Australian Competition Tribunal Decisions

AUSTRALIAN COMPETITION TRIBUNAL

Re EFTPOS Interchange Fees Agreement [2004] ACompT 7

TRADE PRACTICES � exclusionary provisions and price fixing � proposed agreement to set EFTPOS interchange fees at zero � whether proposed agreement should be authorised under s 88 (1) of the Trade Practices Act � whether public benefits arising from proposed agreement outweigh public detriments � whether use of debit cards will increase as a result of proposed agreement � whether debit cards and credit cards substitutable products � whether savings will be passed on to cardholders by issuers � whether costs will be passed on to merchants and what impact this will have on consumers � whether proposed agreement will enhance competition in issuing and acquiring markets � whether economic justification for negative interchange fees has ceased � whether interchange fees a payment for services

WORDS AND PHRASES � �interchange fees� � �two-sided market� � � access regime�

Trade Practices Act 1974 (Cth) ss 88, 90, 101(1), 102(1)

Re Media Council of Australia (No.2) (1987) ATPR 40-774 cited

Re 7-Eleven Stores Pty Ltd (1994) ATPR 41-357 cited

Re Pathology Services [2004] ACompT 4 at [93] cited

Howard Smith Industries Pty Ltd v Adelaide Steamship Industries Pty Ltd (1977) 28 FLR 385 at 405 cited

Queensland Co-operative Milling Association (1976) 8 ALR 481 at 484 followed

RE: EFTPOS INTERCHANGE FEES AGREEMENT

NOS 7 and 9 of 2003

JUSTICE HEEREY (DEPUTY PRESIDENT)

MR ROBIN DAVEY

MISS M M STARRS

25 MAY 2004

BRISBANE (HEARD IN MELBOURNE)

IN THE AUSTRALIAN COMPETITION TRIBUNAL
File Nos 7 and 9 of 2003

RE EFTPOS INTERCHANGE FEES AGREEMENT

AUSTRALIAN AND NEW ZEALAND BANKING GROUP LIMITED, AUSTRALIAN SETTLEMENTS LIMITED, BANK OF QUEENSLAND LIMITED, BANK OF WESTERN AUSTRALIA LIMITED, BENDIGO BANK LIMITED, CASHCARD AUSTRALIA LIMITED, COMMONWEALTH BANK OF AUSTRALIA, CREDIT UNION SERVICES CORPORATION (AUSTRALIA) LIMITED, NATIONAL AUSTRALIA BANK LIMITED, ST GEORGE BANK LIMITED, SUNCORP-METWAY LIMITED, WESTPAC BANKING CORPORATION

APPLICANTS FOR AUTHORISATION

AUSTRALIAN RETAILERS ASSOCIATION, AUSTRALIAN INSTITUTE OF PETROLEUM LTD, AUSTRALIA POST, BUNNINGS PTY LTD, CALTEX AUSTRALIA PETROLEUM PTY LTD, FRANKLINS PTY LIMITED, MCDONALD�S AUSTRALIA LIMITED, AND WOOLWORTHS LIMITED

APPLICANTS FOR REVIEW (NO 7 OF 2003)

COLES MYER LTD

APPLICANT FOR REVIEW (NO 9 OF 2003)

RESERVE BANK OF AUSTRALIA

AUSTRALIAN COMPETITION AND CONSUMER COMMISSION

INTERVENORS

TRIBUNAL:

JUSTICE HEEREY (DEPUTY PRESIDENT)

MR ROBIN DAVEY

MISS M M STARRS

DATE OF DETERMINATION:

25 MAY 2004

WHERE MADE:

BRISBANE (HEARD IN MELBOURNE)

THE TRIBUNAL DETERMINES THAT:

The determination of the Australian Competition and Consumer Commission dated 11 December 2003 is set aside.

IN THE AUSTRALIAN COMPETITION TRIBUNAL
File Nos 7 and 9 of 2003

RE: EFTPOS INTERCHANGE FEES AGREEMENT

APPLICANTS FOR AUTHORISATION

APPLICANTS FOR REVIEW (NO 7 OF 2003)

COLES MYER LTD

APPLICANT FOR REVIEW (NO 9 OF 2003)

RESERVE BANK OF AUSTRALIA

AUSTRALIAN COMPETITION AND CONSUMER COMMISSION

INTERVENORS

TRIBUNAL:

JUSTICE HEEREY (DEPUTY PRESIDENT)

MR ROBIN DAVEY

MISS M M STARRS

DATE:

25 MAY 2004

PLACE:

BRISBANE (HEARD IN MELBOURNE)

DETERMINATION

Contents

Paragraph

1. Introduction ……………………………………………………… [1]

1.1 Statement of issues ………………………………………………… [1]

1.2 The parties ………………………………………………………… [12]

1.3 The Joint Study …………………………………………………… [14]

1.4 The Banks� initial reaction to the Joint Study………………………… [15]

1.5 The Banks� change of position …………………………………… [17]

2. The Proposed Agreement ………………………………………… [18]

3. Statutory criteria ……………………………………………… [22]

4. The Banks� case for authorisation …………………………… [27]

5. The ACCC ………….……….………………………………….. [28]

6. History of the EFTPOS system ………………………………. [30]

7. Contractual arrangements …………………………………… [46]

8. Technical arrangements …………………………………… [49]

9. Fees payable by cardholders …………………………………… [54]

10. Cardholder benefits of credit and debit cards compared ……… [61]

11. Costs and revenues of debit and credit cards ………………… [69]

12. Credit card fees …………………………………………………… [71]

13. RBA credit card reforms ………………………………………… [72]

14. Usage trends ……………………………………………………… [73]

15. A paucity of hard information …………………………………… [75]

16. Overseas experience ……………………………………………… [79]

17. Theoretical framework of a payment system ………………… [82]

17.1 A double sided market � costs and benefits ……………………… [82]

17.2 Are negative interchange fees a payment for services? …………… [85]

17.3 Competition aspects of interchange fees ……………………… [89]

18. Likely effects of the Proposed Agreement ……………………… [93]

18.1 Issuers …………………………………………………………… [93]

18.1.1 Will there be fee reductions? ……………………………………… [93]

18.1.2 Will there be signals to cardholders? ……………………………… [114]

18.2 Cardholders ……………………………………………………… [116]

18.2.2 Most currently pay no fees anyway ……………………………… [116]

18.2.3 Non-price factors affect choice …………………………………… [119]

18.2.4 What do cardholders themselves think? ……………………… [120]

18.3 Acquirers ………………………………………………………… [122]

18.4 Merchants ………………………………………………………… [128]

18.4.1 Abandonment of EFTPOS ……………………………………… [129]

18.4.2 Pass on to consumers ……………………………………………… [130]

18.4.3 Surcharge ……………………………………………………… [131]

18.4.4 Variation of terms ………………………………………………… [133]

18.4.5 General price increase ………………………………………… [134]

18.4.6 Disincentive for investment ……………………………………… [135]

19. The proposed access regime ……………………………………… [137]

20. Alleged benefits ………………………………………………… [148]

21. Conclusion ……………………………………………………… [153]

1. Introduction

1.1 Statement of issues

1 For some years in Australia there has operated a payment system known as EFTPOS, an acronym for electronic funds transfer at point of sale. Consumers who hold a cheque or savings account at a bank or other financial institution (called �the issuer�) receive a programmed plastic card called a debit card. When purchasing goods or services from a merchant the cardholder produces the debit card. The authority of the cardholder to use the debit card is confirmed by entry of his or her personal identification number (PIN) on a terminal at the point of sale. The merchant then transmits details of the sale over a secure electronic network to the merchant�s bank (referred to as �the acquirer�). The acquirer then by electronic message checks the cardholder�s account with the issuer and a message for authorisation or decline of the proposed purchase is returned by the acquirer to the merchant. The whole process is completed within a few seconds. The issuer debits the cardholder�s account and the acquirer credits the merchant�s account, usually via overnight settlement of all transactions processed by the merchant. Often a merchant will provide cash (referred to as �cash out�) to the cardholder, and sometimes even though no purchase is made.

2 EFTPOS facilities are available at some 430,000 retail outlets throughout Australia. There are over 100 issuers and in the order of ten acquirers.

3 The EFTPOS system can only operate by virtue of a complex technical network and a web of contractual arrangements which have come into being over the years. This latter consists of numerous agreements (�bilateral agreements�) between an issuer on the one hand and an acquirer on the other. A diagram appearing in [47] below shows the network of bilateral agreements.

4 Two other features should be mentioned. Some issuers and acquirers do not participate in the EFTPOS system directly but through another institution known as a gateway. Also one large retailer, Coles Myer Ltd, operates as a merchant principal, that is to say it adopts a role akin to that of an acquirer.

5 Under the bilateral agreements issuers pay to acquirers a fee known as an interchange fee. These are flat fees per transaction. They range from 18 cents to 35 cents and average about 21 cents. They are referred to as negative interchange fees; fees flowing from acquirer to issuer are termed positive. Currently EFTPOS negative interchange fees total about $170 million per year.

