The Art, Science and Law of Funding Services of General Economic Interest

The compulsory nature of an assigned public mission makes the services covered by the mission different from those provided by market operators for profit.



The justification of government intervention for the purpose of ensuring delivery of adequate services of general economic interest [SGEI] and the design of measures that compensate providers of SGEI can be very tricky. This article reviews three recent measures that imposed public service obligations [PSO] in three different sectors [air transport, health and energy]. The corresponding Commission decisions highlight the common features of any public intervention that aims to support SGEI. These measures are also instructive because they reveal the problems in defining correctly the scope of the PSO, selecting the SGEI providers and preventing overcompensation.

The problem with these three decisions is that they do not explain sufficiently the reasoning of the Commission on the very important issues of the definition of the PSO, the selection of the SGEI provider and the avoidance of overcompensation. Naturally, the Commission has to grapple with whatever information is submitted by Member States, regardless of its quality. But the Commission can do a better job at explaining its own assessment. As noted recently by the General Court, the text of a decision does not have to reveal amounts which are business secrets in order for the reader to understand fully the reasoning of the Commission.

Measure 1: Kalmar Oland airport [SA.43964][1]

The airport is located in the southeast of Sweden and serves the island of Oland. In 2015, it was used by about 223,000 passengers. It was entrusted with a public service mission which was limited only to servicing airlines. Other commercial activities such as parking or shopping were excluded from the public service mission.

The Commission considered the airport to be an undertaking and the public service compensation [PSC] it received to be State aid. The PSC did not satisfy the 4th Altmark condition as the airport was not selected via a competitive selection procedure, nor was it shown to be efficient by industry standards. Interestingly, Kalmar airport appeared to be more efficiently run than other similar airports. Yet for the Commission this was not sufficient as there was no proof that the comparator airports were “well run and adequately provided”, as required by the 4th Altmark condition [paragraphs 29-31 of the Commission decision].


However, the Commission found the PSC to be compatible with the internal market because it complied with the requirements of the SGEI Framework. According to the Aviation Guidelines [section 4.2], aid to airports with more than 200,000 passengers is assessed on the basis of the SGEI Framework [the SGEI Decision (2012/21) was not applicable because it covers airports with fewer than 200,000 passengers].

Compliance with the SGEI Framework was achieved for the following reasons. First, the airport was entrusted with a genuine PSO. According to point 72 of the Aviation Guidelines, there is a genuine public need when “part of the area potentially served by the airport would, without the airport, be isolated from the rest of the Union to an extent that would prejudice its social and economic development.” But it also requires that “such an assessment should take due account of other modes of transport, and in particular of high-speed rail services or maritime links served by ferries.” Sweden argued that there was no high-speed train connection to Kalmar, that businesses could not survive in the area without the airport, that residents needed to have access to an air ambulance and that the drive to the nearest alternative airports would be about three hours which was considered to be too long. In relation to the last point, the Commission observed that travelling up to 1.5 hours to an airport is an “acceptable level of connectivity” [paragraph 64].

Second, the PSO did not concern the development of air transport services, which is not allowed by point 73 of the Aviation Guidelines, but rather the maintenance of the airport.

Third, the scope of the PSO was limited only to the services provided to airlines. It excluded other commercial services provided to passengers in shops, restaurants or parking.

Fourth, the PSO was entrusted through an official act which defined the content, duration and territory of the PSO, the parameters for calculating the PSC and the arrangements for avoiding overcompensation.

Fifth, the duration of entrustment was 10 years which was considered to be reasonable given the longer amortisation period of many of the assets of the airport.

Sixth, the entrustment of the public service mission to Kalmar airport fell outside public procurement rules because it was wholly owned by the public authority which imposed the PSO. Kalmar airport was considered to be an “in-house” operator [paragraph 92]. In relation to this issue, the Commission decision cited the judgment of the Court of Justice in case C-107/98, Teckal, which developed the concept of in-house operator. In that case the Court established that a legal person is an in-house operator when “(50) […] the local authority exercises over the person concerned a control which is similar to that which it exercises over its own departments and, at the same time, that person carries out the essential part of its activities with the controlling local authority”. This is not the first time that the Commission finds that direct awards comply with the concept of in-house operator [see, for example, the many entrustments to postal operators]. But the decision is silent on why Kalmar airport was considered to “carry out the essential parts of its activities” with the local public authority that owned it. An in-house operator, for example, as an IT department that supports a public authority’s computers, or a catering unit that operates staff restaurants. Kalmar airport provided services to the general public. This is one of the unexplained issues in a number of Commission decisions.

