 August 2011, 10 August 2011, 14 December 2011 and 19 December 2011 which refute the approach of the WIK report and present their own method for estimating a reasonable profit benchmark.

(269)

In a position paper sent on 5 August 2011, the Belgian authorities essentially argued that the sample used in the WIK study is not representative in terms of activity, size, risk profile and reference period. Belgium therefore claims that the sample is very hard to legitimise in the light of the criteria set out in paragraph 18 of the SGEI Framework.

(270)

In the same position paper, Belgium also argues that it makes sense to differentiate between the profit margins for activities which present very different levels of risk: if press distribution and letters are activities which have a different risk profile, the Belgian authorities argue that different levels of reasonable profit should be used.

(271)

In addition, the Belgian authorities claim that DPLP’s operations benefit from productivity gains and that this should be considered by the Commission when establishing the rate of reasonable return, as set out in paragraph 18 of the SGEI Framework. The Belgian authorities point out that DPLP has become profitable as a result of numerous reforms, after having been historically loss-making. DPLP’s costs have been cut, in particular by closing post offices. The 50 % share held by private investors since 2006 has created a new culture geared towards efficiency. According to the Belgian authorities such efficiency gains should be taken into account in determining the reasonable rate of return that can be granted to DPLP.

(272)

On 1 March 2011, the Belgian authorities sent the Commission a report prepared by Charles River Associates (CRA), which critically evaluates the WIK study and proposes an alternative method for benchmarking the reasonable profit of DPLP.

(273)

Regarding the sample composition, CRA argues that all three WIK samples include many small firms whose risk profiles are not comparable to the risk profile of a large postal operator. Thus, CRA criticises WIK’s conclusions that the provision of letter services by the incumbent postal operator in question is low risk and that the companies included in the benchmark groups are good comparators.

(274)

In particular, CRA points to the relationship between a firm’s cost structure (in terms of the proportion of fixed and variable costs) and the business risk it faces. CRA argues that a postal operator entrusted with a USO has larger fixed costs and, therefore, faces a relatively higher risk of a downturn in demand compared with a company with lower fixed costs. In addition, CRA points to the difference in the ROS between the small and big companies in the WIK groups (79), which suggests, in CRA’s view, the existence of systematic differences between the small parcel firms and the larger firms, such as UPS and FedEx, which might be explained by the different cost structure.

(275)

CRA also analyses in more detail the characteristics of some of the small parcel firms included in the WIK Benchmark I sample and concludes that their risk profile is not comparable to that of a large postal operator. CRA also considers that increasing the size of this already barely comparable sample by adding even less comparable companies is not appropriate.

(276)

CRA dismisses WIK’s argument that European postal incumbents should not be included in the sample because their profits have been constrained by price regulation, rather than by competition, leading to the risk that price regulation has allowed these incumbents to earn excessive profits. CRA argues that in dismissing other postal incumbents as good comparators, WIK relies on the theoretical possibility that these incumbents might have been allowed to charge excessive prices without, however, supporting its argument with empirical evidence. In this regard, CRA points out that the type of price regulation mechanisms applied in a number of Member States, such as price caps (80), would suggest that the high profits observed are the result of productivity gains rather than high prices.

(277)

CRA argues that the natural method for calculating a reasonable profit margin for DPLP from its public service obligations would be based on the available data on profit margins for the letter operations of other western European postal operators. This method is all the more natural because the public services at issue are primarily letter services and the data on profit margins from the letter-related operations of other western European postal operators are readily available and do not suggest too lenient price regulation.

(278)

Specifically, CRA argues that the analysis should be based on the EBIT margins (ROS) realised on letter operations by western European postal operators that report segment data showing EBIT margins for their letter operations or (if segment data on EBIT margins are not reported) the company-wide EBIT margins of western European postal operators that derive more than 75 % of their revenue from letter services. CRA excludes from its sample eastern European operators on the basis of their relative inefficiency compared with postal operators in western Europe.

(279)

As a result, CRA bases its analysis on a sample of 11 western European postal operators during the period 2002-2009 (73 observations in total). The mean ROS in this sample is […] %, the median is […] % and the 75th percentile is […] % (81). CRA also calculates the trimmed mean ([…] %) and the Winsorised mean ([…] %), which, like the median, are not affected by extreme values (82).

(280)

As regards establishing a benchmark for the reasonable profit margin, CRA argues that there are methodological problems with relying on average values. CRA explains that because of variations in observed profit margins in the sample, the critical threshold for establishing excessive profitability should be higher than the mean or any other measure of central tendency and could, for example, equal the 75th percentile. Such an approach, according to CRA, would be consistent with the practice in transfer pricing investigations of relying on a range (for example, an interquartile range or other percentiles).

(281)

On 14 Dec 2011, the Belgian authorities presented to the Commission a document summarising DPLP’s views on the risks faced by postal operators and proposing a methodology with a view to developing a tool to measure the risk premium which, together with the risk-free return on sales, would provide a reasonable ROS rate.

(282)

The document does not explain how the risk-free return on sales (expressed in terms of ROS) could be determined in the present case and does not provide a clear view of what precisely the reasonable ROS should be on the basis of the proposed methodology. Instead, the analysis focuses on the assessment of the various risks and the calculation of a risk premium that reflects all types of risks.

(283)

In particular, the Belgian authorities focus on three broad risk categories – sectoral risk, service-related risk and contractual risk – and on the typical types of risks in each category to which postal operators are exposed. The methodology for the assessment of the risk premium provides for: (i) the definition of a ‘maximum risk premium’ (83); (ii) the splitting of the relevant period into contract periods to capture the risks related to each contract; (iii) the use of a simple scale to grade each individual risk for each contract period (as low, medium or high); (iv) the conversion of the individual risk scores into an overall score for each of the three broad risk categories; and (v) the conversion of the overall scores into risk premiums, the sum of which gives the total risk premium.