6 On 20 February 2003 various issuers and acquirers entered into a multi-lateral agreement the effect of which is to abolish these negative interchange fees or, as it is said, to charge and receive fees of zero cents per transaction, for a period of three years.

7 This multi-lateral agreement is expressed to be conditional on the grant of authorisation under the Trade Practices Act 1974 (Cth) (TPA) and until then does not have any force or effect. We shall refer to it as �the Proposed Agreement�.

8 On 8 August 2003 the Australian Competition and Consumer Commission (ACCC) issued a draft determination refusing authorisation but, on 11 December 2003, reversed this position and granted an authorisation pursuant to s 88 of the TPA.

9 The critical factor from the ACCC�s point of view was that Australian Payments Clearing Association Limited (APCA), the industry body which oversees and manages the development and management of Australia�s payment clearing arrangements (other than credit card schemes), is in the process of preparing an access regime. In granting authorisation, the ACCC also noted a submission from the Reserve Bank of Australia (RBA) to the effect that it was watching the process closely and, were it to falter, would seriously consider designating the EFTPOS system under Pt 3 of the Payments System (Regulation) Act 1998 (Cth) with a view to imposing an access regime under s 12 of the Act. In light of these submissions, the ACCC was satisfied that access reform would occur in the short to medium term and it was able to attach a lesser weight to the anti-competitive detriments that it perceived might otherwise arise as a result of the Proposed Agreement.

10 On the acquirer/merchant side of the EFTPOS system, fees for transactions flow between merchant and acquirer and are usually ad valorem. A larger merchant that has its own point of sale network terminals and communications infrastructure may be paid a network access fee by an acquirer for use of the infrastructure. That is, the acquirer may share its interchange fee with the larger merchant or in some other way offset the per transaction fee that would be otherwise payable by the merchant. A smaller merchant will pay an acquirer:

(a) an equipment rental fee for the supply of a terminal;

(b) for associated stationary supplied by the acquirer;

(d) an additional per transaction fee where the merchant provides cash out to a cardholder.

11 Under the Proposed Agreement acquirers would lose some $170 million per annum. The Banks admitted that the acquirers would, to a substantial extent, seek to recoup that loss from merchants. Accordingly many merchants vigorously opposed the authorisation. Pursuant to TPA s 101(1) this Tribunal has now been asked to set aside the determination of the ACCC under s 102(1).

1.2 The parties

12 The applicants for authorisation are Australia and New Zealand Banking Group Limited (ANZ), Australian Settlements Limited (ASL) which represents building societies, Bank of Queensland Limited, Bank of Western Australia Limited, Bendigo Bank Limited, Cashcard Australia Limited (Cashcard), Commonwealth Bank of Australia (CBA), Credit Union Services Corporation (Australia) Limited (CUSCAL) which represents credit unions, National Australia Bank Limited (NAB), St George Bank Limited, Suncorp-Metway Limited and Westpac Banking Corporation (Westpac). All of these, except Cashcard, are parties to the review application before the Tribunal. We shall refer to them collectively as �the Banks�, a term which, although not totally accurate as a matter of description, is convenient and also reflects the commercial reality that by far the bulk of issuing and acquiring in the EFTPOS system is carried on by banks, and in particular the four major banks ANZ, CBA, NAB and Westpac.

13 The applicants for review in Tribunal proceeding No. 7 of 2003 are Australian Retailers Association, Australian Institute of Petroleum Ltd, Australia Post, Bunnings Pty Ltd, Caltex Australia Petroleum Pty Ltd, Franklins Pty Limited, McDonald�s Australia Limited and Woolworths Limited. We shall refer collectively to all these applicants for review as �the Merchants�. In No. 9 of 2003 the applicant for review is Coles Myer Ltd. The ACCC and the RBA were granted leave to intervene.

1. 3 The Joint Study

14 In October 2000 the RBA and the ACCC published a study entitled �Debit and Credit Card Schemes in Australia: A Study of Interchange Fees and Access� (the Joint Study). In its Executive Summary (at (i)) it was said the study was concerned with one aspect of the Australian payments system, the networks for automated teller machines (ATMs), credit cards and debit cards, and concentrated on two aspects of these card networks, namely interchange fees and the conditions of entry into the industry. In relation to debit card payment networks the Executive Summary said (at (iv)):

�Acquirers earn revenues from interchange fees of around $0.20 per transaction and revenues from merchants� service fees of around $0.12 per transaction. They incur costs of around $0.26 per transaction, giving a mark up of revenues over costs of 23 per cent. This mark-up is much lower than in credit card acquiring although infrastructure and procedures are very similar. The major reason is that large merchants have invested in their own acquiring infrastructure and have negotiated arrangements to share interchange fees with their financial institution.

The payment of a debit card interchange fee to acquirers is an arrangement unique to Australia. In other countries, the payment is to the card issuer or there are no interchange fees at all. The study did not find a convincing case for an interchange fee in the debit card payment network in Australia, in either direction.�

1.4 The Banks� initial reaction to the Joint Study

15 Initially the Banks, and in particular the major banks, strongly supported the current EFTPOS interchange fee and opposed the recommendations of the Joint Study for a zero fee. For example the CBA stated (submission to RBA and ACCC, October 2000):

�The interchange fee runs in the opposite direction [to credit card interchange fees] because merchants provide a service to bank customers to access their funds at the merchant (analogous to ATM services and interchange fee flows) … The current structure of debit charge interchange payments clearly values the benefit received by different participants and has contributed to the investment necessary to foster the world class system that we enjoy today.�

16 NAB commissioned a report from Professor Stephen King and Associate Professor Joshua Gans, economists from the University of Melbourne, which was forwarded to the RBA with a letter dated 5 December 2000. The report commented that the methodologies used by the Joint Study:

�…assume that interchange fees for any payment system should be set at zero if possible and deviations from zero should only be based on individual participants� abilities to recovery [sic] their costs. They provide no justification for the use of this benchmark, and do not show why this benchmark would be economically efficient or socially desirable.

All this adds up to an interchange fee that is �negative�; being paid from card issuers to acquirers. The Joint Study is perplexed by this. However when you think about it, as merchant benefits are lower relative to customer benefits, there are fewer issuer risks and there is potentially more acquirer competition, it is not surprising that a fair cost-sharing bargain might involve issuers compensating acquirers in this instance.�

ANZ in its response to the Joint Study dated 29 January 2001 made the point:

�In general, neither the size or direction of the interchange fees can be determined a priori. Both will depend on the relative magnitudes of the costs in the systems and the amounts that can be directly recovered from users. These in turn will depend on the state of development of the networks and the wider payment services markets in which they sit and can change over time. Hence freezing the direction of or level of the interchange fees � setting them at zero, for example � could be highly arbitrary or distorting. In this regard international comparisons may be misleading; for example, Australia has a single EFTPOS network unlike the fragmented situation in North America.�

1.5 The Banks� change of position

17 By 20 February 2003 the Banks had entered into the Proposed Agreement. The reason for this change of heart seems to have been the threat by the RBA to exercise its powers to designate the EFTPOS system under Pt 3 of the Payment Systems (Regulation) Act 1998 (Cth). In the case of the CBA for example, it told an EFTPOS industry working group that it was �cognizant of published regulatory positions�.

2. The Proposed Agreement

18 The Proposed Agreement would reduce the existing negative interchange fees to zero by:

(a) amending existing bilateral agreements between the parties (cls 2.1 and 2.2(a));

(b) having parties use their reasonable endeavours to amend contracts with non-parties (cl 2.2(b));

(c) allowing non-parties to accede to the proposed agreement by Deed Poll and thereby have the right to obtain zero interchange from other parties (cl 4).

19 The parties may amend the Proposed Agreement if at least 75 per cent of the parties in number agree, and any amendment which is not unanimous will not come into force for at least 120 days (cl 3.3).

20 Apart from the interchange fees, the existing bilateral agreements will continue unaffected. Unless terminated earlier, the Proposed Agreement is to continue for the term of any authorisation granted in respect of it (cl 1.4). The authorisation in fact granted was until 31 December 2006.

21 The parties agree that they will monitor the EFTPOS interchange fees and the impact of the changes �on an ongoing basis and meet at least annually� (cl 3.1) and will conduct a review of the EFTPOS interchange fees every three years or at any time there is a material change in circumstances and at least 25 per cent of the parties in number request such a review (cl 3.2).

3. Statutory criteria

22 The Banks brought two applications for authorisation under TPA s 88(1). The first (A30224) sought authorisation for those parts of the Proposed Agreement that may constitute an exclusionary provision in contravention of s 45(2)(a)(i). The second (A30225) sought authorisation of the Proposed Agreement because it would or might have the purpose, or likely effect, of fixing, controlling or maintaining the price for goods or services (s 45(2)(a)(ii) and s 45A(1)).

