JUDGMENT OF THE GENERAL COURT (Third Chamber)
13 July 2022 (*)
(State aid – Agriculture – Lease contract for agricultural land in Estonia – Decision declaring the aid incompatible with the internal market and ordering its recovery – Advantage – Determination of the market price – Private operator principle – Complex economic assessments – Judicial review – Taking into account of all relevant factors – Duty of diligence)
In Case T‑150/20,
Tartu Agro AS, established in Tartu (Estonia), represented by T. Järviste, T. Kaurov, M. Valberg and M. Peetsalu, lawyers,
applicant,
v
European Commission, represented by V. Bottka and E. Randvere, acting as Agents,
defendant,
THE GENERAL COURT (Third Chamber),
composed of G. De Baere, President, V. Kreuschitz (Rapporteur) and K. Kecsmár, Judges,
Registrar: S. Jund, Administrator,
having regard to the written part of the procedure,
further to the hearing on 13 January 2022,
gives the following
Judgment
1 By its action based on Article 263 TFEU, the applicant, Tartu Agro AS, seeks the annulment of Commission Decision C(2020) 252 final of 24 January 2020 on State aid SA.39182 (2017/C) (ex 2017/NN) (ex 2014/CP) – Alleged illegal aid to AS Tartu Agro (‘the contested decision’).
Background to the dispute
The applicant and the leasing of the agricultural land
2 The applicant is a limited liability company, which, in 1997, succeeded Tartu Riigimajand, a State-owned agricultural property producing milk, meat and cereals. On 2 October 2001, the Republic of Estonia, in a restricted call for tenders, sold all of the applicant’s shares to OÜ Tartland, which merged with the applicant in 2002.
3 On 16 November 2000, following a restricted call for tenders (‘the call for tenders’), the applicant and the Republic of Estonia, represented by its Ministry of Agriculture, concluded a 25-year lease contract (‘the lease agreement’) for plots of agricultural land comprising a total of 3 089.17 ha located in the municipality of Tähtvere, in Tartu county, and owned by the Republic of Estonia (‘the lease at issue’).
4 The lease agreement provided for rent of 10 000 Estonian kroon (EEK) (approximately EUR 639) per annum, or EEK 3.24/ha (EUR 0.20/ha); the applicant also had to cover the costs of maintaining and improving the plots, including annual investments in the drainage systems amounting to at least EEK 400 000 (approximately EUR 25 565, or EUR 8.28/ha), the costs of maintaining the land and improving the soil quality, amounting to at least EEK 3 981 000 (around EUR 254 432), including the costs of crop protection (EEK 820 000), mineral and organic fertilisers (EEK 3 100 000), liming (EEK 20 000) and maintenance of the roadsides (EEK 41 000), and the payment of all taxes.
5 The lease agreement also included an amendment clause and was amended several times. Three of those amendments related to the increase in the annual rent to EEK 80 000 (approximately EUR 5 113) on 14 January 2005, to EEK 250 000 (EUR 15 978) on 21 March 2007, and to EEK 416 000 (EUR 26 626) or EEK 136/ha (EUR 8.69/ha) on 12 May 2009. Those rent increases took effect retroactively on 1 January 2005, 1 January 2007 and 1 January 2009, respectively.
The administrative procedure before the Commission
6 On 28 July 2014, the European Commission registered a complaint lodged on 24 July 2014 alleging that the Estonian Ministry of Rural Affairs (formerly the Estonian Ministry of Agriculture) had granted unlawful State aid to the applicant.
7 The Commission forwarded the complaint to the Estonian authorities on 14 August 2014 and requested them to submit information and comments. The Estonian authorities provided the requested information on 3 October 2014.
8 By decision of 27 February 2017, the Commission initiated the formal investigation procedure provided for in Article 108(2) TFEU in respect of the lease at issue (OJ 2017 C 103, p. 4; ‘the opening decision’), calling on interested third parties to submit their comments.
9 In the opening decision, the Commission stated that it could not exclude that the lease at issue might constitute State aid for the purposes of Article 107(1) TFEU. First, it was possible that the applicant had received an advantage, in particular in the form of rent below the market price, which appeared to be evidenced by an independent report submitted by the Estonian authorities. In that regard, the Commission also questioned the transparent, non-discriminatory and unconditional nature of the call for tenders. Second, the advantage was selective in that the agricultural land in question was leased only to the applicant. The Commission considered the other conditions of Article 107(1) TFEU to have been met.
10 On 21 April 2017, the Estonian authorities submitted their comments. The Commission also received comments from two interested third parties, namely the complainant and the applicant, which were subsequently forwarded to the Estonian authorities on 10 and 12 May 2017, respectively. The Estonian authorities filed their comments on the complainant’s comments on 28 June 2018.
11 On 11 and 19 June 2017, the complainant submitted additional comments with five annexes. On 3 July 2017, the Commission forwarded those comments to the Estonian authorities, which submitted their comments on 21 July 2017.
12 By letter of 30 August 2017, the applicant contacted the Commission, which replied on 11 September 2017. The complainant contacted the Commission by letters of 9 January 2018, 30 January 2019 and 14 July 2019, to which the Commission replied on 7 February 2018, 1 March 2019 and 17 July 2019, respectively.
13 On 7 February 2019, the Commission organised a teleconference with the Estonian authorities and sent them a request for further information on 15 February 2019, to which they replied on 17 April 2019.
14 On 24 January 2020, the Commission adopted the contested decision, concluding that the lease at issue met all the conditions of Article 107(1) TFEU and therefore constituted State aid (paragraph 153 of the contested decision).
Forms of order sought
15 The applicant claims that the Court should:
– annul the contested decision in its entirety;
– order the Commission to pay the costs.
16 The Commission contends that the Court should:
– dismiss the action;
– order the applicant to pay the costs.
Law
17 In support of the action, the applicant formally raises eight pleas in law. Since the first formally raised plea concerns solely admissibility considerations, the other seven formally raised pleas should be examined.
18 In essence, the applicant submits that the Commission committed (‘manifest’) errors of law, procedure and fact vitiating its assessment relating to (i) the compliance of the call for tenders with the requirements guaranteeing market conditions; (ii) the consistency of the rent provided for in the lease agreement with the market price; (iii) the determination of the amount of the advantage; (iv) the classification as new aid; (v) the obligation to repay a part of the aid; (vi) the obligation of the Republic of Estonia to recover the aid; (vii) the incompatibility of the aid with the internal market.
19 The General Court considers it appropriate to begin by reviewing the legality of the contested decision as regards the assessment of the conformity of the rent provided for in the lease agreement with the market price and the determination of the amount of the advantage.
20 By its third and fourth formally raised pleas in law, the applicant disputes, in essence, that the condition relating to the existence of an economic advantage, as provided for in Article 107(1) TFEU, is met in the present case. More specifically, it submits that the Commission incorrectly concluded, in paragraphs 131 to 148 of the contested decision, that the lease at issue conferred an advantage on the applicant on the ground that the rent paid by the latter was below the market price during the period from 2000 to 2017 and, in paragraphs 154 to 165 of that decision, that that advantage corresponded to the difference between the average amount resulting from the annual estimates of the market rate for the rent of the land in question and the amount of the actual rent paid for that land, plus half of the amount of the annual investments made in the drainage system and the land tax paid on behalf of the owner.