(284)

Moreover, examples are provided to illustrate that DPLP faces above-average risks. Thus, the Belgian authorities maintain that DPLP faces higher volume risk as it derives a higher proportion of its revenue from its mail segment and is characterised by higher cost rigidity on account of a higher-than-average proportion of fixed costs. Another type of risk which is considered higher for DPLP is security risk. According to the analysis, this is because DPLP is one of the few European postal operators to deliver cash payments of pensions to households.

(285)

On the basis of the assessment of the different risks, the Belgian authorities conclude that the fourth management contract shows a high level of all three risk types, while the first, second and third contracts are characterised by medium-level risk (the only exception being sectoral risk, which is considered low for the first contract). The Belgian authorities provide further examples of how the above ranking of the different risks can be converted into a risk premium on the basis of different values for the ‘maximum risk premium’.

(286)

The Deloitte study referred to in recital 257 above is consistent with the WIK study on several points:

(287)

As regards the set of comparable companies, the main difference between the WIK and Deloitte studies is the size of the companies considered. Deloitte only includes in its sample companies with annual revenue higher than EUR 100 million (84). This restriction is based on Deloitte’s analysis of the particular characteristics of the activities related to universal service provision and the associated business and market risks (85).

(288)

In particular, Deloitte considers that companies in the parcel and postal sector that operate large, complex collection and delivery networks can be reasonably compared to DP. The economies of scale, scope and density prevalent in postal networks and the high proportion of operating costs relative to capital costs, as well as the associated risks (for example, volume risk, risk of e-substitution), mean that large multi-product businesses in the postal and parcel sector are more relevant comparators than small mail businesses. Freight transport and logistics companies would also be – to a more limited degree – suitable comparators because they employ similar assets.

(289)

In order to illustrate the importance of company size, Deloitte compares its sample with that of WIK benchmark group II. Deloitte illustrates that the average ROS in the WIK sample differs significantly between groups of companies of different size (proxied by the number of employees) (86). Deloitte argues that in its own sample (based on larger firms) there is, on the contrary, no strong relationship between scale and ROS.

(290)

As a result of the sample selection process, Deloitte bases its analysis on seven postal and parcels companies, 18 logistics companies and 19 freight transport companies. As regards the postal and parcel sector, the small set of seven comparable companies includes big multinational express parcel operators, such as UPS, FedEx, and TNT Express, as well as a few other postal operators, mostly from the United States and the United Kingdom (87). In principle the data set covers the period 1990-2007, however data for the early years is very limited and, therefore, Deloitte considers the results for these years less robust (88).

(291)

Deloitte also examines the advantages and disadvantages of several alternative profitability measures. It dismisses the use of the IRR in the case in question because of the practical difficulties in obtaining adequate cash-flow data. It also rejects the use of the ROCE or the ROE because of, for example, differences among companies in the accounting treatment of costs and in the capital structure (the latter being important for the ROE). ROA is also deemed inappropriate because it focuses on total assets. Thus, Deloitte proposes to use the ROS but emphasises that, as the ROS is not a direct measure of return on capital, it can be relied upon for benchmarking purposes as long as the companies compared have similar capital intensity.

(292)

Deloitte nevertheless calculates the ROCE (as well as the market-value ROCE and the ROA) for a small subsample of six companies, but concludes that the range and variation of the ROCE might reflect a number of factors, which requires cautious interpretation and undermines its usefulness as a benchmark. Among these factors Deloitte emphasises the importance of capital intensity. Deloitte argues that, in terms of capital intensity, DP’s universal service is more similar to that of, for instance, UPS and Fedex than to that of freight operators like Kuehne&Nagel or Wincanton.

(293)

Deloitte further highlights the significant variation in companies’ profitability over time, as well as within and between the different sectors. This variation, in Deloitte’s view, would warrant the use of ranges of reasonable profit rather than a single average figure. Such an approach would also be appropriate because of the need to avoid penalising above-average performance as a result of efficiency gains.

(294)

The results obtained by Deloitte, by sector and for all sectors together, are summarised in Table 22 of the Deloitte study. Thus,

(295)

As a reference benchmark Deloitte proposes a value based on the average annual mean ROS of 7,9 % (90); to reflect the observed variation of the ROS, Deloitte further derives a range of reasonable ROS based on the interquartile annual ranges (91).

(296)

Choosing a good sample of comparable companies for postal incumbents is a complex task. It goes without saying that care must be taken to properly select a suitable group of companies as a benchmark in terms of business activities and risk.

(297)

The WIK and Deloitte studies both focus on commercial companies with activities comparable to those of DPLP, while excluding postal incumbents from the comparator sample. CRA, by contrast, uses a comparator sample that consists only of other European postal incumbents.

(298)

The Commission shares the views of both WIK and Deloitte that the profits of postal incumbents (which CRA uses for its benchmarking) are not determined under normal market conditions, but largely result from the regulatory choices of other Member States. As such, they may not adequately reflect a proper return benchmark. At the same time, given that the CRA sample focuses on postal incumbents in other European countries, it appears to constitute the sample of companies most similar to DPLP. For this reason, the Commission believes it is appropriate to take the CRA sample fully into account, in particular in order to cross-check the benchmark reasonable profit levels (ROS) established on the basis of the WIK and Deloitte studies.