23 The test for the two authorisations sought are expressed somewhat differently in the TPA. In the case of an exclusionary provision of a contract, arrangement or understanding, s 90(8) provides that the ACCC shall not make a determination granting authorisation unless it is satisfied in all the circumstances that the proposed provision would result, or be likely to result, in such a benefit to the public that the proposed contract or arrangement should be allowed to be made.

24 In relation to the authorisation of a provision of a proposed contract or arrangement or understanding other than an exclusionary provision (ie in the present case application A30225), s 90(6) provides that the ACCC is not to grant authorisation unless it is satisfied that the provision of the proposed contract etc would result, or be likely to result, in a benefit to the public and that that benefit would outweigh the detriment to the public constituted by any lessening of competition that would be likely to result if the proposed contract were made.

25 Until recently the generally accepted view has been that notwithstanding the difference in language the two tests in practical application are essentially the same: Re Media Council of Australia (No.2) (1987) ATPR 40-774, Re 7-Eleven Stores Pty Ltd (1994) ATPR 41-357. However in Re Pathology Services [2004] ACompT 4 at [93] it was suggested that the test under s 90(6) was limited to a consideration of the detriments arising from a lessening of competition. Senior counsel for the Merchants and senior counsel for the ACCC formally submitted that the view expressed in Pathology Services was incorrect. Senior counsel for the RBA submitted to the contrary. However no counsel developed his submissions and it was common ground that the correctness or otherwise of the observations in Pathology Services is not relevant for the disposition of the present case. We therefore say no more about the question.

26 The �likely outcome� does not require satisfaction at or above 51 per cent but rather that there must be �a tendency or a real possibility of a particular result�: Howard Smith Industries Pty Ltd v Adelaide Steamship Industries Pty Ltd (1977) 28 FLR 385 at 405.

4. The Banks� case for authorisation

27 The Banks submitted that the Proposed Agreement would remove a pricing distortion and lead to an increased use of EFTPOS, which is a lower cost and therefore more efficient payment system. The Banks also submitted that the Proposed Agreement would lower barriers to entry in the markets for issuing and acquiring and thereby enhance competition. The RBA generally supported the Banks� position.

5. The ACCC

28 The ACCC�s position was the same as that contained in its final determination, that is to say although the benefits and detriments were �finely balanced� and weighed against authorisation, what made the difference was the �certainty� of access reform with the consequence that new entrants to the markets would increase competition.

29 The ACCC made submissions and adduced evidence for our assistance along the lines indicated by the Tribunal in Queensland Co-operative Milling Association (1976) 8 ALR 481 at 484. We have found the ACCC�s submissions and evidence helpful although, as will be seen, we do not accept everything it has put before us. The ACCC did not, as suggested by senior counsel for the Merchants, adopted a partisan role. Not everything the ACCC has put has been couched in terms of anodyne neutrality, but its role cannot be so restricted.

6. History of the EFTPOS system

30 EFTPOS was introduced into Australia in 1984 by the major banks. ATMs had already been introduced in the early 1980s. As with ATMs, the aim of the banks was to reduce costs by moving customers across to electronic banking and out of retail branches by reducing the number of over-the-counter manually processed transactions and indeed the number of branches.

31 This strategy has been highly successful. In the period 1990 to 2003 bank branch numbers declined by 30 per cent, ATMs increased by more than 300 per cent and the number of EFTPOS terminals increased by more than 2500 per cent. It is a matter of notoriety that profits of banks, and in particular the major banks, increased greatly over the 1990s.

32 The ATM transaction card was conceived as replacement for the traditional paper passbook, which was extremely expensive to issue and process, relying as it did on largely manual processes at retail bank branches.

33 When EFTPOS was introduced it was necessary for the system to be PIN based in order to allow ATM cards to be used for purchases at point of sale. This added value to the ATM cards already held by cardholders. As was said by Mr Charles Gove, the expert on payment systems called on behalf of the Merchants:

�EFTPOS was therefore introduced as a cost reduction and branch replacement strategy for financial institutions. Along with other electronic banking streams, it still performs this function. As the RBA states:

�EFTPOS terminals are another partial substitute for access to a branch�.�

The passage quoted by Mr Gove is from the RBA Bulletin, Bank Branch Trends in Australia and Overseas, November 1996, at 5.

34 Mr Gove also said that, to achieve universal acceptance, banks needed to make a significant effort to introduce EFTPOS at merchant points of sale, particularly high volume, repeat purchase merchants such as service stations and supermarkets.

35 EFTPOS did not get off to a good start. There was no arrangement for interchange of transactions between banks. Therefore a merchant with, say, only a Westpac card terminal could only accept EFTPOS cards issued by Westpac. If the merchant wished to accept all EFTPOS cards it was necessary to have four terminals connected to four different banks. This was prohibitively expensive, very complicated for retail staff and required an amount of counter space that was usually not available. Compounding the problem was the fact that among the customers of any one retailer there were inevitably customers of all four major banks.

36 Moreover the technical equipment in early terminals was very expensive and was described by one witness from a major retailer as �notoriously unreliable�.

37 To overcome these problems merchants lobbied the major banks to modify their systems in order to allow the acceptance of all EFTPOS cards through a single terminal. As a result, interchange links were progressively implemented from 1985. This made it possible to accept all EFTPOS cards from the major banks through a single card terminal.

38 When these interchange links were established, negative interchange fees were negotiated. Mr Gove says that this was:

�… to generate income to acquirers from all debit card issuers in order to achieve some return on the considerable investment required to implement and maintain a secure, on-line, real time acquiring infrastructure. This was the same rationale as was already employed by the banks for ATMs, where the owner of the ATM is paid a negative interchange fee by the EFTPOS card issuer.�

As already noted, the EFTPOS system used the same cards as the ATM system.

39 The cost of providing terminals, PIN pads and communication networks for merchants proved to be much more expensive than the major banks had originally anticipated. The larger retailers all had different business requirements and wanted their card payment systems to be customised to meet their needs. Many merchants also wanted the card terminals integrated with their cash registers and other internal accounting systems. The Banks were not able to meet these differing requirements.

40 Mr Gove says that this problem was resolved when major merchants decided to introduce their own equipment to achieve improved levels of reliability, acceptable customer service and some level of integration with internal systems. As a result, in the early 1990s EFTPOS gained momentum due to investment in infrastructure and strong marketing programs by a number of major retailers and oil companies.

41 The commercial consequence of this was that once major retailers had committed to install their own networks they were in a position to negotiate what for ease of reference may be called a fee (sometimes it took the form of rebates or credits in relation to other transactions) from acquirers to compensate them for the investment in EFTPOS infrastructure and processing costs and the corresponding reduction in the acquirers� costs. Acquirers also benefited from lower unit transaction costs for the existing transactions, due to the high volume of transactions being delivered to them.

42 The first merchant to negotiate a fee from an acquirer was Coles Myer. Coles Myer first reached agreements with smaller financial institutions and then, in mid 1991, with one of the major banks (NAB). This example was followed by others including Woolworths who negotiated to receive an EFTPOS transaction fee from NAB in early 1993 on the basis that all PIN pads, point of sale equipment and communication costs would be paid for by Woolworths. Mr Gove said:

�Over time it has become standard practice for acquirers to share the interchange fee with merchants who own their infrastructure and deliver significant transaction volumes.�

43 The next major growth in EFTPOS usage came as a result of CBA�s program Comm 2000 which was a low cost card terminal and integrated PIN pad designed for small merchants. This strategy proved very successful for CBA and created a new merchant segment and, with the competitive response from other banks, led to the number of terminals increasing dramatically from 1995. Mr Gove produced a graph showing the relative growth in ATMs and EFTPOS and declining bank branch numbers as follows:

44 In 1990 there were slightly over two EFTPOS terminals for every bank branch in Australia. The figure is now 89 terminals for every bank branch.

45 Today more than half of all EFTPOS transactions are processed through terminals and PIN pads owned by merchants.

7. Contractual arrangements

46 According to the Joint Study there were 39 bilateral agreements reported. Of these, 18 had interchange fees of between 18 to 20 cents, 16 between 21 to 25 cents, while four were between 26 to 30 cents and one was above 31 cents.

47 Apparently the fee was usually the same regardless of which of the parties to the agreement was the issuer or the acquirer. Some banks are net issuers and some net acquirers. The difference may be quite substantial if the experience of the only bank for whom information was available is typical. The Bank of Queensland is a net issuer by a margin of three to one. The following diagram produced by Mr Peter Smith, the Chief Executive Officer of APCA, shows the network of bilateral agreements.

48 The only bilateral agreements produced to the Tribunal were two entered into by the Bank of Queensland, one with NAB and one with ANZ. They were tendered as confidential exhibits. The Bank of Queensland also has a gateway agreement with ANZ but that was not tendered. Consistently with the obligation of confidentiality, we think we can at least say that both of the Bank of Queensland bilateral agreements provide for annual review of fees. Mr Keith Wilson, the witness from the Bank of Queensland, agreed in evidence that there was no reason that would have prevented his bank from renegotiating the level of interchange fee payable under those agreements.