21 More specifically, the applicant states that the Uus Maa report, on which the Commission relied for the period from 2000 to 2014, is unreliable since, first, it sets out price ranges for the rent in each period, in which the maximum amount is sometimes more than twice the minimum amount, and, second, the actual prices can vary by 20% from the amount indicated. It criticises the fact that the Commission determined the market price on the basis of the arithmetic mean of the minimum and maximum values of those price ranges. According to the applicant, both the minimum and the maximum amount within a price range, and even rent varying by 20% from that amount, can be considered to be in line with the market price. By relying on the arithmetic mean of that imprecise appraisal, the Commission failed to comply with its obligation to analyse and state reasons. Data from Statistics Estonia, used as a basis for comparison for the period from 2015 to 2017, do not constitute an expert assessment and do not take into account the specific characteristics of the land in question.
22 The applicant submits that, in order to determine whether or not the Estonian State adopted the conduct of a private operator, the Commission had to take account of the situation and economic considerations at the time the lease agreement was concluded in 2000. In its submission, at that stage, the accession of the Republic of Estonia to the European Union was uncertain, the market for agricultural land was depressed, and the value of land and rental prices were low.
23 The applicant submits that all investments in land improvements, as well as expenditure for maintaining the land and other qualitative improvements, as provided for in the lease agreement, should be included in the rental price. According to the Uus Maa report, neither imposing such obligations nor concluding long-term lease agreements covering such a large area were standard practice in 2000. Therefore, there are no comparable contracts and the rental price necessarily includes the value of those obligations.
24 According to the applicant, the Commission has failed to comply with its obligations as regards both the burden of proof and the procedural requirements. In particular, the applicant claims that the Commission misinterpreted the facts and economic data and that it assessed the evidence in a selective manner. According to the Uus Maa report, the land south of Tartu required land improvement systems at an additional cost, the construction of such systems was an investment in the owner’s assets, the obligations as regards soil fertility were not standard practice, no new lease agreement on the market had been as favourable to the owner as the one at issue, private landlords had a lower actual annual net income due to income tax and land taxes, and the lease agreement was not comparable to an ordinary lease agreement.
25 The applicant submits that the Commission failed to take account of the fact that, according to the data set out in the Uus Maa report, the utilisable area of the land in question represented only 2 833.596 ha, or 83% of the total surface area of that land. In the applicant’s view, the actual rent paid was therefore 16.7% higher than that provided for in the lease agreement. In addition, a significant proportion of the land in question consisted of roads and other land that cannot be cultivated.
26 The applicant adds that the analysis by AS Pindi Kinnisvara (‘the Pindi Kinnisvara analysis’), which was submitted by the Estonian authorities during the administrative procedure and which the Commission failed to take into account, also shows that the rent was in line with market conditions.
27 The Commission disputes the applicant’s arguments.
28 The Commission submits that, before making any adjustments, it compared the rent paid by the applicant in the years in question with the average annual rent following from the Uus Maa report and the data from Statistics Estonia. That report provided plausible estimates of rents charged in Tartu county during the period from 2000 to 2014. The Commission states that it relied on the average rent estimate set out in that report and that it was customary to use such estimates and price ranges when making comparisons. The land in question was no different from the average land in Estonia and a comparison with the lowest price in those price ranges, as proposed by the applicant, underestimated the advantage. The Commission thus submits that it adopted a prudent and balanced approach by relying on the average rents and objective data. Furthermore, the report was produced by the Estonian authorities and the applicant was unable to provide more reliable evidence.
29 The Commission states that the data from Statistics Estonia, as used for the period from 2015 to 2017, constitute reliable evidence. It claims that, in the contested decision, it explained that the data were not the result of an expert assessment of the rental price of the land, but reflected the average rental prices reported during that period. It states that those data were the only reliable information at its disposal and that the results obtained were consistent with those for the period from 2000 to 2014. Neither the applicant nor the Estonian authorities have demonstrated that the land in question has a lower value than the average value of agricultural land in Estonia. Nor does the applicant’s financial position indicate a lower value. The Commission adds that the fact that the statistical data cannot take into account all the specific characteristics of a given plot does not mean that it has used them incorrectly.
30 According to the Commission, it took into account the context at the time the lease agreement was concluded. It also submits that the applicant mentions no deficiencies in that respect in the contested decision. The Commission claims that, although the market prices were initially low, it showed, in the contested decision, that the applicant’s rent had remained below those prices throughout the period from 2000 to 2017, despite a series of increases.
31 The Commission disputes the applicant’s assertion that the rent should include all investments made for the land in question. Investments in improving and maintaining the land and improving the soil quality are operating expenses incurred during the annual agricultural production cycle for the benefit of the operator. The Commission states that the applicant did not find itself in any particular circumstances which had to be taken into account since, first, those characteristics were reflected in the rental market prices and, second, all agricultural land in Estonia required massive investment during the period in question. It submits that improving fertility (i) is inherent in the nature of an agricultural holding, (ii) does not show that the value of the land in question was significantly different from that of the average agricultural land in Estonia, and (iii) does not explain why the rent was below the market price. According to the Commission, a lower rent would only have been justified if the soil fertility had been less than the average fertility or if it had deteriorated. As for investments in the drainage system, it states that it acknowledged that the Estonian State, as the owner, had been relieved of certain maintenance costs and benefited from an increase in the value of the land. Since those investments exceeded the legal requirements and benefited also the applicant, the Commission submits that taking into account half of their amount represents a prudent estimate. According to it, that also demonstrates that it took into account improvements to the land, in particular soil fertility.
32 The Commission states that the burden of proof argument is based on unfounded and, accordingly, ineffective allegations. It submits that the Pindi Kinnisvara analysis constitutes an ex-post evaluation which was not presented during the administrative procedure and which cannot undermine the evidential value of the Uus Maa report, as presented by the Estonian authorities in the same procedure and based on data for the relevant period. By contrast, in the rejoinder, it states that it was not the applicant but those authorities that provided the analysis during that procedure. It states that, since the analysis was commissioned by the applicant, those authorities were asked to provide an independent assessment, namely the Uus Maa report. Therefore, it was no longer necessary to take a position on that partial analysis, which was considered to be withdrawn by the Estonian authorities. At the hearing, in response to oral questions put by the General Court, the Commission explained that from the Estonian authorities’ letter of 7 October 2015 and subsequent exchanges with them, the Commission had concluded that the analysis in question had been withdrawn. According to the Commission, that analysis was thus no longer included in the administrative file at the time the contested decision was adopted and it was not necessary to refer to it in that decision. The General Court took note of those explanations in the minutes of the hearing.
33 The Commission submits that, since the assessment of the advantage is a complex economic assessment, it has discretion as to the use of the most appropriate evidence and method of analysis. As for the adjustments made to the initial results of the comparison, it states that it explained its approach and relied on the evidence to make only those adjustments that were necessary. The applicant has failed to prove that any of those actions were inappropriate, nor has it demonstrated the existence of a manifest error of assessment.