(299)

The WIK and Deloitte studies both acknowledge the need to have a sufficiently large sample to draw relevant conclusions for the purpose of benchmarking. For this purpose, they consider both a (relatively) narrow sample of companies in the postal and parcels sector and wider samples including companies in the logistics and freight transport sectors. The Commission recognises that a sufficient sample size is important for statistical purposes. In relation to CRA’s position in this regard, which questions the usefulness of increasing the sample size by adding less comparable companies, the Commission takes the view that the wider samples identified by WIK and, in particular, Deloitte on the basis of the adopted sample selection methodologies are sufficiently comparable to serve as meaningful benchmark groups for benchmarking DPLP’s profit.

(300)

The WIK study attempts to establish a benchmark ROS on the basis of three different comparator groups. The Commission is also of the opinion, however, that the WIK study can be criticised on two distinct, yet related, issues (92): (i) WIK’s use of samples that also include a large number of smaller companies (93) which are clearly not comparable to incumbent postal operators such as DPLP; and (ii) WIK’s assumptions regarding the level of risk faced by the postal operator under investigation.

(301)

The Commission considers that the scale of operations, as well as the cost structure of postal incumbents such as DPLP (in conjunction with the type of regulatory mechanism in place), are essential for assessing the benchmarking of profit levels.

(302)

First, in the presence of significant fixed costs (related, for example, to capital costs or other fixed costs linked to the network), the largest companies benefit more from economies of scale and density than small ones. By comparing the average ROS for large and small companies in WIK’s sample, the Deloitte II study shows convincingly that the average ROS increases with the size of the companies (measured by the number of employees) (94). Indeed, the average first quartile value (Q1) is very low in the three WIK samples (e.g. only 0,75 % for the WIK sample II over the period 1998-2007) (95), so that it would appear that there is a certain over-representation of the (smallest) firms that are loss making and struggling to stay in the market. By including so many small firms in its samples (including in its main WIK II sample), WIK may have underestimated the benchmark ROS that would be more relevant for larger and better-established firms such as DP and DPLP.

(303)

Second, while WIK correctly refers to the operator’s low risk profile on account of the monopoly character of its operations, it does not appear to have paid sufficient attention to the significant fixed costs (relative to variable costs) arising from the USO, nor to changes in the risk profile that may have resulted from changes in the regulatory regime faced by the operator over time, particularly between the different regulatory periods (e.g. the different management contracts, in the case of DPLP).

(304)

The proportion of fixed costs may be high not only because a firm is very capital intensive (a common reason why fixed costs may be high), but also because it cannot adapt its overall costs in view of the obligation to maintain a large distribution network (related to the USO) and because it is unable to reduce labour costs in view of existing labour contracts. The high proportion of fixed costs (96) further calls into question the inclusion of small companies whose cost structure (ratio of fixed to variable costs) appears to differ significantly (97).

(305)

As set out in the introduction, the Commission is also of the opinion that the extent to which a postal incumbent is exposed to risk depends to a large extent on the characteristics of the regulatory mechanism established by the public authority to finance the USO. In particular, the transition to a longer-term price cap for stamp prices (as experienced by DPLP in 2006) entails a significant increase in the size of the volume risk. The existence of such risk weakens WIK’s assumption about the inherently low-risk nature of DPLP’s activities, at least for the period covered by a price cap.

(306)

Given the considerations in recitals 286 to 295, the Commission believes that, for the purpose of benchmarking DPLP’s profit, the Deloitte samples improve upon WIK’s samples in so far as they are limited to companies with a bigger minimum size (annual turnover exceeding EUR 100 million).

(307)

At the same time, it should be noted that, in some respects, the Deloitte approach has certain potential weaknesses and its conclusions should be interpreted cautiously.

(308)

In particular, the Deloitte samples appear to comprise many firms with a capital intensity higher than that which applies to mail operators (98). From a methodological point of view, the use of ROS as a tool for profit benchmarking requires a comparable sample of firms with similar levels of capital intensity (defined here as the capital to sales ratio). The working hypothesis is that companies with the same ROS will have a return on capital employed (ROCE (99)) that is inversely proportional to their capital to sales ratio. Accordingly, the use of ROS as a benchmark indicator of profitability requires identical capital intensity for the companies in the benchmark group. Although reliable estimates of capital intensity are difficult to obtain (100), it would appear that the mail and press distribution activities are less capital intensive than the express parcel delivery activities of big multinational companies such as UPS, TNT, FedEx and DHL (101). Consequently, the reasonable profit indicator established on the basis of the Deloitte samples may overstate the benchmark profit that would be appropriate for companies with the same capital structure as DPLP (102).

(309)

In view of the numerous uncertainties of this benchmarking exercise, the Commission deems it justifiable to establish a reasonable range of ROS on the basis of Deloitte’s estimations of the median ROS for the wider all-sectors sample (5,4 %) as well as the median ROS for the sample of companies active in the postal and parcel sector (7,4 %) (103). The resulting range [5,4-7,4 %] would, in the Commission’s view, constitute a ‘default range’ for reasonable profit, at least for those periods in which DPLP faced a significant degree of risk, in the light of the compensation contract in force, and strong incentives to become more efficient (104).

(310)

The Commission considers the range of [5,4-7,4 %] suitable as a default, both for DPLP’s mail delivery activities and for its press distribution activities, given that these are closely related and have very similar business characteristics.

(311)

It is true that DPLP has faced little or no competition in the areas concerned (mail and press distribution). On this basis, the question arises whether it is in fact appropriate to benchmark its profit with the median profit of samples of companies actually facing competition and whether a more restrictive (i.e. lower) benchmark, e.g. the first quartile (25th percentile), would not be sufficient for a company in DPLP’s position. However, in the Commission’s view, such an approach would not take sufficient account of the significant variation in profit levels, both across firms and over time. Furthermore, removing possible efficiency gains by imposing such a strict cap on the ex-post ROS achieved by the company should be avoided.

