Cost of capital = risk-free rate + capital risk premium

Cost of capital = risk-free rate + beta coefficient of the assets* [market risk premium]

Where the beta coefficient of the assets represents the volatility of the profitability of the assets in question compared with that of the market.

1. (a)

   the part of the capital supporting non-marketable risk insurance (first exclusion criterion);

2. (b)
   the part of the capital that was already supporting the insurance of marketable risks within the ONDD and which was simply transferred to Ducroire/Delcredere with the corresponding insurance business. This was solely a change in the legal form of a pre-existing economic activity together with the relevant capital. Based on the approach followed in the decision on the creation of the Banque Postale58, such a transfer of capital does not confer a new advantage on the economic activity in question and cannot therefore be considered aid as such (second exclusion criterion).

In line with the approach developed in the opening decision, the Commission ought to begin by applying these two exclusion criteria.

However, Belgium advanced an argument based on the second exclusion criterion which, if it were correct, would make it possible to exclude immediately any categorisation of the whole of Measure 3 as aid. The Commission will therefore demonstrate in the first instance that this argument is invalid, before applying the two exclusion criteria set out in the opening decision.

In response to the opening decision, Belgium invoked to its advantage and expanded on the second exclusion criterion according to which the capital which, within the ONDD, already supported the existing marketable business just before its transfer to Ducroire/Delcredere on 1 January 2005 does not constitute aid. More precisely, Belgium considers that all of the initial capital of EUR 150 million allocated to Ducroire/Delcredere constituted the economic capital associated with the short-term insurance business (marketable and non-marketable) transferred. It considers that the creation of Ducroire/Delcredere with its initial capital of EUR 150 million was simply a change in the legal form of an existing business and of the relevant capital. This meant that none of initial capital allocation constituted aid. Consequently, Belgium maintains that there is no need to verify how this capital was used subsequently by Ducroire/Delcredere (i.e. whether it was used for marketable or non-marketable risks — first exclusion criterion) given that it could not constitute aid.

The Commission therefore rejects Belgium's assertion that all of the initial capital of an amount of EUR 150 million constitutes the capital which was simply transferred and which was already allocated within the ONDD to the activity transferred and therefore cannot be considered an advantage. The Commission considers that the difference between the EUR [45-75] million and the capital actually subscribed to the tune of EUR 150 million therefore constitutes additional capital representing a new advantage for this activity.

In the alternative, the Commission would observe that the internal memos of 2004 which underpinned the ONDD's decision to grant Ducroire/Delcredere EUR 150 million of capital (greatly exceeding the EUR [45-75] million allocated to the commercial account) show that the ONDD considered different scenarios for the growth of business within the future Ducroire/Delcredere. The need for EUR 150 million of capital was therefore determined on the basis of hypotheses regarding the future growth of the activities of Ducroire/Delcredere. It can therefore be concluded that the approach of the Belgian authorities is akin to excluding from the scope of the analysis to determine the existence of an overall advantage the entirety of the capital which the new entity could be assumed to require in order to support its future development. This approach cannot be accepted. Only the capital from which the activities transferred within the existing legal entity already formally benefited (short-term insurance activities within the ONDD) does not constitute a new advantage since the activities in question already benefited therefrom. Conversely, any additional capital constitutes a new advantage since the activity in question did not benefit from this capital before the transfer. The decision to create Ducroire/Delcredere and allocate to it a capital of EUR 150 million cannot therefore be considered a simple change of legal form of an existing activity with a simple transfer of capital allocated to the activity transferred. Therefore, the Belgian authorities' assertion that from 2004 it could be anticipated that Ducroire/Delcredere would need a capital of EUR 150 million in 2009 on the basis of the expected growth of its business is irrelevant. The Belgian authorities do not in any way demonstrate that this capital of EUR 150 million was already allocated to short-term activities within the ONDD before their transfer to Ducroire/Delcredere and that this capital of EUR 150 million does not therefore constitute a new advantage for these activities.

1. (a)

   First, the financial projections existing in 2004 covered only the period 2005-07. Accordingly, even if the capital required to support the growth of business was a relevant criterion for the present analysis (which the Commission contests), it cannot be accepted that it be based on the expected level of activity in 2009 since this level of activity had not even been estimated in 2004.

2. (b)
   Secondly, the methodology of the Solvency II Directive accompanied by internal modelling for political risks, proposed by Belgium to estimate the capital required, cannot be accepted. This method was absent from the 2004 internal memos which formed the basis for the decision taken by the ONDD Board of Directors on 20 April 2004 to establish Ducroire/Delcredere and endow it with a certain amount of capital since this method was non-existent at the time (it is not due to enter into force until 1 January 2014). It was applied by the ONDD for the first time in 2011 and is still being developed. Consequently, the internal modelling for political risks has not yet been validated by the Belgian insurance regulator, contrary to the requirements of the Solvency II Directive63. The use of a methodology which did not exist in 2004, such as Solvency II, for which numerous calculation parameters and an implementation timetable have not yet been finalised (they are still at QIS stage) is inappropriate. To consider that the economic capital required by Ducroire/Delcredere when it was set up in 2005 to continue the activity pre-existing within the ONDD must be determined on the basis of the requirements of the Solvency II Directive which did not even exist at the time and for which the regulators would in any case have planned transitional arrangements, seems excessive. Moreover, to draw any other conclusion would be tantamount to contradicting the validation of the level of capital (solvency margin)64 of the commercial account (EUR [45-70] million) by the regulator, who at the time had considered this capital sufficient in the light of the prudential rules in force.

As for using the Cooke ratio to determine the capital required, the Commission criticised, in its opening decision, the appropriateness of applying banking rules (such as the Cooke ratio) to credit insurers, given the many differences between the risks assumed by credit insurers and by banks (see recital (90) of the opening decision). In this regard, Belgium has not proved that the Basel rules (Cooke ratio) — which, in law, apply exclusively to the banking sector — would in practice be applied by credit insurance bodies or be recommended by rating agencies or insurance supervisory bodies.

Regarding the application of Article 8 of the 1939 Law on the ONDD, the Belgian authorities have confirmed that this law does not apply to Ducroire/Delcredere. Moreover, this article concerns only the ONDD's own-account activities guaranteed by the State and activities for the account of the State.

Although Belgium refers to the Standard & Poor's method, which existed at the time to determine the capital requirements of credit insurers, it should be noted that the Belgian authorities have not submitted any capital estimate using this method. Nor was the method used by the ONDD in 2004.

In conclusion, even supposing that the capital required to support the expected growth of Ducroire/Delcredere's business was a relevant criterion for the present analysis (a fact contested by the Commission), the Belgian authorities have not provided any fresh evidence to demonstrate conclusively that the capital requirement relating to short-term risks transferred to Ducroire/Delcredere had to be greater than the existing capital on the ONDD commercial account at the end of 2004.

In this section, the Commission has demonstrated that it was necessary to reject the Belgian authorities' assertion that the allocation of an initial capital of EUR 150 million to Ducroire/Delcredere in 2004 constituted only the transfer of the capital which had already been provided to support the activity relating to short-term risks within the ONDD. The Commission has in fact concluded that the capital which already supported the activity relating to short-term risks within the ONDD before the transfer of this activity to Ducroire/Delcredere amounted to EUR [45-75] million and that therefore additional capital of EUR [75-100] million was allocated to this activity.

It is now appropriate to examine the application of the two exclusion criteria described in the opening decision. The second exclusion criterion will be dealt with before the first.

As also described above, the commercial account, which consisted of a capital of EUR [45-75] million at the end of 2004, did not contain separate accounts making it possible to identify precisely the capital supporting only the activity relating to marketable risks. It is therefore necessary to estimate the part of the EUR [45-75] million which can reasonably be regarded as supporting the activity relating to marketable risks.

The Commission also observes that the Belgian authorities and the ONDD/Ducroire/Delcredere consistently maintained during the procedure that non-marketable risks required, for a given sum insured, more capital than marketable risks. Therefore, the amount of EUR [10-25] million represents more of a ceiling than a floor. The Commission considers, however, that it is reasonable to retain this figure as it is in line with the approach followed by the ONDD itself in 2004.

By analogy, the same reasoning must be applied to the amount of the capital associated with risks relating to debtors located in Romania and Bulgaria. In 2007, non-marketable risks relating to debtors located in Romania and Bulgaria were recategorised as marketable risks owing to the accession of these States to the European Union. The capital which already supported these risks when they were not marketable cannot therefore constitute a fresh advantage merely on the ground that these risks became marketable subsequently. It is therefore appropriate to exclude this capital also from categorisation as state aid.

Therefore, the capital relating to non-marketable risks which became marketable is excluded from the analysis of the existence of an advantage. In their comments of 16 May 2012, the Belgian authorities estimated the capital associated with risks relating to debtors located in Romania and Bulgaria at EUR [0-2] million, based on the assumption that these risks were marketable in 2005. On the other hand, they did not provide an estimate of the capital associated with these risks on 31 December 2006. Since the risks in question were indeed not marketable in 2005, the Belgian authorities' estimate of the capital allocated to these risks is likely to be too low. The separate accounts recreated retrospectively by the ONDD as part of this procedure (comments of 16 May 2012), show that the capital allocated to non-marketable risks was EUR [0-5] million less in 2007 than in 2006. This decrease can be explained by the change in the categorisation of risks relating to Romania and Bulgaria (considered marketable risks from 2007), but also by other factors, such as the change in policy with regard to reinsurance. On the basis of all of the information available, it can reasonably be considered that the capital supporting risks relating to Bulgaria and Romania before they became marketable amounted to EUR [0-5] million.

In the light of the foregoing, the Commission concludes that an amount of EUR [10-25] million already supported marketable risks before 1 January 2005 and that the transfer of this capital with the activities in question cannot therefore constitute an advantage. Similarly, an amount of EUR [0-5] million already supported risks which became marketable on 1 January 2007 (Romania and Bulgaria) before they became marketable.

As described above, EUR [45-75] million of capital was already supporting short-term risks before their transfer to the ONDD (see recital (120)). Of this EUR [45-75] million, it has been concluded above that one can reasonably estimate the share of marketable risks at EUR [10-25] million. Consequently, one can reasonably estimate the share of non-marketable risks at EUR [35-50] million (including EUR [0-5] million for risks relating to Bulgaria and Romania which became marketable in 2007).

For the same reasons as those set out in recital (127), since the transfer of the whole of the EUR [45-75] million (including the share of non-marketable risks estimated at EUR [35-50] million) does not constitute an advantage, the latter is excluded from the analysis of the existence of aid.

In conclusion, of the initial capital of EUR 150 million allocated to Ducroire/Delcredere, only EUR [75-100] million constitutes additional capital, that is to say fresh capital. Only the granting of this EUR [75-100] million could therefore constitute an advantage (see recital (120)).

As was pointed out in recital (65) of the opening decision, Member States are free to support the non-marketable risk insurance business, since the Commission considers that there is no market for these risks. There can therefore be no distortion of competition vis-à-vis other insurers. However, the Commission notes that Ducroire/Delcredere did not keep separate accounts for marketable and non-marketable risk activities. Formally, no capital was allocated to the non-marketable risk activity. No part of the EUR 150 million of capital was earmarked for the financing of non-marketable risks. Accordingly, a purely formal approach could result in none of the additional capital of EUR [75-100] million being excluded. Such an approach would be all the more justified as the obligation to introduce separate accounts is expressly provided for in point 4.3 of the Communication on export-credit insurance which has been in force since 1998. However, it is clear that a large part of the activities of Ducroire/Delcredere concerns the insurance of non-marketable risks. Ducroire/Delcredere could not have insured these risks without adequate capital. Therefore, in addition to the EUR [45-75] million (including EUR [35-50] million for non-marketable risks), the Commission can agree to exclude also from categorisation as state aid that part of the capital which can reasonably be demonstrated to have been used to support the activity of insuring non-marketable risks.

In the opening decision, Belgium was invited to develop a method for determining the part of the capital of Ducroire/Delcredere which can be regarded as supporting the marketable risk activity and that supporting the non-marketable risk activity. Following this request from the Commission, the Belgian authorities submitted, for the period 2005-2011, a separate balance sheet and profit and loss account for the marketable risk and the non-marketable risk activity. The balance sheet for non-marketable risks at 31 December 2011 included EUR [70-110] million of capital and that for marketable risks EUR [40-80] million of capital. Although the separation of the accounts is based on certain hypotheses or data which are partially verifiable by means of retrospective reconstruction, the levels of capital for the marketable and the non-marketable risk activities at 31 December 2011 are considered by the Commission to be reasonable.

The Commission agrees, therefore, to exclude from categorisation as state aid EUR [70-110] million because it supports, de facto, an activity which, according to the Communication on export-credit insurance, is not subject to competition.

Thus, the capital supporting short-term non-marketable risks rose from EUR [35-50] million at the end of 2004 to EUR [70-110] million at the end of December 2011, i.e. an increase of EUR [35-60] million. Since the amount of EUR [35-50] million of capital linked to the non-marketable risk activity includes the EUR [0-5] million for risks relating to Romania and Bulgaria, recategorised as marketable in 2007, the gross increase in capital supporting non-marketable risks is EUR [35-65] million. In other words, of the additional capital of EUR [75-100] million, EUR [35-65] million supported, de facto, the non-marketable risk activity.

An amount of EUR [70-110] million has been excluded from categorisation as aid because it has been possible to regard this amount as supporting the non-marketable risk activity as at 31 December 2011 (first exclusion criterion). An amount of EUR [10-25] million has also been excluded from categorisation as aid because it has been possible to regard it as not constituting an advantage given that it already supported marketable risks before those risks were transferred to Ducroire/Delcredere on 31 December 2004 (second exclusion criterion). Furthermore, an amount of EUR [0-5] million must also be excluded since it already supported risks relating to Bulgaria and Romania before they became marketable on 1 January 2007.

In conclusion, only EUR 36,6 million of the initial capital allocation of EUR 150 million might constitute an advantage and state aid within the meaning of Article 107(1) of the TFEU.

Another way of presenting the result of EUR 36,6 million is to subtract, from the initial capital of EUR 150 million granted to Ducroire/Delcredere, (1) the EUR [45-75] million in support already provided for short-term risks before their transfer to Ducroire/Delcredere (including the EUR [0-5] million for risks relating to debtors established in Romania and Bulgaria which became marketable on 1 January 2007); and (2) the additional EUR [35-65] million which supported, de facto, non-marketable risks. This leaves us with an amount of EUR 36,6 million.

In the alternative, the Commission would make the following observations. The Belgian authorities assert that, with the exception of approximately EUR 7 to 13 million (see Table 2), the entire capital of EUR 150 million was implicitly allocated in 2004 to non-marketable risks. On the basis of this first assertion, Belgium maintains that it was only later (in 2007 and 2008) that a part of the capital initially allocated to non-marketable risks was transferred to marketable risks. Belgium maintains, therefore, that only these capital transfers could constitute aid and that the private investor test must be applied to these amounts at the time of their internal transfer and not at the time of the initial capital allocation in 2004. This reasoning cannot be accepted.

When Ducroire/Delcredere was set up and in the years that followed, the EUR 150 million of capital was never formally allocated to the activity relating to marketable risks and to that relating to non-marketable risks respectively. This capital could be used freely to support the activity relating to marketable risks or that relating to non-marketable risks (in undefined proportions) depending on market opportunities and the strategic choices of Ducroire/Delcredere. The capital allocation between the activity of marketable risks and that of non-marketable risks was recreated only retroactively within the framework of the present procedure. The 2004 documents, as submitted by the Belgian authorities, did not show any allocation of capital between the different activities. The Commission cannot therefore agree to exclude from the analysis of the existence of aid a part of the EUR [75-100] million of additional capital of Ducroire/Delcredere on the sole basis that at that time this amount might have been used to support the non-marketable risk activity. As already indicated, only that part of the additional capital of EUR [75-100] million which can be shown to have actually supported the non-marketable risk activity can be excluded from the analysis of the existence of aid.

Purely in the alternative, the Commission would observe that, even if one were to accept Belgium's argument that part of the initial capital of Ducroire/Delcredere was allocated to non-marketable risks and therefore could not come within the scope of Article 107(1) of the TFEU, this argument would apply to the amount of EUR 36,36 million only if Belgium managed to demonstrate that the capital allocated to non-marketable risks was higher than EUR [70-110] million. (To calculate the EUR 36,6 million, an amount of EUR [70-110] million supporting the activity relating to non-marketable risks has already been deducted from the EUR 150 million.) It is clear that such an amount was never allocated to non-marketable risks in 2004, as internal documents indicate that EUR 100 million of capital was sufficient to cover all short-term risks (both non-marketable and marketable) until 2007.

In the present case, an analysis must be carried out as to whether the supplementary capital of EUR 36,6 million granted in 2004 would have been sufficiently profitable to convince a private market economy investor. However, the profitability of this supplementary capital cannot be examined separately from the capital as a whole as the supplementary capital results from an artificial division of the capital of EUR 150 million subscribed in 2004. It is impossible to identify any precise flow of income coming from this EUR 36,6 million. In other words, no profit from a specific activity was attributed to this EUR 36,6 million since separate accounts did not exist. The Commission therefore considers that, in order to apply the private investor test correctly, the expected profitability of the entire capital of EUR 150 million must be verified to see whether it was sufficient. If not, the conclusion will have to be drawn that the EUR 36,6 million constitutes an advantage.

Purely in the alternative, another possibility would be to apply a pro rata method to the overall profit generated, which would produce the same result.

In the recitals that follow, the Commission will demonstrate that the expected profitability of the EUR 150 million was insufficient to convince a private investor to make such an investment. In the alternative, it will show that, even if one considers that the marketable risk activity benefited virtually from a capital of EUR [45-65] million (EUR 36,6 million plus EUR [10-25] million) and from the expected profits from this specific activity, the expected profitability of this activity was also insufficient.

The Belgian authorities based their analysis of profitability on the business plan of 28 September 2004. Ducroire/Delcredere was in fact incorporated on 23 September 2004. However, it did not actually start operating until 1 January 2005, given that the short-term risk portfolio continued to belong to the ONDD until 31 December 2004, the date on which it was transferred. It can therefore be considered that, until shortly before that date, the ONDD could have gone back on its investment decision by not transferring the risks in question and by liquidating the newly created legal entity. It is therefore acceptable to take into account the ONDD's business plan of 28 September 2004 as requested by the Belgian authorities.

According to the projections in scenario B of the business plan of 28 September 2004, the ONDD was counting on an ROC for all activities of 1,3 % to 1,9 % for 2005, 2006 and 2007 in line with the ‘6 % growth process’ scenario in the business plan (the ROC before the equalisation provision was 2,8-4,3 %) (see Table 4). It must be explained that this profitability is based on a capital of EUR 100 million and does not take account of the EUR 50 million not paid up, which makes the profitability result more favourable than it is in reality. The expected profitability rate in the financial projections of 20 April 2004 was not higher (see Table 3).

There was no serious study in 2004 that could lead one to consider that this low initial profitability would be compensated by an increased profitability in later years. The Commission would also point out that the financial projections available at the time showed that performance remained at the same level between 2005 and 2007, or at least that there was only sluggish growth (see Table 4), therefore there was nothing to indicate that results would improve rapidly in the years beyond 2007. In this connection, it must be noted that the conversion of Ducroire/Delcredere into a subsidiary cannot be regarded as being the same thing as a start-up since it was envisaged at the time that it would continue with the activities previously carried on by the parent company, and the assets and liabilities were simply transferred from the ONDD to Ducroire/Delcredere for this purpose. This conclusion is also based on the way in which financial projections were produced at the time. It emerges that the ONDD, making financial projections for the future Ducroire/Delcredere in 2004, based them on the past performance of the activity in question (short-term risks, including marketable risks) as carried on by the ONDD. Assumptions regarding premiums and costs were also based mainly on historical financial data from the previous five years, apart from a few adjustments linked to changes in market circumstances, such as a reduction applied to premiums for risks that became marketable in 2004 following the accession of ten new Member States to the European Union.

Moreover, it is clear from internal memos that it was envisaged that the entire paid-up capital of EUR 100 million should be absorbed mainly through the gradual growth until the end of 2007 (see recital (61)) of the activities performed by the ONDD in the past, in other words organic growth of existing activities (no expansion into other, more profitable activities was planned in 2004).

No plans were drawn up at the time for the use of the subscribed capital that had not been paid up (EUR 50 million), hence no projection was produced on the expected profitability of this capital. Moreover, neither the acquisition of KUP nor a broader acquisition strategy was taken into account when transferring the capital to Ducroire/Delcredere.

The projections for the period 2005-2014 were produced only ex post, after the Commission had launched the formal investigation procedure. The only ex ante financial projections produced by the ONDD were confined to a three-year period (the first three years of operation of the future Ducroire/Delcredere, namely 2005, 2006 and 2007), a relatively short period from the point of view of a private investor.

The Commission concludes that the expected profitability of the capital was insufficient, and that therefore the EUR 36,6 million constitutes an advantage for Ducroire/Delcredere's marketable risk activity.

In the alternative, the Commission will analyse the expected profitability of the marketable risk activity in the recitals that follow.

Contrary to the ONDD's calculations of profitability, the estimation of expected profitability, as indicated in the previous recital, is based on the entire estimated capital for marketable risks, including the non-paid-up part of the capital. The Belgian authorities' argument (see recital (79)) whereby the part of the capital subscribed in 2004 but paid up in 2009 should not be taken into account in the estimated profitability calculations before 2009, cannot be accepted. While this capital was not paid up in 2004, it could in fact be called upon at any time since it had been subscribed when Ducroire/Delcredere was set up. A private investor in a market economy would demand remuneration for the risks incurred in relation to this investment, since he could lose part or all of his investment in the event of bankruptcy. Thus the calculations produced by the Belgian authorities cannot be accepted.

Even using an investment rate of 3,5 %, as suggested by the Belgian authorities (see recital (80)), the expected profitability would not change significantly and would not achieve the estimated capital cost calculated by Belgium itself.

The estimated profitability (as referred to in recital (168)) takes account of contributions to the equalisation provision, contrary to what is advocated by the Belgian authorities. Since the constitution of this provision, required by the Belgian prudential authorities, is intended to ensure a balance in results over time and to cover future potential losses resulting from future activities, it is more appropriate to take account of it when calculating profitability. The Commission's approach is in line with the accounting approach, where such a provision is regarded as a cost that reduces profit. Even if the provision was not taken into account (as a cost when estimating profit), the expected profitability would not equal the capital cost estimated by the Belgian authorities themselves (in this Decision, the Commission does not comment on the validity of the Belgian authorities' estimate of the capital cost since the reasoning in this Decision is not based on that figure).

It seems, therefore, that, even when the elements put forward by the Belgian authorities are applied (elements which the Commission regards as wrong), the expected profitability of the marketable risk activity was insufficient to convince a private investor in a market economy to proceed with such an investment.

In the alternative, as regards the method advocated by the Belgian authorities in their reply to the Commission's request for information in December 2007, in order to estimate profitability or the ROC ratio, the Commission would point out that this ratio was not used in the ONDD's ex ante financial projections (business plan). Moreover, such a ratio highlights the technical income on turnover, does not take account of the capital invested and does not state the profitability in relation to it. Nor does it take account of the result of purely financial activities, which, however, forms part of the accounting profit that can be distributed to shareholders. A private equity investor (such as a shareholder) would not therefore have used this rate only (which does not indicate the return on equity or on invested amounts) in order to assess the profitability of the investment he was considering.

In conclusion, the EUR 36,6 million constitutes an advantage which the ONDD granted to Ducroire/Delcredere in the form of capital which it could not have obtained on the market on the same conditions.