Commission Decision of 21 October 2008 on a State aid scheme implemented by Italy to remunerate Poste Italiane for distributing postal savings certificates (C 49/06 (ex NN 65/06)) (notified under document number C(2008) 5585) (Only the Italian text is authentic) (Text with EEA relevance) (2009/554/EC)

Commission Decision

of 21 October 2008

on a State aid scheme implemented by Italy to remunerate Poste Italiane for distributing postal savings certificates (C 49/06 (ex NN 65/06))

(notified under document number C(2008) 5585)

(Only the Italian text is authentic)

(Text with EEA relevance)

(2009/554/EC)

THE COMMISSION OF THE EUROPEAN COMMUNITIES,

Having regard to the Treaty establishing the European Community, and in particular the first subparagraph of Article 88(2) thereof,

Having regard to the Agreement on the European Economic Area, and in particular Article 62(1)(a) thereof,

Having called on interested parties to submit their comments pursuant to the provisions cited above(1) and having regard to their comments,

Whereas:

I. PROCEDURE

(1) By letter dated 30 December 2005, Associazione Bancaria Italiana (ABI) complained to the Commission against various measures which it alleged benefited the banking activities of Poste Italiane SpA (PI), which are operated through a fully integrated division of PI called Bancoposta.

(2) According to the complainant, PI was being overcompensated for the distribution of postal savings financial products on behalf of Cassa Depositi e Prestiti SpA (Deposits and Loans Fund — CDP). A second advantage was said to consist in the exclusive distribution of these products, which steered customers away from similar financial products distributed by commercial banks.

(3) By letter dated 7 February 2006, the Commission put a number of question to the Italian authorities relating in particular to postal savings. After requesting an extension of the deadline for a reply, Italy replied to the questions by letter dated 21 April 2006. The Commission put further questions by letter of 29 June 2006; Italy replied on 31 July 2006 and on 9 August 2006.

(4) A meeting took place with the Italian authorities and representatives of PI on 30 March 2006.

(5) On 20 April 2006 the Commission wrote to ABI seeking clarification of a number of points; ABI replied by letter dated 8 June 2006.

(6) By letter dated 22 November 2006, the Commission informed Italy that it had decided to initiate the procedure laid down in Article 88(2) of the EC Treaty in respect of the measure.

(7) The Commission’s decision to initiate the procedure was published in the Official Journal of the European Union (2). The Commission called on interested parties to submit their comments.

(8) The Italian authorities submitted their comments by letter of 2 February 2007.

(9) ABI replied by letter of 13 March 2007. By letter of 20 March 2007 the Commission forwarded ABI’s comments to the Italian authorities so as to enable them to reply. The Italian authorities replied by letter dated 18 April 2007.

(10) On 20 March 2007, a meeting took place between ABI and Commission staff.

(11) By letter of 23 March 2007, the Commission requested additional information from ABI; ABI replied by letter of 18 May 2007. By letter of 23 May 2007, the Commission forwarded the observations received from ABI to the Italian authorities, who replied by letter dated 22 June 2007.

(12) On 25 January 2008 the Commission requested additional information from the Italian authorities, who replied on 11 March 2008, 13 May 2008 and 17 June 2008.

(13) In January 2007, the Commission launched a public tender (2007/S 29-034352) for a study of the remuneration for the placement of postal savings certificates. In the case of one of the bidders a potential conflict of interest arose, and on 29 June 2007 the Commission wrote to the Italian authorities and ABI asking them to comment. Italy replied by letter of 5 July 2007.

(14) On 9 November 2007, the validity of the bids having expired, the Commission launched a new procurement procedure for the study (COMP/2007/D3/23), and on 20 December 2007 the Commission concluded the study contract with an independent expert (the expert). The expert submitted the final report of the study to the Commission on 21 June 2008.

II. THE BUSINESS OF CDP AND PI/BANCOPOSTA — THE RELEVANT MARKETS

II.1. CDP

(15) CDP was a State agency until late 2003, when it was converted into a limited company (SpA). Since then, although 30 % of its share capital has been transferred to 65 banking foundations, CDP has remained under the control of the State(3).

(16) CDP’s object is to foster the development of public investment, local utility infrastructure works and major public works of national interest.

(17) Since 2003, the activity of CDP has been formally structured around two distinct business areas:

• The first area, known as the ‘separate management area’ (gestione separata), manages the funding of investment by the State, the regions, local authorities, public agencies and other public-law bodies. State-guaranteed postal savings provide the main source of funding for these objectives,

• The second area, known as the ‘ordinary management’ area (gestione ordinaria), manages the funding of works, plants, networks and equipment, the provision of public services and drainage. For these purposes CDP may raise funds through the issue of securities, borrowing and other financial operations, without State guarantees.

II.2. PI

(18) PI is the universal postal service provider in Italy, which fulfils the universal postal service obligation(4) according to the Italian legislation(5) and the regulations governing the universal postal service. The service of general economic interest with which PI is entrusted does not include financial services.

(19) Besides providing the universal postal service, PI offers integrated communication, logistic and financial products and services throughout Italy. The main figures are the following(6):

Table 1
Financial figures for PI’s main areas of business
Year 2006
Total staff (annual average)	151 470
Regional areas	9
Branches	140
Post offices	13 893
Main Financial Figures of Poste Italiane Group
(EUR million)
Total revenue	17 055,6
Sales and service revenues	15 932,2
Of which:
from postal services	5 339,4
from financial services	4 382,5
from insurance services	5 993,6
from other services	216,7
Other revenue	1 123,3
Net profit	675,7
Postal Services
Products and services	volumes (number of items)
Mail (standard, priority, registered, insured, legal documents, other recorded mail)	3 522 792 200
Commercial mail (Postatarget, catalogues, unaddressed mail, etc.)	1 887 699 700
Periodicals (printed, gifts, books etc.)	1 216 045 800
Telegrams, fax, telex	17 442 800
Express delivery (Poste Italiane and SDA)	46 284 600
Parcels	16 052 000
Postal savings
Savings books, savings certificates and postal current accounts: total	EUR 282 408 million
Life assurance policies: policies written	EUR 5 989 million
Conto BancoPosta: number of current accounts	4 880 000
Carta Postepay: number of cards issued	2 801 000

(20) According to the annual report, postal services accounted for 33,5 % of the group’s total turnover in 2006, financial activities accounted for 27,5 %, and insurance business for 37,6 %. Financial services contributed 82 % to the operating profit of the PI group, and insurance services 18,7 %. Postal services showed an operating loss of EUR 4 million.

(21) Before December 2003, PI was wholly owned by the Italian State. In December 2003, the State transferred 35 % of the share capital in PI to CDP. PI is likewise a private-law company controlled by the State.

II.2.1. Postal services

(22) According to a recent study(7), before the implementation of the first Postal Directive(8) the Italian postal market was relatively open. Operators other than PI were already free to distribute direct mail and hybrid mail. Some local mail operators were delivering letters as subcontractors to PI. After the implementation of the Directive, the delivery of hybrid mail became part of the reserved area, and the subcontracting relationships were ended. Outgoing and incoming international mail falls entirely within the area reserved to PI. Since 1 January 2003, Italian legislation has incorporated the second Postal Directive, which provides for the full opening of the EU postal market to competition by 1 January 2009: it limits the reserved area to letters up to 100 g and three times the basic tariff for priority mail(9). From 1 January 2006, the reserved area is limited to letters up to 50 g and 2,5 times the basic tariff for priority mail. The reserved area also continues to include upstream services and the conveyance of bulk mail. The postal market is now relatively open to competition de jure, since the delivery of direct mail has been liberalised. The entry regulations are not regarded as severe(10).

(23) On 19 October 2006, the Commission proposed a new Postal Directive completing the single market for Community postal services. That Directive entered into force on 27 February 2008; it provides for the abolition of legal monopolies in postal services by 31 December 2010, though some Member States are permitted to extend the deadline until December 2012(11).

II.2.2. Financial services

(24) Presidential Order (decreto) No 144 of 14 March 2001 lays down rules governing the banking and financial services that can be provided by PI; this business is carried on by Bancoposta. The services comprise: collecting savings from the public in all forms; provision of payment services; foreign exchange intermediation; promotion and placement of loans granted by banks and other authorised intermediaries; performance of some investment services (trading on account of third parties and placement and collection of orders, thus excluding trading on own account and individual portfolio management). PI is expressly prevented from engaging in financing.

(25) PI/Bancoposta can be considered as a deposit institution and a financial intermediary. While it is not a bank, it uses PI’s numerous post-office outlets for its own operations and to provide banking and other financial products.

(26) The 13 893 counters, at least one per municipality on average, make PI the biggest banking network in Italy, and should be regarded as an asset rather than a liability: most of their cost is covered by these financial services, and they are not considered a burden on the universal service provider(12).

(27) The rating Agency Fitch notes in a report published in 2004 that PI/Bancoposta has ‘a capacity to reach the overall [Italian] population that cannot be matched by domestic banks for the foreseeable future’(13). Moreover, Fitch considers that PI has put ‘the development of financial services at the heart of [its] strategy’.

(28) PI offers a wide range of financial services competing with those offered by the banking system:

• direct and indirect collection of savings, and lending,

• payment services,

• placement of financial and investment products.

(29) The aspect at issue in the present Decision is the collection of funds, and particular postal savings, whereby PI raises funds on behalf and for the account of CDP by distributing savings certificates and savings books.

(30) In the last few years, PI has significantly broadened the range of its payment services to customers, adding to the traditional postal instruments (current account deposit receipts and postal money orders) a series of instruments that were formerly offered only by banks (debit and credit cards, credit transfers, or standing debit orders for utility bills)(14).

(31) Finally, PI provides placement services for:

• structured bonds issued by banks,

• insurance policies written by Poste Vita(15);

• investment funds managed by the asset management company Bancoposta Fondi SGR(16);

• loans granted by third parties: personal loans and mortgages are sold on behalf of banks.

III. THE MEASURE UNDER ASSESSMENT

(32) The measure under assessment here is the remuneration paid by CDP to PI for the placement of savings certificates in the period 2000-2006.

III.1. Postal savings

(33) Article 1(1) of the Order (decreto) of the Ministry of Economic Affairs and Finance of 6 October 2004 relating to Decree-law No 269 of 30 September 2003(17) defines ‘postal savings’ as ‘the collection of funds, with an obligation to repay supported by a State guarantee, performed by CDP SpA through Poste Italiane SpA’. The instruments used for the purpose are postal savings books and postal savings certificates.

(34) Article 1(2) of the Order states that postal savings, i.e. the raising of funds by CDP through PI, constitute a service of general economic interest.

III.1.1. Postal savings books

(35) Postal savings books (libretti postali) are fund-raising instruments which are guaranteed by the State and consequently have a low risk profile. They may be either registered or bearer. The main transactions on postal savings books are money deposits and withdrawals.

(36) On 22 November 2006 the Commission decided that the remuneration paid by CDP to PI for the placement of postal savings books over the period 2000-2005 did not constitute State aid within the meaning of Article 87(1)of the Treaty.

III.1.2. Savings certificates

(37) Postal savings certificates (buoni fruttiferi postali) are likewise fund-raising instruments which are guaranteed by the State, and consequently have a low risk profile; they are not negotiable on the market, because by express statutory provision they are not transferable; and they are repayable on demand, a characteristic equivalent to a put option of the US type, which allows the saver to request repayment at a predetermined price at any time during the life of the certificate. Interest is settled in a single payment at redemption, and is subject to a withholding tax of 12,5 %.

(38) Savings certificates are essentially one-coupon bonds, issued ‘on tap’, meaning that the total amount of an issue is not established in advance. The service of placement, management and redemption of savings certificates and other operations relating to them are free of commission and any other charges to subscribers. Since 2003 they have been available in either paper or dematerialised form. Before 2003 only paper certificates were issued.

(39) The four main categories of savings certificate are as follows(18):

• Ordinary savings certificates: these are financial products with a 20-year maturity and a fixed annual yield that grows over time (a step-up structure),

• Indexed savings certificates: these pay an annual fixed rate determined at issue plus a premium at maturity linked to the performance of an equity index(19) (certificates of this kind were first issued in 2003) or to inflation (certificates of this kind were first issued in 2006),

• Fixed-term savings certificates: these have not been issued since the beginning of 2003. They were similar to ordinary certificates, except that they had a maximum maturity of seven years,

• 18-month savings certificates: these were introduced in 2005, to fill a gap at this particular maturity.

Table 2
Stock of postal savings certificates a , by category, in EUR billion
a The stock includes certificates issued by CDP and by the Ministry of Finance (the latter ceased to issue savings certificates with effect from 2001).
b This duration refers to equity-linked certificates. For inflation-linked certificates, see footnote 49.
c The Italian authorities indicate that the average life of this category is not quantifiable because these certificates have been issued only since September 2005.
Category of certificate	Maximum duration	Average duration	31.12.2005	31.12.2006
Ordinary	20 years	[7-10] years	121,1	132,2
Indexed	7 years	[4-7] yearsb	1,6	3,6
18-month	18 months	unknownc	1,9	8,3
Fixed-term	7 years	[4-7 years]	48,6	36,2
Total			173,1	180,6

III.1.3. The funds raised

(40) According to ABI, the amounts held by Italian households in the form of postal savings instruments and the main competing products are as follows:

Table 3
Amounts held by Italian households in the form of postal savings instruments and competing products, in EUR billion
Source: ABI
Financial instrument	Dec. 1999	Dec. 2004
Postal savings certificates	113	160
State bonds	126	203
Euro area government bonds	156	160
Postal savings books	36	60
Bank savings deposits	69	74

III.2. Remuneration paid to PI for placing postal savings certificates

(41) For the service of placing postal savings certificates, PI receives a yearly remuneration, which has been based on successive agreements and amendments since 2000:

• the agreement (convenzione) covering the three years 1999-2001 concluded on 4 August 1999 and cancelled on 27 October 2000,

• the agreement covering 2001, concluded on 10 May 2001,

• the agreement covering 2002, concluded on 26 July 2002,

• the agreement covering the three years 2003-2005 concluded on 23 October 2003; this agreement was amended with respect to the years 2004 and 2005 by two additional agreements between PI and CDP, concluded on 24 December 2004 and 20 October 2005,

• the agreement covering the three years 2006-2008 concluded on 30 March 2006.

(42) Under these agreements the yearly remuneration paid to PI has consisted of three components: i) a commission on the gross amount collected (upfront fee), dependent on the sum collected in savings certificates in the year; ii) a commission for managing the stock of postal savings certificates; iii) other specific components shown in tables 4 and 5.

Table 4
Remuneration for savings certificates under the agreements
a Business secret.
b […].
c […].
d […].
2000	2001	2002	2003	2004	2005	2006
Annual commission on the stock of certificates outstanding: […]a b	Annual commission on the stock of certificates outstanding: […]	Annual commission on the stock of certificates outstanding: […]	Annual commission on the stock of certificates outstanding: […]c	Annual commission on the stock of certificates outstanding: […]	Annual commission on the stock of certificates outstanding: […]	Annual commission on the stock of certificates outstanding: […]
Commission on the value of new certificates subscribed in the year: […]	Commission on the value of new certificates subscribed in the year: […]	Commission on the value of new certificates subscribed in the year: […]	Commission on the value of new certificates subscribed in the year: […]	Commission on the value of new certificates subscribed in the year: […]	Commission on the value of new certificates subscribed in the year: […]	Commission on the value of new certificates subscribed in the year: […].
Other components: […]d	Other components: […]	Other components: […]	Other components: […]	Other components: […]	Other components: […]	Other components: […]

Table 5
Yearly remuneration paid to PI, 2000-2006, in EUR million
	2000	2001	2002	2003	2004	2005	2006
Remuneration for new subscriptions:	57	243	123	302	381	487	460
18-month						13	50
Indexed				8	7	29	59
Fixed-term	25	73	60
Ordinary	32	170	63	293	374	444	351
Remuneration for accounting and management activities:	252	196	194	115	101	96	99
18-month	0,0	0,0	0,0	0,0	0,0	0,0	0,0
Indexed	0,0	0,0	0,0	0,0	0,0	0,0	0,0
Fixed-term	130,6	98,2	94,7	46,9	36,5	26,0	20,6
Ordinary	121,4	98,0	98,9	67,6	64,8	69,6	78,0
Remuneration for paper certificate production				15	16	18	18
Total fees	309	439	316	431	498	600	577

(43) According to CDP’s annual accounts, the remuneration paid to PI varies greatly with the volume of funds collected, and enables CDP to maintain exceptional operational flexibility and an extremely streamlined cost structure.

IV. GROUNDS FOR INITIATING THE PROCEDURE

(44) In the decision opening these proceedings, on 22 November 2006, the Commission said that the legal framework applicable to CDP for the placement of postal savings (postal savings certificates and savings books) did not confer any exclusive right of placement de jure.

(45) But the Commission doubted whether the remuneration paid by CDP to PI for the placement of savings certificates provided an appropriate market reference, and expressed doubts particularly regarding the similar products cited by Italy to show that the remuneration was on market terms. Nor was the Commission aware at that stage of the precise composition and value of the various fees paid to PI in respect of savings certificates since 2000.

(46) Consequently, the Commission had serious difficulties in assessing whether all the components of the remuneration for the placement of savings certificates:

i)

had satisfied the third and fourth tests of the Altmarkjudgment(20) since 2004, and

ii)

had been market-conform in the years 2000-2003.

(47) If the remuneration paid by CDP to PI in respect of savings certificates did constitute State aid, the Commission said it would have serious difficulty in assessing its compatibility.

V. COMMENTS FROM INTERESTED PARTIES

(48) Only ABI submitted comments in the course of the proceedings.

(49) In its complaint of 30 December 2005, ABI indicated that the amounts paid by CDP to PI constituted a remuneration higher than:

• the commissions set by law for placing government securities (BOTs, BTPs and CTZs)(21), and

• the commissions earned by banks for the placement of bonds for third parties.

(50) ABI stated that commissions for placing corporate bonds ranged from 0,5 % to 6 %, depending on a series of factors such as the issuer’s rating, the structure of the issue, the retail network and the degree of complexity of the financial instrument. The commissions sought by banks for distributing bank bonds for the account of third parties averaged between 1 % and 1,5 % of the amount placed.

(51) According to ABI, the remuneration paid by the State to PI for placing postal savings products constituted State aid to PI, to the detriment of banks, which would have been able to provide the same service on better terms at a lower cost to the State.

(52) By letters of 13 March 2007 and 18 May 2007, ABI submitted additional comments.

(53) ABI said that the remuneration for placing postal savings certificates had to be compared to the remuneration payable for placing BOTs or BTPs, depending on the maturity. On savings certificates PI received a remuneration depending on the value subscribed which was higher than the fee for placing State bonds (about 195-160 basis points), but it also received a remuneration based on the outstanding stock of savings certificates.

(54) ABI did not share the conclusions of the benchmark study described in the Commission’s decision of 22 November 2006, for the following reasons:

• Postal savings certificates were issued by CDP, which was a financial intermediary with a rating equal to that of the Italian State (A+, Aa2, AA–). Postal savings certificates consequently had a risk profile far lower than that of corporate bonds, i.e. bonds issued by non-financial companies(22). At the same rating, any corporate bond bore a risk higher than that of State bonds or bonds issued by financial companies. Postal saving certificates consequently could not be compared to corporate bonds, which were not issued by financial intermediaries,

• Ordinary postal savings certificates could not be compared to structured inflation-linked bonds. Structured inflation-linked bonds offered the subscriber a return linked to the performance of a pre-defined index, whereas ordinary postal savings certificates offered a fixed return depending on the period for which they were held, so that the return on structured inflation-linked bonds was potentially higher, for a higher level of risk,

• Contrary to what was said in the benchmark study, 18-month and fixed-term postal savings certificates were not negotiable in a secondary market.

(55) By letter of 18 May 2007, ABI forwarded to the Commission the conclusions of a study carried out with the help of a consultancy firm which sought to identify products that could be compared to postal savings certificates and to analyse the remuneration for their placement. The study confirmed ABI’s positions, and detailed the products comparable to postal savings certificates as follows.

Identification of products comparable to postal savings certificates and remuneration for their placement

(56) Ordinary postal savings certificates were comparable to BTPs. Postal savings certificates and BTPs were the same in terms of maturity, risk, yield and tax rates. The remuneration for placing BTPs was 0,30 %-0,40 % of the amount subscribed.

(57) Contrary to what was said in the benchmark study cited in the Commission’s decision of 22 November 2006, ordinary postal savings certificates were not comparable to:

• corporate bonds with a rating equal to or above BBB: postal savings certificates had a much lower risk profile even when their rating was the same,

• inflation-linked bonds: postal saving certificates had a fixed yield that depended on their maturity and not on the rate of inflation.

(58) As regards the complexity of ordinary postal savings certificates, ABI argued that they fell outside the definition of structured products laid down by the Banca d’Italia. The fact that postal savings certificates were subject to an option equivalent to a put option did not justify a higher remuneration for their placement. This option was rather in the nature of a guarantee to the subscriber, who was entitled to end the investment at any time, and did not affect the structure of the security or the level of risk involved.

(59) 18-month postal savings certificates could be compared to short-term State bonds, either BOTs or CTZs. Contrary to what it was said in the benchmark study, 18-month postal savings certificates were not comparable to corporate bonds with a rating equal to or greater than A+, because they had a far lower risk profile.

(60) Commissions for the placement of BOTs varied from 0,05 % to 0,30 % of the amount subscribed, depending on the maturity (this remuneration was paid by the investor to the operator placing the bond). The remuneration for the placement of CTZs was 0,20 % of the amount subscribed.

(61) Fixed-term postal savings certificates were comparable to State bonds such as BOTs, CTZs or BTPs depending on the residual life of the savings certificate.

(62) Inflation-linked postal savings certificates were comparable to BTPEUR i bonds (medium- to long-term State bonds indexed to the inflation rate of the euro area). They could not be compared to corporate equity-linked bonds, because their risk profile was much lower, and their yield was linked not to a basket of financial instruments, as the yield on equity-linked bonds was, but to the inflation rate.

(63) Commissions for placing BTPEUR i bonds amounted to 0,40 % of the amount subscribed.

(64) Index-linked postal savings certificates were comparable to index-linked bonds issued by financial companies (even though the risk profile was lower in the case of postal savings certificates). They could not be compared to corporate bonds, since CDP was a financial intermediary.

(65) Commissions for placing financial index-linked bonds varied from 1 % to 2,5 % of the amount subscribed.

(66) ABI had initially argued that postal savings certificates could be compared to mutual funds investing in euro area government bonds; ABI’s letter of 18 May 2007 does not return to this argument, though it offers no explanation.

Structure of remuneration

(67) ABI indicated two possible scenarios for placing products comparable to postal savings certificates. Either these products were placed by an operator remunerated solely for placement, and the issuer bore the management costs, or they were placed and managed by an operator remunerated for both activities, in which case it was not easy to distinguish the two components of the remuneration, the placement on the one hand and the accounting and administrative management on the other. The study had shown that market practice did not allow a distinction to be made between the part of the commission that related to placement only, the part that related to accounting and administrative management, and the part that related to the production of paper bonds.

(68) The level of remuneration was affected by the bond’s complexity, the issuer’s rating, the placement network, the maturity, the issuer’s reputation, the issue’s structure and the legal constraints (obligations borne by the issuer under legislation or licences).

(69) Postal savings certificates, which carried a State guarantee, had a risk profile much lower than that of corporate bonds, as could be seen from the fact that the internal rate of return on a State bond was always lower than that on a corporate bond with an equivalent rating. Over the last 10 years, the default rate of Aa-rated sovereign bonds had been 0, whilst the default rate of Aa-rated corporate bonds had been 0,59.

(70) Finally, it was not possible to identify a market remuneration for the production of paper bonds, because financial markets now dealt almost exclusively with dematerialised products.

VI. COMMENTS FROM ITALY

The benchmark study

(71) Italy based its comments on the results of the benchmark study conducted in January 2006 in order to prepare the ground for the agreement covering the three years 2006-2008, which sought to identify a market remuneration for CDP and PI for postal savings products (the benchmark study).

(72) The methodology of the benchmark study worked in six successive steps:

i)

the identification of the main characteristics of each postal savings product,

ii)

the identification of a comparable financial instrument for each postal savings product,

iii)

the definition of a reference sample,

iv)

the identification of the economic components to be compared,

v)

the determination of the rate of remuneration, and

vi)

the comparison of the rate of remuneration with the rate paid by CDP to PI.

(73) The benchmark study found that the level of placement fees was affected by three factors: the risk/return profile, the contractual complexity (whether the security was a structured or ‘plain vanilla’ instrument)(23), and the maturity.

(74) Italy argued, in particular, that the remuneration for the placement of postal savings certificates could not be compared to the remuneration for the placement of State bonds, because:

• the two families of products were intrinsically different: postal investment certificates embedded a put option for the investor (see recital 37), which meant that they were a more complex financial product than ‘plain vanilla’ bonds and could be considered full-scale structured bonds,

• in addition to placement, PI performed the accounting and administrative management of the savings certificates, a service that was not asked of banks that placed State securities.

(75) The main conclusions of the benchmark study were the following:

• The comparables for long-term postal savings certificates were structured instruments like inflation-linked and equity-linked corporate bonds(24) with a rating equal to or higher than Standard and Poor’s BBB or Moody’s Baa2, or mutual funds investing in euro area government bonds. Mutual funds resembled postal savings certificates in that they were issued ‘on tap’ and the subscriber could at any time request repayment of the capital and the accrued interest; another similarity was that the company managing the fund would perform the administrative management of the securities. The study compared different elements of the remuneration for the placement of mutual funds with those of postal savings certificates, and concluded that they were in line with market rates for these comparables(25),

• For 18-month postal savings certificates, no matching comparables were found, and so the benchmark study extrapolated a fee from the shortest-maturity corporate bonds with a minimum rating of Standard & Poor’s A– or Moody’s A3, assuming a linear distribution of placement fees,

• The benchmark study found that postal savings certificates were not comparable to government bonds, because government bonds were ‘plain vanilla’ bonds, whereas postal savings certificates were more complex (see recital 38),

• The benchmark study found that the market rates and the rates applicable to CDP and PI on the basis of a reference sample of corporate bonds were as shown in table 6.

Table 6
Market fees for a reference sample of structured corporate bonds in the period 2006–2008 and fees applicable to CDP and PI (%)
a […].
b […].
Category of certificate		Benchmark		Remuneration for PI
		Min	Max
Ordinary	remuneration on sum collected	1,35	3,05	[…]a
Indexed	remuneration on sum collected	1,8	3,45	[…]b
18-month	remuneration on sum collected	0,48	1,25	[…]

(76) As shown in the table, the benchmark study, after analysing a sample of comparable bonds, concluded that the fees for the placement of postal savings certificates were within a range no higher than that for comparable financial products.

(77) The benchmark study also noted that if PI did not meet the targets set in the agreement for net funds raised in postal savings certificates, the remuneration could be reduced. Since postal savings certificates were repayable on demand, unlike bonds, PI had to manage the system with care in order to meet all the targets set.

(78) For the remuneration of the management of dematerialised postal savings certificates, the benchmark study suggested that the benchmark should be the levels charged to intermediaries by Monte Titoli SpA for central management of securities: the level of this part of the remuneration was market-conform (being in a range of 0,002-0,0023 %, depending on the stock of dematerialised postal savings certificates). In the case of paper savings certificates, however, it was not possible to carry out a comparative analysis of the Italian banking market, because this kind of postal savings certificate was the only one still on the market in paper form.

Comments from Italy following the initiation of the formal investigation procedure

(79) In its letters of 2 February 2007, 18 April 2007, 22 June 2007, 11 March 2008, 13 May 2008 and 17 June 2008, Italy’s main comments were as follows.

(80) Italy argued that the remuneration paid to PI for placing postal savings certificates since the year 2000 was in line with market conditions, and also that it was a maximum which would be paid to PI only if the objectives for the gross sum collected from postal savings certificates were to be achieved. That remuneration was therefore difficult to compare with what was received by other operators for placing similar products.

(81) For the identification of comparable products and their remuneration, Italy referred to the results of the benchmark study.

(82) It said the remuneration for placing postal savings certificates was lower than the remuneration for placing postal savings books, which the Commission’s decision of 22 November 2006 had found to be in line with market conditions. This was because the average life of postal savings certificates was about [7-10] years, and if the 2,63 % remuneration for placing the certificates was distributed over that period the yearly remuneration for placing them was […] %, which was lower than the yearly remuneration for distributing postal savings books, namely 0,90 %. It was true, however, that there were differences between the two products: postal savings books were short-term financial products, whereas postal savings certificates were medium- to long-term ones; and postal savings books were simple products, whereas postal savings certificates were complex structured ones.

(83) Financial products in paper form were extremely rare in the market, but postal savings certificates could be considered comparable to the deposit certificates issued by banks (certificati di deposito — CDs). For these certificates subscribers paid the bank a one-shot fee varying from EUR 4,50 to EUR 9,30 per certificate, regardless of the amount of the deposit and the maturity. The average life of deposit certificates was 24 months, so that credit institutions were collecting an average annual fee of EUR 2,25 to EUR 4,65 per certificate. Taking the average amount on a postal savings certificate to be EUR 2 000, and applying PI’s percentage remuneration (0,11 %) to that amount, the annual fee for a paper certificate was EUR 2,20, which was in line with the commission on deposit certificates.

(84) With regard to the fee for the production of paper certificates, a comparable paper security was the paper certificate representing a holding in a mutual fund, the average cost of which was somewhere between EUR 10 and EUR 20 per certificate. PI’s remuneration of EUR 1,70 per certificate was therefore reasonable.

(85) As far as the margin for the placement of postal savings certificates was concerned, Italy said that these certificates were one of a wider range of financial services offered to PI’s customers in order to achieve the most appropriate asset allocation. The margin on Bancoposta services ([…] % in 2005) was lower than the margin on the comparable services offered by commercial banks (35 %).

(86) Italy said that in order to establish whether or not there was an economic advantage to PI it was not necessary to break the remuneration for comparable products into separate components: it was enough to compare the total remuneration of postal savings certificates to that of the comparable products.

(87) On the question of compliance with the market economy investor principle, Italy stated that the mechanisms for calculating the remuneration had been designed on the basis of the economic rationale underlying such decisions in the market. Although there were no studies dating back to 2000, therefore, the remuneration was nevertheless based on these economic principles, as had subsequently been confirmed by the benchmark study (see tables 4 and 5).

(88) PI kept separate accounts for retail financial services as a whole, and did not distinguish between different financial products for accounting purposes. A comparison of costs and revenues for individual products, including postal savings certificates, was made only for the purpose of management analysis, and was very approximate. That comparison showed that postal savings certificates carried a heavier burden of administration than other products, such as current accounts. For 2005, Italy found that the best proxy for the economic and financial data for postal savings certificates was a gross profitability of […] %(26).

(89) The Banca d’Italia indicated that postal savings certificates qualified as structured financial instruments, for the following reasons: they were securities in the nature of bonds, or at any rate of debt; they were derivatives, in that they incorporated a derivative contract of the option type; and the two components, bond and derivative, were inseparable components of one instrument.

Comments from Italy on ABI’s submissions

(90) Italy did not share ABI’s view regarding the identification of products comparable to postal savings certificates.

(91) Italy argued that the difference in pricing between different issuers having the same risk rating — sovereign (State and international bodies), financial, banking and corporate (industrial in the narrow sense) — was of the order of a few hundredths of one per cent: the difference in the cost of funding between two extreme ranges — sovereign and corporate — was about 5-6 basis points, and the difference between financial and corporate was 2-3 basis points. In other words, between issuers with the same rating, the type of issuer (State, financial institution, bank etc.) was not significant, and did not invalidate the conclusions of the study submitted by Italy. Contrary to what was affirmed by ABI, the benchmark study applied a broad definition of corporate bonds which referred to bonds issued not only by industrial companies but also by financial companies. Moreover, the average rating of the basket referred to by the benchmark study was AA–, higher than the rating of CDP, which issued postal savings certificates (A+), or of Italy, which guaranteed them (A+).

(92) In addition, for the placement of structured bonds issued by Italian banks PI/Bancoposta had received an average commission of 2,9 %, which was in line the remuneration for placing postal savings certificates.

(93) Italy contested ABI’s comparison of postal savings certificates to State bonds, for the following reasons.

• They were issued in different ways, State bonds by an auction system, and postal savings certificates on tap,

• The degree of complexity was different,

• The remuneration for placing State bonds was independent of the amount subscribed; this was not the case for postal savings certificates, which required intense planning and management in PI’s network in order to achieve specific placement targets. Management and organisational efforts and retaining the loyalty of customers were particularly important because investors could withdraw from the investment at any time.

(94) In terms of method and management of placement, therefore, postal savings certificates were not comparable to State bonds, but rather to mutual funds.

(95) As for the fact that the cost of collecting postal savings certificates had by law to be kept in line with the cost of Treasury debt on the market, Italy pointed out that postal savings certificates had a gross yearly return lower than that of State bonds (for the instruments suggested by way of example, with 8-year, 10-year and 20-year maturities, the difference amounted to 180 bps). Italy was of the view that the difference was due to the implied US-style put option. If distribution fees were then added, the cost to the State was in line.

(96) ABI had overlooked the fact that postal savings certificates carried an implicit put option, and could be compared to mutual funds. Postal savings certificates fell within the definition of structured bonds used by the Banca d’Italia.

(97) The commission for placing postal savings certificates also reflected the whole of the work done by PI in order to achieve certain objectives in term of gross subscription and net subscription, both for postal savings as a whole and for the individual products. This remuneration also took account of the customer management carried on by PI after placement and along the whole life of the certificate, specifically because of the investor’s entitlement to ask for the reimbursement of the certificate, and the absence of a secondary market.

(98) Italy contested ABI’s statement regarding the level of commissions for placing index-linked bonds (1 %-2,5 %): the average remuneration received by PI was 3,05 % for placing structured bonds issued by foreign and Italian banks(27), and 2,9 % for placing bonds issued by Italian banks only(28) (as an example, for placing structured bonds issued by Capitalia SpA PI received a fee of 3,212 % of the nominal value subscribed(29).

(99) Italy concluded that ABI’s study was a general one and overlooked the specific features of the placement of postal savings certificates as compared to that of other bonds.

VII. THE EXPERT STUDY

(100) The studies commissioned by PI and ABI followed similar methodologies, but their results were difficult to reconcile. In particular, neither identified products that were perfectly comparable, and in selecting one or other comparator they emphasised different factors affecting placement fees.

(101) The Commission therefore asked an independent expert to conduct a study in order to establish a market remuneration for the placement of postal savings certificates.

Methodology used in the expert study to define the comparators

(102) A fundamental finding of the expert study was that the Italian market for securities was divided into: (i) a first tier, in which issuers targeted institutional investors, yields were at the ordinary level given the default risk and the euro yield curve, distribution was in the hands of large investment banks (or was by open auction, as in the case of most Italian Treasury bonds), and distribution fees were relatively small; and (ii) a second tier, in which distribution took place through banking retail networks targeting individual savers, issuers could obtain better borrowing terms than in the first-tier market, and distribution fees were in the nature of things larger than in the first-tier market. This had direct consequences on the level of distribution fees embedded in financial products: for any given combination of the characteristics that could be expected to influence the distribution fees for the financial product, the simple fact that the product was sold on the second-tier market drove the fees sharply up.

(103) The expert study noted that that the distribution fees for financial products ought to be a function of the effort required of the distribution network for their distribution. It then identified the products most comparable to postal savings certificates by looking at their essential characteristics — risk, liquidity, contractual complexity(30), reputation and legal status of the issuer, maturity, size of the issue, offering structure, minimum investment required, and distribution channel (institutional/retail distribution) — and assigning ratings to these characteristics.

(104) It then compared the rating of each characteristic of each financial product with those of each category of postal savings certificate.

Table 7
Financial products and factors affecting distribution fees
a CCTs (certificati di credito del Tesoro) are seven-year Treasury bonds with a floating interest rate.
Product	Risk	Liquidity	Complexity	Reputation	Duration	Institutional or retail distribution	Offering structure	Minimum Investment required (EUR )
BOTs	low	high	low	High	short	institutional	auction	1 000
BTPs	low to medium	high	low	High	medium to long	institutional	auction	1 000
CCTsa	low	high	low	High	medium	institutional	auction	1 000
Savings deposits	low	high	low	High	short	retail	on tap	none
Time deposits	low	low	low	High	short to medium	retail	on tap	none
Deposit certificates (CDs)	low	low	low	High	short to medium	retail	on tap	≥ 50
Listed plain vanilla bank bonds	low to medium	medium	low	High	medium	retail	public offer	≥ 1 000
Unlisted plain vanilla bank bonds	low to medium	low	low	High	medium	retail	public offer	≥ 1 000
Listed structured bank bonds	high	medium	high	High	medium	retail	public offer	≥ 1 000
Unlisted structured bank bonds	high	low	high	High	medium	retail	public offer	≥ 1 000
Corporate bonds	medium to high	low to medium	low	low to high	medium	institutional	public offer	≥ 1 000
Money market mutual funds	low	high	medium	Medium	not applicable	retail	on tap	≥ 100
Bond mutual funds	medium	high	medium	Medium	not applicable	retail	on tap	≥ 100
Equity mutual funds	high	high	medium	Medium	not applicable	retail	on tap	≥ 100
Equity (listed)	high	medium to high	low	low to high	not applicable	institutional	public offer	≥ 1 000
Ordinary postal savings certificates	low	medium	low	High	long	retail	on tap	≥ 50
Fixed-term postal savings certificates	low	medium	low	High	medium	retail	on tap	≥ 50
Postal savings certificates indexed at maturity	high	medium	high	High	medium	retail	on tap	≥ 250
Postal savings certificates, inflation-linked	medium	medium	high	High	medium	retail	on tap	≥ 250
18-month postal savings certificates	low	medium	low	High	short	retail	on tap	≥ 50

(105) The distribution channel (institutional or retail distribution) was a key point in the identification of comparable products. Postal savings certificates were best compared to retail financial products, because they were typically sold to households. Families held almost the entire stock of postal savings certificates outstanding (about 99 %), and they were not usually bought by institutional investors. This was also the main attribute that meant that State bonds (BOTs, BTPs and CCTs) were not comparable to postal savings certificates.

(106) The expert study found that the product most closely comparable to postal savings certificates was bank bonds (ordinary and fixed-term postal savings certificates being comparable to ‘plain vanilla’bank bonds, and indexed postal savings certificates to indexed bank bonds — see table 14). The second most closely comparable product was mutual funds(31). Both of these comparable products existed only in dematerialised form(32).

(107) The Commission would observe at this stage that although the remuneration has to be assessed for both paper and dematerialised postal savings certificates, the two forms of postal savings certificate share other characteristics, namely average maturity, yield at maturity, and the features linked to PI as distributor.

Methodology used in the expert study to quantify the market remuneration of the comparators

(108) In order to be able to perform the comparison needed to establish a distribution fee for postal savings certificates, the expert constituted a dataset of bank bonds and mutual funds using (i) a large number of prospectuses for bank bonds listed on TLX, the second organised exchange for securities (after Borsa Italiana)(33) and (ii) the websites of major Italian banks, where prospectuses for bond issues are usually available for download. The dataset consisted of 511 bond issues distributed on the Italian market between 2002 and 2008 (162 were issued before 2006). Of total issues in each year the proportion represented by this dataset was as follows: 18 % in 2006, 15 % in 2005, 9 % in 2004, 3 % in 2003, and 0,2 % in 2002.

(109) The expert conducted additional analyses for each comparator using descriptive statistics. In the case of bank-issued ‘plain vanilla’ bonds, for example, which were the product most closely comparable to ordinary and fixed-term postal savings certificates, the expert considered that descriptive statistics were not sufficient to give a proper understanding of the factors determining distribution fees which could be used to judge whether the remuneration for the distribution of postal savings certificates was market-conform. In particular, bank bonds were not homogeneous in all the relevant respects, and might not be homogeneous with the particular class of postal savings certificates to which they were to be compared. The statistical evidence had therefore to be used rationally in order to find out, within a group of similar bonds, what characteristics were really relevant to the measurement of distribution fees. Here the expert made reference to the list of relevant factors that affected distribution fees (see paragraphs 103-104). The effect of these independent variables on the distribution fee for a product like postal savings certificates was estimated by means of regression analysis.

(110) The period covered by the sample of bank bonds prevented any conclusion being drawn regarding the behaviour of distribution fees over time in the years 2000-2003. For statistical reasons, it was not possible to construct meaningful time series for the years 2000-2006(34). The sample of bank bonds allowed conclusions to be drawn only for the period 2000-2006 in general, and not on a year-by-year basis.

(111) The expert was of the clear opinion that a conclusion relating to the whole period was nevertheless relevant, because field interviews helped to establish that there had been no appreciable change in the scale of distribution fees since 2000; some answers even indicated that they had decreased.

(112) The expert also established a dataset of 102 mutual funds(35); on this sample, the only useful data analysis carried out was a matter of descriptive statistics(36). However, a time series could be built for the fees earned by distributors for mutual funds.

(113) Consideration had also to be given to the structure of the remuneration for the distribution of the comparable products:

• the remuneration paid for the distribution of bank bonds consisted of an upfront percentage of the amount subscribed,

• some mutual funds paid the distribution network both upfront fees and annual fees, while others paid annual fees only; in order to obtain a single measure of the fees that would be valid for the purpose of the study, the expert added to the actual upfront fee for each mutual fund an estimated additional upfront fee, equal to the present value of the annual fees to be earned over the expected life of the customer’s investment in the fund.

Establishing figures for comparison with the actual remuneration paid for postal savings certificates by CDP

(114) The overall remuneration that PI received from CDP under the annual agreements for 2000-2006 was made up of a number of components (see also Table 4)(37):

a)

an upfront fee on the subscriptions of postal savings certificates, equal to a given percentage of the amount subscribed;

b)

an annual fee on the volume of postal savings certificates outstanding, equal to a given percentage of the value of the certificates outstanding; this fee was the remuneration for the administrative and accounting management of the certificates; since 2003 it had been different for paper certificates and dematerialised certificates;

c)

in some years, other fixed amounts, notably for achieving given quantitative or qualitative targets.

(115) Under the agreements for the years 2000-2002, the structure of remuneration was quite variable. The overall remuneration structure had become stable since 2003, in the sense that the overall remuneration was made up of the same components, although the figures for each component could vary. Generally, in the period 2000-2002 it was not easy to compare one year with another, and it was not easy to compare the period 2000-2002 period with the period 2003-2006.

(116) To make the remuneration for the distribution of postal savings certificates comparable to the remuneration for similar financial products, the expert converted the dual remuneration schedule of the certificates into an equivalent estimated overall upfront fee which would be directly comparable with the actual upfront fee computed for bank bonds and mutual funds. This single figure for postal savings certificates was obtained for each year by adding to the actual contractually defined upfront fee an estimated additional upfront fee, representing the present value of the remuneration that PI expected to earn in the future on the certificates subscribed in the year from the annual percentage fees on the stock of certificates outstanding. This additional upfront fee was estimated by considering the average life(38) of each class of certificate, and computing the present value, in the year of subscription, of the annual percentage fees that would be earned over the years of the average life of the certificate(39).

(117) In order to compute the present value referred to, the expert found that a further assumption had to be made regarding the annual fees to be earned in future. There were two equally acceptable options:

• One was to assume a constant annual remuneration, the reasoning being that PI could not forecast the variability of annual fees in future, and therefore would each year have computed the equivalent overall estimated upfront fee by assuming that the annual fees would remain fixed at the level contractually agreed in the agreement for that year,

• The second option was to assume a variable annual remuneration, quantifying the remuneration for the distribution of postal savings certificates on the basis of the annual fees set by the successive agreements over the life of the products(40). Here it was assumed that PI could forecast the variability of annual fees in future, and therefore would each year have computed the equivalent overall estimated upfront fee on the basis that the annual fees would have followed the path that can in fact be discerned ex post.

(118) The upfront fees for postal savings certificates recalculated by the variable annual remuneration method appear to be generally lower than those arrived at by applying the constant annual remuneration method. Table 8 illustrates this point for ordinary and fixed-term certificates:

Table 8
Overall estimated upfront fees for postal savings certificates (%)
Ordinary certificates (paper form)
	Constant annual remuneration	Variable annual remuneration
2000	[…]	[…]
2001	[…]	[…]
2002	[…]	[…]
2003	[…]	[…]
2004	[…]	[…]
2005	[…]	[…]
2006	[…]	[…]
Fixed-term certificates (paper form)
	Constant annual remuneration	Variable annual remuneration
2000	[…]	[…]
2001	[…]	[…]
2002	[…]	[…]

(119) For reasons of conservatism the expert relied in most of the study upon the assumption of a constant annual remuneration.

(120) Under that assumption the recalculated upfront fees received by PI for postal savings certificates are the following:

Table 9
Recalculated upfront fees received by PI for postal savings certificates
a These figures do not include the sums of EUR 51,6 million and EUR 25 million paid in 2001 and 2002 respectively for specific tasks having a scope beyond the mere distribution of postal savings certificates. However, the expert carried out a sensitivity check, and concluded that their inclusion would not affect the findings of the report.
	2000	2001a	2002	2003	2004	2005	2006
Paper certificates (all certificates before 2003)
Ordinary	[…]	[…]	[…]	[…]	[…]	[…]	[…]
Fixed-term	[…]	[…]	[…]
18-month						[…]	[…]
Dematerialised certificates (after 2003)
Ordinary				[…]	[…]	[…]	[…]
Indexed (equity-linked)				[…]	[…]	[…]	[…]
Indexed (inflation-linked)							[…]
18-month						[…]	[…]

Comparison with the remuneration estimated on the basis of the sample of bank bonds

(121) The data collected allow an estimate to be made of the spreads of market remuneration for the distribution of postal savings certificates. The estimated market remuneration differs, but does depend on which comparator is taken as the appropriate one.

(122) Table 10 shows the ranges of the overall estimated upfront fees for postal savings certificates (on the basis of the regression analysis on the sample of bank bonds) and the ranges of recalculated actual fees received by PI for postal savings certificates. The range of actual remuneration for postal savings certificates is defined by the minimum and maximum fee observed over the period 2000-2006. The range used for the comparison is defined by the minimum and maximum expected values(41) of fees for postal savings certificates on the basis of regressions of fees on the relevant factors. The table also shows the ranges built by adding one or two standard deviations to the expected minim and maxima(42).

(123) The reference to the classical intervals (mean + 1 standard deviation) and (mean + 2 standard deviations) is useful for assessing the probability of an event: an observation which is less than one standard deviation distant from the expected value has a high probability of being statistically similar to the expected value; when it is between one and two standard deviations distant, its probability is lower, but not abnormally low; when it is more than two standard deviations distant, its probability is quite low, so that it can be termed a ‘rare’ and potentially ‘abnormal’ event(43).

Table 10
Recalculated upfront fee and estimated market remuneration for postal savings certificates based on the sample of bank bonds (%)
a It is important to note that the regression analysis on the sample of both structured bank bonds has an explanatory value significantly lower than in the case of other bank bonds used to estimate the remuneration for the distribution of other categories of postal savings certificate. In this regard, the expert pointed out that in the case of index-linked certificates the results of the simulation appeared difficult to reconcile with fundamental economic mechanisms: it was not apparent why index-linked certificates should command fees so much lower than the average obtained in the equity-linked bank bonds subsample. This point, coupled with the evidence about the low explanatory power of the regressions for the fees of indexed certificates, suggested in the expert’s opinion that more weight should be given in this case to the result of the simple comparison of average values, which gave a more decisive indication of market conformity. The descriptive statistics indicate that the average remuneration is 3,10 % for equity-linked bank bonds and – 2,82 % for inflation-linked bank bonds. Both values are above the maximum remuneration received by PI for index-linked and inflation-linked postal savings certificates.
Form of certificate	Paper	De-materialised	Paper	Demat.	Demat.	Paper	Demat.
Category of certificate	Ordinary	Ordinary	Fixed-term	Equity-linkeda	Inflation-linked	18 month	18 month
Minimum fee for postal savings certificates 2000-2006	[…]	[…]	[…]	[…]	[…]	[…]	[…]
Maximum fee for postal savings certificates 2000–2006	[…]	[…]	[…]	[…]	[…]	[…]	[…]
expected value — lowest estimate	3,05	3,05	1,76	1,69	2,83	0,63	0,63
exp. value + 1 std. dev.	3,4	3,4	2,11	2,51	2,14	0,88	0,92
exp. value + 2 std. dev.	3,75	3,75	2,46	3,32	4,2	1,21	1,21
expected value — highest estimate	4,54	4,54	3,34	1,69	2,83	0,63	0,63
exp. value + 1 std. dev.	5,25	5,25	4,05	2,51	3,51	0,92	0,92
exp. value + 2 std. dev.	5,96	5,96	4,76	3,32	4,2	1,21	1,21

(124) The conclusions drawn by the expert were therefore the following:

a)

for 18-month certificates, dematerialised ordinary certificates and inflation-linked certificates, there was ‘absolutely unambiguous’proof that the conditions were market-conform;

b)

for index-linked certificates and paper ordinary certificates, the evidence was not ‘absolutely unambiguous’, but did clearly tend to show that the conditions were market-conform; in these cases fees for postal savings certificates were, in regression-based simulations, either below the expected value, or within the interval defined by the expected value plus two standard deviations;

c)

for fixed-term certificates, there might have been an abnormal maximum in 2001.

(125) The expert also indicated expected distribution fees for ordinary and fixed-term postal savings certificates when dummy independent variables were included for each distributor, in order to capture another factor, the pricing policies of individual distributors, which was not included in the regression analysis shown in table 10(44):

Table 11
Estimated remuneration of distributor for postal savings certificates
	Plain vanilla bank bonds	Fixed-interest bank bonds
	Panel A: Ordinary postal savings certificates
UBM	3,6 %	3,23 %
PopVicenza	3,06 %	2,65 %
Antonveneta	2,95 %	2,98 %
Carige	2,62 %	2,53 %
IntesaSanPaolo	2,25 %	2,73 %
SanPaoloIMI	1,44 %	2,45 %
	Panel B: Fixed-term postal savings certificates
UBM	2,75 %	1,91 %
PopVicenza	2,21 %	1,33 %
Antonveneta	2,1 %	1,66 %
Carige	1,77 %	1,21 %
IntesaSanPaolo	1,4 %	1,41 %
SanPaoloIMI	0,59 %	1,13 %

Comparison with fees for mutual funds

(126) The overall recalculated upfront fees paid to PI for the distribution of postal savings certificates (point estimate) and for money-market mutual funds are shown in table12.

Table 12
Recalculated upfront fees and estimated yearly market remuneration for postal savings certificates based on the sample of mutual funds
Ordinary postal savings certificates
	Postal savings certificates	Money-market mutual funds
Year		mean	mean + 1 standard deviation
2000	[…]	1,48 %	3,41 %
2001	[…]	1,48 %	3,41 %
2002	[…]	2,13 %	4,06 %
2003	[…]	2,45 %	4,38 %
2004	[…]	2,55 %	4,48 %
2005	[…]	3,38 %	5,31 %
2006	[…]	3,38 %	5,31 %
Fixed-term postal savings certificates
	Postal savings certificates	Money-market mutual funds
Year		mean	mean + 1 standard deviation
2000	[…]	1,15 %	3,08 %
2001	[…]	1,15 %	2,8 %
2002	[…]	1,65 %	3,3 %

(127) Working on the assumption of a constant annual remuneration, the expert computed the annual remuneration for the distribution of postal savings certificates in absolute figures, as compared with the remuneration PI would have obtained if it had been paid the average rates received by distributors of comparable mutual funds. These yearly absolute figures for the remuneration of postal savings certificates and for the remuneration that would have been payable on the basis of the terms for money-market mutual funds are presented in Table 13.

Table 13
Actual and estimated remuneration for postal savings certificates, in absolute figures
	Postal savings certificates	Money-market mutual funds
		mean	mean + 1 standard deviation
2000	361	217	426
2001	439	235	459
2002	316	355	596
2003	416	457	721
2004	482	519	803
2005	582	762	1 062
2006	559	820	1 139
Total 2000-2006	3 157	3 364	5 207

(128) The expert concluded that:

a)

By the first method (table 12), the fees for postal savings certificates nearly always lay between the expected value and the value equal to one standard deviation, if they were estimated on the basis of fees for money market funds, and nearly always below the expected value estimated on the basis of euro government bond funds;

b)

By the second method (table 13), the actual total remuneration obtained by PI was always lower than the remuneration that would have been obtained by applying the average terms for euro area government bond mutual funds. The actual total remuneration was also lower than the remuneration that would have been obtained by applying the average terms for money-market mutual funds in 2003-2006. In the years 2000-2003, it was no lower than the average terms for money-market mutual funds; but it was lower than the remuneration corresponding to a value equal to the expected value plus one standard deviation;

c)

The conclusion that the fees are market-conform, obtained from the comparative evidence on mutual funds’ fees, can be regarded as close to ‘absolutely unambiguous’.

Specific fees for the production of paper certificates

(129) The total fees that CDP paid to PI amounted to EUR 15 million in 2003, EUR 16 million in 2004, and EUR 18 million in 2005 and 2006 (see table 4).

(130) In the analyses outlined above, examining whether or not the fees paid to PI for distributing postal savings certificates were market-conform, the remuneration for the production of paper certificates was not considered. The expert took the view that it was reasonable not to consider this component in the computation of the additional upfront fee, because it related to a highly specific feature of postal savings certificates that was not to be found in comparable financial products in the period 2000-2006. As explained above, Italy compared the compensation PI received for the production of paper certificates to the payment required for issuing paper certificates by mutual funds. Italy observed that the payment required for paper certificates issued by mutual funds amounted to EUR 10 to EUR 20 per certificate, while the fee paid by CDP to PI amounted to EUR 1,70 per paper certificate, and concluded that this did not seem excessive. The expert confirmed that the amounts charged by mutual funds for the production of paper certificates were at the level described by Italy.

(131) Therefore, the fee paid by CDP to PI for the production of paper certificates could be regarded as market-conform.

The expert’s conclusion

(132) The overall conclusion of the expert report is that the remuneration paid for postal savings certificates in the years 2000-2006 was market-conform.

VIII. ASSESSMENT OF THE MEASURE

(133) In order to ascertain whether a measure constitutes State aid within the meaning of Article 87(1), the Commission has in principle to assess whether the scheme:

• is granted by the State or through State resources,

• confers an economic advantage,

• is capable of distorting competition by favouring certain undertakings or the production of certain goods,

• affects trade between Member States.

VIII.1. The context of the assessment

(134) In order to reach a conclusion in the present case, the Commission will have to consider only whether the remuneration at issue conferred an advantage on PI.

(135) As a consequence of the Order of 6 October 2004, the collection of postal savings has become a service of general economic interest(45). From 2004 onward, therefore, the assessment has to be conducted in the framework established by the Altmarkjudgment(46).