(312)

In the case at hand, the Commission also considers it inappropriate to base itself on measures that go beyond the median, such as the third quartile (105). As the approach chosen already establishes a range across different samples, the value added of also considering, for statistical purposes, ranges around the median within the individual samples (the interquartile range commonly used in transfer pricing) appears somewhat limited (106). The Commission therefore takes into account the variation in the median across the different benchmark groups, rather than the variation around the median within a particular sample (107).

(313)

More importantly, however, given that DPLP’s mail segment is less capital intensive than that of the express parcel operators (108), it would appear that using the median of the latter sample already in itself offers some leeway for capturing the variation in profit levels around the median for less capital-intensive companies, such as DPLP, and for sufficiently rewarding efficiency gains where appropriate.

(314)

Several cross-checks for consistency can be performed by comparing the default range mentioned in recital 309 with other available estimates of reasonable profit. As indicated at the start of this subsection, it is useful to perform a cross-check with the median ROS of the CRA study, which is […] % (109). The median ROS calculated on the basis of the WIK benchmark II sample, restricted to companies with a turnover of more than EUR 100 million, gives an ROS of 7,1 %. The Commission concludes that these figures fall within the range of [5,4-7,4 %], thus supporting the claim that this range can be viewed as reasonable.

(315)

Thus, overall, the Commission concludes that a default range of [5,4-7,4 %] would appear to constitute an appropriate benchmark range of profit, at least for the periods when DPLP faced a significant degree of risk and strong incentives to become more efficient.

(316)

In general, the Commission considers that, where the contract between the incumbent and the public authority is such that the public service is connected with a significant degree of risk, the reasonable profit allowed to the incumbent should be commensurate with the corresponding level of risk involved. By contrast, where the contract entails little risk, the level of reasonable profit should be low.

(317)

In the previous section, the Commission explained its approach for establishing a default reasonable profit range of [5,4-7,4 %], which is considered appropriate for contracts involving a significant degree of risk. It is difficult to quantify the appropriate benchmark in terms of a range of reasonable ROS, but it is clear that such an ROS range should be on the left-hand side of the sample distribution of observed profit levels. The Commission is of the opinion that it is justified from this perspective to adopt a reasonable profit range based on the first quartile (25th percentile) instead of the median.

(318)

The use of the 25th percentile as a benchmark for low-risk contracts leads to a range for reasonable ROS of [3,6-4,8 %] based on the different Deloitte samples.

(319)

As before, a comparison with the CRA results for postal incumbents may be a further useful cross-check. The average annual 25th percentile calculated by the Commission is […] %. The Commission concludes that these figures fall within the range of [3,6-4,8 %], thus supporting the claim that this range can be viewed as reasonable.

(320)

In general, the Commission concludes that a default range of [3,6-4,8 %] would appear to constitute an appropriate benchmark profit range for the periods when DPLP faced a low level of risk.

(321)

As explained in the introduction to this section, the risk borne by DPLP depends on the compensation mechanism applied by Belgium. This mechanism is reflected in the different management contracts concluded between DPLP and the Belgian State, and varies over time. Therefore, it is reasonable to consider different ranges for the ‘reasonable ROS’ depending on the compensation mechanism in place, which determines the degree of risk faced by DPLP, but also the extent to which DPLP is incentivised to improve productivity.

(322)

In this context, the Commission wishes to underline that the risk characterisation of the contracts is not only dependent on the mere existence of risk in DPLP’s management contracts, or even the absolute level of risk, but above all on the extent to which the risk is greater or smaller than that faced by the companies in the same sample groups. The same applies to the extent to which DPLP is incentivised to achieve efficiency gains. Given that these sample groups are composed of firms operating in a competitive environment, they normally face significant risk and strong incentives to become more efficient.

(323)

The Commission is of the opinion that the fourth management contract (2006-2010) clearly provides for mechanisms involving long-term price and funding commitments based on ex-ante parameters. This applies both to DPLP’s mail delivery and press distribution activities.

(324)

Specifically, as regards press distribution, Article 9 of the fourth management contract specifies that, over the contract period (standard five years), the weighted average of tariffs for the provision of services belonging to the ‘small users package’ will not increase more quickly than allowed by a price cap mechanism based on the ‘health index’ (a specific consumer price index) and a quality bonus reflecting the percentage of mail items delivered on time. In the Commission’s view, fixing the tariffs over the entire contract period in this way reflects all the characteristics of an incentive contract, exposing the company to a significant degree of risk.

(325)

As regards press distribution, Article 2 of the fourth management contract specifies that, over the contract period, tariffs for press distribution will not increase more quickly than the ‘health index’. In order to cover the net costs of press distribution, taking into account projected efficiencies, Article 13(2) fixes the annual compensation to which DPLP is entitled at EUR 290 613 000 (adjusted for inflation) (110). Article 13(5) sets out that, in case the scope or conditions of the SGEI change, or if there is a significant development of the costs DPLP cannot control, the contracting parties will investigate whether the compensation amount should be adjusted so that it remains in line with costs. Any such adjustment requires a decision by the Council of Minsters. The Commission recognises that this latter clause could be read as an escape clause, reintroducing certain ex-post elements into the contract (which protect DPLP from risk). However, the Commission considers that the language used by the public authorities is sufficiently non-committal and that DPLP cannot therefore count on the clause being used. In general, the Commission considers that the arrangements for DPLP’s press distribution activities imply a long-term commitment as regards the level of compensation.

(326)

Consequently, in the Commission’s view, the fourth management contract (2006-2010) cannot be characterised as involving low risk, but should rather be deemed as involving a significant degree of risk. Overall, the default range of [5,4-7,4 %] would appear appropriate for both types of activities provided for under the fourth management contract.

(327)

By contrast, the earlier management contracts (up to 2005) seem to have been largely based on price and compensation mechanisms involving a low degree of risk for DPLP.