8. Technical arrangements

49 The EFTPOS system is built on a series of bilateral electronic links which have been separately negotiated and implemented over the years. The system, which also covers ATM transactions, is known as the Consumer Electronic Clearing System (CECS) and is managed by APCA which has promulgated the CECS Regulations and Manual (collectively �CECS Rules�).

50 Each acquiring system is slightly different. While each adopts the AS 2805 standard for interchange message formats as mandated by the CECS Manual, the standard permits each acquirer to adopt its own preferences for privately defined data fields and each bank has its own preferences for using these in practice. This means that each electronic link requires its own unique driver software and must be individually developed and tested whenever any changes are implemented, an expensive and time consuming process.

51 A new acquirer seeking to enter the market on equal terms with current acquirers would need to develop, test and implement a separate interchange message format and physical link to each of the major issuers. Mr Gove says that this would be likely to take two years or more to achieve, depending on the level of priority given to the work by the issuers. The technical workload required to connect directly to the major card issuers is the longest lead time and most expensive task facing a new entrant.

52 Mr Gove notes that the new entrant would also have to negotiate a commercial bi-lateral agreement with each of the issuers which would take time but in his view this is a relevantly minor task compared with the cost and time involved in the technical implementations.

53 An alternative for a new entrant is to connect to a gateway provider. The new entrant then only has a single technical link to develop, test and implement, but of course must pay fees to the intermediary.

9. Fees payable by cardholders

54 From the figure provided by Mr Gove the EFTPOS fee payable by a cardholder per transaction is as follows:

Cents

CBA 30

ANZ 50

NAB 60

Westpac 50

Bank of Queensland 65

55 However the practical impact of these fees is greatly affected by the terms of fee free transactions which on a monthly basis are as follows:

CBA 15

ANZ 5-15

NAB 10

Westpac 8

Bank of Queensland 5-15

56 These figures depend on numerous other factors. For example, with ANZ there can be unlimited free transactions for some account types and as low as one free transaction for other accounts. NAB allows $6.00 per month of free fees so the number of free EFTPOS transactions depends on the number of other services conducted, such as bank withdrawals. Mr Gove notes:

�It is very difficult to encapsulate the full range of fees and options in a simple table. All banks have a wide range of options depending on account type and customer category (eg student, pensioner etc). The range of fees and fee structures is complex with the documents describing the fees typically running to twenty pages or more. The … table [extracted above] simply attempts to show some typical fees and exemption limits. Full details of fees, fee structures and exemptions can be obtained on the web sites of the relevant banks.�

57 From this complicated picture however emerges the important fact that, in Mr Gove�s words, �most consumers pay no EFTPOS fees today�.

58 According to the EFTPOS Industry Working Group (EIWG), a group composed mainly of the Banks, in a discussion paper published in July 2002:

�cardholders are not charged for EFTPOS transactions below a threshold (on average eight free EFTPOS, ATM and other electronic transactions per month).�

59 Mr Gove says:

�as the average number of EFTPOS plus ATM transactions per account per month for 2003 was 5.76, it seems likely that many cardholders currently pay no EFTPOS transaction fees.�

60 In a submission to a Parliamentary Joint Standing Committee on Corporations and Securities which issued a Report on Electronic and Telephone Banking in February 2001, ANZ said:

�The incidence of excess withdrawal fees is low, with about 80 per cent of customers in any given month not paying these fees. About 30 per cent of customers are exempt, including full time students, children and customers with home loans.�

10. Cardholder benefits of credit and debit cards compared

61 The holder of a credit card has access to a number of benefits not available to the holder of a debit card.

62 First, the credit card holder is not limited to his or her own funds but can use credit, which is interest free up to 55 days, depending on when a purchase is made in the billing cycle. Thereafter interest is payable, and at a high rate, substantially above overdraft rates. A credit card therefore can be seen as a standing offer of a personal loan.

63 Secondly, the credit card holder may be eligible to participate in a loyalty scheme and receive points which may be redeemed for goods or services and other benefits such as �free� travel insurance when the card is used to pay for an airfare.

64 Thirdly, a credit card can be used for purchases by telephone, mail or internet.

65 Fourthly, in the case of some credit cards there are extended warranties on the purchase of goods.

66 Fifthly, all credit cards provide for a refund of advance payments where goods or services are not received.

67 Sixthly, the financial limits imposed by credit cards are usually higher than for debit cards. Generally there was no limit on the transaction size that can be effected using a credit card, subject to the cardholder�s overall credit limit. By contrast there is usually a maximum limit imposed for any one EFTPOS transaction.

68 Seventhly, credit card statements provide more details of transactions.

11. Costs and revenues of debit and credit cards

69 The only evidence before the Tribunal concerning the respective costs and revenues of debit and credit cards is the 1999 data contained in the Joint Study. These appear as Tables 5.1 and 6.1 as follows. (All figures are a weighted average. For clarity of presentation we have omitted the figures given in the tables for the range in respect of each figure. Because of the weighting process, the components do not necessarily add up to the total.)


Table 5.1: Credit card costs and revenues per transaction ($)
ACQUIRING		ISSUING
COSTS
Operating expenses Of which	0.19	· Production/distribution of cards	0.06
· Staff	0.07	· Authorisation	0.04
· Authorisation	0.04	· Processing	0.17
· Data processing	0.04	· Staff	0.39
· Switching services	0.03	· Interest free period	0.26
		· Fraud	0.07
Overheads	0.24	· Credit losses	0.35
Of which		· Other	0.68
· Depreciation	0.07
· Telecommunications	0.05	Cost per transaction	1.93
· Fraud	0.01
· Other	0.11
Cost per transaction	0.43
Interchange fees paid	1.06

REVENUES
Merchant	1.78	· Interest margin	1.36
Service fees		· Interchange fees	0.95
		· Annual fees	0.33
		· Other	0.05
		Revenue per transaction	2.69


Table 6.1: Debit card costs and revenues per transaction ($)
ACQUIRING		ISSUING
COSTS
Operating expenses Of which	0.08	· Production/distribution of cards	0.06
· Staff	0.04	· Authorisation	0.03
· Data processing	0.01	· Processing	0.03
· Switching services	0.03	· Staff	0.01
		· Fraud	0.01
Overheads	0.18	· Other	0.02
		Cost per transaction	0.15
Of which
· Depreciation	0.06	Interchange fees paid	0.21
· Telecommunications	0.05
· Other	0.07
Cost per transaction	0.26

REVENUES
Interchange fees	0.20	Transaction fees	0.20

Merchant service fees	0.12

70 Senior counsel for the Merchants attacked the conclusion of the Joint Study, based on these figures, that the total resource costs of credit cards are higher than for EFTPOS. It was also pointed out that credit card costs include the costs associated with the provision of credit such as interest free periods and fraud. If the costs associated with the use of credit cards as a payment means were to be compared with EFTPOS costs, the costs would be almost identical. Since credit and debit cards are processed through the same electronic infrastructure, most of those costs would be common. However as the Joint Study has allocated those costs between credit cards and EFTPOS based on the volume of transactions that had the effect of weighting the costs towards credit cards. Also questions of allocation of costs affect the usefulness of the figures to assess the costs and benefits. It was said, with some justification in our view, that no person associated with the preparation of the Joint Study had been called as a witness and the figures were in any event a snapshot now almost four years out of date. Nor had any material been produced by the Banks.

12. Credit card fees

71 There was no up to date evidence produced to the Tribunal concerning the fees for credit cards charged by issuers to cardholders. It is a matter of common knowledge that cardholders frequently pay an annual fee for a credit card, but the size of that fee, and the extent to which it is waived by the Banks in connection with the bundling of services to customers, is not known. Certainly prior to the RBA reforms of credit card interchange fees the interest free period and loyalty awards made credit cards very attractive to cardholders. Credit cards were heavily promoted by banks, which rather suggests they were also profitable.

13. RBA credit card reforms

72 In November 2003 the RBA exercised its powers under the Payment Systems (Regulation) Act to designate credit card systems and impose changes. The RBA did not impose a particular level of interchange fees in respect of credit cards but set a cost based standard which prescribed the maximum which could be charged. Issuers and acquirers were free to negotiate an amount below that cost-based standard. There was no obligation for any pass through. Secondly, the reforms abolished the rules which prohibited merchants imposing a surcharge on transactions with credit cards. Thirdly, there was provision for an access regime. One of the effects of the reforms was to greatly reduce the provision of loyalty programs.

14. Usage trends

73 Up until about 1997 EFTPOS grew at a faster rate than credit cards. From 1997 until mid 2003, credit card use grew faster. This accelerated growth was said to be in part attributable to the introduction of loyalty programs by credit card issuers.

