Aid must be capable of remedying the regional handicap.
A novel aspect of the 2014 State Aid Modernisation was the requirement for ex-post evaluation of large or unusual State aid measures. The purpose of the ex-post evaluation is to determine the effectiveness of State aid. The outcome of the evaluation does not affect the legality of the aid. It is supposed to influence the design of the successor measure. This makes sense. If State aid is proven to be ineffective, it wastes public money that can be put to better use elsewhere. Therefore, future State aid measures will have to be appropriately adjusted to take into account the lessons drawn from the ex-post evaluation.
Although it is hard to provide definitive proof, it is fair to say that when the Commission broached the idea of the ex-post evaluation, most Member States were not enthusiastic, while some of them were outright hostile to it. Yet, measuring the impact of any public policy is at the core of democratically accountable policy making and it is standard practice in several Member States. It appears, however, that the initial reluctance of Member States has declined and since 2014, the culture of ex-post evaluation has taken firm hold.
The latest statistics [May 2017] indicate that so far the Commission has approved the ex-post evaluation plans for 35 schemes accounting for a total of EUR 40 billion of annual expenditure.
|Schemes by Member State||Schemes by category|
|France, Poland, UK
Austria, Czech Rep, Finland, Hungary, Ireland, Lithuania, Spain
Source: Lexxion seminar on ex-post evaluation, May 2017. [Based on data from DG Competition]
Of the 35 plans, 34 were submitted in compliance with the Article 1(2) of the GBER which stipulates the following:
“This Regulation shall not apply to:
(a) schemes […], if the average annual State aid budget exceeds EUR 150 million, from six months after their entry into force. The Commission may decide that this Regulation shall continue to apply for a longer period to any of these aid schemes after having assessed the relevant evaluation plan notified by the Member State to the Commission, within 20 working days from the scheme’s entry into force”.
However, the annual budget of one of the 35 measures did not exceed the threshold of EUR 150 million. Lithuania designed an unusual scheme that focused on a particular sector. This article reviews the Lithuanian measure, examines its novelty and considers the reasons for which the Commission concluded it was compatible with the internal market. The Commission approved it in decision SA.42225. What makes this measure more unusual is that it was not implemented on the basis of the GBER but on the regional aid guidelines.
The main features of the Lithuanian measure
The measure was notified in June 2015, under the Regional Aid Guidelines [RAG]. The aid was granted in the form of an exemption from payment of electricity fee. The beneficiaries were undertakings investing in the development of technology infrastructure projects such as data centres.
The measure aimed at promoting regional development and employment by facilitating the establishment of data centres that could provide high-value data processing and host services which would result in job creation. That is, the aid supported initial investment in a specific type of facilities which could be used for data processing and provision of internet server services (hosting) and related activities. The investment has to take place at designated ITC sites with high reliability of electricity supply.
This was a sectoral measure and normally the RAG allows aid only for multi-sectoral schemes. Lithuania is an assisted area under Article 107(3)(a) TFEU. The maximum aid intensity for large undertakings is 25% gross grant equivalent (GGE).
The scheme was open to both large undertakings and small and medium-sized enterprises [SEMs] which carried out data processing and hosting services on designated sites, and who signed an investment agreement with the government. Lithuania expected that the number of beneficiaries would be no more than 10.
The aid would be granted on a discretionary basis, following a decision of the national authorities. The aid would be in the form of an exemption from payment of the part of the electricity bill which was indicated as the “public interest service fee”. This was the fee intended for public services for the portion of electric power consumed by investors that carried out data processing and provided internet server services (hosting) and with which the government of Lithuania has concluded an investment agreement. The “public interest service fee” is collected from electric consumers.
The amount of the public interest service fee is determined by the Energy Commission and amounts to about 18% of the electricity price. The aid amount could not exceed 25% of the present value of the tangible and intangible assets forming part of the initial investment.
According to the notification, the price of electricity in Lithuania is one of the highest in Europe, equalling EUR 0.986/kWh  whereas the average price of the electricity in the 28 Member States was only EUR 0.0754/KWh. In addition, energy consumption accounts for over 80% of the total operating expenditure of the data centres.
The overall maximum budget foreseen under the scheme in the period until 2020 was EUR 221 million, which was far below the GBER threshold of average annual EUR 150 million.