(328)

As regards the setting of postage stamp prices, the Commission observes that the first three management contracts do not use a price cap mechanism of the type used in the fourth management contract (involving a long-term price commitment). Instead, these contracts refer, in rather general terms, to tariffs based on real costs, established on the basis of adequate accounting principles (111).

(329)

Likewise, as regards press distribution, the Commission observes that, unlike the fourth management contract, the first three management contracts do not involve lump-sum budgets covering the contract period (112). Rather, they refer to compensation on the basis of real costs (113). The Commission notes that the second contract also specified a global cap on compensation, but the actual compensation amounts observed in the period seem to call into question whether the cap was really binding (114). The first management contract also specified an annual compensation amount, but the Commission notes that, at the same time, the contract did not specify a long-term cap on the tariffs for press distribution.

(330)

As a result, the Commission is of the opinion that the range of [3,6-4,8 %], which was deemed reasonable in the absence of significant contract-specific risk, should be applied to both press distribution and mail delivery for the period covered by the first three management contracts.

(331)

The Commission notes that a contract-based approach is also proposed by the Belgian authorities in their submission of 14 December 2011. Although their proposed methodology raises a number of questions concerning the individual risk factors and the way in which they are weighted (115), an analysis by contract period, essentially classifying the risk faced by DPLP using a three-grade scale (low, medium and high), is consistent overall with the approach adopted by the Commission. On balance, the relevant ranking by DPLP of the contracts in terms of riskiness appears consistent with the Commission’s analysis, since the Belgian authorities also consider the fourth contract riskier overall than the first three contracts.

(332)

The Commission concludes that a default range for return on sales (ROS) of [5,4-7,4 %] is an appropriate benchmark range for determining reasonable profit for the periods when DPLP faced a significant degree of risk and for taking account of the strong efficiency incentives.

(333)

The Commission also concludes that an ROS range of [3,6-4,8 %] is an appropriate benchmark range for determining reasonable profit for the periods when DPLP did not face a significant degree of risk.

(334)

In order to take into account the efforts to achieve efficiency gains which the Belgian authorities claim DPLP made, the Commission has decided to take a prudent approach by using the upper bounds of these two ranges for fixing the reasonable profit to be included in the costs of the public service. This approach is also emphasised by the Deloitte study (see recital 293), according to which it is appropriate to use ranges because of the need to avoid penalising above-average performance resulting from efficiency gains.

(335)

The Commission considers that the first three management contracts (1992-1996, 1997-2002 and 2003-2005) are low risk and a reasonable profit rate of 4,8 % ROS can be applied to them, both for mail delivery and for press distribution. The fourth management contract (2006-2010), on the other hand, is deemed to involve considerably more risk and a reasonable profit rate of 7,4 % ROS can be applied to both mail delivery and press distribution.

(336)

The first management contract specifies that the annual State compensation is intended for all public service missions.

(337)

The potential overcompensation is therefore to be calculated by adding together all revenues and costs of all the public service missions.

(338)

The annual State compensation for the public service missions is to be included in revenue. The total amount of annual compensation received by DPLP under the first management contract is EUR 1,5 billion.

(339)

In addition, the following aid measures have to be taken into account:

Transfer of buildings

(340)

The first management contract also provided for a transfer of buildings from the State to DPLP. This transfer of buildings, worth EUR 112,2 million, which did not include the payment of any consideration, is to be included in DPLP’s revenue for the calculation of overcompensation for the period covered by the first management contract. DPLP did not pay any consideration to the State for the use of the buildings in the period preceding the transfer.

(341)

An amount of EUR 62 million was paid by the State to DPLP in 1997 and was booked by DPLP as a capital injection. The Belgian authorities have subsequently admitted that this transfer was in fact a delayed payment of the balance of the annual compensation for 1996 provided for in the first management contract.

(342)

The amount had originally not been paid on time to DPLP in 1996, so the State compensation was lower than the annual amount provided for in the management contract for the year in question.

(343)

For the purposes of controlling overcompensation, the payment of EUR 62 million is included in the SGEI compensation for the period covered by the first management contract.

(344)

The calculation of overcompensation for the first management contract can be summarised as follows:

(345)

For the three subsequent management contracts, the calculation of overcompensation will take into account only the revenues and costs of the SGEIs that, under the provisions of those contracts, could be subject to public service compensation. The corresponding revenues and costs of the other SGEIs that, according to the provisions of these contracts, were not entitled to aid are not taken into consideration. As the universal postal service is not part of the aided activities, its potential net cost will not be taken into account and is to be fully financed by DPLP from its own resources in accordance with the public service mission entrusted to it under the management contracts. The activities that, in accordance with the management contracts, are entitled to the level of compensation established in each contract, include, in particular, press distribution (newspapers and periodicals) and the other public services listed in recital 33 above.

(346)

On account of the low tariffs charged for press distribution, the aided SGEIs are loss-making overall.

(347)

As potential overcompensation of one service may be used to finance the net costs of other undercompensated public service missions, the calculation will add together all revenues and costs of the SGEIs eligible for aid under the management contracts.

(348)

The total amount of annual compensation received by DPLP under these three management contracts is EUR 3,5 billion.

(349)

Apart from the compensation provided for in the management contracts, DPLP benefited, throughout the relevant period under investigation, from exclusive rights granted by the State for the provision of reserved postal services. Although the potential net surpluses from this reserved area do not constitute State resources for the purposes of State aid control and cannot as a result be subject to a recovery order in case of overcompensation, they can and must be taken into account under the State aid rules when assessing the necessity and proportionality of any other State aid granted to the same operator as public service compensation. Therefore, when assessing the compatibility of the compensation granted to DPLP under the different management contracts in the light of Article 106(2) TFEU, the excess profits generated by the Belgian postal operator in the reserved area will be deducted from the costs of any SGEI entitled to public service compensation in accordance with the provisions of those contracts.