Seventh, the method of calculating the PSC complied with the net avoidable cost approach. This approach is based on identification of the likely operations and related costs of the airport in the absence of the PSO. According to the Commission decision, the commercial services not covered by the PSO would not be provided as they depended exclusively on the airport carrying out its public mission [paragraph 99]. It is difficult to believe that a complete closure of all commercial services would be a realistic alternative scenario. As long as the airport remains operational, even at a much smaller scale, passengers would still need some parking facilities or some catering services. Perhaps the airport would have closed down, but the decision does not explicitly say this. At any rate, the Commission accepted that the PSC could be calculated using instead the cost allocation methodology which meant that fixed costs were shared between the PSO and the commercial services. Since the airport was ran on a non-profit basis, all profits from the commercial services were used to reduce the amount of the necessary PSC. It should be noted that in this case the use of the cost allocation methodology increased the amount of costs that had to be borne by the commercial services which were profitable. However, in the end it probably made no difference because the profits from the commercial services were used to reduce the amount of the PSC. Had the fixed costs not been shared, the costs of the commercial services would have been smaller, their profits would be larger and therefore the amount of PSC that would have been needed would be smaller. It should further be noted that although the airport was operated on a non-profit basis, the PSC included a notional profit equal to the swap rate plus 1% [which is the safe-harbour rate defined in the SGEI Framework]. This means that even non-profit undertakings need to earn a surplus in order to cover their financing costs or set aside reserves to be used for future investments.

Eighth, Sweden devised an unusual method for ensuring that the airport was incentivised to carry out its public mission efficiently. At the end of each year, the accounts of the airport and other performance indicators would be compared to those of other airports. If its performance fell short of that of other airports, Kalmar airport would have to provide a sufficient explanation otherwise its compensation would be reduced to an amount that would correspond to the lower costs of comparator airports. Acceptable explanations would concern force majeure or extraordinary events beyond the control of the airport [e.g. adverse weather, increased security, industrial action, etc].

Do you know we also publish a journal on State aid?

EStAL banner

The European State Aid Law Quarterly is available online and in print, and our subscribers benefit from a reduced price for our events.


Measure 2: Klinikum Osnabrueck [SA.36798][2]

After receiving a complaint, the Commission investigated whether a 100% public guarantee for a loan of EUR 28 million and a capital increase of EUR 1 million constituted State aid. The recipient was Klinikum Osnabrueck which is a 100% publicly owned hospital in Lower Saxony, Germany.

Interestingly Germany argued that that the capital increase conformed with the principle of the market economy investor [MEIP] and that both measures complied with the 2012 SGEI decision [Decision 2012/21] which functions like a block exemption regulation.

For its part, the complainant claimed that any SGEI-linked aid could not be compatible with the internal market because the hospital was not efficient.

The Commission dismissed both Germany’s argument and the complainant’s claim. If a capital injection satisfies the MEIP, then the recipient must be profitable. However, SGEI aid can only be granted when the recipient incurs costs that exceed revenue, which means that it makes losses.

Furthermore, SGEI providers do not have to be efficient. If they were efficient by industry standards [i.e. in relation to their competitors] the public funds they receive would not constitute State aid as they would probably satisfy the four Altmark conditions. Also, neither the case law on Article 106(2) TFEU [see, for example, C‑660/15 P, Viasat Broadcasting v Commission; T-137/10, CBI v Commission], nor the Commission’s SEGI rules require that the recipient is efficient. Only the SGEI Framework obliges Member States to incentivise the recipient to improve its efficiency during the period of the assignment of the PSO.

During its investigation the Commission found out that the total amount of the capital increase was EUR 31 million and that it had been implemented in tranches. In addition, five guarantees were granted over a period of time covering a total amount of EUR 63 million. However, the loans that were obtained on the basis of those guarantees until the point of the Commission’s decision were only EUR 15 million. Also, the Commission found out that the City of Osnabrueck had granted various shareholder loans amounting to EUR 18 million.

There was no doubt that the public money put at the disposal of the hospital was State aid. Since the various amounts were granted at different points in time, the Commission assessed their compatibility both with the 2005 SGEI Decision and the 2012 SGEI Decision. Only the compatibility with the 2012 Decision is reviewed here.

The Commission found the various capital injections and guarantees to be compatible aid for the following reasons.

First, the period of entrustment was 10 years which was at the limit of the safe-harbour period laid down in the 2012 Decision.

Second, the Commission found that the entrustment concerned a genuine SGEI. Under German law, public authorities are obliged to offer hospital care as long as other providers cannot do so. This obligation does not apply to private hospitals which can in principle choose the services they provide and to which extent. It is important to note that the text of the Commission Decision indicates the existence of an actual gap only in the provision of emergency services. It does not contain any information that suggests there was a real market gap or an actual failure by private hospitals to cover all medical needs of the local population. The Commission limits its analysis to a discussion of the fact that private hospitals are not under the same obligations as public hospitals. However, this approach deviates from the established approach in other Commission decisions on SGEI which seek to demonstrate actual or real market inadequacy. The Commission also seems to conflate the compulsory nature of a PSO with the existence of a market gap. “(38) […] The complainant however considers that private hospitals in the Osnabrück region provide or can provide the same services as Klinikum Osnabrück. This would put into doubt whether Klinikum Osnabrück performs a genuine SGEI. The Commission notes however that private hospitals are not required to continue to provide all services and that they can choose to stop the provision of loss-making services at any time (if they even provide such services at all). Klinikum Osnabrück on the contrary must continue to provide a full range of care and emergency services even at a loss.”

Interestingly, the Commission in the same paragraph cites the judgment in case, T-17/02, Fred Olsen v Commission (maritime transport) in support of its reasoning. In that case the General Court stated that the presence of competitors providing some services did not prove that the market supplied all services on the same terms as those defined in the public service mission. This emphasis was on the existence of an actual gap in the market. In a similar but more recent case concerning train and bus transport, T-92/11 RENV, Jørgen Andersen v Commission, the General Court ruled that the mere fact that Andersen could provide alternative transportation was not sufficient reason to doubt the existence of a genuine SGEI. This is because the incumbent train operator provided seamless services. In other words, the General Court found that the synergies of interconnection between the different market segments or regions justified the imposition of a PSO that included a part that was served by a private operator. Again, the emphasis was on the actual difference in the services provided by train and by bus. However, in the Osnabrueck case, the Commission does not address the issue of synergies and linkages between different hospital services in order to establish that indeed there is a difference between public and private services. It remains at the very abstract and general level of the obligations of the public hospital.

Third, the act of entrustment defined precisely the geographic area and the contents of the PSO and excluded purely commercial activities which are routinely found in hospitals such as a cafeteria or services not linked to public health such as elective cosmetic surgery.

Fourth, the compensation mechanism defined a limit for the PSC which was the amount of the incurred net cost. However, it is not obvious in the Decision how the parameters of compensation were defined in advance. Nor, is it clear how the Commission quantified the total amount of aid received by the hospital. Presumably, the Commission considered the 100% loan guarantee to grant state aid equal to the total amount of the guaranteed loan.

Fifth, the act of entrustment laid down a procedure for preventing over-compensation through account separation and annual audits. The accounts of the hospital showed that the non-SGEI activities were self-financing so there was no risk that the State aid would cross-subsidise the non-SGEI activities.

Measure 3: Delimara power project [SA.45779][3]

The purpose of the publicly funded project was to increase energy security and improve the efficiency of electricity generation with the construction of a natural gas terminal and a gas-fired plant at the Delimara power station in Malta. The contract for the construction, supply of natural gas and eventual electricity generation was awarded through a competitive bidding process to ElectroGas Malta [EGM] by Enemalta, the state-controlled electricity generator and distributor in Malta. The beneficiary of the aid was EGM. Since Enemalta operates under a PSO, in effect it delegated part of its PSO to EGM. An important component of the contract was the commitment of Enemalta and ultimately of the government of Malta to buy the future electricity produced by EGM and therefore provide a guaranteed stream of revenue to EGM.


Existence of State aid

The Commission first examined whether the contract involved State aid. Certainly, there was transfer of state resources because Enemalta was state controlled and because it acted in this case not in order to make profit but to carry out public policy. And certainly, the measure was selective and did affect trade and distort competition.

With respect to the conferment of an advantage in the meaning of Article 107(1), despite the fact that EGM was selected through a competitive process, the Commission found that “(92) The measures provide an economic advantage to EGM as they ensure a certain IRR and a steady stream of revenues.”

Then it examined whether the Altmark conditions could apply [so that the compensation would not be state aid] but concluded that they did not. “(94) The measure may arguably meet some of the first three criteria, but not the fourth criterion as the outcome of the tender was influenced by the SSA [security of supply agreement] which was added to the contractual structure during the tender procedure. Since that element was not known at the beginning of the procedure, the organised tender procedure did not guarantee the achievement of the least cost to the community. In addition, the Maltese authorities have not provided any information demonstrating that the level of compensation was established according to the costs of a typical, well-run undertaking.” “(95) The Commission therefore concludes that the fourth Altmark condition does not appear to have been complied with in relation to the measures. […] Accordingly, the measures confer on EGM an economic advantage within the meaning of Article 107(1) TFEU.”

The reasoning in paragraphs 92-95 is troublesome. First, the successful bidder of any public contract expects to make profit. Otherwise, no one would bid for public contracts. The Commission claimed in paragraph 92 that the EGM was ensured a certain IRR and steady stream of revenues. Yet, the only evidence in the Decision in relation to this statement is that the Maltese authorities had carried out a financial appraisal of the project and concluded that the expected IRR was 7% which was below the benchmark rate of 8.3%-9.9%. Moreover, while EGM was selected in December 2013, the IRR was calculated in May 2015 and found to be 9.2%. It fell to 7% because of cost overruns [paragraphs 75-77]. Hence, EGM bore the risk of cost fluctuations. There is no indication that the contract guaranteed a fixed rate of IRR. As the Commission itself noted later on in the Decision “(151) […] The fact that the IRR decreased is also an indication that the risk of cost increases is borne by EGM.”

Second, the requirement for a steady stream of revenue is a standard clause in many supply contracts [e.g. obligation to purchase a certain quantity]. EGM invested in the generation of electricity because Malta committed to buying a certain quantity of electricity. This is the “quid” to any contractual “quo”. It also means that if the revenue stream is steady and less risky, then the return on the investment also has to be lower. But it does not mean that there is no risk [such as cost overruns] or that the return necessarily exceeds the risk-adjusted cost of capital. It should also be noted that in paragraph 107, the Commission found that EGM was selected through an “open and non-discriminatory public procurement procedure”, so it is hard to understand how it could have obtained an advantage.

Third, the Security of Supply Agreement [SSA] was inserted in the bidding process after six candidates were shortlisted from an initial 18 expression of interest. The Commission does not explain how the SSA could have prevented the achievement of the least cost to the community after the point of shortlisting. In fact, the absence of the SSA in the initial stages of the process did not deter a total of 18 companies from expressing interest. If the SSA made the project more attractive to less efficient companies, its absence should have deterred them from expressing interest. There is no explanation either how the absence of the SSA may have affected the shortlisting of the six most efficient or best candidates.

In conclusion, this important part of the Decision appears to be based on partly contradictory and partly unexplained reasoning.

Conformity with the 2012 SGEI Decision

Before turning its attention to the specific provisions of the 2012 SGEI Decision, the Commission noted that security of energy supply could be a legitimate SGEI, but certain ways of achieving that kind of security would not be regarded as SGEI. For example, investment in larger capacity could not be classified as SGEI because it is something that the market normally does [paragraph 109]. However, the Commission acknowledged that Malta’s small size and isolated system justified the imposition of a PSO as the market alone could not deliver sufficient quantity of electricity.

The aid was compatible with the provisions of the 2012 Decision for the following reasons.

First, the PSO concerned a genuine SGEI, covering a market gap.

Second, the PSO was imposed through an official act of entrustment which also specified the precise conditions for the PSO and the method for calculation of the compensation. The prices for the delivery of electricity and the supply of gas constituted the method of compensation.

Third, the period of entrustment which was 18 years was shorter than the economic life of the power plant which was estimated to be 25-35 years. [The 2012 SGEI Decision allows entrustment to exceed the safe-harbour period of 10 years if that is necessitated by the economic life of the investment.]

Fourth, EGM was selected after a competitive procedure involving 18 interested firms.

Fifth, the net avoidable cost methodology could not be used to calculate the amount of compensation because all of the EGM’s activities are covered by the PSO. Under the alternative cost allocation methodology, all costs and revenues were eligible and the compensation was the difference between costs and revenues. In this case, however, the amount of compensation was equal to the prices for the delivery of electricity and gas. The Commission concluded that there was no overcompensation as the expected rate of return did not exceed a benchmark rate.

Sixth, the method of compensation incorporated efficiency incentives. “(153) […] In this case, the upfront definition of a fixed compensation level anticipates and incorporates the efficiency gains that EGM can be expected to make over the lifetime of the entrustment act. In particular, the fixed compensation is calculated based on a fixed heat rate. If EGM delivers a more efficient heat rate from Delimara 4, it will increase its profitability.”

Lastly, overcompensation would be avoided by the fixing up front of the compensation in terms of prices that anticipated the expected increases in EGM’s efficiency.


[1] The full text of the Commission Decision can be accessed at:

[2] The full text of the Commission Decision can be accessed at:

[3] The full text of the Commission Decision can be accessed at:



Phedon Nicolaides

Dr. Nicolaides was educated in the United States, the Netherlands and the United Kingdom. He has a PhD in Economics and a PhD in Law. He presently holds positions at the College of Europe and the University of Maastricht. He has published extensively on European integration, competition policy and State aid. He is also on the editorial boards of several journals. Dr. Nicolaides has organised seminars and workshops in many different Member States, and has acted as consultant to several public authorities.