34 The Commission adds that, in 2000, the applicant received agricultural aid covering 2 912.76 ha, which shows the large area of the land comprising various different types of plots with different values. For such land, an average price would thus be representative.
35 It follows from settled case-law that measures which, whatever their form, are likely directly or indirectly to benefit undertakings or must be regarded as an economic advantage which the recipient undertaking would not have obtained under normal market conditions are regarded as State aid (see judgment of 17 September 2020, Compagnie des pêches de Saint-Malo, C‑212/19, EU:C:2020:726, paragraph 39 and the case-law cited).
36 According to settled case-law, the supply of goods or services on preferential terms is capable of constituting State aid for the purposes of Article 107(1) TFEU (see judgments of 11 July 1996, SFEI and Others, C‑39/94, EU:C:1996:285, paragraph 59 and the case-law cited; of 1 July 2010, ThyssenKrupp Acciai Speciali Terni v Commission, T‑62/08, EU:T:2010:268, paragraph 57 and the case-law cited; and of 28 February 2012, Land Burgenland v Commission, T‑268/08 and T‑281/08, EU:T:2012:90, paragraph 47 and the case-law cited).
37 In the case of land leased at an allegedly preferential price, a transaction which is comparable to the sale of land by a public authority to an undertaking, it must be ascertained whether the price paid by the presumed recipient of the aid corresponds to a price it would not have obtained under normal market conditions. In those circumstances, the amount of the aid is equal to the difference between what the recipient actually paid and what it should have paid at the time under normal market conditions (see, to that effect, judgment of 9 December 2015, Greece and Ellinikos Chrysos v Commission, T‑233/11 and T‑262/11, EU:T:2015:948, paragraph 79 and the case-law cited).
38 Therefore, the Commission had to examine whether the rent paid by the applicant under the lease at issue corresponded to normal market conditions.
39 In that context, the Commission had to make complex economic assessments. In the context of a review by the Courts of the European Union of complex economic assessments made by the Commission in the field of State aid, it is not for those Courts to substitute their own economic assessment for that of the Commission (judgments of 22 November 2007, Spain v Lenzing, C‑525/04 P, EU:C:2007:698, paragraph 57, and of 2 September 2010, Commission v Scott, C‑290/07 P, EU:C:2010:480, paragraph 66; see also judgment of 10 December 2020, Comune di Milano v Commission, C‑160/19 P, EU:C:2020:1012, paragraph 100 and the case-law cited). Therefore, the review by the EU judicature is restricted, being limited to determining whether the rules governing procedure and the requirement for a statement of reasons have been complied with, whether the facts are accurately stated and whether there has been any manifest error of assessment or any misuse of powers (see, to that effect, judgment of 9 March 2017, Ellinikos Chrysos v Commission, C‑100/16 P, EU:C:2017:194, paragraphs 18 and 19 and the case-law cited).
40 However, this does not mean that the EU judicature must refrain from reviewing the Commission’s interpretation of information of economic nature. The European Union judicature must establish not only whether the evidence relied on was factually accurate, reliable and consistent, but also whether that evidence contained all the relevant information which must be taken into account in order to assess a complex situation and whether it was capable of substantiating the conclusions drawn from it (judgments of 22 November 2007, Spain v Lenzing, C‑525/04 P, EU:C:2007:698, paragraphs 56 and 57; of 2 September 2010, Commission v Scott, C‑290/07 P, EU:C:2010:480, paragraphs 64 and 65; and of 10 December 2020, Comune di Milano v Commission, C‑160/19 P, EU:C:2020:1012, paragraph 115). Moreover, where an EU institution has a wide discretion, the review of observance of certain procedural guarantees is of fundamental importance. Those guarantees include the obligation for the competent institution to examine carefully and impartially all the relevant elements of the individual case and to give an adequate statement of the reasons for its decision (judgment of 22 November 2007, Spain v Lenzing, C‑525/04 P, EU:C:2007:698, paragraph 58).
41 It is in the light of those principles that the merits of the applicant’s line of argument must be assessed.
42 In the present case, in order to determine whether the lease at issue was in line with market conditions, the Commission conducted a two-part examination. As a first step, it verified whether the amount of the rent in itself was below the market price (paragraphs 132 to 139 of the contested decision). As a second step, it examined whether that amount, plus the additional contractual obligations, provided that they formed part of the State’s rental income, remained below that market price (paragraphs 141 to 146 of that decision).
The conformity of the amount of the rent with the market price
43 In order to compare the amount of the rent with the market price, the Commission relied on the Uus Maa report and the data from Statistics Estonia, both of which were submitted by the Estonian authorities during the administrative procedure. It examined, on that basis, whether the amount of the rent that the applicant paid under the lease at issue, in itself, was at the level of the average amount of rent for agricultural land in Estonia throughout the period of the lease (paragraph 131 of the contested decision).
44 More specifically, first, the Commission noted, in paragraphs 134 to 136 of the contested decision, that, according to the Uus Maa report, the amount of rent for agricultural land in Tartu county was between EUR 6 and EUR 10/ha during the period from 2000 to 2004, between EUR 10 and EUR 20/ha during the period from 2005 to 2009, and between EUR 25 and EUR 60/ha during the period from 2010 to 2014. By contrast, in that first period, the applicant paid rent of EUR 0.20/ha; in that second period, its rent was increased, on 14 January 2005 and 21 March 2007 respectively, to EUR 1.66/ha and EUR 5.21/ha; and in that third period, that rent was increased, on 12 May 2009, to EUR 8.68/ha.
45 Second, the Commission stated, in paragraphs 137 to 139 of the contested decision, that, according to the data from Statistics Estonia, in 2015, the average amount of rent for agricultural land and arable land in Estonia was respectively EUR 52/ha and EUR 55/ha. In Tartu county, the amounts were respectively EUR 63/ha and EUR 65/ha. In 2016, those amounts were respectively EUR 52/ha and EUR 54/ha in Estonia and EUR 61/ha and EUR 61/ha in Tartu county. In 2017, those amounts were respectively EUR 58/ha and EUR 60/ha. There are no data for Tartu county for that year. By contrast, the applicant paid rent of EUR 26.86/ha in 2015, EUR 27.30/ha in 2016 and EUR 27.28/ha in 2017.
46 On the basis of that comparison, the Commission concluded, in paragraph 140 of the contested decision, that the amount of the rent in itself was below the market price throughout the period from 2000 to 2017. It applied a similar approach, in paragraphs 154 to 156 of that decision, for the calculation of the advantage, in respect of which it stated that the annual amount of the rent paid by the applicant should be compared with the average annual amounts resulting from the price ranges used in the Uus Maa report and the data from Statistics Estonia. It concluded, first, that that advantage consisted of the difference between those average amounts and that rent, and second, that the number of hectares for which agricultural support was paid had to serve as a base to calculate the advantage (paragraphs 140 and 165 of that decision).