The promotion of ICT projects was in line with the Lithuanian government’s objective for regional development and with the targets of the Europe 2020 strategy for smart, sustainable and inclusive growth. It was also included in the operational programme for EU structural funds. The measure was expected to attract digital technologies which would reduce disparities between Lithuania and other countries, but also regional disparities within Lithuania.
With respect to jobs, it was expected that the measure would create 305 direct jobs and 243 indirect jobs per year over a period of 12 years.
Lithuania committed to comply with all the requirements of the RAG, to prevent unlawful cumulation with de minimis aid, to notify individually any projects with eligible costs exceeding EUR 100 million and to adjust the aid intensity ceiling for projects with eligible costs above EUR 50 million but below EUR 100 million [given Lithuania’s aid intensity rate of 25%, a project exceeding EUR 100 million would be individually notifiable if the amount of aid would be larger than EUR 18.75 million].
Given the unusual nature of the measure, Lithuania agreed to carry out an ex-post evaluation as required by point 144 of the RAG. This is the first measure to which point 144 applied [the ex-post evaluation that was planned for other regional aid measures was in conformity with Article 1(2) of the GBER].
The Commission had little difficulty finding that the notified measure contained State aid. After all, it is now well understood that exemption from payment of a part of the electricity bill results in transfer of state resources as, in this case, the public service interest fee is collected by network operators which then transfer the fee to a public authority. However, given the specificity of the measure, the assessment of its compatibility was a bit trickier.
Compatibility with the internal market
Contribution to regional objective and need for state intervention
“(60) According to recitals 31 and 33 of the RAG, regional aid schemes should form an integral part of a regional development strategy with clearly defined objectives and should be consistent with and contribute towards those objectives. For aid measures outside an operational programme financed from the cohesion policy funds, Member States can demonstrate that the measure contributes to the objectives of a regional development strategy based on evaluations of past State aid schemes, impact assessments made by the granting authorities or expert opinions.”
“(61) The notified aid scheme is confined to one area of activity, i.e. the beneficiaries of the aid are investors who carry out data processing, web servers (hosting) and related activities in technological development sites. As regards the need for these investment projects and their contribution to regional development, the Commission notes that the aid scheme contributes to the achievement of the strategic aim of the Investment Promotion and Industry Development Programme for 2014-2020 […] to attract investments and increase competitiveness with particular focus on data centres, shared services, manufacturing etc., improving connectivity to the European optic fibre infrastructure and attracting other potential investors.”
At this point, the Commission noted that “(63) European Structural Investment Funds (ESIF) normally do not co-finance fiscal or para-fiscal aid measures, such as the exemption of the public interest service fee. Furthermore, the notified measure is targeting all sizes of undertakings, whereas ESIF operations target SMEs only. Therefore, the notified scheme could not be co-financed under one of the ESIF operational programmes for the period 2014-2020, and this in spite of the fact that the aid scheme positively contributes to one of the pivotal objectives of the broader development strategy agreed with the Commission for the 2014-2020 programming period”. “(64) The Lithuanian authorities provided a study demonstrating that the measure aims to boost the overall economic activity and social development in Lithuania, and in particular to stimulate additional data centre investments, to reduce unemployment and to improve business climate in the most disadvantaged territory in which a data centre would be located.” “(65) Based on the experience of the Lithuanian authorities dealing with companies looking for the location of a data centre, the main factors that may affect the data-centre’s site location are: (1) availability, reliability, and cost of electricity, (2) telecommunication infrastructure, geographic location for a data centre in terms of (3) security and (4) climate and (5) logistics infrastructure. Lithuania scores well on all of the above listed criterion, except on the cost of electricity which is among the highest in Europe.” [Apparently data centres are located in colder climates to avoid overheating of their equipment which themselves generate heat.]
“(67) In view of above, the Commission considers that this is consistent and contributes towards the development strategy of Lithuania as an assisted area.” In addition, “(69) the Lithuanian authorities confirmed that they will request applicants to provide relevant information in the application proceeding allowing the granting authority to select the projects whose will contribute mostly towards the objective of the scheme and thus towards the development strategy of the region concerned.”