74 As a consequence of the RBA reforms, issuers have begun to reduce the benefits offered in connection with credit cards and to alter their fees. There was evidence from merchants to suggest that since mid 2003 the relative rate of growth of credit card use has slowed compared with EFTPOS. Both Mr Nicol of Caltex and Ms Winn of Woolworths said that their recent sales figures showed an increasing rate of growth of debit cards compared with credit cards. The RBA Bulletin shows that for the four month period November-February for 2003/04 as against 2002/03, the number of EFTPOS transactions has grown by 11.1 per cent in number and 14.3 per cent in value. In contrast credit card transactions growth is 9 per cent in number and 9.9 per cent in value.

15. A paucity of hard information

75 Senior counsel for the Merchants, Coles Myer and the ACCC criticised strongly the lack of provision of hard information by the Banks. There was similar criticism of the RBA, although not joined in by the ACCC. In our opinion, these complaints were well founded and were not satisfactorily answered. Coles Myer identified the following information which the Banks could have but did not, put before the Tribunal:

�1. The level of interchange fees paid;

2. The breakdown of what banks are issuing banks and what banks are acquiring banks;

3. The breakdown of the costs of the issuing bank in processing debit cards;

4. The breakdown of the costs of the acquiring banks in processing debit cards;

5. The breakdown of where the acquiring banks actually spend the interchange fee they receive (eg. costs, past [sic] to merchants, profit on acquiring businesses linked to the merchants that they share the interchange fee with);

6. Each banks� consideration of the impact to its business of a change to zero interchange fee;

7. Any analysis done by banks regarding the introduction of a positive interchange fee;

8. Issuing banks� arrangements with cardholders as to who pays fees for accessing debit cards;

9. What plans, if any, the Banks have to pass on the savings in the interchange fees to customers;

10. What plans, if any, Banks have to grow credit card businesses;

11. What plans, if any, Banks have to grow the use of debit cards;

12. The Banks� profit on their credit card businesses;

13. The �profit� generated by the use of debit card;

14. The Banks� commitment to access and how it would look like in practice.�

76 We do not say that each and every one of those matters was necessarily an essential item of information, but nevertheless, looked at broadly, this list of topics gives some indication of the extent of material which could well have been provided, but was not.

77 Even more striking is the Banks� strategy in putting forward Mr Wilson as the only witness who could speak as to the current operation of the EFTPOS system. We make no criticism of the solicitors and counsel for the Banks, who are bound by their clients� instructions. Still less is there any reflection on Mr Wilson himself, who appeared to us to be an honest and competent witness. However, the fact is that the Bank of Queensland is a regional bank, presently confined to that State, and holds about one per cent of cheque and savings accounts in Australia. Even within his own bank Mr Wilson�s responsibility was for the interchange arrangements and the arrangement and processes which support those systems. He has no responsibility with respect to pricing decisions. Within the major banks there must be people with far more knowledge of the operation of the EFTPOS system than Mr Wilson. The strategy of the Banks in this regard calls to mind that of the Duke of Plaza-Toro who �led his regiment from behind, he found it less exciting�.

78 As for the RBA, its senior counsel relied on s 79A of the Reserve Bank Act 1959 (Cth) which by subs (2) makes it an offence for an officer to disclose to any person or to a court (defined to include a tribunal) any �protected information acquired in the course of his or her duties as an officer�. The penalty prescribed is imprisonment for two years. But as senior counsel for the Merchants pointed out, under subs (4) an officer can produce any information if the Governor of the Bank approves and in any event subs (7) provides that subs (2) does not prohibit a person from disclosing information or producing documents if the information is in the form of a summary or collection of information that is prepared so that information relating to any particular person cannot be found out from it. If the RBA had really wanted to provide information to this Tribunal we are satisfied that this could have been done with little risk of any officer of the Bank ending up in jail.

16. Overseas experience

79 Those supporting authorisation asserted that Australia was �unique� in having a negative interchange fee. This is not in fact correct. In New Zealand acquiring banks have established a jointly owned company to undertake switching activities, a substantial component of the acquiring service. Issuers pay a fee to the switching company for that service. Whether or not that is called an interchange fee, it seems to reflect at least in part the present Australian system.

80 In the RBA�s final submissions it was said that Canada and the Netherlands have zero interchange fees. However it was also asserted that the United Kingdom and the United States have positive interchange fees. It was not explained why, if overseas experience is so important, Australia should not adopt a positive interchange fee regime rather than zero since countries with the former have much larger economies than those with the latter.

81 Frankly we do not see a great deal of value in overseas comparisons. The way a banking system operates in a given country is a result of a complex mix of historical, geographical, political, cultural and socio-economic factors. It is not likely to be a profitable exercise to engage in a detailed examination of these with a view to seeing what features are or are not replicated in the Australian experience. For example, it was said that South Africa had a negative interchange fee, then it was said that that was only in respect of service stations and then that proposition was in turn disputed. All in all it did not seem a very useful line of enquiry.

17. Theoretical framework of a payment system

17.1 A double sided market � costs and benefits

82 Viewed in terms of economic theory, issuing and acquiring takes place in what is termed a �two sided market�. On the one side there are issuers and cardholders and on the other side acquirers and merchants. If the system as a whole is to work both sides have to have benefits greater than costs. There may be need for what is called a balancing payment. In what economists in this case spoke of as the seminal work in the area, W F Baxter in his article Bank Interchange of Transactional Paper: Legal Perspectives (1983) 26 Journal of Law and Economics 541 noted (at 575):

�…it was overwhelmingly improbable that the revenue stream [on each side of the system] would equal the costs of the subset of activities that a particular bank was required by the technology of the payment system to perform; thus some redistribution of those revenues [between both sides of the system] was likely to be necessary for the payment system to compete effectively with alternative mechanisms.�

Further, because the fee was a balancing payment Baxter observed (at 575) that:

�(i)t makes no difference when addressing this question in the abstract whether the transfer payment is made by card issuing banks to merchants bank or by merchant banks to card issuing banks.�

83 Mr Jeffrey Rae, an economist called by Coles Myer, pointed out by way of illustrative example that typically in Australia some media services such as suburban newspapers and commercial radio and broadcast television are entirely paid for by advertisers. Readers, listeners and viewers pay nothing but without their participation publishers and broadcasters would be unable to earn revenue from advertisers. Or as Dr Philip Williams, an economist called by the Merchants, put it:

�From an economic perspective, an interchange fee can be used to balance revenues and costs in an open payment system … the purpose of such transfers of funds between issuers and acquires via interchange fees is to provide incentives for a participant to participate even when the participant is not willing to bear all of the costs that it would otherwise incur.�

84 As we understand their cases, the Banks and the RBA, in arguing for authorisation, did not directly rely on the characterisation of negative interchange fees as having a balancing role. Rather their argument was that the historical justification for negative interchange fees was no longer apposite and that they distorted price signals to consumers because the issuer had to recover the interchange fee from the cardholder. Bound up with those questions is the issue whether negative interchange fees represent a payment for services received by issuers.

17.2 Are negative interchange fees a payment for services?

85 Let us suppose a hypothetical situation where X is contemplating opening an account with a branch of, say, CBA. In substance CBA holds out the following promise:

�If you put your money with us, we will keep it safe. We will provide also other benefits, including access to the EFTPOS system. This means that you will be able to get your money when you need it not just from this branch, or any other branch of CBA, but you will be able to use your money to pay for goods and services at 430,000 retail outlets all over Australia. At some of these you will be able to withdraw cash 24 hours a day, and sometimes without purchasing anything.�

86 CBA cannot make good on this promise (or deliver these outcomes, as it would probably say) from within its own organisation. It can only do so by entering into arrangements with acquiring banks. In the course of negotiating such an arrangement, a potential acquirer might say to CBA:

�You are getting the money of customers because they are attracted by the EFTPOS system, which can only operate because we provide services to our merchants, including, in the case of small merchants, the physical infrastructure itself. This has cost us, and will continue to cost, a great deal of money. You therefore ought to make some payment to us for EFTPOS transactions.�

87 In response, CBA might say:

�You should look to your merchants for your costs. They should pay the costs of EFTPOS just like they pay the cost of dealing with other forms of customer payments, such as cash. It is a cost of their business and nothing to do with us.�

88 The respective stances of CBA and the potential acquirer are perfectly rational. One cannot say that one is correct and the other is not. Which argument prevails will depend on the economic imperatives bearing on each party. No a priori rule provides the answer.