47 In that respect, it must be noted that the expert reports drawn up by independent experts after the transaction in question – such as the Uus Maa report, which contains an expert assessment of the land and plots covered by the lease agreement (see paragraph 28 of the contested decision) – may, admittedly, be used to determine whether the rent paid by the applicant in respect of the lease at issue deviates sufficiently from the market rent to justify a finding that there is an advantage (see, to that effect, judgment of 16 September 2004, Valmont v Commission, T‑274/01, EU:T:2004:266, paragraph 45). Similarly, it is not ruled out that the Commission may take into account the average prices of rents reported by Statistics Estonia based on a methodology approved by the statistical office of the European Union (Eurostat) (see paragraph 39 of that decision), since methods other than expert evaluations can also provide prices corresponding to actual market values, to the extent that those methods provide for the updating of the prices where prices are rising sharply (see, to that effect, judgment of 16 December 2010, Seydaland Vereinigte Agrarbetriebe, C‑239/09, EU:C:2010:778, paragraphs 39 and 54), which is necessarily the case for statistical data.
48 However, in the contested decision, the Commission’s comparison of the rent paid by the applicant under the lease at issue with the average amounts from the Uus Maa report and the data from Statistics Estonia is too general and insufficiently nuanced to demonstrate that that rental price did not correspond to that which the applicant would have obtained under normal market conditions (see paragraphs 37 and 38 above). In particular, the comparison did not enable the Commission, in the absence of the margin for variation which must be tolerated, to substantiate, in a sufficiently convincing and coherent manner, the price that is as close as possible to the market value, as required by the case-law (see, to that effect, judgment of 16 December 2010, Seydaland Vereinigte Agrarbetriebe, C‑239/09, EU:C:2010:778, paragraphs 35 and 54). Moreover, in the context of that comparison, the Commission did not take into account all the relevant information available to it or that could have been available to it at the time when the contested decision was adopted (see, to that effect, judgments of 2 September 2010, Commission v Scott, C‑290/07 P, EU:C:2010:480, paragraph 91; of 20 September 2017, Commission v Frucona Košice, C‑300/16 P, EU:C:2017:706, paragraph 70; and of 26 March 2020, Larko v Commission, C‑244/18 P, EU:C:2020:238, paragraphs 41 and 42).
49 This assessment is based on the following factors.
50 In the first place, the Uus Maa report sets out market prices in price ranges. According to pages 101 and 107 of that report, the amount of rent for arable land of average quality, situated in Tartu county, was between EUR 6 and EUR 10/ha during the period from 2000 to 2004, and between EUR 10 and EUR 20/ha during the period from 2005 to 2009 (see paragraphs 31, 32, 134 and 135 of the contested decision; see also paragraph 44 above). For the period from 2010 to 2014, page 101 of the report indicates a first range between EUR 25 and EUR 60/ha, while the table on page 107 thereof contains a second range with amounts between EUR 30 and EUR 50/ha. The Commission used the first range between EUR 25 and EUR 60/ha (see paragraphs 33, 136 and 156 of the contested decision; see also paragraph 44 above), for which the minimum amount of EUR 25/ha is lower than that of the second range, which is EUR 30/ha.
51 The Commission took the view, without offering more detailed explanations, that the arithmetic mean of the amounts included in the price ranges in question should be used as a basis for comparison with the rent paid by the applicant under the lease at issue, both in the context of the finding of the existence of an advantage and in its quantification (see paragraphs 131 and 154 to 156 of the contested decision; see also paragraphs 43 and 46 above). In its written submissions to the General Court and at the hearing, the Commission confirmed that it had relied on that arithmetic mean.
52 As the applicant correctly submits, the fact that the Uus Maa report expressed the market price in the form of price ranges necessarily means that a rent corresponding to the minimum amounts of those price ranges constitutes a rent which is also in line with the market price. This is a fortiori the case since the Commission itself acknowledges that, for property and land, it is customary to use estimates expressed in price ranges as a basis for valuation, which is why it has relied on the figures used in that report. Thus, in the absence of any specific justification to derogate from it, the Commission had, in principle, to conclude that rent corresponding to the minimum amounts of the price ranges adopted in the contested decision, namely EUR 6, EUR 10 and EUR 25/ha (see paragraph 50 above), was in line with that market price during the relevant periods.
53 Contrary to the assertions made in its written submissions and at the hearing, the Commission has failed to demonstrate that the arithmetic mean of the amounts in the price ranges in question was the most representative value and the most suitable calculation method for determining the market price, and that the use of that arithmetic mean was necessary to balance the results obtained over several years and make them more reliable, to avoid underestimating that price or to take account of the average quality of the land in question.
54 As stated in paragraphs 28 and 30 of the contested decision, the Uus Maa report gives an account of the expert assessment of the land and plots covered by the lease agreement and is based on a method consisting in comparing them with other properties with similar characteristics. That comparative method, as explained on page 106 of the report, is common practice for assessing the market price and provides the most reliable market price. The price ranges in question express the amounts of market rent for arable land which, like the land in question, is of average quality and is located in Tartu county (see paragraph 50 above). Therefore, unless otherwise indicated, the view can be taken that those price ranges provide the most credible and suitable estimates of the market rent for the land in question, which the Commission acknowledged in paragraph 133 of the contested decision by stating that ‘the Uus Maa report [provided] plausible, albeit conservative, estimates of the amounts of rent in Tartu county in the period from 2000 to 2014’.
55 Thus, in view of the fact that the price ranges in question, first, take into account the average quality of the land in question, and second, are already the result of a balancing exercise inasmuch as they cover four-year periods and thus contain only aggregated data, it has not been established that the use of the arithmetic mean of those price ranges is such as to provide more reliable results and to give a price which is as close as possible to the market price. On the contrary, such use is likely to lead to imprecise results and a significant overestimation of the market price.
56 For the same reasons, the wide margin of error of the Uus Maa report, as set out on page 107 of the report and noted in paragraph 38 of the contested decision, according to which the actual rent on the market can vary by more or less 20% from the one referred to in the report, was also unable to justify the use of that arithmetic mean, which the Commission claimed neither in the contested decision nor during the proceedings. It is not apparent from any element of the analysis relating to the conformity of the amount of the rent with the market price, as set out in paragraphs 131 to 140 and 154 to 157 of the contested decision, that the Commission took into account that wide margin of error, notwithstanding its scale and its potential impact on the Commission’s assessment. As the applicant submits, it cannot be ruled out that a rent which is up to 20% lower than the lowest price ranges in the Uus Maa report can also constitute a market price.
57 Nor does using the arithmetic mean solve the problem of the price ranges of the Uus Maa report being particularly wide, as the applicant correctly states. The difference between the minimum and maximum amounts in a price range corresponds, for the period from 2000 to 2004 and the period from 2005 to 2009, to approximately 67% and 100% respectively. For the period from 2010 to 2014, for which the Uus Maa report contains differing price ranges (see paragraph 50 above), the difference corresponds to approximately 67% and 140% respectively, depending on the price range. It follows from this that the use of the arithmetic mean of those price ranges necessarily results in considerable imprecisions or even a significant overestimation of the market price.
58 Admittedly, the Commission clarified at the hearing, referring to paragraph 156 of the contested decision, that it had used the data from Statistics Estonia for the period from 2010 to 2014 to take into account the wide price ranges used in the Uus Maa report, which was noted in the minutes of the hearing. However, apart from the fact that that seems to be the case only for the quantification of the advantage in that paragraph and not for the finding of the existence of an advantage in paragraph 136 of the contested decision, in the context of which it compared the rents paid by the applicant with the average amounts included in those price ranges, the Commission thus did not attempt to solve the problem of the particularly wide price ranges for the periods from 2000 to 2004 and from 2005 to 2009.