Leave a Reply

Related Posts

15. Dec 2020
State Aid Uncovered by Phedon Nicolaides

The Law, Economics and Art of Compensation for Public Service Obligations

State aid in the form of compensation for public service obligations must exclude any intangible benefits from the obligation, but may include reasonable profit. The Commission guidelines for State aid in agriculture and forestry, the block exemption regulations covering agriculture, forestry, fishery and aquaculture, and the Regulation on de minimis aid for fishery and aquaculture have been prolonged to 31 […]
08. Dec 2020
State Aid Uncovered by Phedon Nicolaides

A Rare Case of Altmark-compliant SGEI (Part II)

For a service to be in the general economic interest, it must be shown to fill a gap in the market or to offer what the market fails to provide adequately. [In case you have missed part I, you can access it here.] Introduction Services of general economic interest [SGEI] are important for the functioning of European societies. However, the […]
01. Dec 2020
State Aid Uncovered by Phedon Nicolaides

A Rare Case of Altmark-compliant SGEI (Part I)

For a service to be in the general economic interest, it must be shown to fill a gap in the market or to offer what the market fails to provide adequately. Introduction Services of general economic interest [SGEI] are important for the functioning of European societies. However, the granting of compensation to an undertaking for the provision of SGEI must […]
15. Sep 2020
State Aid Uncovered by Phedon Nicolaides

Non-economic Activities and Services of General Economic Interest

Non-economic tasks and economic activities which are inseparable from those non-economic tasks are together, as a bundle, non-economic in nature. Public compensation for the extra costs of services of general economic interest [SGEI] may not cross-subsidise activities that fall outside the scope of the SGEI. Update on Temporary Framework: Number of approved and published covid-19 measures, as of 11 September […]
14. May 2020
Guest State Aid Blog by Wout De Cock
corona virus

Belgium and COVID-19: The European Commission Approves Several Belgian State Aid Measures

We are happy to share with you an update on the Covid-19 measures that have been approved in Belgium. Our guest auhor Wout De Cock is a PhD candidate at the Vrije Universiteit Brussel and part-time teaching assistant at the Katholieke Universiteit Leuven.* Introduction In issue 1/2020 of the European State Aid Law Quarterly, we concluded that the European Commission […]
07. Apr 2020
State Aid Uncovered by Phedon Nicolaides

EU Decisions Override National Decisions on Compensation for Public Service Obligations

The principle of res judicata cannot be used to avoid recovery of incompatible State aid.   An update on State aid measures to counter the impact of COVID-19 As of 3 April 2020, the Commission has approved 29 measures submitted by 16 Member States. Article 107(3)(b) is the legal basis for 26 of those measures, while Article 107(2)(b) is the […]
04. Feb 2020
State Aid Uncovered by Phedon Nicolaides

How to Compensate for the Extra Cost of Public Service Obligations: Step-by-Step Guidance on a Complex Calculation

The “net avoided cost methodology” takes into account only incremental costs. Introduction The two most difficult issues in the design of State aid measures to support services of general economic interest [SGEI] are the proper definition of the public service obligation [PSO] and the identification of the “counterfactual” on the basis of which the public service compensation [PSC] is calculated. […]
23. Jan 2020
State Aid Uncovered by Phedon Nicolaides

Public Subsidies to Households Can Be State aid

Support for individual borrowers can be indirect State aid to banks.   Introduction The fact that a public measure has social objectives and aims to help poor households or disadvantaged persons instead of undertakings does not necessarily remove it from the reach of State aid rules. During the past decade, in response to the economic crisis, a number of Member […]
17. Dec 2019
State Aid Uncovered by Phedon Nicolaides

Can a Tax (rather than a Tax Exemption) Confer a Selective Advantage?

A tax that is levied at one level of government and does not apply to products and activities at a different level of government need not be selective. Introduction A tax exemption normally confers a selective advantage, unless it is justified by the logic of the tax. Counterintuitively, a tax itself can be selectively advantageous if its scope is too […]
30. Oct 2019
State Aid Uncovered by Phedon Nicolaides

The Operator of an Economic Infrastructure is an Undertaking

The assignment of public service tasks to an infrastructure operator is not sufficient to prevent the application of State aid rules. Introduction The characterisation of an entity as an undertaking is activity based, not status based. If it carries out activities for which there is a market, it becomes an undertaking, regardless of providing services which are in the general […]