(350)

It should be noted in this respect that this principle is included in the SGEI Framework, which states: ‘if the undertaking in question holds special or exclusive rights linked to a service of general economic interest that generates profit in excess of the reasonable profit, or benefits from other advantages granted by the State, these must be taken into consideration, irrespective of their classification for the purposes of Article 87 of the EC Treaty, and are added to its revenue’ (116). This principle reflects the necessity and proportionality condition inherent in Article 106(2) TFEU. In fact, any aid for an SGEI paid on top of excess profits resulting from an exclusive or special right granted to the same operator which are already sufficient to cover the costs of another SGEI is not necessary for the provision of that SGEI. The operator making such excess profits already benefits from a State measure allowing it to provide the SGEI under conditions of economic equilibrium and, therefore, no additional compensation is necessary.

(351)

Therefore, any excess profit earned in the reserved area as a result of the exclusive right must be taken into account when assessing the necessity of the SGEI compensation granted to DPLP.

(352)

Accordingly, any profit from the reserved area (USO R) above the level of reasonable profit is included in the analysis of the SGEI overcompensation.

(353)

However, since the reserved area is first and foremost meant to finance the universal service, any loss in the other domains of the universal service will reduce the contribution of the reserved area to the costs of other linked SGEIs.

(354)

Furthermore, in order to determine the genuine excess profit generated by the reserved area, it is necessary to take account of costs incurred as a result of the universal service obligation that are borne by activities outside the scope of the universal service. In line with the Chronopost case law, the costs of the retail network are apportioned in DPLP’s cost accounting among the activities using the retail network, which includes products both within and outside the scope of the universal service. In particular, the obligation imposed on DPLP to maintain a retail network of a certain size and density to provide a universal postal service at an adequate level of quality as defined by the Belgian authorities leads to losses in commercial activities outside the scope of the universal service because a share of the retail network costs is allocated, in line with the Chronopost case law, to the commercial products using the retail network.

(355)

Network constraint imposed by the management contracts: the constraints imposed by the State concern the density and type of outlets (post offices and/or‘points poste’ (service counters)) composing the retail network. These constraints are defined in the successive management contracts. For example, the fourth management contract (2005-2010) provides for:

(356)

The Belgian authorities consider that the specific requirements concerning the number and type of outlets in the DPLP retail network constitute a separate, territorial coverage obligation imposed by the State beyond the requirements linked to the USO. On this basis Belgium considers that all the net costs of that specific territorial coverage obligation should be compensated by the State.

(357)

Quantifying the network constraint: in order to determine the net costs for DPLP arising from the obligation to operate a retail network composed of at least 1 300 outlets, of which a minimum of 589 must be post offices, the Belgian authorities have compared the actual retail network, in which the density and type of outlets are imposed by the State, with a hypothetical retail network that DPLP would operate in the absence of specific constraints. In other words, the hypothetical retail network is what DPLP would have chosen to operate under commercial terms in the absence of the public service obligations imposed by the State.

(358)

The Belgian authorities consider that the hypothetical network would be composed entirely of points poste. There would no longer be post offices. However, in order to be able to continue to serve the same total volume of customers, DPLP would have to drastically increase the number of points poste.

Comparison of DPLP’s current retail network with a hypothetical network in the absence of State obligations

(359)

As can be seen from table above, DPLP’s actual retail network fulfils the requirements of the State: the total number of outlets and the number of post offices are both above the required level. By contrast, the hypothetical network that DPLP would operate if it were allowed to choose its retail network on a purely commercial basis would be significantly different: the total number of points poste would increase, but there would no longer be post offices run by DPLP.

(360)

Reality of the network constraint: it should be pointed out that the State-imposed constraint on DPLP is not clearly established.

(361)

First, the actual network operated by DPLP is larger than that required by the State. Both the total number of outlets and the number of post offices is higher than the minimum required by the State. This could imply that DPLP is not actually constrained by the obligations, but instead chooses freely to operate a network of the current size.

(362)

The Belgian authorities justify this excess number of post offices by the 10 km distance requirement, the need to have a safety margin in case of unexpected closures of points poste, and their intention to increase the required minimum number of post offices to 650 in the next management contract, which would make it counter-productive to close offices to come down to the level currently required, only to re-open them again later. The Belgian authorities argue that, for these reasons, it has not been feasible for DPLP to further decrease the number of post offices to move it closer to the minimum requirement of 589. According to the Belgian authorities, the current network of 713 post offices is therefore de facto constrained by the State requirements despite the nominal excess in the numbers of outlets.

(363)

Second, the method chosen by the Belgian authorities relies upon backward projection of retail network costs from 2009 to 1992. Such backward projection inevitably entails methodological uncertainties. The business model of the points poste has not necessarily been available since 1992. Furthermore, in the past, DPLP’s workforce was mainly composed of civil servants, whose payroll costs were very rigid. As the points poste are not run by DPLP personnel, the DPLP statutory employees no longer working in post offices would have remained on DPLP’s payroll. It is therefore doubtful whether, in the early part of the period under investigation, the potential closure of post offices would have entailed savings in direct payroll costs. By contrast, in 2009, when the workforce was composed to a higher degree of contractual workers, such cost savings could have materialised to some extent. With the gradual transition towards less rigid staffing and the successive improvements in the cost accounting system, the accounting data from more recent years can be considered free from this type of distortion. However, given that there are still over 20 000 civil servants working for bpost, who can neither be employed in points poste nor be dismissed, the replacement of all post offices by points poste remains a purely theoretical and unrealistic option even today.