59 It follows that the Commission has failed to demonstrate that the arithmetic means of the price ranges used in the Uus Maa report on which it relied were capable of determining a price as close as possible to the market price. Nor has it taken sufficient account of the margin of error and the particularly wide price ranges in that report. On the contrary, in view of the above considerations, that approach must necessarily lead to an overestimation of that price.
60 In the second place, regarding the context at the time the lease agreement was concluded, page 101 of the Uus Maa report states that, before the accession of the Republic of Estonia to the European Union in 2004, ‘there were few leases for arable land’, ‘the leasing of land was not very common, lease agreements were rarely concluded’, ‘land was made available free of charge to prevent it from being abandoned’, ‘the value of the land was low, rents were very low and lease agreements were only signed with relatively large landowners’ (see also paragraph 31 of the contested decision). As the applicant submits, the Commission did not take those factors into account in its examination.
61 Having regard to the context in which the lease agreement was concluded, it cannot be ruled out that, in 2000, a rent of EUR 0.20/ha – such as that provided for in that agreement (see paragraph 4 above) – was in line with normal market conditions. This is a fortiori the case since the aim of the agreement – as is clear from page 106 of the Uus Maa report, and in view of the extensive additional obligations it imposes, particularly for the maintenance of land and for improving the soil quality (see paragraph 4 above) – is to preserve the use of the land in accordance with its designated purpose and to maintain soil fertility.
62 It is also apparent from the Uus Maa report, as set out in paragraph 31 of the contested decision, that, in the period from 2000 to 2004, when the lease agreement was concluded (see paragraph 3 above), leases were worded in general terms and did not include a rent increase clause. That suggests that, at the time the agreement was concluded, a private operator, in an arm’s length transaction, whose conduct is one of the factors that the Commission is required to take into account for the purposes of establishing the existence of aid (see, to that effect, judgment of 10 December 2020, Comune di Milano v Commission, C‑160/19 P, EU:C:2020:1012, paragraph 109 and the case-law cited), would not have necessarily required, in a similar situation, a contractual provision providing for an automatic or unilateral increase in rent.
63 Thus, as regards the lease agreement, the Commission did not assess, in the contested decision, whether that agreement contained such a provision. Nor did it assess subsequent amendments to that agreement, but merely noted, in paragraph 27 of the contested decision, that the ‘amount of the rent was adapted thrice’. Therefore, it has not determined whether the Estonian State had, pursuant to that same agreement in its original version and the subsequent amendments thereto, the option of increasing the rent automatically or unilaterally each year so as to adapt it to the market rent.
64 In the absence of such a contractual option, the Commission could not simply compare the rent paid by the applicant under the lease at issue with the market price for the entire period from 2000 to 2017, but had to examine whether, at the time the lease agreement was concluded – namely in 2000 (see paragraph 3 above) – and on each amendment to that agreement, in 2005, 2007 and 2009 (see paragraph 5 above), the rent was below a price that the applicant would have obtained in normal market conditions. It follows from this that, also for that reason, the Commission could not merely use the arithmetic mean of the amounts included in the price ranges set out in the Uus Maa report, which were expressed in four-year periods and did not enable the market price to be determined for a given year. Furthermore, although it correctly stated, in paragraphs 135 and 136 of the contested decision, that the rent increases took place respectively on 14 January 2005, 21 March 2007 and 12 May 2009, it failed to take account of the fact that those increases took effect retroactively on 1 January 2005, 1 January 2007 and 1 January 2009 respectively (see paragraph 5 above).
65 In the third place, as regards the size and use of the land in question, the Commission stated that the lease agreement did not specify that certain parts of the land were unsuitable for agriculture (paragraph 157 of the contested decision). It considered that the non-arable land could be useful in other ways, that it was difficult to find reliable indices for the market price for renting forest land, grassland and other land and that, consequently, the number of hectares for which agricultural support was claimed from the European Union, which represented between 95% and 97% of the land in question during the period from 2004 to 2018, should serve as a basis to calculate the advantage (paragraph 158 of the contested decision).
66 In that respect, the Uus Maa report contains, first, on pages 9 and 10 thereof, a table summarising the amendments to the lease agreement, including concerning the size of the land in question, and second, on pages 12 to 99 thereof, a section on a detailed description of each property and each plot of land covered by the lease agreement, including their size, qualitative distribution, use, characteristics, fertility, any restrictions and the types of access. That table and that detailed description were intended to allow the closest rental price to the market price to be determined on the basis of the size and specific characteristics of the land in question.
67 Thus, that enabled the applicant to calculate, in the table produced as Annex A.22.12 to the application, the exact proportion of arable land, which, as it correctly claims, amounted to 2 833.596 ha, which corresponds to approximately 83% of the land in question.
68 For the remaining part of the land, that allowed the exact proportions of natural grassland, forests, residential land, farms, submerged land or other land to be calculated and assessed in the light of the information from the Uus Maa report on the development of the rental market in Tartu county. Thus, the Commission should have considered the fact, set out on page 101 of that report, that no rent was paid for the ‘other land’ composed of peat land, polder land and floodplain land in the periods from 2000 to 2004 and from 2005 to 2009, and that it was only in the period from 2010 to 2014 that a rental market emerged for that other land, for which the rent was significantly lower than for arable land inasmuch as it amounted to around EUR 10/ha. It should also have taken into consideration the fact that, in 2013, as stated on page 102 of that report in support of the data from Statistics Estonia, ‘in general, the price of arable land [was] higher than the price of permanent grassland’. In addition, the detailed description of the land in question in that report would have allowed the Commission to take into account that, for small plots, lower rents were paid during the period from 2005 to 2009 (page 101 of the report in question).
69 By contrast, in determining the size and use of the land in question, the Commission did not take into account the relevant information from the Uus Maa report, on which it nevertheless based a substantial part of its assessment of the conformity of the rent paid by the applicant under the lease at issue with the market price for the entire period from 2000 to 2017 (see paragraphs 131 to 136 and 154 to 156 of the contested decision; see also paragraph 44 above), and to which the applicant explicitly referred during the administrative procedure to demonstrate that 16.7% of the land in question could not be used for agricultural production purposes (paragraph 84 of the contested decision). In view of the importance of that report for determining the size and specific characteristics of the land in question, the Commission was not exonerated from assessing the evidential value of that information (see, to that effect and by analogy, judgment of 16 September 2004, Valmont v Commission, T‑274/01, EU:T:2004:266, paragraph 53) and from taking it into account in its examination. Therefore, it could not merely use, as it appears to believe, an average price for all the plots and their specific characteristics, taken together. Nor could the Commission merely use information derived from agricultural aid, for which it specified neither the methods of calculation nor the conditions under which it was granted. Therefore, since it ignored the specific characteristics of the land in question and the fact that arable land made up only 83% of that land, it inevitably overestimated the market price of that land.