17.3 Competition aspects of interchange fees

89 The fact that interchange fees fulfil a balancing function in a two-sided market does not mean they are exempt from competitive pressures. This is not a case where of necessity there is competition at the ends but not upstream. In defending the confidentiality of the Bank of Queensland�s bilateral agreements with the ANZ and NAB, senior counsel for the Banks said:

�One knows from the evidence that the acquiring market is highly competitive, and it may be that an acquirer, for example ‑ just to take the example we've been talking about ‑ would seek to get the Bank of Queensland business and, knowing a bit more about their current arrangements, could reconfigure an offer that it would otherwise make to have a better chance of getting somewhere.�

90 Senior counsel for the Merchants observed that if there no competition between issuers and acquirers in this market then the Bank of Queensland�s claim for confidentiality in its agreement had �no commercial substance�. As he said, if other banks knew what the Bank of Queensland was paying NAB or ANZ for the benefit of accessing the system and for the service of reaching merchants and being able to ensure that their customers could present their Bank of Queensland card, other banks would try and steal the Bank of Queensland away from them by offering a better price. That was, he said, the sort of competitive tension that does in fact go on between issuers and acquirers.

91 Another example given by senior counsel for the Merchants was that a major bank, being both an issuer and acquirer, might decide that its issuing division might reduce its fees paid to the acquiring division and use the saving to invest in a new retail product. The acquiring division may have some technological improvement in switching systems which can compensate for the loss in income without passing on the costs to merchants. So the bank can take the commercial risk, without waiting for other issuers

92 One of the arguments in support of authorisation is that there has been, so to speak, an ossification of interchange fees; there has been little change over the years. Even if this be so, it seems an odd solution to a perceived problem of insufficient competitive activity to authorise an agreement deemed by law to be anti-competitive. It is said to be too hard for banks to move individually, therefore you have an agreement which means they will not be able to move at all.

18. Likely effects of the Proposed Agreement

18.1 Issuers

18.1.1 Will there be fee reductions?

93 The Joint Study showed that issuers incur higher costs in providing credit card services compared with debit cards because of the cost of the interest-free period, bad debts and fraud, but nevertheless earn more net revenue when their customers use credit cards rather than debit cards. A J P Morgan study tendered by Coles Myer, Banking Sector, the Payment System in Australia, 29 May 2003, analysed the impact on major banks of credit card reforms and concluded that even on a worse case scenario the major banks will still make substantial profits.

94 Therefore it is a fact of commercial life that the Banks will continue to have an incentive to encourage credit card use as against debit card use. As Coles Myer points out, it would be contrary to the Banks� shareholders� interests to do otherwise.

95 To the extent that a cost is removed the tendency, other things being equal, is to apply the money saved in other profit-generating activities or to pass it on as dividends. More specifically, the Banks have provided virtually nothing by way of concrete commitment to this Tribunal, or the ACCC, for the passing through to cardholders of savings from a zero interchange fee.

96 On 27 February 2003 the ACCC wrote to the solicitors for the Banks asking, amongst other things:

�1.4 If authorisation were to be granted to the proposed EFTPOS reforms, and with all other factors remaining constant, what effect would this have on the debit card transaction fees charged to [the Banks�] consumers?

1.4.1 If the anticipated effect is to reduce current debit card transaction fees, what is the value of the anticipated reduction?

1.4.2 How will this reduction be reflected in fees charged to [the Banks�] customers?�

97 In response some of the Banks provided vague promises which were mostly, initially at least, tendered before us as confidential exhibits, in itself something of an irony given the importance attached elsewhere in the Banks� case to transparency of signals.

98 The first piece of evidence relied on by the Banks in final submissions was a statement of CUSCAL (the representative of credit unions) to the effect that it:

�…wished to reiterate that its card issuing members had indicated that they intended to pass on the benefit of the interchange fee cost savings to members.�

99 The statement of CUSCAL comes from a document produced to this Tribunal entitled �Record of Predetermination Conference� (the Record). The Record was prepared by the ACCC and records a conference held in Canberra on 1 September 2003 chaired by Commissioner McNeill of the ACCC. This was after the ACCC�s draft determination and before the final determination. Commissioner McNeill noted at the outset that the conference was requested by the Banks pursuant to TPA s 90A.

100 According to the Record there were in attendance representatives of the major banks, Ms Jane Nash (ANZ), Mr Chris Campbell (Westpac), Mr Doug Johnson (CBA) and Ms Martha Georgiou (NAB). In respect of each of the first three named it is said that the representative �declined Commissioner McNeill�s invitation to provide a separate submission at this time�. Ms Georgiou was recorded as stating that NAB supported the submissions made by Mr Woodward but that NAB �did not wish to provide a separate submission at this time�. Mr Stuart Woodward is an officer of the CBA and spoke on behalf of the applicants for authorisation. His statement is attached to the Record. It argues in favour of zero interchange and criticises the draft determination. In a seven page submission the only thing that Mr Woodward says about fee reduction is:

�We believe that there is likely to be a pass through of savings by issuers which will be a benefit resulting from the proposed agreement. It might not be a complete pass through of the savings, but the direction is certainly clear, compared with the status quo.�

101 The second piece of evidence relied on by the Banks in final submissions comes from the Record in the form of a statement by Mr John Toms on behalf of ASL (which represents building societies). In its final submissions the Banks quoted from Mr Toms� statement as follows:

�With relief from EFTPOS interchange fees, building societies will be able to pass on some cost savings to cardholders. Mr Toms expects that the EFTPOS reform will lead to a reassessment of current fee structures to cardholders. In particular, the proportion of costs currently passed on as a fee to cardholders would be passed on as a saving if interchange fees were reduce to zero.�

102 Vague as this passage is, Mr Toms went on to say by way of further qualification, in a passage not quoted in the Banks� final submissions:

�However the pass through is complex given that there are currently fee free transactions.�

103 Thirdly, the Banks then referred to responses to the ACCC in March 2003, that is to say twelve months before the present hearing. All of these were initially submitted to the Tribunal as confidential.

104 The first is from Westpac as follows:

�1.4 If EFTPOS interchange fees were set at zero and all other factors remained constant, overall cardholder debit card transaction fees would be reduced.

1.4.1 [Confidential].

1.4.2 To achieve this, we estimate that EFTPOS fees on our consumer accounts would be reduced from 50c to 25c per transaction.�

105 In final submissions senior counsel waived confidentiality for 1.4 and 1.4.2 but retained confidentiality for 1.4.1 which gives a dollar figure for total reduction of costs to cardholders.

106 The second is from another major bank, the identity of which remains confidential. It said that if the fees were reduced then the price to customers would be reduced on a per transaction basis by a given amount.

107 The third was from a third major bank. The initial confidentiality claim was waived, save that the bank was not to be identified other than as a major bank:

�It is likely that there would be a decrease in transaction prices for cardholders conducting EFTPOS debit transactions. The size and extent of any change would be reflective of the context of the overall transaction banking relationship, the structure of the transaction account service, the cost of service provision, customer usage patterns and the competitive market environment.�

108 Finally there was quoted in the Banks� final submissions what was said by Suncorp-Metway in a letter to the ACCC dated 19 March 2003. The passage quoted was immediately preceded by a passage which is underlined:

�SUN has not yet developed its position with respect to the anticipated effect of debit card transaction fees charged to consumers. Calculating the SUN position is a complex undertaking given the range of some products and range of transaction types making up the various fees charged to consumers. However SUN would take steps to ensure it gives due consideration to reduced costs of providing debit card facilities to its customers and where supportable would pass benefits to consumers in some form, for example, through lower fees or enhanced services. SUN would also give due care and attention to competitor reactions.�

109 While all economists who gave evidence agreed with the general proposition that the proposed change in interchange fees is likely to be passed on (at least to some extent) to cardholders, the evidence presented leaves us quite unable to make any worthwhile finding as to the quantification of that extent (quite apart, as we shall see, from the further question as to the effect any such reduction would have on cardholders� decisions).

110 Of the four major banks only two gave any information expressed in actual monetary terms. Those figures were twelve months out of date. Coles Myer in its final submissions pointed to other statements by banks which were not referred to in the Banks� �sample of responses�. We think it is fair to say that none of these took the matter any further.

111 As to the Banks� submission that pass on will occur �because it is expected by the RBA� we note that the RBA does not regulate individual contractual relationships between cardholders and their banks and has not proposed to guarantee pass through in the present case.

112 As to the effect of competition in ensuring pass through, we accept the ACCC�s submission that card issuing and merchant acquiring are currently highly concentrated markets with the major banks dominating. The ACCC submitted that it was:

�… concerned that current competition between card issuers may not ensure the lasting pass through of savings to cardholders.�

113 We think that is right, particularly when one bears in mind the fact that competitive pressure to pass on savings to cardholders will depend on cardholders� likely reaction to price changes. As to this, as will be seen, there is little or no evidence. We note that neither the expert on payment systems called by the Banks (Mr White) nor the economist Dr Abraham have received any instructions from their clients as to plans to pass through cost savings arising from zero interchange fees.

18.1.2 Will there be signals to cardholders?

114 Just as there was very little in the way of concrete evidence as to what savings might be passed through to cardholders, there was even less as to the nature of the signalling to cardholders.

115 Since such pass through as may occur is likely to be small in amount and in any event may take the form of other adjustments to the complex bundle of bank charges, any signalling of the change is not likely to be attention-grabbing and conduct-altering.

18.2 Cardholders

18.2.2 Most currently pay no fees anyway

116 All economists who gave evidence agreed with the proposition that the proposed arrangements are only likely to have the effect of increasing demand by cardholders for EFTPOS if they decrease the price that cardholders pay for marginal transactions.

117 In their final submissions the Banks relied on the following passage from the evidence of Mr Paul Wood of MoneySwitch, a potential entrant into the acquiring market:

�Based on my experience and my research, I would anticipate that cardholders would rapidly shift to using the full (expanded) free entitlement. So if, for example, there are at present seven free transactions available, and the number was doubled as a result of the pass through, I would expect the usage of debit cards to double.�

118 This seems to us a non sequitur. As Dr Williams explained, consistently with the views of all the other economists, choice is only likely to be affected at the margin (or, as Dr Williams put it, when the cardholder is �up against the barrier�). If a cardholder usually has five transactions a month, with a fee free level of seven, the decision whether to go to a sixth or seventh transaction is not likely to be affected by the lift of the fee free level from seven to fourteen. The importance of this for present purposes lies in the fact that, as has been noted already, there is a very high level of EFTPOS users who under current arrangements pay no fees anyway. Moreover, a cardholder�s use of EFTPOS is inextricably linked to, and limited by, the funds in his or her account. Those persons who get close to or exceed their fee free transaction limit each month may well also be close to, or about to exceed, the limit of their funds. An offer of additional fee free EFTPOS transactions to a person in those circumstances would not increase their use of EFTPOS. The Banks advanced no information, which should have been readily available to them, addressing the link between consumers� use of EFTPOS and their funds. What use is there in sending a signal to a consumer to make additional use of EFTPOS if the consumer has no funds to allow that additional use?

18.2.3 Non-price factors affect choice

119 The benefits which credit cards confer and debit cards do not have already been mentioned. These are all non-price factors which might dictate a choice of credit card over debit card (or over cash for that matter).

18.2.4 What do cardholders themselves think?

120 In this case we have heard evidence from four large merchants, three small merchants, an association of merchants, (only) one bank officer, five economists and two payment systems experts. Virtually all of this evidence is directed in one way or another towards the critical issue of the likelihood of change of behaviour by consumers in debit card usage. Is there not an odd gap? There has been no evidence from persons who are actual or potential holders of dedit cards as to how their usage might be affected by any changes.

121 In final submissions senior counsel for Coles Myer put it that, in summary:

�… there is an absence of firm evidence that issuing banks intend to introduce positive price signals to encourage the use of debit [cards]. Of equal importance is the fact that there is a total absence of evidence before the Tribunal that the issuing banks would seek to encourage the use of EFTPOS in preference to the use of credit [cards]. Finally, even if (contrary to their own commercial interests) banks did so, there is no evidence to support the view that consumers would change their behaviour. Many consumers would notice no difference, and even if they did, a significant proportion would maintain current habits for non-price reasons.�

We think those submissions are well-founded in the evidence, and we accept them.

18.3 Acquirers

122 All economists agreed that the Proposed Agreement is likely to increase the cost to merchants of accepting EFTPOS. Moreover there was substantial agreement by them to the further proposition that the increased cost to merchants of using EFTPOS is likely to discourage them (or at least some of them) from accepting EFTPOS for low value transactions or to reduce the functionality of EFTPOS to cardholders in other ways.

123 The evidence indicates that, understandably, acquirers will seek to recover from merchants the lost revenue of negative interchange fees. Whatever the defects of the evidence as to the voluntary reduction of fees charged to cardholders by issuers, there is no suggestion at all of voluntary reductions by acquirers vis-�-vis merchants. And this is likely to be so even with those larger merchants who currently receive fees from their acquirers. The evidence is that it is likely that such fees will no longer to be paid by the Banks to the participating merchants but rather a positive charge will be imposed on the merchants.

124 The Banks argued that those larger merchants which had installed their own infrastructure had by now recovered the costs and that therefore there was no longer any justification for a negative interchange fee (which currently in part funds the payments or allowances paid by acquirers to large merchants). We are not aware of any economic principle to the effect that where a firm has invested in a capital asset and rented that asset out, once the cost of the asset has been recovered the firm should continue to make the asset available to the tenant free of charge. Certainly the Banks themselves show no signs of applying such a principle with regard to the smaller merchants to whom they rent EFTPOS infrastructure.

125 The ACCC submitted that increases to the merchant service fees are unlikely to be significantly constrained by competitive forces currently present between acquirers. In its determination in support of that conclusion the ACCC noted that acquiring is highly concentrated with a small number of large institutions dominating. The major banks provide about 85 per cent of merchant credit and debit card acquiring services although this share has been falling gradually in recent years. The ACCC was concerned that acquirers seeking to recoup the cost increases resulting from the abolition of negative interchange fees were unlikely to be significantly constrained by competition, in particular in relation to small business banking customers. This is one of the reasons why the ACCC considered that access reform was essential to its grant of authorisation.

126 In its determination the Commission noted (par 6.63) that according to a KPMG report Small Business Banking in Australia (2003), 83 per cent of Australia�s 1.2 million small businesses still bank with a major bank with only eleven per cent switching banks in the previous two years. Of those small businesses that did switch, 61 per cent changed to a major bank rather than a smaller institution. The Small Business Banking Issues Paper (September 2003)compiled by the Financial Services Consumer Policy Centre concluded that the level of competition in the small business banking market is �very poor�. This was because in FSCPC�s view a small number of large institutions dominate the small business banking market and fail to compete with each other on either price or service. This lack of mobility and �stickiness� of small business banking customers made it hard for small players and new entrants to compete for reasonable market share, even though they might have superior products.

127 As to larger businesses, the parties before this Tribunal included Coles Myer, Woolworths and Caltex which collectively handle approximately 38 per cent of EFTPOS transactions in Australia. The evidence before the Tribunal from those firms and also Australia Post was that they have been �categorically� advised in discussions with banks that if zero interchange fees are introduced the merchants will have to pay fees to the acquiring banks.

18.4 Merchants

128 On the assumption that acquirers will seek to recover the lost average cost of 21 cents per transaction from their merchant customers, the evidence as to the likely consequences is broadly as follows.

18.4.1 Abandonment of EFTPOS

129 It does seem clear there is unlikely to be any significant abandonment of EFTPOS. It is a popular and well-recognised system which consumers expect merchants to provide.

18.4.2 Pass on to consumers

130 Merchants are likely to pass on all, or a substantial part, of the increased fees they are charged by their acquirers. This could take the form of surcharge on EFTPOS transactions or a general price increase across the merchant�s range of goods or services or a variation of terms such as a higher minimum level for EFTPOS transactions.

18.4.3 Surcharge

131 As to surcharge, the evidence is equivocal. Mr Nicol of Caltex said that independently owned Caltex sites may surcharge. Coles Myer would consider the imposition of a surcharge and Mr Michael Kovas, the owner of two small supermarkets in New South Wales, stated that he would be most likely to introduce a surcharge equivalent to the increased fees amount, although this could be a problem if other small retailers did not impose one too. However, in his opinion other small retailers may also not be operating sufficient margins to absorb the additional cost. Ms Wynn from Woolworths said her company had never surcharged the customer but she was not in a position to say if a decision to surcharge was even in contemplation. Mr Charles Gunnee, owner of a card and gift shop in Bankstown, New South Wales, said he would need to pass on any increase in merchant service fees. He had not yet decided the form this would take. He was constrained by a particular feature of his business, viz that 90 per cent of cards and associated products such as wrapping paper and ribbon come pre-priced by manufacturers. Mr Timothy Williams, who owns three card and gift shops (a field of retail activity perhaps over-represented in the Merchants� case) has not decided what he would do but his understanding of other small retailers� experience was that surcharging credit card transactions met with �considerable customer resistance�. Mr Gove thought a surcharge was �unrealistic� and that he knew of no instances anywhere in the world where there was surcharge on debit card use.

132 The material available to us did not enable a firm finding about this although we are inclined to think that the Banks are overstating matters in their submission that it is �most unlikely� that retailers would impose a surcharge. The RBA did point to the lack of surcharging since its reforms made surcharge for credit card transactions lawful. It says that other than �isolated instances in niche markets� little surcharging has occurred. The niche markets include that in which Qantas participates, a fairly large niche. Moreover during the course of the hearing it was publicly announced by Telstra that it would be imposing credit card surcharges.

18.4.4 Variation of terms

133 As for variation of terms, such as a higher minimum level for EFTPOS transactions, that may result in an increased use of cash, a lower cost payment system and perhaps, as the Banks suggest, a more efficient alternative, although it is not clear that it will inevitably be a cheaper alternative vis-�-vis EFTPOS in small value transactions. However variation of terms would also be subject to competitive pressures.

18.4.5 General price increase

134 All in all we find the most likely consequence (without excluding at all the possibility of other consequences) is a general price increase. To the extent that an increased cost is borne by all competitors in a market, especially where margins are small, the more likely it is that competitors in that market will raise prices rather than absorb the cost.

18.4.6 Disincentive for investment

135 Notwithstanding that abandonment of EFTPOS can be ruled out, the increased cost for merchants arising from the Proposed Agreement is likely to operate as a disincentive to undertake future investment in the upgrading of the system. An example was provided by Mr Nicol of Caltex, who said that his company had decided that it would not proceed with the introduction of EFTPOS readers at petrol pumps in a zero interchange fee environment.

136 True it is, as the Banks submitted, EFTPOS terminals serve other purposes; they are used for the processing of credit cards and are often integrated with merchants� internal accounting systems. Thus there are reasons why merchants would want to upgrade their infrastructure regardless of what happens to interchange fees. It is not possible to come to a firm conclusion on this particular issue, depending as it does on commercial reactions at some unspecified future time to proposed technological innovations of unspecified cost. Perhaps the forecast of Ms Winn of Woolworths is the best that can be made. She said:

�…the choice as to whether [Woolworths] upgrades to a Rolls Royce or a tractor in terms of future support of the EFTPOS network will come down to the business case surrounding that. So the answer will be, we will probably continue to support EFTPOS but the way in which we do it may change over

time.�

137 We think there is likely be a disincentive for investment to some extent. The evidence however did not allow us to make a finding whether such disincentive amounted to a public detriment.

19. The proposed access regime

138 As already noted, the prospects of an access regime lowering barriers of entry to the acquiring and issuing markets was the critical factor as far as the ACCC was concerned.

139 The Banks submitted that the ACCC erred in this view. They pointed to the evidence of Ms Rhonda Smith, the economist called by the ACCC who agreed that there could potentially be public benefit without the access reform, depending on the balance of detriments and benefits. This was because access reform was not necessary for the interchange fee to be passed through to cardholders, although it may add to pass through.

140 All parties agree that access reform is desirable because it will enhance competition, and as a consequence efficiency, in the issuing and acquiring markets. However it seems preferable, all other things being equal, to achieve that objective by access reform rather than an agreement which is per se unlawful.

141 The evidence as to proposals for access reform were given by Mr Peter Smith of APCA. Its work on the access regime started in April 2003. An EFTPOS Access Working Group was established and after preparation of a Parameters Report a workshop was conducted on 29 January 2004. With the assistance of solicitors Gilbert and Tobin a draft options report was being prepared and at the time Mr Smith gave evidence (16 April 2004) this was expected by the end of the month. As it turned out, the draft report was produced to us on 29 April. A number of issues remain open. APCA is, in Mr Smith�s words, �still wrestling with a number of issues� in relation to physical connectivity.

142 At the moment we are unable to say what an access regime would look like in any detail or how it would work.

143 Coles Myer pointed out that an enhanced access regime is unlikely to make much difference for new entrant issuers since they would primarily be concerned with gaining approval as an Authorised Deposit Taking Institution under the Banking Act 1959 (Cth) and complying with the Australian Prudential Authority�s standards. So the likelihood of the access regime leading to increased competition among issuers, and the sending of signals to debit cardholders, is negligible.

144 For present purposes one question is how any access regime would affect, one way or the other, the negative interchange fee. Given some sort of access regime, is the prospect of a successful entry likely to be helped by there being one zero interchange fee as opposed to the present position of multiple fee arrangements which have to be individually negotiated? We think that it is unlikely that a zero interchange fee would facilitate entry to any significant extent.

145 Mr Smith disavowed any suggestion that a zero interchange fee is a necessary condition of an EFTPOS access regime. All he would say is that bilaterally negotiated interchange fees would �add a layer of complexity to an access regime because they would replace a constant (ie zero interchange fees) with a variable (ie negotiated interchange fees)�. A new entrant would of course have to negotiate bilateral agreements with other parties in the network. Those agreements involve many features. With access reform there may be less features to negotiate. But to remove simply one feature and thus �one less hassle� seems unlikely to make the difference between entry and rejection. In any event, the removal of one issue which has to be negotiated may mean the removal of the one feature which a new entrant may be able to exploit.

146 Finally, and most importantly, it is clear that technical barriers are much more important. The access regime being developed by APCA will not address the significant barrier of non-standard electronic message formats. The APCA Board has determined that changes to the physical and technological architecture is not to be considered by the Working Group. There is a project for the review of the CECS Manual, but this rules out �any new technical or operating requirements that are not current practice�. Whether or not the non-standard message format would be removed as part of the review was disputed.

147 Mr Gove�s evidence as to the technical difficulties facing a new entrant have already been referred to ([50]-[52] above). This is confirmed by the evidence of Mr Paul Wood whose firm MoneySwitch is a potential new entrant. He believed that interconnection under the current technology system would take six to twelve months or longer. His attempts to negotiate with banks to date have been unsuccessful and he has been effectively rebuffed.

20. Alleged benefits

148 The fundamental benefit claimed in support of the authorisation is that increased use of EFTPOS will mean greater use of a lower cost means of payment and therefore more efficient payment systems as a whole.

149 However, all five economists who gave evidence agreed with the proposition that in assessing the efficiency with which resources are allocated among products with different characteristics, one cannot say that all resources should be allocated to those products whose production involves the least cost. In any case, on the evidence we have, it is not even clear that EFTPOS has the lowest cost, as we have noted in relation to the Joint Study ([70] above). By the same token, the Banks� argument that there are currently distorted price signals has not been established.

150 As Dr Williams points out, it should not be surprising that credit card transactions involve significantly more cost compared with debit card transactions. They have different characteristics, primarily caused by the credit card combining a loan with a payment mechanism. The extra costs identified in the Joint Study are incurred because of the difference in the characteristics of the product that is offered to the cardholder and the merchant. Moreover, those differing characteristics of credit transactions compared with debit transactions are likely to affect consumers� choice and willingness to pay for the credit transaction.

151 Secondly, again accepting Dr William�s evidence, as a matter of allocative efficiency (as distinct from productive efficiency) one cannot assess relative efficiency simply by comparing costs. It is not necessarily more efficient to use a cheaper product as opposed to a more expensive product if they confer different benefits. One of the benefits of free markets is that they provide a range of products in accordance with the preferences of consumers, thus encouraging greater use of a less costly product at the expense of a more costly product where the two products have different characteristics. That the differing characteristics are likely to affect the relative value of the products to consumers does not encourage efficiency and is not necessarily a public benefit from an economic perspective.

152 The point may be illustrated by the reference to the use of cash as a payment system. Depending on the circumstances, cash is probably the cheapest, certainly from a merchant�s point of view. It is certainly cheaper from everyone�s point of view than credit cards or cheques. But would it be regarded as a public benefit if an otherwise anti-competitive agreement led to increased use of cash? Plainly not. Cash has obvious disadvantages, including the fact that sometimes people do not have any, and risks of robbery, embezzlement and the like.

21. Conclusion

153 We do not agree with the ACCC that the public benefits and detriments of the Proposed Agreement are finely balanced. Any public benefits are clearly outweighed by the detriments.

154 We are not satisfied, on the evidence available to us, that the Proposed Agreement would result in significant increased use of EFTPOS. This is because there is no satisfactory evidence to show

(a) the extent of pass on of benefits to cardholders;

(b) signalling of any such benefits to cardholders; and

155 Encouraging a switch from credit cards to debit cards is not warranted on allocative efficiency grounds. They are simply different products.

156 In any event, a switch to debit cards is now occurring as a result of the RBA reforms.

157 There is real public detriment in the likelihood of a flow on of costs to consumers generally. The Proposed Agreement is likely to have the effect of passing on to the general body of consumers an annual cost of $170 million, or a substantial part thereof. This cost has up until now, following freely negotiated agreements, been incurred within the banking system and recovered from bank customers. We see little public benefit in allowing this change to come about by the means of a per se unlawful agreement.

158 In our opinion the authorisation should be set aside.

I certify that the preceding One hundred and fifty-eight (158) numbered paragraphs are a true copy of the Determination herein of the Australian Competition Tribunal.

Associate:

Dated:

Counsel for the Banks	A C Archibald QC and P Zappia

Solicitors for the Banks	Mallesons Stephen Jaques

Counsel for the Merchants	N J Young QC and M O�Bryan

Solicitors for the Merchants:	Minter Ellison

Counsel for Coles Myer Ltd	J E Middleton QC and K Knights

Solicitors for Coles Myer Ltd	Corrs Chambers Westgarth

Counsel for the Reserve Bank of Australia	T F Bathurst QC and A Payne

Solicitors for the Reserve Bank of Australia	Clayton Utz

Counsel for the ACCC	C M Maxwell QC and J Beard

Solicitor for the ACCC	Australian Government Solicitor

Date of Hearing:	14, 15, 16, 26, 27, 29, 30 April 2004

Date of Determination:	25 May 2004