(136) In the opening decision the Commission established that the recipient, PI, had public service obligations to discharge, and that these obligations were clearly defined. The parameters on the basis of which the compensation was calculated had been established in advance in an objective and transparent manner, notably by means of agreements between CDP and PI. It remained to be established whether the compensation might exceed what was necessary to cover all or part of the costs incurred in the discharge of the public service obligations, taking into account the relevant receipts and a reasonable profit; and in the absence of a public procurement procedure, the level of compensation needed had to be determined on the basis of an analysis of the costs that would be incurred by a typical undertaking, well run and with proper resources, taking into account the receipts and a reasonable profit for discharging the obligations.

(137) If the remuneration paid by CDP was market-conform, it would provide an appropriate estimate of the level of the costs that would be incurred by a typical well-run undertaking, taking into account the receipts and a reasonable profit for discharging the obligations. In that event the remuneration would satisfy the four tests in Altmark, and there would be no State aid(47). Lastly, if the yearly remuneration paid to PI on postal savings books was market-conform before the service of general economic interest was entrusted to PI and CDP, it is obvious that no advantage was conferred on PI, and that this remuneration did not constitute State aid either.

(138) To analyse whether the remuneration paid by CDP was market-conform, the Commission will follow the main stages of the methodology used by the benchmark study, a methodology which the Commission endorsed in the opening decision and which was likewise followed in the other studies conducted subsequently:

a)

first, the financial instruments most comparable to the various categories of postal savings certificate have to be identified (section VIII.2); it is of the utmost importance here that the essential factors affecting the level of distribution fees should be identified and properly taken into account;

b)

second, the market remuneration paid by market participants for the placement of these comparables has to be established and compared with the remuneration paid by CDP (section VIII.3).

VIII.2. Definition of the comparators

VIII.2.1. Validation of the expert study by the Commission

(139) The three studies differ essentially with regard to the comparators identified and the underlying determinants of the remuneration influencing the choice of comparables. The following table presents the comparators identified by the three studies:

Table 14
The comparators identified by the three studies
a Very little public data is available, since the issuers are not obliged to disclose the fees paid for the distribution of CDs, and this comparator has not been used extensively in the expert study.
Product	Benchmark study	ABI study	Expert study
Ordinary postal savings certificates	Structured corporate bonds with a rating of at least BBB (S&P’s) or Baa2 (Moody’s)	BTPs of similar maturity	Bank-issued plain vanilla bonds (fixed-rate and floating-rate)
Fixed-term postal savings certificates	Structured corporate bonds with a rating of at least BBB (S&P’s) or Baa2 (Moody’s)	BOTs, CTZs and BTPs depending on the residual life of the postal savings certificate	Bank-issued plain vanilla bonds
Indexed postal savings certificates (inflation-linked and equity-linked)	Structured corporate bonds with a rating of at least BBB (S&P’s) or Baa2 (Moody’s)	For inflation-linked certificates: BTPEUR i indexed; for equity-linked certificates: index-linked financial bonds with a rating of at least A+ (S&P’s) or Aa2 (Moody’s).	Bank-issued indexed bonds (equity-linked or inflation-linked as the case may be)
18-month postal savings certificates	Structured corporate bonds with a rating of at least A– (S&P’s) or A3 (Moody’s)	Short-term BOTs and CTZs	Deposit certificates (CDs)a, with useful indications provided by shorter-maturity fixed-interest bank bonds, and money-market mutual funds
Comments on other comparators	Italy considers postal savings certificates to be comparable to mutual funds investing in euro area government bonds	ABI considered initially that postal savings certificates were comparable to mutual funds investing in euro area government bonds. In the ABI study, however, comparability between postal savings certificates and mutual funds is no longer mentioned.	Useful indications for the determination of the market level of the remuneration for the distribution of postal savings certificates (except for indexed certificates) might be provided by the observation of the remuneration for the most closely comparable types of mutual funds, notably money-market mutual funds.

(140) The Commission considers that the methodology used by the expert (see in particular table 7 above) and its outcome are fully appropriate. In particular, the comparators identified by the expert address the concerns of both the Italian authorities and ABI as detailed below. All factors relevant for defining the essential characteristics of the relevant financial products are assessed and weighed in a reasonable manner. Moreover, the expert study identifies a very significant factor influencing distribution fees which has been overlooked by both the Italian authorities and ABI, namely the distribution network. The expert noted here that in distributing products targeting retail investors, at least in Italy, it was possible to offer investors a lower return for the same level of risk than would have to be offered on similar products targeting institutional investors. In this case, the higher level of distribution fees earned by the retail network, in comparison to what can be obtained by distributors placing with institutions, can be thought of as a form of sharing the surplus generated(48).

(141) While no perfect comparators for postal savings certificates can be found, the analysis presented in the expert study will allow a market remuneration to be estimated for products with characteristics similar to those of postal savings certificates (for instance by means of regression analysis).

VIII.2.2. The Commission’s opinion on the arguments of the Italian authorities

(142) The benchmark study and the Italian authorities have frequently asserted that the put option embedded in postal savings certificates makes them a more complex financial product than ‘plain vanilla’bonds, and that they have to be compared to complex products like structured bonds.

(143) Some instruments, such as ‘plain vanilla’bonds and ordinary shares, have relatively straightforward structures and payoffs, while others, such as structured bonds and bonds with option-like provisions, have complex structures and payoffs that are contingent on the behaviour of underlying assets, or depend on a series of events. Explaining and, especially, selling a product with a complex structure to retail investors is a demanding task for retail networks, which have to invest resources to educate their workforce and allocate valuable resources and time to the process.

(144) Moreover, there is no generally accepted official statement issued by any regulatory body with binding legal effect that provides an exact definition of what is a ‘plain vanilla’ or a ‘structured’ bond(49). In addition, if a broad definition of ‘structured’ bonds is adopted, many financial products would lie between strictly ‘plain vanilla’ and strictly ‘structured’. The expert noted that a puttable bond like the postal savings certificate cannot be considered equivalent to a typical structured bond. The structure of the return to the investor on a postal savings certificate was simple, and small savers in Italy had been well aware for several decades of the significance of the possibility of early redemption of the invested capital and accrued interest, so that no special effort on the part of the distributor was required in order to explain the structure of the product.

(145) The expert therefore adopted the definition whereby a bond is ‘structured’ when its return is linked to trends in given parameters, according to a mechanism set out in the terms and conditions of the issue.

(146) The Commission shares the expert’s opinion that postal savings certificates cannot be considered ‘structured’ products on the usual definition by reason merely of the put option embedded in them.

(147) To conclude, a comparison between postal savings certificates and structured instruments, as proposed by Italy, is not appropriate, except in the case of indexed certificates(50).

VIII.2.3. The Commission’s opinion on the arguments of ABI

(148) ABI has frequently suggested a comparison with short-term and long-term State securities. This comparison is not appropriate, because Treasury products are distributed essentially to institutional investors, while bank bonds and postal savings certificates are distributed according to the typical process of the retail segment of the market. Moreover, the level of remuneration charged by banks for the distribution of Treasury bonds to retail investors is not a market price resulting from the interplay between supply and demand for retail distribution services, but is fixed unilaterally by the State, and has the force of law. This fee consequently cannot be taken as the market level of remuneration for distribution services in the Italian market(51).

(149) The Commission considers that the comparators identified by the expert address ABI’s concerns in that i) bank bonds are issued by financial companies, and hence are more closely comparable to postal savings certificates issued by CDP; ii) State securities and bank bonds have similar risk profiles; and iii) the bank bonds considered comparable to ordinary and fixed-term postal savings certificates are not structured bonds but ‘plain vanilla’ bonds.

VIII.2.4. Conclusion

(150) In conclusion, the Commission takes the view that the comparators relevant for purposes of the present Decision are those identified by the expert study, i.e. bank bonds in the first place, and then mutual funds with similar characteristics.

VIII.3. Estimated market remuneration for postal savings certificates

VIII.3.1. Methodology for evaluating market remuneration

a) Postal savings certificates

(151) The Commission considers that the expert’s methodology and calculations regarding equivalent upfront fees for postal savings certificates are appropriate. In particular,

• the split of the remuneration according to its nature (annual, upfront, others) is appropriate, and

• the expected life, the computation of present value and the discount rate for all products accepted by the expert are based on reasonable assumptions.

(152) In conclusion, the upfront fees arrived at by the expert are correct.

b) Comparables

(153) The Commission takes account of the limitations of the available data, and agrees with the expert that on the basis of economic literature and market research there is no financial product that is perfectly comparable to postal savings certificates in terms of all the relevant factors identified. The Commission considers that the substantial volume of quantitative and qualitative data collected and used by the expert is adequate for the purposes of the present Decision(52).

(154) The Commission agrees that a conclusion regarding the whole period (2000-2006) is of relevance, because field interviews conducted by the expert established that there had been no appreciable change in the size of distribution fees for bank bonds since 2000 (some answers even indicated that they had decreased in size). Moreover, the negotiations between the parties have resulted in significant changes in the terms of remuneration over the years (for instance, the annual management fee has fallen both as a percentage and as an amount). The Commission is of the view that the fee negotiations must take account of a multiannual perspective. It is therefore important that the reasoning should cover a multiannual period.

(155) Furthermore, the Commission agrees with the expert that since the remuneration structures for postal savings certificates and all the comparable financial products are different, the rate of remuneration for the distribution of the different products needs to be expressed in homogenous terms. In this regard, since for bank bonds the only fees paid are distribution fees, the Commission considers it right to compute equivalent upfront fees for mutual funds.

(156) The Commission also considers that the regression analysis carried out on the basis of the sample of bank bonds appears generally to be the best method in order to establish the market remuneration for the distribution of postal savings certificates. The Commission agrees with the expert that the descriptive statistics for mutual funds, identified as comparables by both ABI and Italy, can be used to draw meaningful conclusions. The Commission accordingly finds that the means and expected values computed by the expert are acceptable.

(157) A technical difference has emerged between the Commission and the expert regarding the ranges to be applied to the expected values (standard error/standard deviation) in order to ascertain whether remuneration is market-conform. However, the Commission will not need to examine this question further here, because the ranges will not be needed in its assessment.

VIII.3.2. Analysis of market conformity

(158) It has to be assessed whether the following products were market-conform over the relevant period:

Table 15
period of subscription of the various types of postal savings certificate
Category of certificate	Dematerialised form	Paper form
Ordinary	2003-2006	2000-2006
Fixed-term	Not available for subscription	2000-2002
18-month	2005-2006	2005-2006
Indexed: equity-linked	2003-2006	Not available for subscription
Indexed: inflation-linked	2006	Not available for subscription

a) Dematerialised certificates

(159) As regards dematerialised postal savings certificates a comparison of the values in tables 9 and 10 shows the following.

• Ordinary certificates: over the relevant period, the upfront equivalent of the certificates is 2,64 %, while the lowest estimate of the expected value is 3,05 %. The remuneration is therefore market-conform,

• 18-month certificates: over the relevant period, the upfront equivalent of the certificates is 0,65 %, while the expected value is 0,63 %. The remuneration for the certificates is so close to the expected value that the Commission can reasonably conclude only that it is market-conform(53). Besides, the Commission considers, as the expert has shown, that the most similar bank product is the deposit certificate (CD). Despite the small size of the sample of these deposit certificates (which have no upfront distribution fee, and annual fees of 0,60 %), the Commission feels that for purposes of this Decision it is possible to ascertain the level of the distribution fee. If an annual fee of 0,60 % on an 18-month security(54) is calculated back to present value in the years 2005 and 2006, the equivalent upfront fee obtained is equal to 0,863 %, which is higher than the estimated value of the upfront fee for 18-month postal savings certificates. The remuneration is therefore market-conform(55),

• Index-linked certificates: over the relevant period, the upfront equivalent of the certificates is 2,64 %. However, because of the limited explanatory power of the regression analysis on the sample of structured bank bonds (see footnote 55), and given that mutual funds are not an appropriate comparable for index-linked postal savings certificates, the Commission shares the expert’s view that the comparison has to rely on the results of the descriptive statistics. These results show that:

  i)

  the remuneration for inflation-linked postal savings certificates in the only year at issue — 2006 — was below the average remuneration for inflation-linked bank bonds (2,82 %, from 42 observations);

  ii)

  the remuneration for index-linked postal savings certificates in the years at issue (2003-2006) was below the average remuneration for index-linked bank bonds (3,10 %, from 109 observations).

  The remuneration is therefore market-conform.

b) Paper 18-month certificates

(160) Paper 18-month postal savings certificates are remunerated at the same rate as dematerialised ones (because there is no annual management fee for this type of certificate). This category of certificate is therefore also market-conform.

c) Paper ordinary certificates and paper fixed-term certificates

(161) It only remains to be ascertained whether the remuneration for paper ordinary postal savings certificates for the years 2000-2006 and for paper fixed-term certificates for 2000-2002 is market-conform.

(162) The Commission notes that the expert also compares paper postal savings certificates to dematerialised comparables, because he could not identify suitable comparables in paper form among the products available on the market.

(163) The Commission is of the view that the specific tasks that PI performs in the case of paper bonds are different from the simple production of paper certificates, for which a separate fee is paid (see paragraphs 130-132).

(164) It must be pointed out that the difference between the overall fees for paper and dematerialised postal savings certificates derives only from the management component of the fee(56).

(165) PI does carry out specific tasks connected to the management of the stock of paper certificates. Ordinary postal savings certificates can be ‘bought’in paper form any day in any Italian post office, by anyone who presents themselves in the post office with an identification document and with the amount in cash they wish to invest in the certificates. It is not necessary to have a current account with PI, because paper savings certificates can be bought for cash. The counter staff will ask the customer to fill in some forms, will enter the customer’s name on the certificates, and will hand over the certificates to the customer. For the redemption of the certificates the procedure is much the same. Managing these steps requires an organisational effort that is not needed, or is at least easier, in the case of dematerialised postal savings certificates and bank bonds.

(166) Thus the larger remuneration on the annual stock of paper certificates seems to be justified by the specific tasks that PI performs in respect of these paper certificates, so that a comparison between the overall estimated upfront fees for paper ordinary postal savings certificates and bank-issued bonds would be biased against PI, and a determination of the question whether the remuneration for postal savings certificates is market-conform ought to be based mainly on the evidence concerning dematerialised certificates.

(167) The Commission agrees with the expert that PI performs specific tasks in the case of paper postal savings certificates that go beyond compensation for the material production of the paper certificates(57). The additional tasks justify higher annual fees. On average over the period, the annual remuneration paid to PI for the specific additional tasks of managing paper certificates was around 0,14 % of the stock of paper certificates (see table 4)(58). The costs incurred in the management of paper certificates cannot be quantified precisely, but the Commission takes the view that the small extra remuneration identified here is not excessive, considering the work and organisational effort required for the management of paper certificates (described in paragraph 166). In 2006, the management fees paid to PI amounted to just EUR 7 126 per post office branch. In 2000-2001, when specific qualitative targets were negotiated between the parties(59), the management fees paid to PI amounted respectively to EUR 18 064 and EUR 14 231 per post office branch(60). The Commission is of the view that these fee levels are proportionate to the management of paper certificates, and that they are market-conform. The Commission also notes that the trend of this annual remuneration has been downward, and that the volume of paper certificates is expected to diminish over time.

(168) Fixed-term postal savings certificates were never available in dematerialised form: here the Commission has computed the fees paid to PI excluding the remuneration for management tasks. Over 2000-2002, the figure obtained is 1,72 % on average, taking a conservative approach(61), which is below the lowest estimated market remuneration (1,76 % — see table 10) and below several estimates for the remuneration of fixed-term postal savings certificates, taking account of the pricing policies of different distributors (see table 11, panel B). The Commission can therefore conclude that the remuneration paid for paper fixed-term postal savings certificates was market-conform.

(169) These considerations lead to the conclusion that the remuneration for paper ordinary and fixed-term postal savings certificates are market-conform.

(170) The Commission accordingly concludes that the remuneration paid was market-conform in the years 2000-2006 for all paper postal savings certificates.

Alternative method

(171) As an alternative, in this specific case, the Commission might have taken the view that the fees paid to PI were market-conform as long as they were below the highest estimate of the expected value.

(172) If that view were to be taken, table 10 shows that the remuneration for paper ordinary certificates and paper fixed-term certificates would be market-conform: the highest estimate of the expected value for paper ordinary postal savings certificates is 4,54 % (the highest limit of the expected value), while the maximum fee over 2000-2006 is 3,07 %. The respective values for paper fixed-term postal savings certificates are 3,34 % and 3,09 %.

VIII.3.3. Specific observations

a) Variable annual remuneration methodology

(173) Using the variable annual remuneration methodology, the Commission would have arrived at the same conclusions.

(174) The variable annual remuneration methodology is acceptable, because the Commission considers that the annual negotiations between PI and CDP took account of the variability of annual fees.

(175) If the Commission had followed that approach, its conclusions on market conformity would have been even stronger.

b) The remuneration for the production of paper certificates

(176) The expert observes that it can be stated with absolute certainty that paper circulation is specific to postal savings certificates, with the single exception of units in mutual funds, where it is still legally permitted. ABI likewise acknowledges that financial markets now deal almost exclusively in dematerialised products. ABI does not present any arguments or data suggesting how it should be determined whether this remuneration is market-conform, nor does it contest the calculation of a comparable remuneration for the production of paper certificates by means of a comparison with the paper certificates issued by mutual funds. The Commission therefore agrees with the expert’s conclusion on this question (see paragraphs 130-132).

VIII.4. Reply to remaining comments by Italy and ABI

VIII.4.1. Italy

(177) With regard to the arguments put forward by Italy concerning the market conformity of the margin for placing postal savings certificates (paragraphs 85-88), the Commission would observe that the simulated profit margin indicated by Italy for postal savings certificates is not sufficient to demonstrate the market conformity of distribution fees for the certificates, since, as the Italian authorities acknowledge, PI does not keep separate accounts for separate financial products, so that for purposes of this Decision it is not possible to assemble reliable information regarding the revenue and costs generated by the certificates.

(178) The comparison drawn between PI/Bancoposta’s overall profit margin and the margin of all comparable activities of commercial banks is even less specific.

(179) The Commission consequently considers that no conclusion can be drawn on the basis of this argument regarding the market conformity of distribution fees for postal savings certificates, which are only one of the services provided by Bancoposta, and which have to be compared to activities and products of commercial banks that are not comparable.

(180) With regard to the comparison advanced by the Italian authorities between the distribution fees for postal savings certificates and the fees for postal savings books, and their extrapolation of the allegedly market-conform remuneration for postal savings certificates over their average life (see paragraph 83), the Commission considers in the light of the analyses carried out by the expert that the assumption of a linear effect of the maturity on the distribution fees is much too simplistic. In addition, it is more appropriate to compare postal savings certificates with a financial instrument issued by market operators.

VIII.4.2. ABI

(181) ABI argues that the internal rate of return of State bonds is always lower than of equally rated corporate bonds, and that this is due to a difference in risk even where the financial instruments concerned have the same rating; but the Commission considers that the difference in returns is better explained by differences between distribution networks (retail and wholesale) rather than by a difference in level of risk.

VIII.5 Separate accounting

(182) According to the Community framework for State aid in the form of public service compensation(62), where an undertaking is entrusted with the operation of several services of general economic interest, the undertaking’s internal accounts must make it possible to ensure that there is no overcompensation at the level of each service of general economic interest.

(183) It is clear from the information provided by Italy that in PI’s internal accounts there is no proper separation of the costs and receipts associated with the service of general economic interest relating to postal savings and the costs and receipts associated with other services.

(184) The Commission would therefore remind Italy that it has an obligation to discharge in this regard.

IX. CONCLUSIONS

(185) The market remuneration paid by Italy to PI for the distribution of postal savings certificates is an appropriate estimate of the level of the costs that would be incurred by a typical well-run undertaking for discharging the same obligations, taking into account the receipts and a reasonable profit for discharging them.

(186) Since the service of general economic interest was entrusted to PI and CDP in 2004, the four Altmark tests have been satisfied. The remuneration paid to PI for distributing postal savings certificates between 2004 and 2006 did not constitute State aid.

(187) Before the service of general economic interest was entrusted to PI and CDP, the remuneration paid to PI for distributing postal savings certificates was market-conform. As no advantage was conferred on PI, the remuneration did not constitute State aid.

(188) The Commission takes the view that the remuneration paid to PI for distributing postal savings certificates in the years 2000-2006 does not constitute State aid,

HAS ADOPTED THIS DECISION:

(1)

OJ C 31, 13.2.2007, p. 11.

(2)

See footnote 1.

(3)

Under Article 5 of Decree-law No 269 of 30 September 2003, converted into statute by Act No 326 of 24 November 2003, the shares in CDP were assigned to the State. Foundations and other public or private parties may hold shares which together must not amount to more than a minority of the whole.

(4)

The universal service comprises the conveyance of items of correspondence and addressed printed matter weighing up to 2 kg and postal packages of up to 20 kg, and services relating to registered items and insured items.

(5)

Legislative Order No 261 of 22 July 1999, published GU 182, 5.8.1999, and Ministry of Communication Order of 17 April 2000, published GU 102, 4.5.2000.

(6)

Source: PI’s website, February 2008.

(7)

ECORYS-NEI, Development of Competition in the European Postal Sector, July 2005.

(8)

Directive 97/67/EC of the European Parliament and of the Council of 15 December 1997 on common rules for the development of the internal market of Community postal services and the improvement of quality of service (OJ L 15, 21.1.1998, p. 14).

(9)

Directive 2002/39/EC of the European Parliament and of the Council of 10 June 2002 amending Directive 97/67/EC with regard to the further opening to competition of Community postal services (OJ L 176, 5.7.2002, p. 21).

(10)

See footnote 7.

(11)

Directive 2008/6/EC of the European Parliament and of the Council of 20 February 2008 amending Directive 97/67/EC with regard to the full accomplishment of the internal market of Community postal services (OJ L 52, 27.2.2008, p. 3).

(12)

Study by PriceWaterhouseCoopers, The Impact on Universal Service of the Full Market Accomplishment of the Postal Internal Market in 2009, Final Report, May 2006.

(13)

FitchRatings, special report, The European Regulated Mail Sector: Tomorrow’s deliveries, 9 July 2004.

(14)

In some cases, such as debit cards and standing debit orders, the service is provided by PI itself; in other cases, PI/Bancoposta distributes third-party products (such as credit cards, which it distributes on behalf of banks).