(364)

The Belgian authorities underline, however, that potential redundancy costs in the past would have been offset by savings in operating costs and the sale of unnecessary buildings. As regards the reliability of the data for the years 1992-2002, the Belgian authorities state that the costs of the retail network have probably been underestimated, as some items that could have been considered network costs have been grouped in other cost centres. Consequently, the weaknesses in the backward projection method would be offset by the other, conservative assumptions made by the Belgian authorities when calculating the cost in the past years.

(365)

Existence of a specific obligation for territorial coverage: even if the existence of the network constraint were accepted, the Belgian authorities’ claim that it results in a separate territorial coverage mission does not seem justified because there do not appear to be specific provisions in the management contracts defining such an obligation that would be distinct from and go beyond the requirements of the USO and other public services obligations. On the contrary, the management contracts between DPLP and the Belgian State specify that the sole purpose of the network constraint is to ensure the universal service and the other SGEIs (distribution of newspapers, magazines, postal accounts, postal orders, etc.); and the network costs necessary for the provision of these services are a priori already booked in the accounts of these different services (119). In addition, the management contracts do not provide for any mechanism to compensate for the costs of a territorial coverage obligation. The Commission notes that the requirement to define parameters in advance for calculating compensation was introduced in the 2005 SGEI Framework and that overcompensation for SGEI costs incurred by DPLP has been identified in the period after 2005 (see recital 386).

(366)

It should be noted that, in accordance with Article 3 of Directive 97/67/EC, Member States are required to impose an obligation upon postal incumbents to provide a universal service at an appropriate level of quality, including appropriate geographic coverage. However, this territorial coverage, which also exists in the present Belgian case, is only one of the multiple aspects of the USO. In the management contracts entrusting DPLP with the universal service mission, the Commission cannot identify any reference to an additional mission requiring the postal incumbent to operate a larger network of offices beyond the territorial coverage imposed by the USO and other related public services to meet objectives of territorial cohesion or development. There is no doubt that each post office run by DPLP for universal service reasons contributes to territorial cohesion (120). However, legally speaking, the obligation to operate it is linked to the need to guarantee a sufficient degree of quality in the universal postal service and other related services. In order to invoke such a complementary territorial mission, in the light of the relevant case law, it would be necessary for DPLP to be entrusted with a separate and clearly defined mission to operate offices beyond the size of the postal network fixed for universal service reasons. This entrustment would have to specify the obligations, the ensuing costs and the system for financing the net costs of the additional territorial mission.

(367)

The Commission cannot find any of these elements in the management contracts between the State and DPLP. The Belgian authorities and DPLP are in fact trying artificially to justify a posteriori the existence of a distinct territorial mission by arguing that the territorial coverage set out in the management contracts is necessary on two counts: both for USO purposes and for an alleged territorial mission beyond the USO. However, the relevant contracts lack any provisions on the alleged separate territorial mission requiring DPLP to operate a larger network than that laid down for universal service provision.

(368)

It should be noted that the artificial nature of the claims made by the Belgian authorities and DPLP is further evidenced by their own reasoning. The Belgian authorities, with a view to diminishing the level of potential overcompensation, would like the Commission to take into consideration, as part of the costs of an additional territorial mission beyond the scope of the USO, the difference between the real cost of a fully-fledged office and the cost of a hypothetical point poste that would replace it. This reasoning does not rely upon a hypothetical larger network that DPLP would operate to meet its territorial objectives, but only identifies the extra cost of the universal service coverage, which is not entitled to aid under the management contracts. This absence of a territorial objective is confirmed by the fact that the Belgian authorities would have to substantially increase the number of points poste to provide the universal service since it is not possible to do it with an equivalent number of points poste.

(369)

Nevertheless, it can be observed that, although it does not appear possible to accept the claim by the Belgian authorities of a real and distinct territorial coverage obligation, the network constraint imposed by the Belgian State to operate a network capable of delivering the universal service and the other public services not only concerns those services, but also induces additional costs for the commercial services.

(370)

In line with activity-based costing and in accordance with the Chronopost case law, all the activities using the retail network contribute towards its common costs, as explained in recital 354. Consequently, all the products, including the commercial products, contribute to the network costs. The costs deriving from the State-imposed requirement to operate post offices are a burden on the commercial products, which appear to be consistently loss making as a result, even though they are sold at market price. DPLP sells its commercial services through the network because they finance a share of the fixed costs of the network and therefore contribute indirectly to the universal service.

(371)

From an economic viewpoint, it would seem sensible, when assessing overcompensation to DPLP, to take into account the losses on commercial products induced by the USO and other public services obligations. While DPLP has an economic incentive to sell commercial products through the retail network, the consequence, from a cost-accounting perspective, is that a proportion of the network costs that would otherwise be charged to the public service is instead allocated to the commercial products, which appear loss making as a result.

(372)

Accordingly, the calculation of overcompensation in the present case can summarised as follows:

(373)

The compatibility of the 1997 capital injection should be examined as part of the assessment of the compatibility of the annual SGEI compensation since it was not a separate recapitalisation, but a payment related to compensation for the public service missions for the period 1992-1996.

(374)

As the 2003 capital injection does not constitute State aid within the meaning of Article 107(1) TFEU, the issue of compatibility of the aid does not arise.

(375)

As the 2006 capital injection does not constitute State aid within the meaning of Article 107(1) TFEU, the issue of compatibility of the aid does not arise.

(376)

It is worth noting that the State guarantee is not included in the management contract as a form of compensation for the public service obligations, and therefore its compatibility cannot be assessed in the light of Article 106(2) TFEU. As no other basis for compatibility has been brought forward by the Belgian authorities nor detected by the Commission, the measure constitutes operating aid that is incompatible with the internal market and, consequently, must be recovered.