70 In the fourth place, the Commission relied on the data from Statistics Estonia to determine the market price for the period from 2015 to 2017 (see paragraphs 137 to 139 of the contested decision) and, according to the Commission, for the period from 2010 to 2014, in order to correct ‘the wide range of amounts included in the Uus Maa report’ (see paragraph 58 above).
71 In that respect, it must be pointed out that, as the Commission itself acknowledges, the data from Statistics Estonia do not constitute an expert assessment of the rental price of the land, but are average prices for the rent of agricultural land which do not take into account the specific characteristics of the land in question (paragraph 133 of the contested decision).
72 In the absence of price ranges, such as those used in the Uus Maa report, it is in particular not possible to ascertain the lowest market price and to verify whether the rent paid by the applicant under the lease at issue differs from it. Nor is it possible to compare that rent with the market price in Tartu county in the years from 2010 to 2012 and in 2017, since for those years, there are only data for Estonia as a whole (see paragraph 39 of the contested decision). Added to this are the imprecisions on account of the fact that, for the period from 2015 to 2017, the Commission compared – in paragraphs 137 to 139 of the contested decision – that rent only with the average amounts for rent of agricultural land and arable land, without taking account of the fact that, as is apparent from the detailed description in that report (see also paragraphs 66 to 69 above), a part of the land in question consisted of grassland, for which the data from Statistics Estonia tend to indicate lower amounts (see paragraph 39 of the contested decision).
73 It follows that the Commission has failed to demonstrate that the data from Statistics Estonia on which it relied enabled it to calculate the price that is as close as possible to the market value and to establish that the applicant paid, for the lease at issue, a rent below that price during the periods from 2010 to 2014 and from 2015 to 2017.
74 Even if the Commission did not have better information at its disposal, as it seems to maintain, it could not simply rely on the data from Statistics Estonia to substantiate its conclusions.
75 In view of its central and exclusive responsibility to ensure, subject to review by the EU courts, observance of Article 107 TFEU and the implementation of Article 108 TFEU, the Commission is obliged to verify, with the help of experts where appropriate, whether a State measure involves an advantage which does not correspond to normal market conditions (see, to that effect, judgments of 16 September 2004, Valmont v Commission, T‑274/01, EU:T:2004:266, paragraph 72 and the case-law cited; of 9 December 2015, Greece and Ellinikos Chrysos v Commission, T‑233/11, EU:T:2015:948, paragraph 91; and of 16 March 2016, Frucona Košice v Commission, T‑103/14, EU:T:2016:152, paragraphs 164 to 179). In the present case, the Commission neither sought the help of experts nor requested additional information from the Estonian authorities for the periods for which, in its view, the Uus Maa report could not serve as an exclusive basis for comparison.
76 Moreover, it must be borne in mind that, in view of the detailed challenge by the applicant and those authorities, the burden of proving the existence of ‘State aid’ for the purposes of Article 107(1) TFEU and, accordingly, also of proving that the condition of granting an advantage to the beneficiaries is fulfilled, lies with the Commission (see judgment of 4 March 2021, Commission v Fútbol Club Barcelona, C‑362/19 P, EU:C:2021:169, paragraph 62 and the case-law cited). Even supposing that the Commission were faced with a Member State which does not fulfil its duty to cooperate and has not provided the Commission with all the information requested, it must base its decisions on reliable and coherent evidence which provides a sufficient basis for concluding that an undertaking has benefited from an advantage amounting to State aid and which, therefore, supports the conclusions which it arrives at. Given that the aim of the recovery of the aid at issue from the beneficiary is to eliminate the distortion of competition brought about by a certain competitive advantage and, thus, to re-establish the status quo before the aid was granted, the Commission cannot assume that an undertaking has benefited from an advantage constituting State aid solely on the basis of a negative presumption, based on a lack of information enabling the contrary to be found, if there is no other evidence capable of positively establishing the actual existence of such an advantage (see, to that effect, judgment of 26 March 2020, Larko v Commission, C‑244/18 P, EU:C:2020:238, paragraphs 69 and 70 and the case-law cited). Therefore, the Commission cannot justify the imprecisions in the data from Statistics Estonia on the sole ground that the applicant and the Estonian authorities have failed to demonstrate that the value of the land in question was lower than that of the average agricultural land in Estonia.
77 In the fifth place, it must be borne in mind that, during the administrative procedure, the Estonian authorities produced not only the Uus Maa report and the data from Statistics Estonia, but also, by letter of 2 October 2014, the Pindi Kinnisvara analysis, to which, however, no reference is made in the contested decision.
78 According to the Pindi Kinnisvara analysis, which is dated 12 April 2013 and comprises a single page, rental fees in Tartu county on that date predominantly ranged between EUR 30 and EUR 50/ha for existing agreements and between EUR 50 and EUR 100/ha for new or extended agreements; they could exceed EUR 100/ha in areas with good logistics and stronger competition. Those figures tally with those from the Uus Maa report, which indicates, for the period from 2010 to 2014, according to the respective price ranges, amounts between EUR 25 and EUR 60/ha and EUR 30 and EUR 50/ha (see paragraph 50 above), and with the data from Statistics Estonia, according to which the average amount for rent of agricultural land in that county was EUR 50/ha in 2013 (see paragraph 39 of the contested decision).
79 As the applicant correctly submits, the Commission failed to take into account the Pindi Kinnisvara analysis in the contested decision. In its defence, the Commission stated that that was an ‘ex-post evaluation which [had] not [been] submitted to the Commission during the administrative procedure’. It was only in its rejoinder that the Commission acknowledged that the Estonian authorities had submitted that analysis to it during the administrative procedure, while claiming that it was not necessary to take a position on that analysis on the ground that ‘the Estonian authorities [had] considered it to be withdrawn and replaced by the Uus Maa report’. At the hearing, in response to an oral question put by the General Court, the Commission stated, referring to footnote 6 to the rejoinder, that it had, on account of the Estonian authorities’ letter of 7 October 2015 and subsequent exchanges with those authorities, considered that analysis to be withdrawn. As a result, it was no longer part of the administrative file at the time the contested decision was adopted. Those statements were noted in the minutes of the hearing (see also paragraph 32 above).
80 In that respect, it should be noted that, by letter of 7 September 2015, bearing reference agri.ddg4.i.2(2015)4096993, the Commission inter alia requested the Estonian authorities to produce an assessment, by an independent expert, of the rental prices for comparable land to the land in question. Those authorities replied to that request by letter of 7 October 2015, the relevant section of which, as reproduced in footnote 6 to the rejoinder, reads as follows: ‘we note that, in order to dispel any doubt about the expert, we have launched a call for tenders, reference number 167431 and, after the award and performance of the contract, we will probably be able to provide you with the requested expert’s report’. By letter of 16 December 2015, the Republic of Estonia submitted the Uus Maa report to the Commission. That letter is accompanied by the following comment, as reproduced in footnote 6 to the rejoinder: ‘Please find enclosed the expert’s report requested in paragraph 5 of your letter (reference agri.ddg4.i.2(2015)4096993)’.