(15)

Poste Vita SpA is a wholly-owned subsidiary of PI.

(16)

Bancoposta Fondi SGR SpA is a wholly-owned subsidiary of PI.

(17)

Published GU 241, 13.10.2004.

(18)

See recital 32 of the decision initiating the proceedings. In July 2006, CDP began issuing savings certificates for minors, which mature at the holder’s eighteenth birthday. This category of certificate is outside the scope of the opening decision and of the present decision. In any case, the stock of savings certificates for minors is marginal compared to the total stock of certificates: it represented 0,3 % of the total stock on 31 December 2007.

(19)

Dow Jones Euro Stoxx 50. See CDP annual report 2004.

(20)

Judgments in Case C-280/00 Altmark Trans GmbH and Regierungspräsidium Magdeburgv Nahverkehrsgesellschaft Altmark GmbH [2003] ECR I-7747 and Joined Cases C-34/01 to C-38/01 Enirisorse SpA v Ministero delle Finanze [2003] ECR I-14243.

(21)

BOTs (buoni ordinari del tesoro) are short-term zero-coupon Treasury bonds; BTPs (buoni del tesoro poliennali) are long-term Treasury bonds; and CTZs (certificati del tesoro zero-coupon) are short-term zero-coupon Treasury bonds (maximum two-year maturity). The commissions for placing these securities are set by law.

(22)

According to ABI, corporate bonds exclude bonds issued by financial companies.

(23)

‘Plain vanilla’ is an expression commonly used to mean ‘the most basic or standard version of a financial instrument, usually options, bonds, futures and swaps’ — www.investopedia.com, quoted by the Commission’s independent expert.

(24)

Italy uses the general term ‘corporate bonds’, but the reference sample employed in the benchmark study is made up mainly of bonds issued by banks (see p. 7 in the Italian authorities’ letter of 22 June 2007). The term ‘corporate’ is thus used in a broad sense, to refer to bonds issued not just by industrial corporations but also by banks.

(25)

The simulation was based on the placement commission and the annual management commission returned to the distributor in a year, multiplied by the notional period for which the fund would be held, which was taken to be equal to the average life of postal savings bonds. For bond investment funds the subscription fee varied between 0,70 and 2,20 %, and the management fee between 0,80 and 1,05 %, though some mutual funds applied only a management fee.

(26)

On the basis of PI’s separation of accounts and the average values for financial services, the profitability ratio for 2005 was […] %. In view of the additional costs of managing postal savings certificates, Italy indicated that a reasonable proxy for the profitability ratio of postal savings certificates was […] %.

(27)

Average fee calculated for the period November 1999–December 2007 on a total volume of EUR 33 billion.

(28)

Average fee calculated for the same period on a total volume of EUR 22 billion.

(29)

The value of the issue was EUR 971,2 million.

(30)

‘Plain vanilla’ or structured instrument.

(31)

In theory the product most similar to 18-month postal savings certificates was deposit certificates. But the expert was able to find only one issuer of deposit certificates which distributed its products through other bank networks and disclosed the distribution fees it paid.

(32)

Among the financial products in circulation in Italy the expert could not identify any comparables in paper form.

(33)

TLX SpA organises and manages markets for the trading of financial instruments aimed at the investment needs of non-professional investors.

(34)

A time series is a sequence of values, measured typically at successive times, and spaced at intervals of time that are often uniform. A time series of data on comparables would have permitted a comparison between the real annual fees for postal savings certificates and the corresponding figures for the comparators.

(35)

These funds were being distributed at the end of 2007.

(36)

Regression analysis would be meaningless here, because the mutual funds were by contractual design homogenous with respect to the variables employed in the regression analysis for bank bonds.

(37)

The remuneration for the production of paper certificates was not included in the analysis at this stage. The remuneration for the production of paper certificates is a very specific feature of the services provided by PI, and will be assessed separately (see paragraphs 130-132).

(38)

The expert used the terms ‘average life’ (durata media) and ‘expected life’ (durata prevista) as synonyms. In this Decision the term used is ‘average life’.

(39)

The expert used the values provided by Italy. When these data were lacking (for inflation-linked and 18-month certificates), the expert made reasonable estimates (at 7 years and 18 months respectively). The discount rate in the computation of the present value of the annual fees was the rate for Treasury bonds of similar maturity.

(40)

The rates in the 2006 agreement were used from 2006 onward.

(41)

In statistics the expected value (or mathematical expectation, or mean) of a random variable is the sum of the products of the value of each possible outcome multiplied by the probability of that outcome. Thus, it represents the average amount one ‘expects’as the outcome of a random trial if identical odds are repeated many times. The minimum and maximum expected values here refer to two best performing regressions, as the expert study indicates. In particular, low estimates relate to the sample of fixed-rate bank bonds, a sample which allows an important variable of yield at maturity to be used in the regression analysis. The higher estimates derive from the regression analysis based on the sample of all ‘plain vanilla’bank bonds, which includes both fixed-rate and variable-rate bonds.

(42)

In probability and statistics, the standard deviation is a measure of the dispersion of a set of values. It is defined as the root mean square (RMS) deviation of the values from their mean, or as the square root of the variance. The standard error of a method of measurement or estimation is the estimated standard deviation of the error in that method. Specifically, it estimates the standard deviation of the difference between the measured or estimated values and the true values (source: Wikipedia).

(43)

When comparing the actual outcome with the ‘classical intervals’ of mean plus or minus one or two standard errors, it is usually assumed that the underlying distribution is symmetric and normal. In this case, we can say that the outcome has a probability of 68,26 % of falling in the interval defined by the mean plus or minus one standard error, and a probability of 95,44 % of falling in the interval defined by the mean plus or minus two standard errors. However, the relevance of the reference to the interval defined by the mean plus or minus one or two standard errors is not limited to the case of the normal distribution, so that is possible in general to say that an event falling within one standard deviation of the mean can be considered — roughly speaking — ‘ordinary’ (this interval will be employed in the analysis in the next section).

(44)

In regression analysis, a ‘dummy variable’(also known as an ‘indicator variable’ or just a ‘dummy’) is one that takes the values 0 or 1 to indicate the absence or presence of some categorical effect that may be expected to shift the outcome. Use of dummy variables usually increases model fit (coefficient of determination), but at a cost of fewer degrees of freedom and loss of generality of the model. Too many dummy variables result in a model that does not provide any general conclusions (source: Wikipedia).

(45)

However, the Commission does not consider that the successive remuneration agreements between CDP and PI entrust a service of general interest by themselves.

(46)

Judgments in Case C-280/00 Altmark Trans GmbH and Regierungspräsidium Magdeburgv Nahverkehrsgesellschaft Altmark GmbH [2003] ECR I-7747 and Joined Cases C-34/01 to C-38/01 Enirisorse SpA v Ministero delle Finanze [2003] ECR I-14243.

(47)

See also the opening decision of 22 November 2006, paragraph 88.

(48)

The Commission would observe that it is commonly accepted among financial specialists that because of market inefficiencies distribution fees are different at wholesale and retail level. Under present market conditions, the difference between the fees goes to the intermediaries, and not to the issuers (see Hayes, Roger G., ‘The Path to Bond Market Efficiency: How increased retail distribution can lower borrowing costs’, in Government Finance Review, 1.6.2003.

(49)

One commonly employed financial glossary on the web (InvestorWords.com) defines a structured note as ‘a debt security with one or more special features, such as making payments based on an underlying index. For instance, a structured note is a bond which, instead of paying the typical interest payments, will use an index, such as the S&P 500, to determine the amount of the interest payment. This type of debt security is complex and is used primarily by sophisticated investors’. The OECD adopts this definition: ‘Structured bonds have characteristics that are designed to attract a certain type of investor and/or take advantage of particular market circumstances. However, structuring securities to appeal to a particular type of investor risks the possibility of loss of liquidity if the market moves in such a way as to make the structured features of the issue no longer attractive. Typically the structured features are achieved through the use of derivatives — for instance, a credit-linked note is a bond with an embedded credit derivative’.

(50)

The note from the Banca d’Italia on the definition of structured bonds does not change this assessment. The Commission is of the view that the real problem is not one of terminology, choosing the right adjective for a given noun, but rather a problem of substance, finding the right comparable for a specific financial product. Even if it were to be accepted that postal savings certificates were structured bonds, they would in any event have to be compared to structured bonds with similar characteristics, bonds with a US-style put option embedded, and nothing more.

(51)

Even if Treasury bonds were to be taken to be the closest comparable to postal savings certificates, on the same footing as bank-issued bonds, the expert concluded that their inclusion would not affect the empirical assessment of the market remuneration for the distribution of postal savings certificates. The Commission shares that view.

(52)

The expert collected a large number of prospectuses for bank bonds listed on TLX, the second organised exchange for securities in Italy (after Borsa Italiana), which is managed by TLX SpA. He also consulted the websites of the major Italian banks: Mediobanca, Intesa San Paolo, Unicredit, UBI Banca, Banca Carige, Cassa di Risparmio di Pistoia, Banca Popolare di Verona, Banca Antonveneta, Banca Popolare di Milano, Banca delle Marche, Banca IMI, Mediocredito del Friuli Venezia Giulia, Cassa di Risparmio di Parma e Piacenza, Banca Popolare di Vicenza, and Monte dei Paschi di Siena, and PI’s own website. He included in the database all euro-denominated bank bond issues for which the prospectus showed the distribution fees. The dataset consists of 511 bond issues distributed on the Italian market. Reporting distribution fees was made compulsory only recently (by the Prospectuses Directive, incorporated into Italian law in March 2006, with effect from November 2006), and the database is composed mainly of bonds issued after 2006. However, he was able to find distribution fees for 162 bank bonds issued before 2006. Overall, the dataset is composed of issues made by 42 issuers and distributed through 28 banking networks.

The expert also conducted a number of field interviews in order to gain insights into the factors that might be relevant in fixing distribution fees.

(53)

In addition, even if the standard deviation computed by the expert were not appropriate, it would be very likely indeed that the estimated market value plus one corrected standard error would be above the remuneration paid to PI.

(54)

Assuming six-monthly payment of a fee of 0,30 %.

(55)

An upfront fee of up to around 0,7 % would be market-conform.

(56)

The subscription-related part of the fee is the same for paper and dematerialised certificates.

(57)

The direct additional expenses incurred in the production of paper certificates are paid for separately — see section VIII.3.3.b.

(58)

The management fees were not calculated in the same way over the period 2000–2006, but this does not change the assessment made by the Commission.

(59)

The Commission is satisfied with the robustness test conducted by the expert regarding the amounts paid in 2001 and 2002 for specific tasks going beyond the mere distribution of postal savings certificates.

(60)

There were 13 950 post office branches in 2000 and 13 787 in 2001 (source: PI’s annual reports).

(61)

The management fee used in this computation is 0,13 %, which is the lowest of the management fees paid to PI over the period 2000–2002, and certainly cannot be regarded as excessive. The discount rate used by the Commission here is the average of the annual five-year discount rates. An alternative approach would be to use the real management fee paid to PI year by year over the life of the fixed-term bonds. That methodology would lead to an even clearer conclusion that the fees paid for paper fixed-term postal savings certificates were market-conform.

(62)

OJ C 297, 29.11.2005, p. 4.