(377)

As already mentioned, the corporate tax exemption has not conferred an advantage on DPLP and therefore does not constitute State aid within the meaning of Article 107(1) TFEU.

(378)

The property tax and local tax exemptions constitute State aid within the meaning of Article 107(1) TFEU, but are to be considered existing aid as they were established before the entry into force of the Treaty. In addition, it is to be noted that these fiscal measures are not included in the management contracts as a form of compensation for the public service obligations.

(379)

Consequently, the tax exemptions do not affect the calculation of overcompensation, which is based on the actual costs borne by DPLP (which already include the effect of those tax exemptions).

(380)

Therefore, the tax exemptions will be dealt with separately in an existing aid procedure as provided for by Articles 17-19 of Council Regulation (EC) No 659/1999 of 22 March 1999 laying down detailed rules for the application of Article 93 of the EC Treaty (121).

(381)

The amount of overcompensation can be calculated on the basis of the assessment set out in subsection 6.3.3.2. In line with the public service financing mechanism established by the Belgian authorities in the management contracts, overcompensation will be calculated separately for each period of entrustment.

First management contract: 1992-1996

Second management contract: 1997-2002

Calculation of surplus from excess profit from USO reserved area (OSU R)

Overcompensation calculation:

Third management contract: 2003-2005

Calculation of surplus from excess profit from USO reserved area (USO R)

Overcompensation calculation:

Fourth management contract: 2006-2010

Calculation of surplus from excess profit from USO reserved area (USO R)

Overcompensation calculation:

Other measures

(382)

As agreed with the Belgian authorities (122), the ‘most objective and balanced’ method for defining the time period to be taken into account in calculating overall overcompensation in the present case is the ‘contract-by-contract’ approach. The Commission does not see any problem in accepting the Belgian authorities’ request. In fact this method ensures greater predictability and legal certainty for the contracting parties and thereby tends to reduce the risk of overcompensation. Consequently, each period will be considered separately and no over-/undercompensation will be carried over from one period to the next.

(383)

The Belgian authorities, however, argue that, before the 2005 SGEI Framework entered into force, the Commission had consistently applied a global method for calculating over-/undercompensation. On that basis, the Belgian authorities have specifically requested a pooling of all years before 2005 (123) in one global calculation of over-/undercompensation.

(384)

The Commission observes that, since all three management contracts before 2005 result in undercompensation, the pooling of all years would have no consequence on the amount of overcompensation to be recovered.

(385)

Furthermore, the Commission notes that, since the adoption of the 2005 SGEI Framework, the ‘parameters for calculating, controlling and reviewing the compensation’ have to be defined in advance in the entrustment act(s) (124). Since the fourth management contract does not provide for compensation for any costs incurred outside its period of validity (2006-2010), undercompensation identified for periods not covered by the contract cannot be offset against possible overcompensation granted to DPLP.

(386)

The Commission concludes that DPLP has been overcompensated by the amount of EUR 415 million for discharging its public service mission, notably during the period of application of the fourth management contract. Since overcompensation constitutes incompatible aid, it must be recovered.

(387)

The pension relief, the annual public transfers, the transfers of buildings, the 1997 capital injection, the tax exemptions, and the State guarantee constitute aid measures within the meaning of Article 107(1) TFEU because they relieved DPLP of costs that are normally borne by private undertakings.

(388)

The 2003 and 2006 capital injections do not constitute aid within the meaning of Article 107(1) TFEU because the State acted in conformity with the MEIP.

(389)

The Commission finds that Belgium unlawfully implemented the measures in breach of Article 108(2) TFEU, with the exception of the 2003 capital injection, which was duly notified to the Commission before being executed, and the tax exemptions, which constitute existing aid.

(390)

The pension relief has not placed DPLP in a more favourable situation than its competitors with regard to social security contributions. The measure can therefore be declared compatible pursuant to Article 107(3)(c) TFEU.

(391)

The annual compensation granted under the fourth management contract constitutes incompatible aid pursuant to Article 106(2) TFEU to the extent that it overcompensated DPLP.

(392)

The State guarantee constitutes incompatible aid.

(393)

The property tax exemption and local tax exemption constitute existing aid and will be dealt with separately in an existing aid procedure in accordance with Articles 17-19 of Regulation (EC) No 659/1999.

(394)

In accordance with Article 5 of the SGEI Framework, this Decision is without prejudice to the Union provisions in force in the field of public procurement and competition,

(a)

the total amount (principal and interest) to be recovered from the beneficiary;

(b)

a detailed description of the measures already taken and those planned to comply with this Decision;

(c)

documents demonstrating that the beneficiary has been ordered to repay the aid.

‘3.

In addition, but separately from the USO, DPLP was and still is entrusted with certain specific non-mail SGEIs, such as home-delivery of pensions and allowances for disabled persons, delivery of election documents and, most significantly, distribution of newspapers and periodicals. These SGEIs have, by and large, generated net costs and these net costs have, for the most part, been covered by State compensation pursuant to successive management contracts concluded with the State.

4.

Accordingly, Belgian postal legislation has always carefully distinguished between the USO and other (non-mail) SGEIs, with respect to their financing. Thus, successive management contracts between the State and DPLP have provided for periodic State compensation for the net costs of (non-mail) SGEIs, with the specific exception of the USO.’

In contrast, a specific statutory provision deals separately with the possible financing of the USO burden, until recently through the possible creation of a compensation fund and, as of 2011, by providing for the possibility of a direct subsidy under conditions that mirror those in Annex I to Directive 97/67/EC, as amended by Directive 2008/6/EC. However, this compensation fund has never been activated, nor has the Belgian State given any direct subsidy in respect of a loss-making USO.