81 It follows from this that neither the letter from the Estonian authorities of 7 October 2015 nor the letter from those authorities of 16 December 2015 refers to the Pindi Kinnisvara analysis, but merely responds to the Commission’s request of 7 September 2015 to provide an independent expert assessment of rental prices for comparable land to that in the use of the applicant, a request complied with by the submission of the Uus Maa report. Contrary to what the Commission claims, it cannot be inferred from those letters that that analysis was withdrawn by those authorities, but only that it was supplemented by that report.
82 Consequently, although the Commission based its assessment of the Uus Maa report on the ground that the Pindi Kinnisvara analysis had been commissioned by the applicant, the fact remains that that analysis was still available to it at the time the contested decision was adopted. In those circumstances, the Commission was not exonerated from assessing the evidential value of that analysis and, where appropriate, from taking it into account in its examination (see, to that effect and by analogy, judgment of 16 September 2004, Valmont v Commission, T‑274/01, EU:T:2004:266, paragraph 53).
83 Moreover, it should be borne in mind that the Commission is required, in the interests of sound administration of the fundamental rules of the FEU Treaty relating to State aid, to conduct a diligent and impartial examination of the contested measures, so that it has at its disposal, when adopting the final decision, the most complete and reliable information possible for that purpose (judgments of 2 September 2010, Commission v Scott, C‑290/07 P, EU:C:2010:480, paragraph 90, and of 26 March 2020, Larko v Commission, C‑244/18 P, EU:C:2020:238, paragraph 67). In particular, it is obliged to examine carefully and impartially all relevant elements submitted to it. Having regard to that duty of diligence, which is, in accordance with that case-law, a necessary prerequisite to enable the EU Courts to ascertain whether the elements of fact and of law, on which the exercise of the Commission’s broad discretion depends, were present (see, to that effect, judgments of 21 November 1991, Technische Universität München, C‑269/90, EU:C:1991:438, paragraph 14; of 20 September 2007, Fachvereinigung Mineralfaserindustrie v Commission, T‑375/03, not published, EU:T:2007:293, paragraph 90; and of 16 September 2013, ATC and Others v Commission, T‑333/10, EU:T:2013:451, paragraph 84), during its examination the Commission could not disregard the Pindi Kinnisvara analysis, the relevance of which was recalled in paragraph 78 above.
84 In the light of all the foregoing, it must be concluded that, in the context of the comparison of the amount of the rent to the market price and the calculation of the advantage, the Commission did not give sufficient consideration, in particular, to the minimum amounts included in the price ranges set out in the Uus Maa report and their margin of error, to the context at the time of the conclusion of the lease agreement, including the initial absence of rent increase clauses, to the information in that report on the size and use of the land in question and to the imprecisions in the data from Statistics Estonia. Furthermore, in that context, it erred in failing to take into account the Pindi Kinnisvara analysis, which the Estonian authorities submitted to it during the administrative procedure.
85 Consequently, the examination of the conformity of the amount of the rent in itself with market conditions (paragraphs 132 to 140 of the contested decision) and the part of the assessment relating to the quantification of the advantage relating to that amount and to the area of the land in question (paragraphs 154 to 159 of the contested decision) are vitiated by a manifest error of assessment and a breach of the Commission’s duty of diligence.
The taking into account of the additional contractual obligations
86 It must be borne in mind that the lease agreement provided for, in addition to payment of the rent, additional contractual obligations which concerned (i) annual investments in the drainage systems, (ii) expenditures for the maintenance of the land and for improving the soil quality, including expenditures such as those for crop protection, mineral and organic fertilisers, liming and maintenance of roadsides; and (iii) payment of all taxes (see paragraph 4 above; see also paragraph 141 of the contested decision).
87 In the contested decision, the Commission agreed to take into account, as income of the Estonian State, only half of the annual investments in the drainage systems and the annual amounts of property tax paid by the applicant (see paragraphs 147 and 165 of the contested decision).
88 In essence, regarding the annual investments in the drainage systems, the Commission took the view, first, that the Estonian State was relieved of some maintenance costs that it would otherwise have had to bear as owner of the land. Second, in view of the long duration of the 25-year lease agreement, the applicant, as the user of the land, also benefited from those investments, which contributed to more efficient use of the land and the average amount of which, at EUR 91 163 per year, was neither required by law nor imposed by the State (paragraphs 143, 144 and 162 of the contested decision). As regards the expenditure incurred for the maintenance of the land and for improving the soil quality, amounting to approximately EUR 255 444, the Commission concluded that the expenditure was of interest to the applicant as lessee (paragraphs 145 and 163 of the contested decision). As regards the taxes, it found that the applicant had paid taxes on behalf of the Estonian State as the owner of the land, with the result that the State did not have its own expenses on that item (paragraphs 146 and 164 of the contested decision).
89 In the light of those considerations, the Commission concluded, in paragraph 147 of the contested decision, that, even if the amount of the rent was increased by half of the investments in the drainage systems and the taxes paid by the applicant, the rental income of the Estonian State remained lower than the market price throughout the period from 2000 to 2017.
90 In the first place, the Commission has not based its analysis of the additional contractual obligations on any independent expert assessment – unlike its analysis relating to the conformity of the amount of the rent with the market price (see paragraphs 131 to 136 and 154 to 156 of the contested decision; see also paragraphs 43 to 46 above). In particular, it referred neither to the Pindi Kinnisvara analysis nor to the Uus Maa report, which the Estonian authorities submitted to it during the administrative procedure (see paragraph 77 above).
91 First, with regard to the Pindi Kinnisvara analysis, it is stated therein that it is not customary, in lease agreements concluded for the purpose of making available agricultural land, to impose obligations on the lessee to make specific investments, so that such an exorbitant contractual obligation affects the rate of net rent in an inverse proportion to the amount of the contractual obligation per unit of leased land.
92 Second, in relation to the Uus Maa report, it is stated, on page 106 thereof, that its objective was, inter alia, ‘to establish whether, in lease agreements for agricultural land, it [was] customary to impose on lessees obligations relating to the construction and maintenance of land improvement systems as well as various fertilisation measures’. Thus, on the same page, that report contains an assessment of additional obligations customary in lease agreements and compares them with those of the lease agreement. The results of that appraisal are set out in paragraphs 34 to 37 of the contested decision, although the Commission did not take them into account in its assessment of the additional obligations set out in that agreement in paragraphs 143 to 146 and 161 to 164 of the contested decision.
93 It is apparent from this assessment of the Uus Maa report that, in the early 2000s, lease agreements were rather rudimentary, were expressed in general terms and no specific obligations were imposed on lessees (see also paragraph 34 of the contested decision). Later, essentially from 2005 onwards, the obligation to pay property tax and other taxes related to rent, notably the obligation to pay income tax, was added to lease agreements. The improvement of the land and the maintenance of the roadsides would have been, in those agreements, one of the lessee’s obligations (see also recital 35 of the contested decision). According to that assessment, the construction or replacement of drainage systems is an investment in the owner’s assets and means a significant burden in addition to paying rent. With the development of the leasing of agricultural land, clauses relating to the rational use of land and prudent agronomic practices were included in the agreements in question from 2010 to 2011 (see also paragraph 36 of the decision).
94 As regards the lease agreement, the Uus Maa report states that it was not customary to add to lease agreements obligations such as those imposed on the applicant, in particular in a monetarily measurable form. The length of the lease and the number of hectares leased out are also unusual. It infers from that that it was particularly important to preserve the use of the land for the designated purpose and to maintain its fertility, while noting that certain clauses of that agreement restrict the applicant’s freedom to choose which crops to grow (see also paragraph 37 of the contested decision).
95 The Commission failed to take into account the relevant information from the Pindi Kinnisvara analysis and from the Uus Maa report, despite the fact that it had that information at its disposal at the time the contested decision was adopted and that information concerned the determination of the value of the additional contractual obligations. In so far as, according to the case-law cited in paragraphs 82 and 83 above, it was not exonerated from assessing the evidential value of that information and from examining it diligently and impartially, the applicant’s argument that the Commission only selectively took account of the expert opinions is well-founded.
96 In particular, for the purpose of calculating the advantage, the Commission should have given more serious consideration – in the light of the relevant information from the Pindi Kinnisvara analysis and the Uus Maa report – to the value of the applicant’s investments and their proportion of the total rent. In that context, it failed in particular to give sufficient consideration to the investment in the drainage systems, which, according to that report, first, constitutes an investment in the owner’s assets, and second, represents a significant burden for the lessee in addition to paying rent (see paragraph 93 above).
97 In the second place, the conclusion in paragraphs 144 and 162 of the contested decision that half of the investments in the drainage systems had to be added to the State’s income is not based on any explicit and verifiable quantified transaction, but appears to be based on a vague estimate by the Commission. In so far as the Commission considered that those investments exceeded the contractual or legal requirements (paragraphs 143 and 144 of the contested decision), it failed to indicate which part complied with those requirements and which part exceeded them. Similarly, it could not take into account the fact that the applicant also benefited from those investments owing to the long duration of the 25-year lease (paragraph 144 of the contested decision), without indicating the period that exceeded that for which those investments could be accepted.
98 Therefore, the Commission’s estimate was not able to attribute, with sufficient accuracy, a value for the additional contractual obligations in question (see, to that effect and by analogy, judgment of 9 December 2015, Greece and Ellinikos Chrysos v Commission, T‑233/11, EU:T:2015:948, paragraph 131 and the case-law cited), and, consequently, did not enable the price that is as close as possible to the market value to be determined, as required by the case-law cited in paragraph 48 above.
99 Moreover, the fact that the applicant made investments exceeding those required by the lease agreement or by law does not, as such, show that those investments did not benefit the assets of the owner of the land in question. On the contrary, to the extent that that agreement expressly provides for minimum amounts for the additional contractual obligations (see paragraph 4 above), investments exceeding these minimum requirements are also likely to be in the interests of the lessor, namely the Estonian State.
100 In the third place, the Commission did not examine, in its analysis of the additional obligations taken into account (recitals 141 to 147 of the contested decision), whether a private operator, in an arm’s length transaction and in a situation as close as possible to that of the Estonian State, would have imposed the additional contractual obligations in question. According to the case-law cited in paragraph 62 above, the Commission, with which the burden of proof lies, was required to carry out such an examination, as was the case in similar cases (namely those which gave rise to the judgments of 28 February 2012, Land Burgenland v Commission (T‑268/08 and T‑281/08, EU:T:2012:90); of 28 February 2012, Grazer Wechselseitige Versicherung v Commission (T‑282/08, not published, EU:T:2012:91, paragraph 126); and of 22 May 2019, Real Madrid Club de Fútbol v Commission (T‑791/16, EU:T:2019:346)), in which the Commission examined whether a private operator had adopted the conduct of the Member State in question.
101 Given that a private operator is guided by prospects of profitability in the longer term (see, to that effect, judgments of 12 December 2000, Alitalia v Commission, T‑296/97, EU:T:2000:289, paragraph 84, and of 13 December 2018, Ryanair and Airport Marketing Services v Commission, T‑165/16, EU:T:2018:952, paragraph 249 and the case-law cited), and that it is apparent from the Uus Maa report that, first, at the time the lease agreement was concluded, the land was generally made available free of charge in order to prevent it from being abandoned (see paragraph 60 above) and, second, the lease agreement was intended to preserve use of the land for its designated purpose and to maintain its fertility (see paragraph 94 above), it is not excluded that such an operator would provide, as the applicant claims, for additional contractual obligations so as not to have to make the necessary investments itself, such as those in drainage systems, for the maintenance of the land and for improving soil quality, which, moreover, increases the value of land over the long term.
102 In the absence of an examination of those elements and a comparison of the economic rationality of the conduct of the Estonian State with that of a private operator in similar circumstances, the Commission could not reasonably consider, in paragraphs 147 and 165 of the contested decision, that the amount of the rent to be taken into account should be increased by only half of the investments in the drainage systems and the taxes paid by the applicant.
103 In the light of all the foregoing, it must be concluded that the Commission erred in failing to consider the total amount of investments in the drainage systems as an integral part of the rental income of the Estonian State.
104 It follows that, as regards the relevant points in time – namely in 2000, the year the lease agreement was concluded, in 2005, in 2007 and in 2009, when that agreement was amended to increase the rent (see paragraph 64 above) – the Commission was required to assess whether that rental income was in line with the market price, in particular in the light of the relevant information from the Uus Maa report.
105 Accordingly, the analysis relating to the taking into account in the rent of the additional obligations set out in the lease agreement (paragraphs 141 to 147 of the contested decision) and the part of the assessment relating to the quantification of the advantage which concerns those obligations (paragraphs 161 to 164 of the contested decision) are vitiated by manifest errors of assessment and a breach of the Commission’s duty of diligence.
Conclusion
106 It is appropriate to conclude, first, that the Commission breached its duty of diligence, and second, that both the examination of the conformity of the amount of the rent in itself with the market conditions (paragraphs 132 to 140 of the contested decision) and of the taking into account in the rent of the additional contractual obligations (paragraphs 141 to 147 of the contested decision) are vitiated by manifest errors of assessment which also affect the assessment relating to the quantification of the advantage (paragraphs 154 to 165 of the contested decision).
107 Accordingly, the third and fourth formally raised pleas in law must be upheld and the contested decision must be annulled in its entirety, without the need to rule on the other pleas in law formally raised by the applicant.
Costs
108 Under Article 134(1) of the Rules of Procedure of the General Court, the unsuccessful party is to be ordered to pay the costs if they have been applied for in the successful party’s pleadings. Since the Commission has been unsuccessful, it must be ordered to pay the costs, including those relating to the proceedings for interim measures, in accordance with the form of order sought by the applicant.
On those grounds,
THE GENERAL COURT (Third Chamber)
hereby:
1. Annuls Commission Decision C(2020) 252 final of 24 January 2020 on State aid SA.39182 (2017/C) (ex 2017/NN) (ex 2014/CP) – Alleged illegal aid to AS Tartu Agro;
2. Orders the European Commission to bear its own costs and to pay those incurred by Tartu Agro, including those relating to the proceedings for interim measures.
De Baere | Kreuschitz | Kecsmár |
Delivered in open court in Luxembourg on 13 July 2022.
[Signatures]
Table of contents