Appropriateness of the scheme
“(73) According to section 3.4 of the RAG, the notified aid measure must be an appropriate policy instrument to address the policy objective concerned. When introducing a scheme outside an operational programme financed from the cohesion policy funds, Member States must indicate why regional aid is an appropriate instrument to tackle the common objective of equity or cohesion.”
“(74) The main handicap of Lithuania as a location for data centres is the relatively high price of electricity, which is an important location factor for investment in this sector. The measure will directly remedy the competition distortion as the aid takes the form of an exemption from payment of the public interest service fee.” “(75) In view of above, the Commission accepts that the measure is an appropriate instrument to achieve the regional development objective in the region concerned.”
After explaining the nature and importance of the incentive effect, the Commission decision only mentioned that “(79) the Lithuanian authorities committed to carry out a credibility check of the counterfactual and confirm that regional aid has the required incentive effect. A counterfactual is credible if it is genuine and relates to the decision-making factors prevalent at the time of the decision by the beneficiary regarding the investment.”
“(81) Pursuant to recital 82 of the RAG, in case of schemes applicable to SMEs the maximum aid intensities from the regional aid map serve as safe harbours. As long as they are not exceeded the criterion “aid limited to the minimum” is deemed to be fulfilled and therefore the aid is proportional.”
“(83) The aid granted to large undertakings under the scheme will be awarded on the basis of the “net-extra cost approach”, calculated in accordance with recitals 79-80 of the RAG. The maximum aid intensities allowed as per the applicable regional aid map are used as a cap to the net-extra cost approach.”
Avoidance of undue negative effects
“(89) In line with the requirement set out in recital 120 of the RAG, the Commission notes, according to a study provided by the Lithuanian authorities, that the measure is addressed at undertakings operating in a global market (i.e. the data-centre market), which will continue to increase, with more than 60 new large data centres expected to be established in Western Europe by 202048.” “(90) The Commission notes that the aid granted under the measure is designed to remedy a very specific handicap of the Lithuanian economy, which undermines its attractiveness for investments in data-centres, i.e. its electricity price, which is significantly higher than the average of the 28 EU Member States. The nature and size of the aid is therefore designed to address a specific competitive disadvantage of the country for this specific sector. The measure is limited to the amount necessary to compensate for the additional cost of investing in Lithuania as compared to a counter-factual investment elsewhere. Moreover, data submitted by Lithuania show that the measure will have only a marginal impact on the geographical distribution of activities in the sector in the EU, as the Lithuania’s market share in terms of ICT/data processing and hosting technologies sector is insignificant. In 2013, Lithuanian market share was equal to 0.2% of the EU 28 market in terms of turnover under the NACE Code 63.11 Data processing, hosting and related activities. In addition, the Lithuanian authorities estimate that, as a result of the measure, Lithuania will attract only a relatively small share of the expected increase in the capacity of the sector, which would result in a relatively small increase in Lithuania’s market share in the sector (namely from 0.2% in 2012 to some 0.8% by 2020.”
The Commission concluded that there would be no undue distortion of competition. However, it should be noted that the explanation given by the Commission in paragraph 90 quoted above does not really correspond to the logic of preventing excessive distortions by not aiding firms which already have large market share.
When assessing the Lithuanian measure in the context of the regional aid guidelines, it looks a very well designed scheme, expertly presented and justified to the Commission. But when it is examined in a broader policy context, a number of troubling questions arise. First, if Lithuanian electricity is already very expensive by European standards, why has the government of Lithuania decided to incorporate in the price of electricity a public interest service fee? If the purpose of the fee is to pay for public services, a less distortionary tax should have been preferred.
Second, if the locational handicap of Lithuania is the high price of electricity, why not address this issue directly, either through investments in generating capacity or simply reduction of taxes on electricity?
Third, is high electricity price a real market failure? Electricity is a scarce resource, like any other resource. The fundamental role of prices is precisely to signal relative scarcity. Skewing the consumption of electricity in favour of data centres will simply penalise other sectors of the economy which consume electricity. Instead of inducing savings in electricity, because it is expensive, the tax exemption will encourage more consumption of the relatively scarce resource.
Fourth, will the ex post evaluation plan also examine the impact on the rest of the economy? This is a textbook case that demonstrates that showing that an aid measure is effective is not the whole story. It is also necessary to identify the cost to the rest of the society from the channelling of resources to the aided activity.
 The full text of the Commission decision can be accessed at: